Health informatics : a systems perspective [Second ed.] 9781640550056, 1640550054

Table of contents :
Cover
Brief Contents
Detailed Contents
Preface
Chapter 1 - Health Systems Informatics: A Transformational Science
Chapter 2 - Knowledge-Based Decision Making
Chapter 3 - Health Professions, Patients, and Decisions
Chapter 4 - The Coming of the Corporation: Transforming Clinical Work and Processes
Chapter 5 - Predictive Analytics in Knowledge Management
Chapter 6 - Clinical Decision Support Systems in Medicine
Chapter 7 - Nursing Informatics
Chapter 8 - E-Health and Consumer Health Informatics
Chapter 9 - Precision Medicine
Chapter 10 - Information Systems as Integrative Technology for Population Health
Chapter 11 - Global Heatlh Systems Informatics
Chapter 12 - Controlled Terminology and the Representation of Data and Information
Chapter 13 - Information Management Technology
Chapter 14 - The Role of People and Information in Delivering Patient-Centered Care
Chapter 15 - Valuation and Financing of Healthcare Services and Information Technology Infrastructure
Chapter 16 - Data and Information Security in the Healthcare Enterprise
Appendix: Professional Societies Accrediting Agencies, and Additional Insights in Health Informatics
Glossary
Index
About the Authors/Editors
About the Contributors

Citation preview

HAP/AUPHA Editorial Board for Graduate Studies Carla A. Stebbins, PhD, Chairman Rochester Institute of Technology Kevin Broom, PhD University of Pittsburgh Erik L. Carlton, DrPH University of Memphis Daniel Estrada, PhD University of Florida Edmond A. Hooker, MD, DrPH Xavier University LTC Alan Jones, PhD, FACHE US Army Christopher Louis, PhD Boston University Peggy J. Maddox, PhD George Mason University Donna Malvey, PhD University of Central Florida Brian J. Nickerson, PhD Icahn School of Medicine at Mount Sinai Stephen J. O’Connor, PhD, FACHE University of Alabama at Birmingham Maia Platt, PhD University of Detroit Mercy Debra Scammon, PhD University of Utah Tina Smith University of Toronto James Zoller, PhD Medical University of South Carolina

Health Administration Press, Chicago, Illinois Association of University Programs in Health Administration, Washington, DC

Your board, staff, or clients may also benefit from this book’s insight. For information on quantity discounts, contact the Health Administration Press Marketing Manager at (312) 424-9450. This publication is intended to provide accurate and authoritative information in regard to the subject matter covered. It is sold, or otherwise provided, with the understanding that the publisher is not engaged in rendering professional services. If professional advice or other expert assistance is required, the services of a competent professional should be sought. The statements and opinions contained in this book are strictly those of the author(s) and do not represent the official positions of the American College of Healthcare Executives, the Foundation of the American College of Healthcare Executives, or the Association of University Programs in Health Administration. Copyright © 2019 by the Foundation of the American College of Healthcare Executives. Printed in the United States of America. All rights reserved. This book or parts thereof may not be reproduced in any form without written permission of the publisher. 23 22 21 20 19

5 4 3 2 1

Library of Congress Cataloging-in-Publication Data Names: Brown, Gordon D., editor. | Pasupathy, Kalyan S., editor. | Patrick, Timothy B., editor. Title: Health informatics : a systems perspective / [edited by] Gordon D. Brown, Kalyan S. Pasupathy, Timothy B. Patrick. Description: Second edition. | Chicago, Illinois : Health Administration Press (HAP) ; Washington, DC : Association of University Programs in Health Administration (AUPHA), [2019] | Includes bibliographical references and index. Identifiers: LCCN 2018026342 (print) | LCCN 2018027304 (ebook) | ISBN 9781640550063 (ebook) | ISBN 9781640550070 (xml) | ISBN 9781640550087 (epub) | ISBN 9781640550094 (mobi) | ISBN 9781640550056 (alk. paper) Classification: LCC R858 (ebook) | LCC R858 .H3478 2019 (print) | DDC 610.285—dc23 LC record available at https://lccn.loc.gov/2018026342

The paper used in this publication meets the minimum requirements of American National Standard for Information Sciences—Permanence of Paper for Printed Library Materials, ANSI Z39.48-1984.  ∞ ™ Acquisitions editor: Jennette McClain; Project manager: Andrew Baumann; Manuscript editor: Jane Calayag; Cover designer: James Slate; Layout: Cepheus Edmondson Found an error or a typo? We want to know! Please e-mail it to [email protected], mentioning the book’s title and putting “Book Error” in the subject line. For photocopying and copyright information, please contact Copyright Clearance Center at www.copyright.com or at (978) 750-8400. Health Administration Press Association of University Programs A division of the Foundation of the American   in Health Administration   College of Healthcare Executives 1730 M Street, NW 300 S. Riverside Plaza, Suite 1900 Suite 407 Chicago, IL 60606-6698 Washington, DC 20036 (312) 424-2800 (202) 763-7283

To students who have the vision and courage to lead profound change in the health system. To Kathleen for her values, high standards, and genius for working with and teaching children. —Gordon D. Brown To my beloved wife Jocey and my son Neal, who have each had a considerable effect on my perspective on life. —Kalyan S. Pasupathy To Lillian, my wife, colleague, and friend. —Timothy B. Patrick

BRIEF CONTENTS

Preface.....................................................................................................xvii Chapter 1. Health Systems Informatics: A Transformational Science....1 Gordon D. Brown Chapter 2. Knowledge-Based Decision Making..................................21 Gordon D. Brown, Kalyan S. Pasupathy, and Mihail Popescu Chapter 3. Health Professions, Patients, and Decisions......................49 Gordon D. Brown Chapter 4. The Coming of the Corporation: Transforming Clinical Work Processes..............................................................73 Gordon D. Brown Chapter 5. Predictive Analytics in Knowledge Management...............97 Gordon D. Brown, Kalyan S. Pasupathy, and Mihail Popescu Chapter 6. Clinical Decision Support Systems in Medicine...............121 Pavithra I. Dissanayake and Karl M. Kochendorfer Chapter 7. Nursing Informatics.......................................................147 Carol G. Klingbeil, Pei-Yun Tsai, and Timothy B. Patrick Chapter 8. E-health and Consumer Health Informatics....................167 George Demiris and Blaine Reeder Chapter 9. Precision Medicine.........................................................191 Timothy B. Patrick and Aurash A. Mohaimani Chapter 10. Information Systems as Integrative Technology for Population Health.......................................................207 Julie M. Kapp Chapter 11. Global Health Systems Informatics.................................227 Gordon D. Brown

vii

viii

B rief Co n t ents

Chapter 12. Controlled Terminology and the Representation of Data and Information..................................................251 Timothy B. Patrick and Carmelo Gaudioso Chapter 13. Information Management Strategy.................................269 James D. Buntrock Chapter 14. The Role of People and Information in Delivering Patient-Centered Care................................291 Naresh Khatri Chapter 15. Valuation and Financing of Healthcare Services and Information Technology Infrastructure.......................315 Kalyan S. Pasupathy and Gordon D. Brown Chapter 16. Data and Information Security in the Healthcare Enterprise............................................341 Dixie B. Baker and Timothy B. Patrick Appendix Professional Societies, Accrediting Agencies, and Additional Insights in Health Informatics.............369 Timothy B. Patrick Glossary..................................................................................................375 Index......................................................................................................381 About the Authors/Editors.......................................................................411 About the Contributors............................................................................413

DETAILED CONTENTS

Preface.....................................................................................................xvii Chapter 1. Health Systems Informatics: A Transformational Science....1 Gordon D. Brown Learning Objectives...........................................................1 Key Concepts.....................................................................1 Introduction......................................................................1 Complex Adaptive Systems in Healthcare...........................2 Bioinformatics....................................................................4 Medical Informatics............................................................5 Public Health Informatics...................................................7 Health Systems Informatics and Transformational Change...............................................8 Management Information Science....................................12 Conclusion.......................................................................13 Chapter Discussion Questions..........................................14 Case Study: Electronic Health Records: Where Does the System End?..........................................................15 Additional Resources........................................................18 References........................................................................18 Chapter 2. Knowledge-Based Decision Making..................................21 Gordon D. Brown, Kalyan S. Pasupathy, and Mihail Popescu Learning Objectives.........................................................21 Key Concepts...................................................................21 Introduction....................................................................21 Definition and Use of Knowledge in Decision Making......22 Knowledge Organizations................................................24 Clinical Knowledge Management.....................................25 Patient-Centered Care......................................................33 Transformational Strategy.................................................36 Knowledge Socialization...................................................37 Knowledge Brokering.......................................................38 ix

x

Det a iled Co n te n ts

Conclusion.......................................................................40 Chapter Discussion Questions..........................................41 Case Study: Knowledge Management in Accountable Care Organizations..................................41 Additional Resources........................................................44 References........................................................................44 Chapter 3. Health Professions, Patients, and Decisions......................49 Gordon D. Brown Learning Objectives.........................................................49 Key Concepts...................................................................49 Introduction....................................................................49 Transformation of the Clinical Function...........................50 The Science of Clinical Decision Making..........................54 Conclusion.......................................................................66 Chapter Discussion Questions..........................................66 Case Study: Redesigning Futures: The First-Ever Engineering-Driven College of Medicine.....................67 Additional Resources........................................................69 References........................................................................70 Chapter 4. The Coming of the Corporation: Transforming Clinical Work Processes..............................................................73 Gordon D. Brown Learning Objectives.........................................................73 Key Concepts...................................................................73 Introduction....................................................................73 Traditional Corporate Structures as the Logic for Clinical Information Systems.......................................75 Standardization of Clinical Work Processes.......................78 Integrated Systems Perspectives........................................83 Conclusion.......................................................................90 Chapter Discussion Questions..........................................90 Case Study: Not All Innovation Is Created Equal in the Transition to Value-Based Care....................................91 Additional Resources........................................................93 References........................................................................93 Chapter 5. Predictive Analytics in Knowledge Management...............97 Gordon D. Brown, Kalyan S. Pasupathy, and Mihail Popescu Learning Objectives.........................................................97

D etailed C ontents

Key Concepts...................................................................97 Introduction....................................................................97 Data Mining and Analytics...............................................98 Database Types and Their Impact on Data Mining.........101 Data-Mining Methods....................................................104 Dynamic Systems Modeling............................................106 Conclusion.....................................................................113 Chapter Discussion Questions........................................114 Case Study: Analytics for Disease Management and Wellness..............................................................115 References......................................................................118 Chapter 6. Clinical Decision Support Systems in Medicine...............121 Pavithra I. Dissanayake and Karl M. Kochendorfer Learning Objectives.......................................................121 Key Concepts.................................................................121 Introduction..................................................................121 Definition......................................................................122 History and National Policies.........................................122 CDSS Types...................................................................124 Effective Characteristics..................................................128 Design and Implementation...........................................129 Challenges and Barriers..................................................131 Clinical Domain Examples..............................................134 Conclusion.....................................................................136 Chapter Discussion Questions........................................137 Case Study: Effective CDSS Implementation..................137 Additional Resources......................................................140 References......................................................................140 Chapter 7. Nursing Informatics.......................................................147 Carol G. Klingbeil, Pei-Yun Tsai, and Timothy B. Patrick Learning Objectives.......................................................147 Key Concepts.................................................................147 Introduction..................................................................147 Informatics, Nursing, and the Transformation of Clinical Care..............................................................149 Roles of Nurses in Informatics........................................150 Nursing Work and Information System Applications.......151 Quality and Safety of Care..............................................154

xi

xii

Det a iled Co n te n ts

Consumer Engagement..................................................157 Nursing Education and Research....................................160 Research and Practice.....................................................160 Conclusion.....................................................................161 Chapter Discussion Questions........................................161 Case Study: A Question of Evidence...............................162 Additional Resources......................................................164 References......................................................................164 Chapter 8. E-health and Consumer Health Informatics....................167 George Demiris and Blaine Reeder Learning Objectives.......................................................167 Key Concepts.................................................................167 Introduction..................................................................167 Review of Patient-Centered Systems...............................169 Social Media and Consumer Health Informatics.............171 Challenges in E-health Applications................................175 Success Factors for E-health...........................................177 Conclusion.....................................................................180 Chapter Discussion Questions........................................181 Case Study: Blue River Home Care................................182 References......................................................................185 Chapter 9. Precision Medicine.........................................................191 Timothy B. Patrick and Aurash A. Mohaimani Learning Objectives.......................................................191 Key Concepts.................................................................191 Introduction..................................................................191 Precision Medicine and Genomic Science.......................192 Precision Medicine Initiatives.........................................193 Precision Medicine and Popular Culture.........................194 Precision Medicine and Big Data....................................194 Precision Medicine and Scientific Reproducibility...........197 Conclusion.....................................................................199 Chapter Discussion Questions........................................200 Case Study: Whose Body?...............................................200 Additional Resources......................................................203 References......................................................................203

D etailed C ontents

Chapter 10. Information Systems as Integrative Technology for Population Health.......................................................207 Julie M. Kapp Learning Objectives.......................................................207 Key Concepts.................................................................207 Introduction..................................................................207 Status of Population Health in the United States............208 Population Health Provisions in the Affordable Care Act.............................................209 Difference Between Public Health and Population Health.....................................................211 Population Health Management in the United States.....212 Population Health as a System........................................215 Integrating Healthcare and Public Health Through Systems Design...........................................220 Conclusion.....................................................................221 Chapter Discussion Questions........................................222 Case Study: Pemiscot County.........................................222 Additional Resources......................................................223 References......................................................................223 Chapter 11. Global Health Systems Informatics.................................227 Gordon D. Brown Learning Objectives.......................................................227 Key Concepts.................................................................227 Introduction..................................................................227 Comparative Analysis of Health Systems Informatics: Design and Function.................................................228 Restructuring Health Systems According to the Logic of Health Systems Informatics....................232 Knowledge-Based Health Systems Design.......................236 Development of Global Health Systems: Collaborative Systems................................................237 Global Health Policy and Population Health..................243 Conclusion.....................................................................245 Chapter Discussion Questions........................................245 Case Study: Envisioning a Global Community................246 Additional Resources......................................................248 References......................................................................248

xiii

xiv

Det a iled Co n te n ts

Chapter 12. Controlled Terminology and the Representation of Data and Information..................................................251 Timothy B. Patrick and Carmelo Gaudioso Learning Objectives.......................................................251 Key Concepts.................................................................251 Introduction..................................................................252 Health Informatics and Representational Science............252 Surrogate Representations and Controlled Terminology Components.......................253 Two Basic Uses of Controlled Terminology....................257 Metadata and Metadata Schemata...................................258 Common Data Elements and Repositories......................260 Interoperability and the Terminology Problem...............261 Terminology Mapping....................................................262 Quality Measurement, Variation, and Coding.................263 Conclusion.....................................................................264 Chapter Discussion Questions........................................264 Case Study: A Problem of Display Codes........................265 Additional Resource.......................................................266 References......................................................................266 Chapter 13. Information Management Strategy.................................269 James D. Buntrock Learning Objectives.......................................................269 Key Concepts.................................................................269 Introduction..................................................................269 Data as Assets.................................................................270 Steps in Strategy Development.......................................272 Data Movement.............................................................280 Data Representation.......................................................281 Data Accessibility...........................................................283 Other Considerations.....................................................283 Conclusion.....................................................................286 Chapter Discussion Questions........................................287 Case Study: Guiding a Merger........................................287 Additional Resources......................................................288 References......................................................................288

D etailed C ontents

Chapter 14. The Role of People and Information in Delivering Patient-Centered Care................................291 Naresh Khatri Learning Objectives.......................................................291 Key Concepts.................................................................291 Introduction..................................................................291 Major Features of the Health Services Delivery Process........................................................292 Proactive Work Behaviors in Health Services Delivery Process........................................................293 HRM Capabilities..........................................................295 HIT Capabilities............................................................299 Is a Homegrown System Better Than an Outsourced System?.....................................................................304 Conclusion.....................................................................305 Chapter Discussion Questions........................................305 Case Study: University Hospital......................................306 Additional Resources......................................................308 References......................................................................308 Chapter 15. Valuation and Financing of Healthcare Services and Information Technology Infrastructure.......................315 Kalyan S. Pasupathy and Gordon D. Brown Learning Objectives.......................................................315 Key Concepts.................................................................315 Introduction..................................................................315 Financing Models...........................................................316 Lack of Coordination and Information Sharing..............320 Redesigning Structure and Financing.............................321 Valuation of Healthcare Services.....................................326 Valuation of IT Infrastructure.........................................328 Conclusion.....................................................................335 Chapter Discussion Questions........................................336 Case Study: Med City’s Diabetes Management Care Group...............................................................336 Additional Resources......................................................338 References......................................................................338

xv

xvi

Det a iled Co n te n ts

Chapter 16. Data and Information Security in the Healthcare Enterprise............................................341 Dixie B. Baker and Timothy B. Patrick Learning Objectives.......................................................341 Key Concepts.................................................................341 Introduction..................................................................342 Health and Financial Risks Related to Data and Information Security...................................342 Resources for Tracking Breaches.....................................343 Technical Issues in the Protection of Data and Information................................................344 Fair Information Practices..............................................347 Security Implementation in EHR Technology.................348 New and Developing Opportunities and Challenges.......355 Authentication Redux....................................................359 Conclusion.....................................................................361 Chapter Discussion Questions........................................362 Case Study: Heinz Children’s Health.............................363 Additional Resources......................................................366 References......................................................................367 Appendix Professional Societies, Accrediting Agencies, and Additional Insights in Health Informatics.............369 Timothy B. Patrick Professional Societies......................................................369 Accreditation and Other Organizations..........................372 Additional Insights.........................................................373 References......................................................................374 Glossary..................................................................................................375 Index......................................................................................................381 About the Authors/Editors.......................................................................411 About the Contributors............................................................................413

PREFACE Could a hurricane in Puerto Rico affect clinical practice in Washington, DC, eight weeks later? Yes. In fact, it did just that. Before Hurricane Maria nearly destroyed the island in the fall of 2017, Puerto Rico supplied more pharmaceutical products to the US market than did any other state or territory—nearly $40 billion worth. These products included intravenous (IV) bags that contain saline solution to which drugs are added later or that are preloaded with a mixture of medications. Plants in Puerto Rico that manufactured the IV bags were shut down in the aftermath of the hurricane, leading to a shortage of these bags in US hospitals. Patients at MedStar Washington Hospital Center in Washington, DC, for example, who typically received IV medications, were administered the pill forms of the drugs instead (Kodjak 2017). All parts of the world, along with their endeavors and challenges, have become increasingly interconnected. What happens in one system (e.g., community, territory, country) affects in some ways the activities and outcomes in other systems. As evidenced by the long-range effects of the catastrophe in Puerto Rico, the health system is affected by environmental, infrastructure, manufacturing, economic, and many other systems. We believe that information technology (IT) informed by these interconnected systems (what we refer to as health systems informatics) is necessary to properly support clinical care, clinical decision making, and healthcare management. Health systems informatics has the power to enable the transformation of the US health system and individual healthcare organizations into entities characterized by information sharing, coordinated care, patient centeredness, and evidence-based clinical decisions. In this second edition of Health Informatics: A Systems Perspective, we once again cover both conceptual and physical IT systems that interact with and affect healthcare processes and outcomes. All chapter and case study authors come from both academic and practice settings and represent a wide range of training and experience in the health informatics field. Such a diversity imparts a balanced theoretical and practical perspective to this book. This book examines health systems informatics in the context of clinical decision making across the health professions (chapters 2, 3, and 6), knowledge management (chapter 5), interactions and interdependencies among the xvii

xviii

Prefa c e

health professions (chapters 1, 3, 7, and 14), developments in IT and data representation (chapters 12 and 13), cybersecurity (chapter 16), population health and global health (chapters 10 and 11), management and corporate systems (chapters 4 and 15), Big Data (chapters 9 and 16), advances in science and scientific medicine (chapter 9), healthcare financing and valuation (chapter 15), and the role of patients and e-health (chapters 3 and 8). Each chapter offers the following: • Learning Objectives that list the main takeaways from the discussion • Key Concepts that list the major topics explored and terms used • Sidebars that present extra information, examples, scenarios, or opportunities for critical thinking and application • Terminology definitions on the page • Chapter Discussion Questions that serve as a framework for reviewing, conceptualizing, or articulating the concepts • A Case Study that translates the theories into real-world situations • Case Study Discussion Questions that challenge the reader’s understanding and judgment • Additional Resources that point to websites, books, and journal articles relevant to the concepts discussed • References that include both current and classic publications A glossary, an appendix (Professional Societies, Accrediting Agencies, and Additional Insights in Health Informatics), and an index round out the book. Writing this second edition with a systems perspective was a daunting but rewarding task. We hope you find the culmination of our work to be beneficial and valuable to your studies and career. Gordon Brown, Kalyan Pasupathy, and Timothy Patrick

Reference Kodjak, A. 2017. “Hurricane Damage to Manufacturers in Puerto Rico Affects Mainland Hospitals, Too.” Published November 15. http://wuwm.com/post/ hurricane-damage-manufacturers-puerto-rico-affects-mainland-hospitals-too.

Prefac e

Instructor Resources This book’s instructor resources include the authors’ responses to the chapter and case study discussion questions; PowerPoint slides to supplement classroom discussions and lectures; and suggested activities for exploring chapter topics, including data sets. For the most up-to-date information about this book and its instructor resources, go to ache.org/HAP and browse for the book by title or author names. This book’s instructor resources are available to instructors who adopt this book for use in their course. For access information, please e-mail [email protected].

xix

CHAPTER

HEALTH SYSTEMS INFORMATICS: A TRANSFORMATIONAL SCIENCE

1

Gordon D. Brown

Learning Objectives After reading this chapter, you should be able to do the following: • Understand the concept of open systems theory. • Conceptualize health systems informatics and differentiate it from bioinformatics and biomedical informatics as an analytical framework. • Explain the transformative power of information technology. • Discuss the differences in concept but interdependencies in function between management information systems and health systems informatics. • Apply clinical information technology to process improvement and system transformation.

Key Concepts • • • • •

Complex adaptive systems Conflict between business and clinical functions Health systems informatics and biomedical informatics Management information systems Transformational change

Introduction Health systems informatics strives to align the disparate components of a healthcare organization—professional, financial, and organizational—to achieve optimal system performance. How did these components become so

Health systems informatics Application of multidisciplinary sciences to transform (not just automate) the structure and behavior of systems, organizations, and individuals who interact to provide personalized care

1

2

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

dysfunctional in an advanced nation such as the United States? The great irony is that we built the US health system that way, and we continue to maintain it that way. Health systems informatics assumes a larger integrated systems perspective, according to systems theory (Encyclopedia.com 2001): As a way of looking at things, the “systems approach” in the first place means examining objects or processes, not as isolated phenomena, but as interrelated components or parts of a complex. An automobile may be seen as a system; a car battery is a component of this system. The automobile, however, may also be seen as a component of a community or a national transportation system. Indeed, most systems can be viewed as subsystems of more encompassing systems.

This chapter examines the interdependencies among clinical, organizational, financial, and individual (patient) functions and the role of health systems informatics in guiding and connecting these functions to achieve optimal performance. Health systems informatics provides the decision-making logic that serves as the basis for designing, financing, managing, and evaluating the healthcare organization to improve performance. It enables and requires the transformation of the roles and behaviors of health professionals but does not diminish them. It views the clinical function from the perspective of complex adaptive systems. The result of these efforts is transformational change, wherein each function is fundamentally realigned to serve the more encompassing system.

Complex Adaptive Systems in Healthcare Complex adaptive system Organization with a large number of interdependent parts or agents that have their own pattern relationships, present interaction complexity, and are self-organizing but can adapt to their environments and help create those environments

Health systems informatics enables the conceptualization of an integrated clinical perspective, based on the theory of complex adaptive systems. Complex adaptive systems are characterized by large numbers of interdependent parts or agents—each with its own pattern relationships and interaction complexities—that adapt to and create their environments through coevolution (Akgün, Halit, and Byrne 2014; Birdsey, Szabo, and Falkner 2017; Lee and Mongan 2009). By nature, dynamic systems are transformational in that they call for new delivery models, professional roles, organizational structures, and system designs, but they are difficult to analyze. If systems reject or are slow to react to new enabling technologies, they will underperform and their very viability might be threatened. A systems perspective or systems theory recognizes the following: 1. The health of a population is not determined primarily by its health system, no matter how structured, but by its community and lifestyle.

C h a p te r 1:   H e a l th Sy s te ms I nfor m atic s: A Transfor m ational Sc ienc e

2. Simple generalizations about how services should be provided or how change should be made—such as whether privatizing services will make them more efficient—are rejected. A systems perspective does not narrowly focus on legacy pieces of a system—such as how physicians should be educated, healthcare services financed, or information systems structured—but broadly considers how the overall system should function to obtain superior results. 3. The functions and strengths of both the private (investor-owned) and public sectors, as well as the nonprofit or “plural” sector that draws unique strengths from the private and public sectors (Mintzberg 2015), must be considered. The plural sector has long been the organizational basis for much of the US health system and might offer even greater potential for the future. A full exploration of the private, public, and plural sectors is beyond the scope of this book, but they are fundamental to an understanding of health system effectiveness and efficiency. The design of health systems is not determined by traditional or prescriptive structures and roles of organizations, financing, health professionals, or information systems. Each of these functions is subordinate to optimal outcome measures of clinical quality, continuity, patient satisfaction, efficiency, and population health. This assumption is made more complex because no template or single pathway to achieving optimal performance exists. Systems theorists use the term equifinality to suggest that any given end can be achieved by many different means. However, the means do not define the system and are accountable to the desired end state. Each region and country envisions its own system structure; each will be different and each can be optimal. Information technology (IT) can provide the systems architecture that enables the structure of functions, but it cannot dictate them. Each function must pursue its own design, measured against the crucible of optimal system performance. IT can enable a national or even global health information exchange, consistent with a patient-oriented system. Such exchange is a goal, but each system must craft its own structure and not impose any preconceived or central planning (teleologic) design on the professions, organizations, or other functions. Central control that imposes rigid designs on health system structure and function has proven to be ineffective (Garrety et al. 2016). Commercial and business applications have demonstrated the ability to facilitate the accomplishment of corporate goals while allowing local autonomy and freedom to innovate. Leading transformational change in clinical decision making is not based on a knowledge of computers, information science, or medical informatics. Although these sciences are important, they traditionally have been applied within health professions, organizations, and systems that are themselves

3

4

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

obsolete. Health systems informatics assumes a broader focus, including providing the science for guiding how organizations, systems, and professional and patient roles can be structured to improve system performance. Each function has its own identity and integrity. Although interdependent, functions are neither defined by nor subordinate to another function but to optimal system performance only. All must be aligned, or the system will be dysfunctional. For example, organizations might try to enforce the use of clinical guidelines and protocols through bureaucratic rules and sanctions based on organizational logic and reward. Neither organizations nor professions possess the dominant logic for system structure; this structure is based on optimal clinical quality, system efficiency, and population health. Patients should have access to information and participate in the clinical decision process, but they cannot determine the design or content of clinical decision support systems (CDSSs). Increased information sharing and participation of patients in clinical decisions will transform the health system. In addition, the role of the health professions in society is essential and protected, which is justified by their contribution not to the profession but to what society determines as optimal system performance (as discussed in chapter 3). Such protection assumes that as society changes, the professionals will change. Maintaining the historical domains and decision-making context of the medical profession, for example, violates the physicians’ protected role in society. However, a changing role for physicians does not necessarily mean a diminished role; in fact, it might be enhanced. The structure of healthcare organizations has long been detached from the clinical function and operates under the principle that organizations could not interfere with the practice of medicine and the autonomy of physicians. A landmark case occurred in the 1930s when Drs. Ross and Loos were removed from the Los Angeles County Medical Association. The association also wanted these physicians’ licenses revoked for violating professional ethics by engaging in the corporate practice of medicine (Starr 1982, 299–304). The Ross-Loos Group had established a group practice and the first managed care plan in the United States—a prepaid health plan that emphasized prenatal care and childhood immunization. To enable the health system to harness the power of IT, it must undergo an equally disruptive transformation, like the one Drs. Ross and Loos had taken. In this book, we explore the field of health systems informatics as a transformational science.

Bioinformatics The term informatics is credited to A. I. Mikhailov, of the scientific information department of Moscow State University, who first used it in his 1968 book Oznovy Informatiki (Foundations of Informatics) (Collen 1995). It is adopted

C h a p te r 1:   H e a l th Sy s te ms I nfor m atic s: A Transfor m ational Sc ienc e

from the Russian term informatik or informatikii, defined as a study of the “structure and general properties of scientific information and the laws of all processes of scientific communication” (Collen 1995, 39). This definition establishes informatics, at its root, as the study of linguistics applied broadly to scientific language. As such, the field of informatics combines basic science with computational science, particularly computer science. Bioinformatics can be defined as a form of computational linguistics—the statistical or rule-based modeling of scientific information. In 1976, the Oxford Dictionary defined informatics as the “discipline of science which investigates the structure and properties of scientific information, as well as the regularities of scientific information activity” (Collen 1995, 39). The focus of bioinformatics is on the management, analysis, and interpretation of data from biological experiments and observational studies (Moore 2007). The sequence analysis of the three billion chemical base pairs that make up human DNA would not be possible without complex algorithms and powerful computers. The standard language and the volume of data were a perfect match for the computer, and the level of analyses grew with the rapid increase in computer memory and processing speed. One might regard bioinformatics, with its focus on computational biology, as peripheral or unrelated to the topic of health systems informatics. Yet, IT enables bioinformatics to transcend the laboratory, informing clinical decision making as well as individual patients and consumers. Clinical bioinformatics has emerged as a field of translational science that integrates genomics and proteomics data with clinical data to provide molecular diagnostics, pharmacogenomics, and evidence-based clinical outcomes. Bioinformatics continues to evolve by incorporating diverse technologies and methodologies from disparate fields to apply advanced computational and informational tools to biomedical research (Mattick et al. 2014; Sarkar et al. 2011).

Medical Informatics A 1999 report of the Biomedical Information Science and Technology Initiative (BISTI), formed by the National Institutes of Health, described the field of informatics applied to healthcare and labeled it biomedical informatics (Friedman et al. 2004). The term was broadly applied to include bioinformatics, imaging informatics, clinical informatics, and public health informatics (exhibit 1.1). Imaging and clinical informatics have generally been included in the description of medical informatics, and we use the term medical informatics as the inclusive term. Although nursing informatics is considered its own area of scientific exploration (chapter 7), it is discussed here under the inclusive heading of medical informatics because it draws on the same informatics core

5

Bioinformatics Discipline that combines the biological sciences (microbiology, biochemistry, physiology, genetics) with computational fields (e.g., statistics, computer science)

Medical informatics Discipline that deals with the structure and properties of clinical information generated from clinical trials and medical records; generally includes imaging informatics and clinical informatics

6

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

EXHIBIT 1.1 Biomedical Informatics

Biomedical Informatics Methods, Techniques, and Theories

Basic Research

Applied Research

Bioinformatics

Imaging Informatics

Clinical Informatics

Public Health Informatics

Molecular and Cellular Processes

Tissues and Organs

Individuals (Patients)

Populations and Society

Source: Friedman et al. (2004). Reprinted with permission from Hanley & Belfus, Inc.

competencies that are applied to clinical practice. Although the BISTI report included public health informatics in its paradigm because it also draws on the same core competencies, we discuss it separately given its primary focus on population health and not medical care. The BISTI report was an important contribution because it delineated a core body of knowledge for informatics applied to the range of areas. Our focus builds on it to explore how this technology helps enable the transformation of these areas. The underlying theories, techniques, and methods that serve as the core competencies of medical informatics are algorithms, data structures, database design, ontology/vocabulary, knowledge representation, programming languages, software engineering, modeling, and simulation (Friedman et al. 2004). Medical informatics originated from the clinical area of pathology, which had developed standardized language applied to large data sets, requiring the distribution of standard test results to a range of clinical services. Physicians demanded tests that used standardized measures and processes so that the results could be interpreted on the basis of good science. Imaging informatics also benefited from standardized measures and language but lagged as a result of the limitations of the computer to store and process large amounts of data. In contrast, clinicians in other specialties valued flexibility in language (e.g., text over drop-down lists), tailoring medical records to individual choice and clinical decisions based on individual judgments and not the evidence derived from CDSSs.

C h a p te r 1:   H e a l th Sy s te ms I nfor m atic s: A Transfor m ational Sc ienc e

The step from measuring and reporting laboratory findings to developing the electronic medical record (EMR) was a major advance. The idea of an EMR has been around for half a century, but it proved to be a complex challenge in part because the structure of clinical information lacked basic vocabularies and data standards essential for a unified language. Automating the medical record was greatly hampered by the lack of computing power and a common clinical vocabulary. The first initiatives in EMR development focused on enabling clinicians to record and retrieve clinical data to frame a diagnosis and treatment plan. The EMR could locate and present clinical information about previous conditions, tests, diagnoses, and treatment protocols. The operative question was whether the EMR would be developed by individual clinicians, departments, organizations, or the health system. Lacking a clear conceptual framework or vision of future applications, healthcare entities pursued all four approaches, resulting in EMRs that were not compatible—even within institutions. The valued priority was initially to maintain organizational autonomy and not integration. Developing a common clinical language was thus made more complex by the proliferation of different systems with different vocabularies and syntax. The computer has become a tool for health professionals to record, store, retrieve, process, distribute, and integrate clinical information. As the health professions maintain and advance their own informatics perspectives, they inherently embrace collaboration and develop a greater team orientation, both of which are essential for improving clinical care. Digitizing information brought about changes in “cognitive and human factor interfaces,” but these changes were limited to clinical decision making within fundamentally traditional roles. This form of change might be characterized as evolutionary or transactional as opposed to transformational or innovative (Havighurst 2008; Herzlinger 2006; Stange, Ferrer, and Miller 2009). The availability of electronic information enabled not only greater and better information to be processed in a more readable form but also clinical evidence to be displayed and shared to support decisions that allow health professionals to better serve patients. Through this process, disparate EMRs became more standardized and integrated, enabling them to share clinical information and to access evidence-based clinical guidelines, thus transforming them into electronic health records (EHRs).

Public Health Informatics The first major conceptual development in informatics was in public health; although it used the same logic, it occurred long before the computer was envisioned. In the early 1800s, public health workers in many countries saw the need for a common vocabulary for classifying diseases and causes of death

7

8

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

that would enable the establishment of surveillance programs locally, nationally, and even internationally. The first initiative to standardize clinical information was the development of the International Classification of Diseases (ICD), first by the International Statistical Congress in 1853 and later by the World Health Organization (2018). Surveillance as an outcome measure is still a public health focus in the United States and most countries, as evidenced by the presence of many registries and surveys of illness and health status of populations. Registries are federally mandated and center on particular diseases (e.g., hemophilia), groups of similar diseases (e.g., cancer), or specific exposure (e.g., toxins in hazardous-waste sites). Each registry requires a degree of standardization to provide summary data on incidence and prevalence. Outcomes data, however, are difficult to link to individual clinical decisions. The aggregation of data, and data’s use as outcomes information, will be facilitated by advanced IT. In the future, public health registries can be generated from data in EMRs, making current efforts obsolete. A major transformation will be the integration of public health data systems with medical systems, which will enable medical care to be linked to health risk and population health. This transformation will result in the medical care system assuming shared responsibility for population health by focusing on individual health and wellness. Population health will be an important factor in the transformation to health systems informatics (chapter 10). Increasingly, health system leaders are envisioning IT as the integrating architecture for professionals, organizations, and patients within the larger regional, national, or international context. This vision is difficult to realize because of the complexity of the task and because clinicians, managers, and policymakers were trained within the silos of their own profession or discipline. As a result, efforts have been focused on the institutional level—that is, the strategy of integrating medical records within institutions. Integrating records across institutions has been pursued primarily through acquisition, leasing, or some other way the institutions are integrated. The challenge now is integrating IT across disparate professions and organizations. A common architecture for integrating these public–private, decentralized, and disparate systems of care is needed to meet this challenge, but an old saying must be heeded: “You cannot build a skyscraper by nailing together dog houses.” Health systems informatics includes and builds on biomedical informatics using new architecture that is based on system optimization.

Health Systems Informatics and Transformational Change Health systems informatics starts with the desired system outcomes and then considers the structure of each function of that system and the changes to those

C h a p te r 1:   H e a l th Sy s te ms I nfor m atic s: A Transfor m ational Sc ienc e

functions necessary to achieve the desired outcome. The focus is thus on the performance of the overall system and how each function—the professions, organizations, financing, IT, public policies, and so on—is aligned to achieve optimal performance. IT does not assume any preconceived structure but provides the logic for the system. This architecture, in turn, is based on open systems theory. Open systems theory views the organizational and clinical functions in terms of their relationship with and contribution to overall system performance. Healthcare organizations and professionals have historically pursued a closed-system focus, as have many business enterprises, assuming the independence of internal functions and adopting rationalistic approaches on the basis of optimizing individual functions. Open systems are considered not only in relation to their environment but also in relation to internal components because interactions between components affect the system as a whole. Open systems theory was developed by biological scientists such as Ludwig von Bertalanffy, who observed that (1) biological and social systems function in relationship to their environment in that they receive inputs, transform those inputs, and export outputs to maximize the system and (2) the functions are internally interdependent such that the interactions among them affect the system as a whole. This concept introduced a new science for understanding the structure and function of organizations and other social systems (Kast and Rosenzweig 1972). Introduced earlier, equifinality is the belief that there is no one best way to structure a system as long as the means deliver optimality. An underlying assumption of health systems informatics is that implementation of IT does not cause change in the health system but enables its transformation. The health system has lagged most business sectors in fully using advanced IT, instead concentrating its efforts on automating existing processes. Health system change will be realized through the vision and innovation of health leaders who recognize and embrace the transformative power of health systems informatics.

Extended Model The model developed by Friedman and colleagues (2004) based on the BISTI report (exhibit 1.1) is extended here to include complex adaptive systems (exhibit 1.2). This extension does not imply that the original model is incorrect but rather that it now includes the concepts of industrial engineering science, organizational

9

Open systems theory Theory that views organizational and clinical functions in terms of their relationship with and contribution to overall system performance

Opportunities for Interprofessional Education and Interoperability A cross-cutting theme in this book is the alignment of multiple systems—specifically, clinical teams (including patients) working together, knowingly or unknowingly, on clinical care processes and decision making—to provide the best possible patient care. This theme highlights two fundamental concepts—interprofessional education (IPE) and interoperability. As you read this book, think of opportunities for IPE and circumstances that signal the need for interoperability of semantics, systems, data, information, policies, persons, and other resources of the health system.

10

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

EXHIBIT 1.2 Health Systems Informatics

Health Informatics Methods, Techniques, and Theories

Basic Research

Applied Research

Bioinformatics

Imaging Informatics

Clinical Informatics

Public Health Informatics

Molecular and Cellular Processes

Tissues and Organs

Individuals (Patients)

Populations and Society

Health Systems Informatics Complex Adaptive Systems

Source: Adapted from Friedman et al. (2004).

theory and behavior, and systems theory. The exhibit 1.1 model correctly depicts the linear relationship of the core body and application of science with the four domains—bioinformatics, imaging informatics, clinical informatics, and public health informatics. For example, in clinical informatics, the EMR requires doctors to use the computer for charting and informing their decisions, thus changing behaviors but not the fundamental structure of the clinical process. When health professionals apply the methods, techniques, and theories of complex adaptive Transformational Change in Banking systems (exhibit 1.2), decisions Understanding the extended biomedical informatics model (exhibit and work processes are trans1.2) requires an understanding of the difference between automating formed. The extended model, processes and transforming processes. An example is the automated expressed in a nonlinear manteller machine (ATM) in the banking industry. ner, adds a conceptual dimenBefore the ATM emerged, personal banking was conducted in sion. Health systems informatics large banks staffed by numerous tellers working side by side behind does not merely add a dimension counters. Withdrawing cash from a checking account entailed going in itself. However, the addition to the bank (typically located downtown) between the hours of 9:00 of complex adaptive systems a.m. and 4:00 p.m., filling out a withdrawal slip, and getting in line expands biomedical informatfor the next available teller. The teller received the withdrawal slip ics and enables health systems and dispensed the appropriate amount of cash, usually counting out informatics to substantially alter the money twice to avoid an error. all other relationships, reflecting Initially, the ATM was designed to be placed behind the its disruptive, transformative counter to support the teller. The teller would take the customer’s nature and the profound change credit card, swipe it through the ATM, receive the money, and count it brings to health system struc(continued) ture and function.

C h a p te r 1:   H e a l th Sy s te ms I nfor m atic s: A Transfor m ational Sc ienc e

11

Focus Health systems informatics expands the focus from the clinical decision to the structure of the delivery system. The clinical function cannot be substantially transformed within a system whose structure is based on a different logic. Assumptions about traditional structures of clinical processes, hospitals, clinics, financial services, and health policies are made obsolete through transformational change in healthcare organizations. Each of these functions must be transformed, based on the logic of optimal system performance enabled by the power of IT (Glushko 2013). Health systems informatics constitutes transformational change based on higher-order goals and differs from transactional change, which is focused on tactics, incremental processes, and performance metrics. A microprocess can be changed, even radically, but that change is not transformational because transformation is a macro systems concept. Health systems informatics integrates the science of IT, clinical science, engineering, and the social sciences to conceptualize the design of a healthcare organization that is seamless from the patient perspective and that ensures choice, continuity, quality, and efficiency. Envisioning such a

(continued from previous page)

it out to the customer. One reason banks closed at 4:00 p.m. before the advent of ATMs was so that tellers could count the cash and complete the accounting to close the books for the day. The ATM replaced that process; it not only counted the money but also provided a simultaneous accounting of all transactions. Thus, the ATM was conceived as transactional but not transformational—it automated the work process but did not transform it. A bright, visionary clerk at a bank asked why the ATM was not placed in front of the counter to allow the customers to swipe their card and directly receive their cash and receipt. This suggestion was met by opposition, with the bank president and vice presidents arguing that customers would not trust a machine to count their money correctly, although engineers emphasized that machines were actually more accurate than people in performing this task. In addition, the skeptics believed money transactions were a personal experience, so customers would not be comfortable interacting with a machine; plus, it would destroy the banker–customer relationship. The clerk was persistent, however, and the change-averse president reluctantly agreed to test the restructuring of the work process by placing an ATM in front of the counter. Customers flocked to the machine because of the convenience it offered. After the ATM was moved from behind the counter, it began appearing outside the bank—at commercial establishments and neighborhood branches—and was available 24/7. The transformation of the banking industry was on! Today, ATMs are everywhere, including in shopping malls and hospitals, and even on street corners. They also have become travel-friendly; international travelers, for example, can withdraw money from their bank account and receive the local currency at the current exchange rate in the country they are visiting. The ATM proved to be a technology that at first changed processes but ultimately transformed functions, jobs, institutions, and customers in the banking industry. Health leaders and practitioners, like the reluctant bankers, primarily envision electronic systems as automating functions and processes (transactional change) rather than as transforming the entire enterprise. Although extremely disruptive to the current way of doing things, the application of electronic systems or IT will be transformative in the long run. Introducing such innovation is the work of both organizational leaders and health professionals.

12

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

system may be difficult given that we are captive to traditional models, which have proven to fall significantly short of the aforementioned standards. IT can serve as the architecture and logic for developing systems that are tailored to individual conditions and values but still serve essential societal goals, such as maintaining local hospitals and clinics and providing flexible health insurance plans. Thus, many elements of the existing system will be maintained, but their function will be fundamentally changed and aligned with a higher-order goal based on outcomes. The integration of the logic of the clinical and business functions is difficult given that these functions are a clash of history, values, and culture. Health systems informatics ushers in change that will be highly disruptive to the healthcare organization and, as such, requires visionary, innovative, and highly skilled leaders. The appropriateness of such change is measured by whether it adds value to improved organizational performance in clinical care quality, operational or business efficiency, and patient satisfaction. The challenge is determining whether organizational leaders have the competencies and commitment to bring about this profound change.

Management Information Science Management information science (MIS) is the application of advanced IT to the functional areas of both product and service organizations, including accounting, finance, marketing, strategy, purchasing, supply, and operations. A well-developed information system that is deeply ingrained in organizational cultures, MIS is fundamentally different in logic from health systems informatics but is nonetheless an inherent component of it.

Business Versus Clinical Functions The classic or historical view of a healthcare organization is that the business function is different from the clinical function in logic and design. Thus, the responsibility for the clinical function (assumed by health professionals) is separate from the operation of the business function (overseen by managers and top executives) (Brown, Stone, and Patrick 2005, 31–50). MIS in healthcare organizations, as in corporations, was applied first to manage payroll, billing, and accounting; then to human resources; and then to the supply chain. These business functions, unlike medical or clinical functions, are well suited for electronic data processing because they possess universal vocabularies and standards and large databases. Generally accepted standards for measuring and reporting accounting and finance information, for example, were developed in the 1930s with the establishment of the Committee on Accounting Procedures. These accounting standards were not initiated with the computer in

C h a p te r 1:   H e a l th Sy s te ms I nfor m atic s: A Transfor m ational Sc ienc e

13

mind but with the aim of communicating with external stakeMissed Technological Opportunities holders, including the public, The town of Barclay, Iowa, was founded in 1857 on a busy rural crosswhich is the antithesis of the roads by a visionary named James Barclay. Several general stores, two approach of the health profeshotels, a blacksmith forge, a drugstore, a jewelry store, two physician sions. Given a standardized offices, a sawmill, a post office, and a cemetery soon were built one language, computers greatly after another. Barclay’s population was thriving and growing, and its facilitated the measuring, anafuture was bright. It was located two miles north of and between the lyzing, storing, and reporting larger towns of Independence and Waterloo, which were 35 miles functions for the purpose of apart. With a good road network, the town’s location seemed ideal. both internal operations and In the early 1860s, the Dubuque and Pacific Railroad initiated external communications. a survey to build a railroad across Iowa from Chicago. The company Because the initial applicontacted James Barclay to gain the right of way through the town cation of IT in health systems with tracks that would form a slight arc between Independence and was solely in accounting and Waterloo. Construction of a water tower and train depot would ensure finance, MIS became functhat both goods and people were serviced. He refused right-of-way tionally structured under the permission, believing that railroads conflicted with his vision of a finance department. This inirural town because they were noisy, dirty, and disruptive to a peaceful tial focus was the genesis of community. The company instead built a direct rail line with a water the conflict in IT when it was station between Independence and Waterloo, on a prairie two miles extended to support both MIS south of Barclay. A community sprang up around the water station and the clinical IT functions. and formed the town of Jesup. Thus, the business function has Within a few years, nothing but the cemetery was left at an inherent corporate orientaBarclay. Oops! tion, whereas the clinical function maintains an individual or professional perspective. This clash of structures and cultures is still a factor in the health system and directly affects the architecture, use, and effectiveness of IT. The two structures will remain in systems of the future, but each must change.

Conclusion Health systems informatics in the US health system reflects the traditions, values, and strengths of the local and diverse health system because, as the saying goes, “all healthcare is local.” These values cannot and should not be changed by the imposition of an externally mandated system architecture for integration and uniformity. Such a mandated approach would fail in the United States, as it did in other countries (Garrety et al. 2016). As a result, the transformation of healthcare to an integrated and uniform information system is and will continue to be slow and messy. Traditional structures, services, and markets will continue

14

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

to be pursued as innovation is initiated within new structures and strategies. This complexity has been identified by Hwang and Christensen (2007) as disruptive innovation. They identify the complexity of the process as integrating an “enabling technology” with an “innovative business model,” all within a “coherent value network.” This is a complex transformation process indeed. Winston Churchill once said, “Americans can always be counted on to do the right thing but only after they have exhausted all other possibilities.” Certainly, we have a few possibilities yet to consider. Efforts are under way to learn how data from disparate institutions and registries can support the exchange of information and how automated systems can enable knowledge generation, transfer, and application to improve evidence-based decision making. Such systems will be local but will have the capacity to interface regionally, nationally, and internationally. The challenge is keeping healthcare local and responsive to individual needs and values while enabling broad access and maintaining the system’s knowledge-based design. Keeping it local does not mean keeping the system as it is. The health system must be transformed in a purposeful manner. Change will be disruptive, but the health system cannot use its current design as a justification for avoiding the chaos and disorder of transformation. Healthcare organizations and professionals must fundamentally change because the technology exists to design systems that provide significantly better quality, continuity, and patient satisfaction at a much lower economic cost. These values provide a mandate for system transformation that is unlike the changes of the past. The transformation of the US health system enabled by advanced IT is the subject of this book.

Chapter Discussion Questions 1. Describe the US health system as a complex adaptive system. 2. How does health systems informatics differ conceptually from bioinformatics and biomedical informatics? What is the scientific base for each? 3. What is transformational change, and why is health systems informatics considered to be transformational? 4. Describe the key functions of the US health system, and analyze them in the context of open systems theory. 5. In what ways was the development of the EMR disruptive to, yet an adaptation of, traditional clinical practice in hospitals?

C h a p te r 1:   H e a l th Sy s te ms I nfor m atic s: A Transfor m ational Sc ienc e

Case Study  Electronic Health Records: Where Does the System End? Brian K. Hensel You work for a leading electronic health record (EHR) company and support one of its biggest clients—an academic medical center. You are part of a team that helped this medical center achieve the highest level (stage 7) of the Healthcare Information and Management Systems Society’s electronic medical record adoption model. The not-for-profit medical center owns a not-for-profit continuing care retirement community (CCRC), which operates independent living duplexes, assisted living apartments, and a nursing home with a separate wing offering Medicare-certified skilled nursing facility beds for post-acute rehabilitation and recovery. Your company wants to develop a long-term care EHR, and the medical center has agreed to do so for its nursing home, which would serve as the alpha site, but with an important stipulation: The EHR must support and facilitate the nursing home’s recently adopted Vision of Care approach. One year ago, the medical center created and filled a new position— vice president (VP) of post-acute, long-term, and palliative care services. This position reports directly to the CEO and is charged with leading a range of mostly nonhospital services and integrating these services across the system. Services reporting to the VP include the CCRC; palliative care services at the medical center’s flagship hospital; and a system-owned hospice, home health agency, and adult day service program. Early on, the new VP began meeting with the nursing home’s resident and family advisory council. Discussions at these meetings became the impetus for developing the nursing home’s Vision of Care approach. The nursing home’s reputation was considered better than that of local competitors, but residents and family members of the advisory council expected more. The VP was gaining valuable and actionable feedback and decided to bring in faculty consultants familiar with long-term care from the local state university. These consultants organized the feedback from the advisory council and provided terminology that synthesized the needs expressed. They then worked with the nursing home’s staff to form an implementation task force to determine what could be done to better meet the council’s needs. Using the consultants’ report, the VP and the nursing home’s leadership worked with the staff to develop the Vision of Care approach.

Vision of Care Vision of Care responds to needs expressed by the advisory council and includes concepts from research literature that compares long-term care (where residents (continued)

15

16

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

live) with other types of care, such as hospitals and the physician offices that patients visit. Person-centered care is emphasized instead of patient-centered care, and so is holistic, multidimensional care that includes but goes beyond the physical dimension of the medical model to assess and address psychosocial and spiritual needs. Protection of each resident’s dignity is also a core goal. Within this framework, the Vision of Care vividly describes care that is personalized, connected to the community, and noninstitutionalized.

Personalized Care Two dimensions of personalized care are identified. The first dimension is for staff to get to know each resident in ways other than the medical descriptor in the resident’s chart. Meaningful, personal interactions between residents and staff are based on shared knowledge. Charlie, a resident on the advisory council, voiced, with others nodding in agreement, “I don’t think nurses, who see me every day, really know anything about me or about my life before coming here.” Ann, whose father has latter-stage Alzheimer’s, added, “They don’t know Dad was a mechanic and could fix about any car you can name. Or that he was the best fast-pitch softball pitcher in the whole state!” The implementation task force recommended the creation of short digital stories of every resident, by digital media students from the university. Nursing home staff would be required to view these three- to five-minute stories. Cues such as a picture of Dad in his softball uniform would be placed in resident rooms to help jog the staff’s memory. The second dimension is for staff to get to know, in a more complete way and across dimensions of care, the personal needs and preferences of residents including food likes and dislikes and other important details, such as routine or favorite activities. Jaylen described how his mom loved to sit on the front porch in the afternoon when she lived at home: “She misses that. Unless I’m here to help her, it seems sitting outside is out of the question, except for maybe official, planned outings.”

Community-Connected Care Care that is connected to the community includes bringing both the community to the residents and residents into the community. To nursing home residents, “community” represents not only the immediate surrounding community but their larger outside world. And, consistent with personalized care, each resident’s community is different. Advisory council discussions revealed a longing by some local residents to attend community events. “I know I can’t get around like I did when we lived in the [independent living] duplex, or even the [assisted living] apartment, but I miss knowing about and going out to local events, like the yearly barbershop quartet concert—something

C h a p te r 1:   H e a l th Sy s te ms I nfor m atic s: A Transfor m ational Sc ienc e

we never missed.” Hailey, who moved her father from another state to be closer to her, shared, “Dad misses Friday night fish-frys at the VFW. Having a beer with other vets, including the younger ones. Though he appreciates the work and kindness of staff here, he doesn’t really much enjoy making crafts and such during activities hour.”

Noninstitutionalized Care Institutionalization depersonalizes people. It is marked by a loss of control and choice. One resident lamented, “The nursing home’s schedule rules my day. I have to eat when they say, whether I’m hungry or not. Then I take naps, go to bed, get up, take baths—all on someone else’s schedule.” Noninstitutionalized care provides greater control and choice. Such care is demonstrated by fluid, continuous response to real-time choices by residents.

Your and the EHR Team’s Charge The VP is experienced in long-term care and knows that the Vision of Care approach is ambitious and requires innovation and investment. The implementation task force is charged with operationalizing this vision of personalized, community-connected, noninstitutionalized care in a larger framework of person-centered, multidimensional care that promotes dignity. The task force’s focus is on staffing and other resources, processes, training, policies, and metrics. It wants the EHR to integrate the recommended resident digital stories; more broadly, however, it wants the EHR to support and actively facilitate the Vision of Care approach. The VP wants to know how your team will design the EHR and how the team will use linked communication technologies for staff, residents, and family to address the following questions.

Case Study Discussion Questions 1. How will you integrate, in the EHR, the nonmedical information of personalized care that promotes personal interaction and accounts for the personal needs and preferences of individual residents? Who on the staff would be included in learning the digital stories? 2. How will the EHR support personalized, community-connected care for individual residents? How does the concept of health information exchange relate to that of the hospital EHR? 3. How will the EHR support noninstitutionalized care, including fluid, continuous staff response to real-time choices by residents? 4. How will the EHR use linked communication technologies for staff, residents, and family to actively facilitate the resident choice and staff response in question 3? What linked communication technologies will be used?

17

18

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Additional Resources American Medical Informatics Association (AMIA): www.amia.org. Stead, W. W. 2005. “Challenges in Informatics.” In Building a Better Delivery System: A New Engineering/Health Care Partnership, edited by P. P. Reid, W. D. Compton, J. H. Grossman, and G. Fanjiang, 193–94. Washington, DC: National Academy of Engineering and Institute of Medicine.

References Akgün, A. E., K. Halit, and J. C. Byrne. 2014. “Complex Adaptive Systems Theory and Firm Product Innovativeness.” Journal of Engineering and Technology Management 31 (1): 21–42. Birdsey, L., C. Szabo, and K. Falkner. 2017. “Identifying Self-Organization and Adaptability in Complex Adaptive Systems.” Presentation at the 2017 IEEE 11th International Conference on Self-Adaptive and Self-Organizing Systems, Tucson, Arizona, September 18–22. Brown, G. D., T. T. Stone, and T. B. Patrick. 2005. Strategic Management of Information Systems in Healthcare. Chicago: Health Administration Press. Collen, M. F. 1995. A History of Medical Informatics in the United States, 1950 to 1990. Indianapolis, IN: American Medical Informatics Association. Encyclopedia.com. 2001. “Systems Theory.” Accessed July 20, 2017. www.encyclopedia. com/social-sciences/encyclopedias-almanacs-transcripts-and-maps/systemstheory. Friedman, C. P., A. B. Altman, I. S. Kohane, K. A. McCormick, P. L. Miller, J. G. Ozbolt, E. H. Shortliffe, G. D. Stormo, D. T. Szczepaniak, and G. Williamson. 2004. “Training the Next Generation of Informaticians: The Impact of ‘BISTI’ and Bioinformatics—a Report from the American College of Medical Informatics.” Journal of the American Medical Informatics Association 11 (3): 167–72. Garrety, K., I. McLoughlin, A. Dalley, R. Wilson, and P. Yue. 2016. “National Electronic Health Record Systems as ‘Wicked Projects’: The Australian Experience.” Information Polity 21 (4): 367–81. Glushko, R. J. 2013. The Discipline of Organizing. Cambridge, MA: MIT Press. Havighurst, C. C. 2008. “Disruptive Innovation: The Demand Side.” Health Affairs 27 (5): 1341–44. Herzlinger, R. E. 2006. “Why Innovation in Health Care Is So Hard.” Harvard Business Review 84 (5): 58–66. Hwang, J., and C. M. Christensen. 2007. “Disruptive Innovation in Health Care Delivery: A Framework for Business-Model Innovation.” Health Affairs 27 (5): 1329–35.

C h a p te r 1:   H e a l th Sy s te ms I nfor m atic s: A Transfor m ational Sc ienc e

Kast, F. E., and K. E. Rosenzweig. 1972. “The Modern View: A Systems Perspective.” In Systems Behaviour, edited by J. Beishon and G. Peters. New York: Oxford University Press. Lee, T. H., and J. J. Mongan. 2009. Chaos and Organization in Health Care. Cambridge, MA: MIT Press. Mattick, J. S., M. A. Dziadek, B. N. Terrill, W. Kaplan, A. D. Spigelman, F. G. Bowling, and M. E. Dinger. 2014. “The Impact of Genomics on the Future of Medicine and Health.” Medical Journal of Australia 201 (1): 17–20. Mintzberg, H. 2015. “Managing the Myths of Health Care.” World Hospital and Health Services 48 (3): 4–7. Moore, J. H. 2007. “Bioinformatics.” Journal of Cellular Physiology 213 (2): 365–69. Sarkar, I. N., A. J. Butte, Y. A. Lussier, P. Tarzy-Hornoch, and L. Ohno-Machado. 2011. “Translational Bioinformatics: Linking Knowledge Across Biological and Clinical Realms.” Journal of the American Medical Informatics Association 18 (4): 345–57. Stange, K. C., R. L. Ferrer, and W. L. Miller. 2009. “Making Sense of Health Care Transformation as Adaptive-Renewal Cycles.” Annals of Family Medicine 7 (6): 484–87. Starr, P. 1982. The Social Transformation of American Medicine. New York: Basic Books. World Health Organization. 2018. “History of the Development of the ICD.” Accessed January 14. www.who.int/classifications/icd/en/HistoryOfICD.pdf.

19

CHAPTER

KNOWLEDGE-BASED DECISION MAKING

2

Gordon D. Brown, Kalyan S. Pasupathy, and Mihail Popescu

Learning Objectives After reading this chapter, you should be able to do the following: • Apply the principles of knowledge management to clinical decision processes. • Define the essential features of a knowledge-based decision support system. • Identify the types and levels of knowledge related to knowledge-based decision support. • Understand the structure and function of electronic medical records, electronic health records, and personal health records. • Describe the assumptions that different clinical decision processes make about the structure and financing of health systems.

Key Concepts • • • • • •

Knowledge management Explicit, implicit, latent, and tacit knowledge Evidence-based clinical decisions Knowledge socialization Knowledge brokering Community of practice

Introduction Harnessing all the relevant knowledge assets of an organization and system and then deploying them to achieve optimal system performance has been referred to as knowledge management. The core function in knowledge management is information technology (IT), which captures, compiles, transmits,

Knowledge management Harnessing all the relevant knowledge assets of an organization and system and then deploying them to achieve optimal system performance

21

22

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

and applies available knowledge to clinical, operational, and strategic decision making. The concept of knowledge management is complex in any system, but is particularly complex in highly professionalized systems such as health. However, knowledge management has been widely used in business and commercial industries, and findings can be used to guide its development in the healthcare sector (Kothari et al. 2011). Several forces contribute to the dynamics and complexity of knowledge accumulation and management in health systems. Basic and clinical sciences that inform clinical decisions have grown exponentially and are increasingly being integrated. Similarly, advances in information, systems, and social sciences are expanding our understanding of the interdependencies among clinicians, institutions, and social sectors and the ways to manage them (Best et al. 2016). However, the complexity and dynamics of the healthcare environment make capturing and presenting all available knowledge difficult. Knowledge that informs clinical and organizational decisions consists of scientific and experiential knowledge. IT enables knowledge capture and management. The typical strategy of a well-led healthcare organization is to enlist all available assets to improve outcomes and the organization’s competitive positioning. One asset is the institutional knowledge detailed in medical records, documents, rules, policies, strategies, and the like, along with the unarticulated wisdom and experiences of professionals and other domain experts in the health system. These knowledge assets—particularly in highly professionalized service sectors such as healthcare—complement and potentially outweigh other assets, such as land, labor, and capital (Herremans et al. 2011). Information and knowledge, unlike all other assets, are not depleted as they are consumed but instead increase in value and can be leveraged. When considering the value of IT, leaders must look at its fluidity, which enables the knowledge assets of multiple professionals and organizations to be strategically aligned. Multiorganizational and multiprofessional collaboration is possible with loosely configured IT systems that are flexible enough to be tailored to different applications. Information-enabled systems have considerable potential to improve healthcare services and bring about superior outcomes because they are science based, highly integrated, and tailored to individual patient conditions. This chapter discusses knowledge management within professions, organizations, and systems.

Definition and Use of Knowledge in Decision Making Knowledge management includes both clinical and organizational functions related to clinical knowledge management, knowledge socialization, and knowledge brokering. This systems view is well established in science and has been articulated by Lee (2000) as follows:

C h a p ter 2:   Knowledge- Based D ec ision Making

23

A discipline that promotes an integrated approach to identifying, managing, and sharing all of an enterprise’s information needs. These information assets may include data bases, documents, policies, and procedures as well as previously unarticulated expertise and experience resident in individual workers.

Harnessing all the knowledge assets of an organization requires the deployment of advanced IT and constitutes a core function of health systems informatics. Health systems informatics draws on the clinical, engineering, and social sciences, each of which brings knowledge whose application is enabled and advanced by IT.

Types of Knowledge The science of clinical decision making relies on explicit knowledge, which is derived from research and science and presented as decision support measurements, relationships, and evidence. Knowledge can be generated from data files or the minds of individuals and expressed when needed. The art of clinical decision making also relies on implicit, latent, and tacit knowledge stored in the creative capacity of individuals, which cannot be captured in text or data files. Explicit knowledge (exhibit 2.1) is the focal area of clinical decision support systems. Having a clinical decision support system (CDSS) increases the capacity to search, retrieve, and structure information that draws from a vast reservoir of scientific evidence. Much scientific evidence can be structured and presented within the clinical decision process, although only a fraction of the explicit knowledge available is currently accessed, analyzed, and applied. In addition, considerable knowledge exists in the minds of highly trained and experienced professionals, some of which is accessed through interactions between professionals and by nature cannot be incorporated into the CDSS. Information systems are primarily decision support and not decision making, although

Communication, representation, analysis Reasoning, dialogue, learning

EXHIBIT 2.1 Application of Information and Knowledge

Documented Explicit Expressed Implicit Latent Tacit

Knowledge management

Documented—Stored in files, documents, records, etc. Expressed—Conversations, symbols, and other representation Implicit—Not expressed but communicated and understood Latent*—Underdeveloped and not communicated Tacit*—Hidden from consciousness Explicit

*Turns into explicit knowledge when discovered

Clinical decision support system (CDSS) Software that presents users with a knowledge base, patientspecific data, and related information at the point of care to enhance healthcare provision and management

24

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

that distinction is changing somewhat with increased machine learning. Basic, clinical, and engineering sciences bring new knowledge from laboratories to the bedside at such a rapid pace that CDSSs are now essential to clinicians for accessing the best evidence and incorporating it into the decision-making process. Implicit knowledge in clinical teams represents information that is communicated and understood but never expressed. One need only observe a surgical procedure to understand the nonverbal communication that takes place in highly skilled clinical teams. Clinical guidelines and pathways have reduced clinical errors by drawing on available evidence and increasing communication within teams (Haynes et al. 2009). The integration of explicit and implicit knowledge within teams will have implications on system efficiency, patient satisfaction, staffing, and surgical and other procedural outcomes. Latent knowledge and tacit knowledge can be discovered only through human interaction, such as when framing issues, developing mental models, generating solution strategies, and engaging in innovative thinking. This knowledge draws on the science of each profession and the collective latent and tacit knowledge of teams. This process is complex and acknowledges the human, social, and scientific aspects of knowledge where communication becomes highly interactive. The nature of the problem being analyzed determines who is involved in the decision-making process. The challenge for healthcare leaders is not only developing IT to support a knowledge-based system but also transforming the current structure of their healthcare organization. Health systems informatics is the science of developing interactive clinical decision support systems that access and present, at the appropriate point of decision, the best scientific and experiential knowledge.

Knowledge Organizations

Knowledge organization Organization in which leaders engage in innovative and creative pursuits involving health professionals and other knowledge workers

Knowledge organizations are information-oriented organizations that focus on creating change through innovation. This concept was introduced by Senge (1990) and then broadly applied and institutionalized in the 1990s to describe organizations involved in generative learning and not adaptive learning. Through learning, “we expand our capacity to create, to be part of the generative process of life” (Senge 1990, 14). The analysis of organizations suggests that many self-identify as knowledge organizations because of the external pressures they face (e.g., accreditation) or a desire to be trendy, without understanding its basic principles. When knowledge becomes a strategy, it is a powerful asset in an environment that demands clinical quality, efficiency, and accountability. A knowledge organization is characterized by a culture and management structure that is different from the culture and structure of traditional,

C h a p ter 2:   Knowledge- Based D ec ision Making

25

hierarchical institutions. In a knowledge organization, leaders engage in innovative and creative pursuits involving health professionals and other knowledge workers. A control-oriented organization that pursues a knowledge strategy must first understand the power of a commitment-oriented culture and be able to achieve it. Knowledge organizations recognize the power of IT as they seek to transform their mission, structure, processes, and culture. Becoming a knowledge organization is not an end point; rather, it is an aspiration on a journey that is never completed, as new technologies and strategies emerge. Healthcare organizations that adapt health systems informatics to fit traditional clinical and organizational structures without recognizing and committing to its transformative power have not yet started this journey. Considerable evidence exists to support the power of knowledge organizations (Wang, Arnett, and Hou 2016). Computer applications in the healthcare field have been most effectively applied in data processing and deriving useful (and nonuseful) information. The traditional paradigm assumes that organizations have static structures with rapid-processing information systems made possible by sophisticated IT. Health systems informatics transforms the structure of decision processes and the design of the organization itself. This focus emphasizes the importance of conceptualizing the nature of the problem being addressed.

Clinical Knowledge Management Knowledge management includes the way in which clinical knowledge is generated, compiled, systematically reviewed, integrated, and transmitted to clinical decision makers. The term management is used to denote the system capacity to electronically capture, distribute, and apply evidence by the clinician and clinical teams.

Electronic Medical Record The electronic medical record (EMR) automated gathering, integrating, and analyzing patient-specific information (“1” in exhibit 2.2). It provides information—such as medical condition, treatment, and history—that has been compiled and reported in a usable format. As with paper records, the early EMR was merely a repository of test and scan results from the laboratory and radiology and chart notes from the attending clinician. Converting text, image, and even sound files into digital format was a complex process in an EMR system without common clinical vocabularies or standards and with medical records generated and maintained by different departments. In many instances, the record was based on insurance data, collected for billing rather than clinical purposes. Computerization

Electronic medical record (EMR) Person’s electronic clinical information created, integrated, managed, and accessed by authorized clinicians, nurses, coders, and other health professionals

26

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

EXHIBIT 2.2 Clinical Decision Support in Knowledge Systems

System structure Scientific

Training Clinicians

Hospital

C

B

1. Gather, integrate, analyze explicit patient information (EMR)

2. Access clinical evidence for decision support (EHR)

A Experiential

Knowledge

Knowledge accumulation

Outcome–decision relationship

Note: EHR = electronic health record; EMR = electronic medical record.

required standardized language and standardized protocols—both anathema to autonomous clinical decision making. Once digitized, clinical data could be accessed, stored, integrated, analyzed, and reported to anyone with access in the institution. Clinical vocabularies, data standards, and the integration of clinical support in such areas as laboratory, radiology, and pharmacy changed how information was structured—from a focus on individual clinicians to a focus on information exchange within the institution—and thus how information was compiled, represented, and integrated in the clinical record. The use of the EMR in clinical practice has become pervasive, and the amount of available clinical data has steadily increased, making the EMR an even more valuable knowledge resource. The EMR expanded the knowledge base of individual clinicians, who had previously extracted data from individual EMRs and then applied their own personal knowledge (“B” in exhibit 2.2) derived from training and experience. Digitization increasingly enabled data to be converted into information and knowledge, giving caregivers (“A” in exhibit 2.2) additional “knowledge accumulation” and thus an improved basis for making clinical decisions. Digitization can also assist in coordinating care, increasing safety through technology such as barcoding, and developing advanced forms of knowledge-based IT systems. Clinical data are collected not only during the patient encounter but also throughout the process of care. Types of clinical data include EMR,

C h a p ter 2:   Knowledge- Based D ec ision Making

27

administrative, claims, health survey, and patient registry. These data are highly heterogeneous, distributed, and not collected according to a given protocol. EMR data typically comprise demographic information, diagnoses, treatments, prescription drugs, laboratory tests, physiologic monitoring data, imaging studies, and genomics. These data are available in many organizations and are primarily used for administrative reports and quality improvement. Because these data are local, extracting knowledge from them is difficult (the number of records is small), and extracted evidence is typically not generalizable. As a knowledge source, the EMR is minimally invasive to the traditional structure of the clinical decision process, increasing the probability of its acceptance by clinicians. The EMR requires its users to change behavior and the way they use clinical vocabularies, access and apply information, weigh clinical decisions, and achieve outcomes, but the system itself does not change the basic structure of the clinical process. Integrating information in the organization is a disruptive process, but in many cases selecting an EMR system was even more disruptive. Different hospitals select—sometimes intentionally—different vendors, with incompatible systems; in the past, some organizations developed their own EMR. Automating the clinical decision process had an institutional focus. Selecting and investing in an EMR system requires an understanding of its value-added potential. Leaders must carefully consider the effects of IT changes on clinical processes, including the workflow, and on staffing. Acceptance of these changes by staff is an essential measure, but improvement in clinical outcomes is the more relevant metric.

Electronic Health Record The electronic health record (EHR) represents the transformation of the EMR as more clinical information is stored, analyzed, and extracted for decision support (this transition is reflected in exhibit 2.2). In addition to internal decision support information, the EHR imports scientific decision support information. Scientific evidence from statistical analysis of large population studies can be brought to the point of decision making to inform and complement data in the EMR and experiential knowledge. The EHR (“2” in exhibit 2.2) is a major step in developing a knowledge-based system. Scientific inquiry through biological research and clinical trials (“C” in exhibit 2.2) has served as the basis of clinical science since the Flexner Report on medical education was published in 1910 (Cooke et al. 2006). The EHR can access and pre­ sent guidelines, reminders, alerts, protocols, and other decision support tools backed by evidence from clinical trials. Users of the EHR should recognize that the evidence it presents, being population based, might not fit every patient and condition; accordingly, the final medical decision must include clinician judgment. The challenge in introducing CDSSs, such as the EHR, is not only technical or information related but also institutional. Institutional issues

Electronic health record (EHR) Documentation of the clinical workflow; provides alerts, reminders, therapy plans, and medication orders

28

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

include the fear that the system may slow down the decision process or be too prescriptive, and such concerns will affect acceptance by health professionals. Clinical decision making is a complex and dynamic process, requiring evidence and a degree of decision-maker autonomy and professional accountability (see chapter 3 for an in-depth discussion). Two ways in which clinical evidence has been generated in the past 50 years are randomized clinical trials and data-driven trials (e-trials). Physicians, nurses, and other clinicians are trained to respect and use scientific evidence to support their clinical decisions. Using computer-generated clinical guidelines and protocols alters the behavior and role of the clinicians or decision makers, making them more dependent on external guidelines and rules. However, when the application of clinical guidelines leads to a demonstrated improvement in clinical performance, health professionals and organizations have a shared moral obligation—because they are accountable to patients for delivering high-quality and safe care—to use, promote, and contribute to those guidelines and rules. The CDSS has demonstrated its contribution to increasing care quality and safety, and in doing so it has helped transform clinical decision making from an independent activity to a collaborative one supported by IT. A successful CDSS implementation might require changing clinicians’ values, behaviors, and attitudes. Part of this change process involves enlisting clinicians in the design of the CDSS. Training programs designed simply to teach clinicians how to use IT are ill conceived and of limited value. Randomized Clinical Trials The preferred method—often considered the gold standard—to generate clinical evidence is the randomized clinical trial (RCT). RCTs are research studies in which human volunteers undergo an intervention based on a well-defined protocol approved by an institutional review board. The intervention is evaluated on the basis of collected data and health outcomes. Typically, RCT results are disseminated through publication in biomedical journals. To centralize RCT information, the US government created an online portal: ClinicalTrials.gov (Zarin et al. 2016). As of 2016, about 224,000 studies have been cataloged in this portal, which, at a publication rate of about 66 percent (Chen et al. 2016), have generated more than 160,000 journal articles in the past 30 years. The National Library of Medicine (NLM) has led the effort of making the literature available to physicians and researchers. On its PubMed portal, NLM provides more than 27 million citations in multiple languages. Its publication search capabilities are enabled by medical terminologies such as Medical Subject Headings (MESH) and Universal Medical Language System (UMLS) and enhanced by complex algorithms. Accessing and integrating knowledge from clinical trials is an exceedingly complex process given the lack of a unified clinical language (chapter 12). However, the volume of biomedical literature makes it impossible for clinicians to stay informed about best practices. Two

C h a p ter 2:   Knowledge- Based D ec ision Making

important tools that help clinicians find clinical information for improving their practice are the National Guideline Clearinghouse (NGC) and systematic review. The NGC is an initiative of the Agency for Healthcare Research and Quality in partnership with the American Medical Association and the American Association of Health Plans. The NGC’s mission is to provide physicians and other health professionals with detailed information on clinical practice guidelines. However, even if guidelines are readily available, they can be hard to use, partly because of the limited time clinicians have during patient visits. For example, studies have found that participants received only about 55 percent of the care recommended by guidelines (McGlynn et al. 2003). As the use of the EMR spreads, more efforts are being directed at integrating clinical guidelines into this system (De Backere et al. 2012) and building CDSSs to assist clinicians in complying with best clinical practices (Ennis et al. 2015; Rodriguez-Loya, Aziz, and Chatwin 2014). Systematic review is another tool clinicians may use to make sense of often-contradictory clinical evidence (Tranfield, Denyer, and Smart 2003). These reviews attempt to synthetize the published evidence in a systematic, transparent, and reproducible manner to inform the building of clinical guidelines. Moreover, as many organizations and countries may have similar guidelines, systematic reviews may be used to adapt and merge separate guidelines and remove redundancy (Fervers et al. 2006). Currently, systematic reviews are performed by hand. However, thanks to advances in natural language processing, automated summarization of medical documents has become a possible solution for clinical evidence acquisition (Afantenos, Karkaletsis, and Stamatopoulos 2005; Mishra et al. 2014). Importing population-based clinical guidelines and clinical protocols invades and changes the clinical decision process by inhibiting individual autonomy and engendering greater external accountability. The application of clinical guidelines remains institution centered and does not fundamentally restructure the process. Clinical guidelines, protocols, alerts, reminders, and other decision support tools reinforce the tenet that the clinician is the primary locus of clinical decision making. Using decision support tools transports clinicians and institutions into the information age by altering the clinical decision process in a manner that improves clinical outcomes, although it does not fundamentally restructure the process. Artificial Intelligence A massive amount of clinical data gathered in ever-increasing clinical repositories can be used to support traditional evidence, such as clinical trials, but it can also be used for more disruptive technologies, such as artificial intelligence. One of the greatest by-products of machine-learning technologies in healthcare is the image-based specialty, such as radiology, pathology, and ophthalmology. Machine learning has been used for various applications such as fracture detection (Olczak

29

Clinical guidelines Evidencebased clinical information that guides clinicians during clinical encounters; also used to represent alerts and reminders embedded in the EMR

Clinical protocols Evidence-based clinical information that informs clinicians in a clinical process, as opposed to a clinical encounter; similar to critical pathways, a concept developed in industrial engineering

30

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

et al. 2017), knee magnetic resonance imaging interpretation (Hassanpour et al. 2017), and prostate cancer detection (Azizi et al. 2016). Machine learning has already penetrated many commercial radiology products, such as Philips’s IntelliSpace Portal 9.0, that help radiologists detect and diagnose diseases. As one recent review of the use of machine-learning technologies in radiology concluded, “The complexity of creating, training, and monitoring machine learning indicates that the success of the algorithms will require radiologist involvement for years to come, leading to engagement rather than replacement” (Kohli et al. 2017, 754). Supervised machine-learning algorithms need experts for labeleddata training, which requires physician involvement. However, the question is how the clinical workflow is going to change as machine learning becomes more pervasive. Some physicians caution against the use of machine learning because studies of its effectiveness in medicine are lacking. Other concerns related to the use of machine learning are that it may reduce the skills of physicians, focus too much on data and miss the context, or not be able to handle the uncertainty that is intrinsic to medicine, as well as the black box nature of most algorithms (Cabitza, Rasoini, and Gensini 2017). The black box nature of machine-learning algorithms has been a concern since the early days of artificial intelligence in medicine—specifically MYCIN, a rule-based system designed by Shortliffe (1976) to diagnose infectious diseases. In medicine, rule-based systems have always been preferred by clinicians because of their ability to explain decisions and analyze context. However, a new direction in artificial intelligence—explainable artificial intelligence—is gaining the attention of machine-learning researchers (Biran and Cotton 2018). Artificial intelligence is explored in chapter 5.

Health Information Exchange Health information exchange (HIE) Framework that enables the movement of patient health data and information across organizations that are geographically dispersed by using nationally recognized standards

Health information exchange (HIE) is the sharing of EHR data among institutions and clinicians involved in a patient’s care or as a basis for e-trials (“3” in exhibit 2.3). Two ways in which clinical evidence has been generated during the past 50 years are randomized clinical trials and data-driven trials (e-trials). An e-trial of a medical record is conducted like a paper-based clinical conference except that the documents are submitted and viewed electronically. In principle, trials can be carried out in a single institution, although the number of cases is not sufficient to provide statistical confidence in the findings. They can also be carried out among a group of institutions, through HIE, if the EMR data are compatible. HIE allows clinical information located in disparate EMRs to be accessed and transferred electronically (“D” in exhibit 2.3). Direct electronic access to and exchange of information from clinical records, replacing transmissions by hand, fax, or e-mail, is the power of IT. This capability enables greater collaboration and integration of clinical decisions at less cost and inconvenience to patients. HIE itself does not change the basic structure of clinical decision making, being clinician and institution centered and allowing access to treat patients outside the primary system. As the health system becomes more patient

C h a p ter 2:   Knowledge- Based D ec ision Making

System structure ACO

Training

Internet, social media

C

B Community of practice

Scientific evidence D

1. Gather, integrate, analyze explicit patient information (EMR)

3. E-trials (HIE)

2. Access clinical evidence for decision support (EHR)

A

5. Integrated PHR

F

E

Experiential Knowledge

Knowledge accumulation

Outcome–decision relationship

4. Networked HIE, health data vault

Note: ACO = accountable care organization; EHR = electronic health record; EMR = electronic medical record; HIE = health information exchange; PHR = personal health record.

centered, HIE becomes the enabling technology and redefines the function and technology that drive it. This topic is explored later in the chapter. Clinical consultations and referrals are long-established clinical practices, but transferring clinical information electronically is not a trivial chore. Hospital and clinic EHRs are typically not compatible—frequently by design—because they are based on a different logic and language structure. Closed systems evolved because they enable hospitals, clinics, and physicians to remain independent and to control the transfer of patients outside their institutions. Control is based primarily on financial logic rather than the quality or efficiency of patient care. Networked EHRs One value of electronic information is the ease with which it can move across time and space, a fact belatedly accepted by many institutions. Data vaults and cloud computing (“4” in exhibit 2.3) will become the technological vehicles that enable the integration and transmission of information that has traditionally resided in EHRs (Dixon et al. 2013). This logic has served to guide a strategy for developing a comprehensive national HIE and underlies much of the recent policy debate in the United States (Vest 2009). Electronic information systems should be conceptualized as national and global in nature, appropriately addressing issues of security and data standards. As these issues are resolved, the

EXHIBIT 2.3 PatientOriented Knowledge System

31

32

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

EMR will be an integrated record with controlled universal access, and future generations will look back and wonder why it was ever institution centered. Test results, diagnoses, and treatment regimens shared through the HIE have differing levels of quality because exchange members do not share the same level of quality. Thus, the quality of diagnoses and test results from a given institution might not be acceptable to another institution. This variation is an argument against the use of clinical information across institutions but also supports a shared commitment to increase standards and levels of quality within regional clinical networks. A shared interest in and commitment to quality by regional institutions could have a greater effect on quality than accreditation, licensure, certification, and public reporting. Systems thinking aligns institutional goals to optimize the system. Again, the challenge is not clinical or informational but primarily financial and organizational. Each of these is a complex system problem that needs to be resolved, and the knowledge to do so exists. The reality is that the process is messy and complex and will take time to evolve in a decentralized, private–public health system. One model of collaboration among regional centers is to integrate data across centers to generate evidence-based guidelines from Big Data. In 2004, the National Institutes of Health funded the development of a computational platform—Informatics for Integrating Biology and the Bedside (i2b2)—that consists of a data repository and critical components such as security, ontology, and governance layers. By sharing a common i2b2 ontology, different institutions can share data and integrate their i2b2 repositories in a common network called Shared Health Research Information Network (SHRINE; “4” in exhibit 2.3). Although the data in SHRINE are deidentified, individual institutions have a way of identifying their own patients for specific research purposes. An example of SHRINE is the Greater Plains Collaborative (GPC; www.gpcnetwork.org), an i2b2 network of 12 leading medical centers (University of Kansas Medical Center, Children’s Medical Center University of Wisconsin, University of Iowa, Marshfield Clinic Research Foundation, Medical College of Wisconsin, University of Minnesota Academic Health Center, University of Nebraska Medical Center, University of Texas Health Science Center at San Antonio, University of Texas Southwestern, Medical College of Wisconsin, University of Missouri–Columbia, and Indiana University Medical Center) committed to improving healthcare delivery through the adoption of best clinical practices and dissemination of ongoing research findings. The GPC SHRINE contains data related to 16 million patients across eight states. Because of the amount and characteristics of its patient data, the GPC SHRINE was able to participate in the first of a new type of distributed clinical trial—ADAPTABLE—pioneered by Patient-Centered Outcomes Research Institute, where the final consent form and protocol were shaped with input from patients, local institutional review boards, physicians, and study coordinators; this new type of clinical

C h a p ter 2:   Knowledge- Based D ec ision Making

33

trial was referred to earlier as e-trials, and knowledge generated is represented as “E” in exhibit 2.3. Technology that supports the transfer of medical information is a major challenge, but a greater challenge is integrating independent professions and institutions. HIE is recognized as the logic for developing integrated regional information systems and thus integrated regional health systems. This integration will require changing financial and organizational structures as well as HIE. Thus, the constraint is both the manner in which IT systems have been constructed and a broader systems failure. The problem is that the system functions just the way it was built, but the process of transforming it is underway.

Patient-Centered Care The development of patient-centered care introduces a transformational change in how clinicians practice, organizations are structured and function, and services are financed, as well as in the role patients and families play (note the change in “System structure” in exhibit 2.3). The development of the EMR, EHR, and HIE has been evolutionary, but each is transformed in a patient-oriented system. These information systems are enabled by advances in technology as the agent of change, but they are driven by increased accountability for quality, the coordination of services across systems, efficiency, and patient involvement. Patient-oriented care includes an EHR that draws on the best available clinical evidence, an HIE that allows the transfer of clinical information and shared consultations across systems, and the development of a complementary personal health record.

Personal Health Record The personal health record (PHR) (“5” in exhibit 2.3) allows patients to access and manage their own health information by drawing from a number of sources that conform to national data standards (chapter 8). PHRs include access to content communities and other shared sources of experience and information. Patient-centered care spawned new technologies, such as remote patient monitoring and the Internet of Things, that increase the amount of clinical data and alter clinical practice from the subjective human input to the objective sensor data, changing the way care is delivered. Remote patient monitoring brings data to the EHR that differ in nature and quantity from the traditional record, depending on the type of sensors used for monitoring. Physiological measurements such as blood pressure, heart rate, and respiration rate are known to clinicians, whereas others such as amount of restlessness in bed and amount of activity are not. In essence, clinicians will face a wave of new vital signs that they will be forced to comprehend

Personal health record (PHR) Person’s digital medical record that conforms to nationally recognized interoperability standards and is managed, shared, and controlled by the individual

34

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

(Rantz et al. 2015). For example, in an eldercare monitoring project conducted at the University of Missouri by researchers and Foresite Healthcare (www. foresitehealthcare.com), various sensors, such as bed, motion, and gait, have been deployed in the elderly’s aging-in-place apartments (Rantz et al. 2015). The installed sensors collect data, such as heart and respiration rates, movement in the apartment, and walking speed, to generate alerts that are sent to the facility’s clinical personnel. The personnel are readily able to interpret the alerts, but clinicians in general might require training to acquire this new skill. Such transformations are consistent with the rate of increase in chronic care cases associated with the aging of the population and high-tech precision medicine. The current US health system is designed for acute, episodic care, but patients, families, and public agencies are demanding greater accountability from healthcare organizations, including better integration and coordination of care data and information. This shift is associated with changes in societal values, patterns of communication, and expectations about access to information.

Patient-Centered Transformational Change To a degree, advances in informatics have led to an evolutionary change in the US health system. We have examined this transition as the EMR added science-based guidelines to become the EHR, and the EHR developed an HIE network to enable sharing of patient information across systems and to generate knowledge through e-trials. The PHR added knowledge through remote sensors, health and wellness literature, and content communities. Hospitals and clinics have responded by adapting to change and are poised to lead the transformation; exhibit 2.3 depicts this transformation. Patient-centered care is the correct conceptual framework for planning the health system transformation. All institutions have adopted elements of patient-centered care, but few have really conceptualized a system that is truly patient oriented. The patient becomes the central focus of the process, which has been referred to as the “deification” of the customer (Fowler 2003). The conversion to such an information system poses complex technical challenges to create semantic interoperability, usability, and security, but these problems are solvable. The traditional structures of clinical processes, organizations, and financing systems face the greater challenge (Best et al. 2016). Current strategies tend to adapt to changes enabled by IT, but only to the extent that the organization retains power. The United States can learn from other countries that have tested and implemented different information systems designs, by adapting what has been learned in the political and market context (Dixon et al. 2013). A greater complexity in forming such a system is the transformation from a traditional profession- and institution-oriented system to an integrated system with multiprofessional clinical teams and accountable care organizations

C h a p ter 2:   Knowledge- Based D ec ision Making

(ACOs; note the change in “System structure” from exhibit 2.2 to exhibit 2.3). The challenge is in viewing IT as an organizing principle. Interactions between patients and clinicians and among clinical teams inform and transform the decision process. The clinical process does not conform to patients’ desires but enlists them in the care process as co-producers. Decisions are based on evidence that the process achieves improved outcomes, efficiency, and satisfaction. Increased patient compliance is achieved when the structure of the process is patient centered but does not assume patients dictate the process, such as mandating the deletion of clinically relevant information in the EHR (Garrety et al. 2016). Decision making is designed to bring the most comprehensive, integrated scientific information relevant to the case from a range of health professionals combined with the professionals’ experiential knowledge and the patient’s preferences based on affective knowledge. The structure of the clinical process involves multiprofessional teams (knowledge workers), whose composition depends on their specialty and relevance to the patient’s condition, according to the evidence. The team engages in dialogue to sift through the available collection of explicit knowledge to arrive at an evidence-based solution. This dialogue gives the team a chance to draw from each other’s experiential and tacit knowledge. The dialogue should be sufficiently rigorous to enable knowledge to be elicited from all participants. Depending on the case being considered, any given professional might take the lead as the case progresses. The common practice of having a separate and predetermined case manager only adds cost and mandates what is essentially a team function. Patient-centered clinical processes require a strong, team-oriented, dynamic environment with a culture of strong commitment to quality. The current human resources function in healthcare organizations has not created an environment that is supportive of knowledge workers and itself must be transformed (chapter 14). Multiprofessional teams involved in a case need not be confined to a single location or organization. IT enables teams to form communities of practice, which come together to develop the best treatment plan and coordinate the delivery of services. Such an IT system does not require large monolithic institutions but allows loosely structured and dynamic systems tailored to individual patients. These loosely structured systems form communities made up of professionals with high technical skills, possibly in solo practices but bound by a culture of cooperation, mutual respect, integrity, and reward (Faraj, Jarvenpaa, and Majchrzak 2011). If common values of community of practice are violated, the community can easily re-form with members who share a philosophy and values (knowledge socialization). Exhibit 2.3 shows a transformed structure of a clinical decision process—from one with a clinician and institution focus to one with a community orientation. Health system leaders need to develop innovative, disruptive solutions for transforming the health system.

35

36

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

The transformation to a systems perspective necessitates a change in the traditional role of health professionals as independent or autonomous decision makers; such a change may enhance their role. It requires new values, behaviors, and skills necessary to achieve superior outcomes. As a result, the institution will become more oriented to health and wellness, identifying high-risk patients and tailoring clinical, evidence-based interventions. It will have a financing structure defined by its support of the delivery process (a revolutionary concept), and reimbursement will be based on valuation and contribution to the care process and not the cost of services (chapter 15).

Transformational Strategy The delivery of healthcare services that are truly knowledge based and patient oriented is not possible within the structure of the existing system. This transformation is represented under “System structure” in exhibit 2.3, where the traditional structures of hospitals and clinicians (exhibit 2.2) are transformed into ACOs and communities of practice. Traditional organizations are designed to protect and perpetuate legacy structures and functions, sometimes mislabeled as ACOs, and changes achieve only marginal improvement. Innovative strategies must be developed within institutions but outside of existing financial schemes and functional or clinical structures. Innovative frontline clinicians, managers, and staff must be invited to the “commons” and engage in the process of developing new models and transforming old ones. Some organizations (e.g., Garfield Innovation Center at Kaiser Permanente or the Mayo Clinic Center for Innovation) have created centers of innovation to address how system transformation can be introduced into existing operations. IT is essential as the architecture for such a system, but transformation comes through the design of organizations, financing, and clinical networks. A transformational strategy is made more complex because the process transcends institutional and professional boundaries. An integrated system cannot innovate if planned and enabled independently by institutions. This process is more difficult than organizational strategic planning because participants must shed their professional and institutional frames of reference. Why would institutions subordinate their own interests for the collective good? This question is particularly germane to US institutions that are primarily private and operating within a heavily market-oriented health system. It is also a fundamental question for providers as well as for insurance, pharmaceutical, and other suppliers of essential resources and services. Such discussions will call on a new business model, one that aligns a higher order of values and ethics with financial reward. These are strong qualities of health professionals and systems and will be tested. The potential of transformation lies in institutions creating

C h a p ter 2:   Knowledge- Based D ec ision Making

value within a system and being rewarded for value through new functions or products. A regulatory strategy is not assumed but might be necessary to cause disruptive change.

Knowledge Socialization The concept of knowledge socialization recognizes human interaction as a means of creating and reinforcing new work relationships within and across organizations and developing innovative solutions and improved outcomes in the process. Nonaka and Takeuchi (1995, 65) define knowledge socialization as “a process of sharing experiences and thereby creating tacit knowledge such as shared mental models and technical skills.” These problem-solving models transcend the mechanistic models (chapter 3) assumed in applying scientific knowledge to clinical decision making. They recognize the essential nature of developing respect and a common culture among professionals and organizations where they share the responsibility for clinical and financial decisions. Historically, socialization was within the purview of clinical specialties and departments. More recently, the focus has been on interprofessional team communication in organizations and is now interorganizational, which is more complex (MacArthur, Dailey, and Villagran 2016). Integration of services across professionals and organizations is not just about IT but about common goals, rewards, and mutual respect. Socialization skills must be taught in training programs and then transferred into practice in the real world (Ketcherside et al. 2017). Knowledge socialization applies to the range of functions within organizations and systems, including clinical, strategic, operational, and financial. In clinical practice, it serves to conceptualize and reinforce, at the social and behavioral levels, the work of caregiving teams. At one level, information systems can define and inform teams but, in isolation, do not recognize the importance of building multiprofessional teams as a social unit (Farrell, Payne, and Heye 2015). Team socialization in communities of practice is complex because these communities occur outside organizational boundaries, are tailored to an individual patient condition, and thus form and re-form as a unit as conditions change. Teams that are self-organizing have advantages over assigned teams because decision control in forming the team and its function is maintained by the team. The team must then hold itself accountable for the highest evidence-based standards of quality, efficiency, and satisfaction. Knowledge of clinical conditions and treatment approaches in team settings is elicited from each member on the basis of the relevant technical knowledge the team member brings to the case. Relevance can be assessed only by the team—through dialogue with each other—on the basis of each

37

38

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

member’s professional expertise (McDermott, Pedersen, and Körner 2016). In traditional organizational structures, knowledge socialization is reinforced primarily by profession and department and only recently by multiprofessional teams. Even more limiting is that interaction occurs only in real time and within a defined physical space (e.g., team meetings). However, the use of interactive communication software, such as IBM’s Watson Assistant (formerly Watson Conversation), enables team members to engage in dialogue over time and space (IBM 2018). Such interactions are based on the Bohm Dialogue, designed not only to share explicit information but to extract tacit and implicit information through a rigorous, structured process. This technology enables highly skilled professionals to form as interactive teams and function within a distributed community of practice whose culture thrives on teamwork and collaboration. The dialogue includes patients and has the potential to improve decision making, problem solving, patient participation, and patient compliance (and it results in the patient becoming a co-producer). Less attention has been paid to knowledge socialization in the healthcare field applied to professionals in ACOs or communities of practice. The challenge is for organizations that are loosely structured and legally separate to develop a common culture, committed to and rewarded for shared objectives and performance, while maintaining their corporate independence (McDermott, Pedersen, and Körner 2016). A corporate strategy to address these opportunities through acquisitions and mergers is not effective and might negate the intended purpose (Ash et al. 2012).

Knowledge Brokering Knowledge brokering focuses on how professionals and organizations can use their institutional and information expertise in building and supporting a collaborative, integrated health network. Brokering does not focus on clinical information exchange per se but on shared expertise, developing and sustaining a system to support integrated, evidence-based clinical practice. The enabling technology and system design are based on information systems. Expertise about shared CDSSs, interorganizational structures, shared financing, and the like is concentrated within regional centers, thus expanding their shared-knowledge potential from clinical to include organizational. Clinicians as well as local clinics and hospitals in the region are assisted in developing clinical, organizational, and financial relationships, resulting in integrated and dynamic clinical teams. Teams are enabled and supported by the most effective use of clinical evidence contained in electronic systems and by team interactions (Waring et al. 2013). Knowledge brokering is based on developing both electronic interfaces and formal and informal individual and organizational relationships. Social and

C h a p ter 2:   Knowledge- Based D ec ision Making

organizational networks act as intermediaries to identify, access, distribute, and use knowledge across systems (Ward, House, and Hamer 2009). They define how organizations and professionals within systems relate to each other by aligning them to maximize the use of knowledge in the system to provide the highest levels of clinical care. Relationships that transcend organizational boundaries are complex issues in the private, disaggregated systems that characterize the US health system, but this complexity does not necessarily mean moving to a more centralized system or public control (although this is a risk if the system fails to transform itself). Knowledge brokering for clinical care includes subspecialist clinicians in tertiary centers providing clinical consultations to local practitioners using a common patient record and decision support that allows patients to receive seamless specialty consultations within their own community. Such consultations do not replace face-to-face consults but reduce their frequency while increasing quality and satisfaction. This practice enhances not only the quality but also the efficiency of patient care, building strong local medical practices and hospitals. Major obstacles include how subspecialists are paid for valueadded service based on knowledge contribution rather than patient contact; this is a reimbursement system problem, not a clinical function or information system problem, and it should be addressed as such. Some extended form of bundled payment or capitation payment based on value-added service would enable such a relationship to be developed. Reimbursement based on contact time is an obsolete concept. Regional medical centers bring value to local hospitals and clinics by helping them understand, develop, and manage complex CDSSs. Regional centers have expertise in IT, systems engineering, financing, and organization management and strategy, which are valuable to local hospitals and clinics because these systems enable them to survive and thrive. This form of knowledge brokering is essential for small, local institutions and fills a need that is essential to a patient-centered, high-performing health system. Regional centers, with local institutions, can develop executive suite and technical staff expertise and dedicate themselves to supporting local institutions by avoiding referrals and maintaining care within their communities (Bergenholtz 2011). They can benefit from keeping patients in local hospitals and clinics instead of referring them to tertiary care institutions by using a value-added calculation built into the agreement. Local providers would have funds to reimburse the knowledge-broker function because regional institutions would have a shared interest and dedication to keeping services local. These problems are not insurmountable or overly complex in an era when technology is applied to even more complex problems, such as intercelestial travel. It is not that the problems cannot be solved but rather that the focus is on the wrong problem. Healthcare leaders need to bring a systems perspective to problem solving and

39

40

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

not focus on maximizing institutional or individual performance. Changing the system would greatly improve efficiency and effectiveness as well as cost, quality, continuity, and patient involvement—and patients and families would be the beneficiaries. Knowledge brokering assumes that integrated decision support systems provide the logic and value for structuring the delivery of quality and efficient clinical services. Health systems informatics calls for the alignment of organizational structures and strategies, financing, and suppliers to support such a system. Such a system might appear futuristic, but financial and pharmaceutical firms are exploring the design of new business models—in which they are paid not on the basis of sales volumes but on the value added—that promote the delivery of efficient and effective care. These companies envision a future when “outcomes data will drive healthcare” and “pharma will be paid for outcomes, not products” (Studin 2002). In the health system, much of the value-added contribution lies in the knowledge embedded within the information system and in the minds of health professionals. This recognition can serve to realign and restructure many of the current strategies of clinicians, healthcare organizations, financing institutions, suppliers, and information vendors. Information vendors should develop and be rewarded for a business model that delivers value through IT and not for selling hardware or software.

Conclusion Clinical decision processes that are structured to involve all parties, each of whom has scientific and experiential knowledge relevant to the treatment process, are consistent with the principles of knowledge management. The mix of scientific and experiential evidence and the contributions of each health profession to the process will be measured over time against optimal outcomes. Communities of practice can serve as the structural framework for federally mandated ACOs and follow the principles of health maintenance, quality outcomes, patient satisfaction, and system efficiency. Linking organizations and clinicians within the current structures of the clinical decision process will result in suboptimal outcomes. A community of practice operates under the assumption of professional autonomy, and physicians and nurses should take the lead in forming such communities. However, lack of training in the sciences that enable health systems informatics will limit many health professionals from doing so. They need to work with healthcare leaders to create the vision and process for accessing and applying the best clinical and organization knowledge. The community of practice restructures the clinical process and has the potential to improve quality and outcomes. Quality improvement techniques, such as Six Sigma, are frequently applied to departments or microprocesses,

C h a p ter 2:   Knowledge- Based D ec ision Making

which are then bundled to form the clinical work process. However, system failure occurs in the process of bundling and the handoffs between the clinicians and departments involved. Quality improvement applied to the macro or larger process—not individual practices, departments, or institutions—has been shown to work in healthcare, as it has in other service industries (Harris 2006; Margolis et al. 2009; Yang et al. 2012). Comprehending the basics of knowledge management and systems theory is important to understanding the sources of information for clinical decision making. Existing electronic systems help with accessing, integrating, and analyzing clinical information. Although these systems reflect profound advances in health IT, they primarily digitize the existing clinical decision process.

Chapter Discussion Questions 1. Discuss the types of knowledge and how they are applied in clinical decision support. 2. How can tacit and implicit knowledge be generated by teams who do not occupy a common physical location? 3. How would you design a financing system that rewards the professional value contribution of an individual working as a member of a team in a community of practice? 4. In a patient-oriented system, why should sensitive clinical information not be deleted from the EHR? 5. What value-added contributions might tertiary centers bring to an integrated system other than clinical expertise?

Case Study  Knowledge Management in Accountable Care Organizations Gordon D. Brown A large psychiatric specialty group practice in a metropolitan area provides a range of psychiatric services. Since its formation in the 1980s, the practice has grown rapidly and has thrived as a result of the development of managed care plans, including the state Medicaid program that follows a carve-out model for behavioral health services. The psychiatric group is made up of highly respected psychiatrists, who carry out extensive translational research and develop (continued)

41

42

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

evidence-based protocols that are highly respected. They are also skilled at applying clinical guidelines to inform and transform their clinical practice. Several psychiatrists are concerned about the long-term effect of federally mandated accountable care organizations (ACOs). The mandate identifies 65 performance measures that the standard ACO must meet under the Medicare Shared Savings Program. These measures span five quality domains: patient experience of care, care coordination, patient safety, preventive health, and at-risk population/frail elderly health. The only behavioral health measure mandated is in the preventive health domain—a measure for depression screening. Some leaders of the group believe that mental and behavioral health is a highly specialized area and that the practice’s clinical volume will not change significantly. They do not support developing an ACO strategy and have resolved to take a wait-and-see approach. They point out that managed care was promoted in the 1970s by the federal government and then went away as a general policy. Other leaders, including the practice CEO, believe that ACOs do present a long-term threat but also provide an opportunity for the group to transform itself into one that is information driven.

Evidence-Based Strategy The practice organized a multidisciplinary team to explore an ACO strategy. The team comprises two psychiatrists, one psychiatric nurse, the CEO, and a healthcare management intern. The team agrees that it will entertain all ideas and proposals, as well as research the literature to bring the best explicit information and experiential knowledge to the deliberation. Relevant literature topics include ACO basics, knowledge management, disruptive innovation, and managed care organizations’ limited acceptance and success since the 1980s. The management intern, Marjorie, is interested in the medical-offset effect and its potential as a strategic asset. She presents to the team 30 years of critically reviewed research on the concept, including closed clinical trials. Studies on medical-offset measure the impact of providing effective behavioral health services on the utilization of medical care, including physician consults and visits to the emergency department. Marjorie is impressed by the extensive studies that include a wide range of populations and conditions, including Medicaid patients and chronic care diagnoses. The findings consistently demonstrate a savings of 10 to 20 percent through reductions in medical care utilization, particularly specialized services. The psychiatrists meet these studies with skepticism. Although they think the science behind the research is valid, they reason that the practice’s specialty is mental health and not prevention or behavioral health; thus, the studies are not relevant to what they do.

C h a p ter 2:   Knowledge- Based D ec ision Making

In a brainstorming session, the team explores alternative strategies for how the practice might add value to the ACOs that are developing in the area. The first strategy is to embed specialty psychiatric knowledge into the ACO’s decision support system. The second strategy is to develop formal affiliations with as many ACOs as possible to capture their referrals for specialty care. The third strategy is to extend the practice’s decision support protocols to address prevention, early detection, and aggressive management of behavioral health issues. This last idea is suggested by the psychiatric nurse, who points out that the nursing staff and social workers have considerable, but underused, expertise in behavioral health. The psychiatrists on the team worry that developing behavioral health decision support protocols would result in a loss of status for the psychiatrists and thus would be strongly opposed. They note that the specialty practice of psychiatry is the core competency of the group and that the proposed strategy would result in a loss of prestige and reputation for the practice. After considerable discussion and debate, the team agrees to respect and maintain the existing culture and practice of the psychiatrists but to pursue a broader strategy to better position the group for the future. They begin to develop a full proposal for the third strategy.

Case Study Discussion Questions 1. Create a strategy to leverage the knowledge base of the practice against the value of knowledge in the developing ACOs. Consider the following guidelines and questions for this exercise: • What change will be made in who accesses knowledge generated by the practice and how it will be used? • Address each of the five quality domains specified by the ACO mandate, and justify their inclusion or exclusion. What are the implications of each on the structure of the clinical process and on the information system that supports the process? • What new properties of the decision support system must be included in the proposed strategy? How might tacit knowledge in the practice be leveraged by primary care physicians, other specialists, and patients? • What kind of organizational structure would be formed with the ACO? What are the implications of collaborating with more than one ACO? Should the psychiatric group serve as the focal organization for developing an ACO? • What value would be brought to the ACO, and how would it be assessed? How would the practice be paid for its value-added services?

(continued)

43

44

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

2. Comment on the practice’s environment and its readiness for strategic change. Whom might you want to add to the team that is exploring an ACO strategy? 3. In the face of strong evidence, why do you think managed behavioral health has been slow to develop and become a niche industry by carving out behavioral health and managing it separately? What are the reasons for this, and what alternative strategies might be considered?

Additional Resources Center for Healthcare Innovations: www.chisite.org. Centers for Medicare & Medicaid Services. 2011. “Medicare Resource Use Measurement Plan.” www.cms.gov/QualityInitiativesGenInfo/downloads/Resource Use_Roadmap_OEA_1-15_508.pdf. ClinicalTrials.gov: https://clinicaltrials.gov/ct2/about-site/for-media. IBM. 2018. “About Watson Assistant.” https://console.bluemix.net/docs/services/ conversation/index.html#about. Informatics for Integrating Biology and the Bedside (i2b2): www.i2b2.org. Kaiser Permanente Garfield Innovation Center: http://xnet.kp.org/innovationcenter/. Mayo Clinic Center for Innovation: http://centerforinnovation.mayo.edu/transform/. Medical Subject Headings (MESH): www.ncbi.nlm.nih.gov/mesh. Morrisey, J. 2017. “Ripe for Disruption: Technology Is Changing the Way Providers Deliver Care.” Health Data Management July/August, 14–17. National Guideline Clearinghouse: www.guideline.gov. PubMed, National Library of Medicine: www.ncbi.nlm.nih.gov/pubmed/. Roberts, K. M., R. Boland, L. Pruinelli, J. Dcruz, A. Berry, M. Georgsson, R. Hazen, R. Sarmiento, U. Backonja, Y. Kun-Hsing, J. Yun, P. F. Brennan, K.-H. Yu, and Y. Jiang. 2017. “Biomedical Informatics Advancing the National Health Agenda: The AMIA 2015 Year-in-Review in Clinical and Consumer Informatics.” Journal of the American Medical Informatics Association 24 (e1): e185–e190. Shared Health Research Information Network (SHRINE): https://catalyst.harvard. edu/services/shrine/. Universal Medical Language System (UMLS): https://uts.nlm.nih.gov.

References Afantenos, S., V. Karkaletsis, and P. Stamatopoulos. 2005. “Summarization from Medical Documents: A Survey.” Artificial Intelligence in Medicine 33 (2): 157–77.

C h a p ter 2:   Knowledge- Based D ec ision Making

Ash, J. S., D. F. Sittig, K. P. Guappone, R. H. Dykstra, J. Richardson, A. Wright, J. Carpenter, C. McMullen, M. Shapiro, A. Bunce, and B. Middleton. 2012. “Recommended Practices for Computerized Clinical Decision Support and Knowledge Management in Community Settings: A Qualitative Study.” BMC Medical Informatics and Decision Making 12 (6): 1–19. Azizi, S., F. Imani, J. T. Kwak, A. Tahmasebi, S. Xu, P. Yan, J. Kruecker, B. Turkbey, P. Choyke, P. Pinto, B. Wood, P. Mousavi, and P. Abolmaesumi. 2016. “Classifying Cancer Grades Using Temporal Ultrasound for Transrectal Prostate Biopsy.” In Medical Image Computing and Computer-Assisted Intervention—MICCAI 2016  edited by S. Ourselin, L. Joskowicz, M. Sabuncu, G. Unal, and W. Wells, 653–61. Basel, Switzerland: Springer International Publishing AG. Bergenholtz, C. 2011. “Knowledge Brokering: Spanning Technological and Network Boundaries.” European Journal of Innovation Management 14 (1): 74–92. Best, A., A. Berland, C. Herbert, J. Bitz, M. van Dijk, C. Krause, D. Cochrane, K. Noel, J. Marsden, S. McKeown, and J. Millar. 2016. “Using Systems Thinking to Support Clinical System Transformation.” Journal of Health Organization and Management 30 (3): 302–23. Biran, O., and C. Cotton. 2018. “Explanation and Justification in Machine Learning: A Survey.” In Proceedings of IJCAI-17 Workshop on Explainable AI (XAI), 20 August 2017, Melbourne, Australia, edited by D. W. Aha, R. Darrell, M. Pazzani, D. Reid, C. Sammut, and P. Stone, 8–13. Accessed May 8. https://pdfs. semanticscholar.org/d8be/bea1497357b3d2738e0ae5cde9d38ee45d46.pdf. Cabitza, F., R. Rasoini, and G. F. Gensini. 2017. “Unintended Consequences of Machine Learning in Medicine.” Journal of the American Medical Association 318 (6): 517–18. Chen, R., N. R. Desai, J. S. Ross, W. Zhang, K. H. Chau, B. Wayda, K. Murugiah, D. Y. Lu, A. Mittal, and H. M. Krumholz. 2016. “Publication and Reporting of Clinical Trial Results: Cross Sectional Analysis Across Academic Medical Centers.” British Medical Journal 352: i637. Cooke, M., D. M. Irby, W. Sullivan, and K. M. Ludmerer. 2006. “American Medical Education 100 Years After the Flexner Report.” New England Journal of Medicine 355 (13): 1339–44. De Backere, F., H. Moens, K. Steurbaut, K. Colpaert, J. Decruyenaere, and F. De Turck. 2012. “Towards Automated Generation and Execution of Clinical Guidelines: Engine Design and Implementation Through the ICU Modified Schofield Use Case.” Computers in Biology and Medicine 42 (8): 793–805. Dixon, B. E., L. Simonaitis, H. S. Goldberg, M. D. Paterno, M. Schaeffer, T. Hongsermeier, A. E. Wright, and B. E. Middleton. 2013. “A Pilot Study of Distributed Knowledge Management and Clinical Decision Support in the Cloud.” Artificial Intelligence in Medicine 59 (1): 45–53. Ennis, J., D. Gillen, A. Rubenstein, E. Worcester, M. E. Brecher, J. Asplin, and F. Coe. 2015. “Clinical Decision Support Improves Physician Guideline Adherence for

45

46

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Laboratory Monitoring of Chronic Kidney Disease: A Matched Cohort Study.” BMC Nephrology 16: 163. Faraj, S., S. L. Jarvenpaa, and A. Majchrzak. 2011. “Knowledge Collaboration in Online Communities.” Organization Science 22 (5): 1224–39. Farrell, K., C. Payne, and M. Heye. 2015. “Integrating Interprofessional Collaboration Skills into the Advanced Practice Registered Nurse Socialization Process.” Journal of Professional Nursing 31 (1): 5–10. Fervers, B., J. S. Burgers, M. C. Haugh, J. Latreille, N. Mlika-Cabanne, L. Paquet, M. Coulombe, M. Poirier, and B. Burnand. 2006. “Adaptation of Clinical Guidelines: Literature Review and Proposition for a Framework and Procedure.” International Journal for Quality in Health Care 18 (3): 167–76. Fowler, A. 2003. “Systems Modeling, Simulation, and the Dynamics of Strategy.” Journal of Business Research 56 (2): 135–44. Garrety, K., I. McLoughlin, A. Dalley, R. Wilson, and P. Yue. 2016. “National Electronic Health Record Systems as ‘Wicked Projects’: The Australian Experience.” Information Polity 21 (4): 367–81. Harris, C. 2006. Hyperinnovation and Building Innovative Teams. New York: Palgrave MacMillan. Hassanpour, S., C. P. Langlotz, T. J. Amrhein, N. T. Befera, and M. P. Lungren. 2017. “Performance of a Machine Learning Classifier of Knee MRI Reports in Two Large Academic Radiology Practices: A Tool to Estimate Diagnostic Yield.” American Journal of Roentgenology 208 (4): 750–53. Haynes, A. B., T. G. Weiser, W. R. Berry, S. R. Lipsitz, A. S. Breizat, E. P. Dellinger, T. Herbosa, S. Joseph, P. L. Kibatala, M. C. Lapitan, A. F. Merry, K. Moorthy, R. K. Reznick, B. Taylor, and A. Gawande. 2009. “Safe Surgery Saves Lives Study Group: A Surgical Safety Checklist to Reduce Morbidity and Mortality in a Global Population.” New England Journal of Medicine 360 (5): 491–99. Herremans, I., R. G. Isaac, J. B. Kline, and J. A. Nazari. 2011. “Intellectual Capital and Uncertainty of Knowledge: Control by Design of the Management System.” Journal of Business Ethics 98 (4): 627–40. IBM. 2018. “About Watson Assistant.” https://console.bluemix.net/docs/services/ conversation/index.html#about. Ketcherside, M., D. Rhodes, S. Powelson, C. Cox, and J. Parker. 2017. “Translating Interprofessional Theory to Interprofessional Practice.” Journal of Professional Nursing 33 (5): 370–77. Kohli, M., L. M. Prevedello, R. W. Filice, and J. R. Geis. 2017. “Implementing Machine Learning in Radiology Practice and Research.” American Journal of Roentgenology 208 (4): 754–60. Kothari, A., N. Hovanec, R. Hastie, and S. Shannon. 2011. “Lessons from the Business Sector for Successful Knowledge Management in Health Care: A Systematic Review.” BMC Health Services Research 11: 173–200.

C h a p ter 2:   Knowledge- Based D ec ision Making

Lee, J. Sr. 2000. “Knowledge Management: The Intellectual Revolution.” HE Solutions 32 (10): 34–37. MacArthur, B. L., S. L. Dailey, and M. M. Villagran. 2016. “Understanding Healthcare Providers’ Professional Identification: The Role of Interprofessional Communication in the Vocational Socialization of Physicians.” Journal of Interprofessional Education & Practice 5: 11–17. Margolis, P., L. Provost, P. J. Schoettker, and M. T. Britto. 2009. “Quality Improvement, Clinical Research, and Quality Improvement Research—Opportunities for Integration.” Pediatric Clinics of North America 56 (4): 831–41. McDermott, A. M., A. R. Pedersen, and M. Körner. 2016. “Knowledge Integration, Teamwork and Performance in Health Care.” Journal of Health Organization and Management 30 (2): 227–43. McGlynn, E. A., S. M. Asch, J. Adams, J. Keesey, J. Hicks, A. DeCristofaro, and E. A. Kerr. 2003. “The Quality of Health Care Delivered to Adults in the United States.” New England Journal of Medicine 348 (26): 2635–45. Mishra, R., J. Bian, M. Fiszman, C. R. Weir, S. Jonnalagadda, J. Mostafa, and G. Del Fiol. 2014. “Text Summarization in the Biomedical Domain: A Systematic Review of Recent Research.” Journal of Biomedical Informatics 52: 457–67. Nonaka, I., and H. Takeuchi. 1995. The Knowledge Creating Company. New York: Oxford University Press. Olczak, J., N. Fahlberg, A. Maki, A. S. Razavian, A. Jilert, A. Stark, O. Sköldenberg, and M. Gordon. 2017. “Artificial Intelligence for Analyzing Orthopedic Trauma Radiographs.” Acta Orthopaedic 88 (6): 581–86. Rantz, M., M. Skubic, M. Popescu, C. Galambos, R. J. Koopman, G. L. Alexander, L. J. Phillips, K. Musterman, J. Back, and S. J. Miller. 2015. “A New Paradigm of Technology-Enabled ‘Vital Signs’ for Early Detection of Health Change for Older Adults.” Gerontology 61 (3): 281–90. Rodriguez-Loya, S., A. Aziz, and C. Chatwin. 2014. “A Service Oriented Approach for Guidelines-Based Clinical Decision Support Using BPMN.” Studies in Health Technology and Informatics 205: 43–47. Senge, P. M. 1990. The Fifth Discipline. New York: Currency Doubleday. Shortliffe, E. H. 1976. Computer-Based Medical Consultations: MYCIN. North Holland, NY: Elsevier. Studin, I. 2002. “Reframing the Pharmaceutical Manufacturer/Health Plan Relationship in Managed Care.” Managed Care 11 (2): 46, 49–54. Tranfield, D., D. Denyer, and P. Smart. 2003. “Towards a Methodology for Developing Evidence-Informed Management Knowledge by Means of Systematic Review.” British Journal of Management 14 (3): 207–22. Vest, J. R. 2009. “Health Information Exchange and Healthcare Utilization.” Journal of Medical Systems 33 (3): 223–31.

47

48

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Wang, X., D. B. Arnett, and L. Hou. 2016. “Using External Knowledge to Improve Organizational Innovativeness: Understanding the Knowledge Leveraging Process.” Journal of Business & Industrial Marketing 31 (2): 164–73. Ward, V. L., A. O. House, and S. Hamer. 2009. “Knowledge Brokering: Exploring the Process of Transferring Knowledge into Action.” BMC Health Services Research 9: 12. Waring, J., G. Currie, A. Crompton, and S. Bishop. 2013. “An Exploratory Study of Knowledge Brokering in Hospital Settings: Facilitating Knowledge Sharing and Learning for Patient Safety.” Social Science & Medicine 98: 79–86. Yang, H., W. Li, K. Liu, and J. Zhang. 2012. “Knowledge-Based Clinical Pathway for Medical Quality Improvement.” Information Systems Frontiers 14 (1): 105–17. Zarin, D. A., T. Tse, R. J. Williams, and S. Carr. 2016. “Trial Reporting in ClinicalTrials. gov—the Final Rule.” New England Journal of Medicine 375 (20): 1998–2004.

CHAPTER

HEALTH PROFESSIONS, PATIENTS, AND DECISIONS

3

Gordon D. Brown

Learning Objectives After reading this chapter, you should be able to do the following: • Understand the structure and importance of the role of the health professions in society. • Explain how advanced information technology changes the role of the health professions but does not diminish it. • Conceptualize clinical decision support systems that enable judgment and experiential knowledge to be applied to clinical decision making.

Key Concepts • • • • • •

Professional versus individual autonomy Evidence-based clinical practice Clinical decision process and decision support Decision types and contexts Actor–network theory Team decision processes

Introduction Will the traditional role of the health professions survive, or will it be transformed, diminished, or enhanced in an information technology (IT)–driven world? This chapter explores the privileges afforded to health professionals and their clinical obligations to society. The role of the health professions is viewed through advances and transformation in clinical decision making. Clinical decision making and coordination of work are grounded in the long-standing values and behavior of highly trained and dedicated health professionals. Their role changes 49

50

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

over time, in part because of a rapid increase in basic and cliniProfessional Versus Individual Autonomy cal sciences, many of which are There’s a difference between professional and individual autonomy. in themselves transformational. Professional autonomy is the freedom of the profession (as opposed to Changes in the basic sciences policymakers, executives, financial managers, and others) to develop includes breakthroughs in the and use evidence-based guidelines. Physicians enjoy individual autonbiological sciences, the rapid omy in that they can deviate from the guidelines if, in their judgment, development of genetic sciences, doing so is warranted. However, to physicians, being required to follow and the growing importance of “institutional” guidelines, no matter how they are developed, feels like a engineering sciences. Informaloss of individual autonomy. They may fear that once guidelines become tion science provides the archi“institutionalized,” they will further intrude on individual autonomy. This tecture for transforming the fear is warranted in that future information systems will be designed to structure of clinical decision increase integration of the clinical function. The operative question is making and the clinical process. the role professionals will play in leading health system transformation. The traditional, independent role of health professionals in society has defined how (1) health professionals are trained, (2) healthcare organizations are structured and managed, (3) services are financed, (4) policies are set, and (5) systems are evaluated. Health systems informatics fundamentally changes the structures of these systems, including the logic of their design and the basis on which they are evaluated. In a well-designed and well-managed system, there will be a loss of individual decision control as decisions are increasingly based on collaboration and informed by accumulated clinical evidence. There need not be a loss of professional control or flexibility, even though it might look and feel that way to individual professionals if the system is not well designed and managed. One could argue that a loss of professional control will occur as patients increasingly participate in the decision-making process, but engaging patients is part of the professional’s role. It is the professional’s reason for being. A new environment might be characterized as a team structure, drawing on the best scientific evidence with full participation of the patient. The new challenge of the health system is to collect the latest and best evidence available, within a distributed decision-making structure, and use it to inform decisions. There is extensive literature on decision making that has not been applied to clinical decisions. The types of decisions in the healthcare sector may be different from those in other disciplines, but they are based on the same science. The types of decisions made across the clinical spectrum and the ways they are informed are discussed here.

Transformation of the Clinical Function Customizing healthcare services to patient and family conditions has traditionally been carried out by individual health professionals, which is their role in

C h a p te r 3:   H e a l th Professions, Patients, and D ec isions

society. This expectation does not suggest that in a transformed health system, it will be business as usual for health professionals. Redesigning or transforming the health system requires transforming the clinical function. Thus, in an information-driven health system, the challenge is preserving the role of the professions while transforming the clinical function.

Health Professionals The structure of clinical decision making and of the clinical process requires an understanding of the nature of professions in society. The label professional has been inappropriately adopted by many occupations; it is a privileged social role that carries with it a high moral obligation and certain freedoms. Professions are institutions created by society and thus vary from society to society and change over time (Muzio and Kirkpatrick 2011). Professionals have “specialized training and knowledge, ethicality, and importance to society” (Friedson 1994, 19). This training and knowledge give them special privileges and responsibilities, as well as power and legitimacy. Their role changes in response to society’s changing values, demographics, economics, and technology. The role of health professionals is reflected in their oath to society and patients, which recognizes medicine as a calling. The modern Hippocratic oath is as follows (Tyson 2001): I will remember that there is art to medicine as well as science, and that warmth, sympathy, and understanding may outweigh the surgeon’s knife or the chemist’s drug. . . . I will remember that I do not treat a fever chart, a cancerous growth, but a sick human being, whose illness may affect the person’s family and economic stability. My responsibility includes these related problems, if I am to care adequately for the sick.

Likewise, the nursing profession abides by an oath known as the Nightingale Pledge (American Nurses Association 2012): I will do all in my power to maintain and elevate the standard of my profession, and will hold in confidence all personal matters committed to my keeping and all family affairs coming to my knowledge in the practice of my calling.

These oaths state the commitment of professionals to their patients, who have varying physical and medical conditions, socioeconomic statuses, beliefs, values, and traditions. This orientation requires professionals to retain some autonomy—the freedom and flexibility to tailor their response according to their own clinical judgment as well as the patient’s health, desires, social conditions, and values. Working independently and relying on their knowledge base and available information to make clinical decisions characterize the culture of health professionals.

51

52

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

The role of physicians does not only start when they obtain admitting privileges in hospitals; it begins once they are accepted into medical school and continues throughout their training. Medical faculty are granted considerable autonomy in universities to set standards, select and admit students, define the curriculum, and assess student performance. Personal values, attitudes, and behaviors are just as important as clinical knowledge and skills in the medical education process. Changing how physicians make decisions, collaborate within and across specialties, work in teams, and share decision making with patients, starts upon admission to medical school. This change is not easy; as Calvin Coolidge once said, “Changing a curriculum is like moving a graveyard.” This is particularly true in highly professionalized fields. Clinicians are expected to be not only innovative but also responsive to advances in medical knowledge, information technology, and cultural values— a challenge given the speed of change in the field and society. Carle-Illinois College of Medicine (https://medicine.illinois.edu), for example, has caused a disruptive innovation in physician education by adding engineering science, to complement biological and clinical sciences, to its curriculum. Engineering science includes bioengineering science, precision medicine, and clinical medicine systems engineering (Khayal and Farid 2017), which have been increasingly contributing to medicine.

Clinical Decision Process and Decision Support Studies have found that clinical guidelines, based on scientific evidence, result in outcomes improvement—although most are focused on the clinical decision (even clinical specialty) and not the structure of the decision-making process, making it difficult to draw firm conclusions (Hoomans et al. 2011; Latoszek-Berendsen et al. 2010). There is evidence that clinical decision support tools have had a positive effect on clinical outcomes, and students should critically review this vast and varied collection of studies. We explore the literature here as a framework for understanding the relationship between decision process and outcome. Much attention has been paid to the quality of clinical evidence reported in guidelines, physicians’ acceptance of the evidence, and the evidence’s effect on quality (Alonso-Coello et al. 2010; Tricoci et al. 2009). The acceptance of a clinical decision support system (CDSS) depends on the level of science and evidence supporting the guideline it presents, such as closed clinical trials. What is not regularly reported is the structure of the clinical decision process within which the guideline was developed and tested and how it is to be applied (Bates, Kuperman, and Wang 2003). A CDSS sometimes conforms to—and sometimes transforms—the structure of the clinical decision process. Thus, more attention must be paid to the process itself. Such studies require testing different decision models for exploring new approaches to nonlinear interaction among health professionals. Studies should carefully identify the context and process as well as the outcome (Brenner et al. 2016).

C h a p te r 3:   H e a l th Professions, Patients, and D ec isions

Clinical decision processes are becoming increasingly complex, involving multiprofessional teams. Hafferty and Levinson (2008) suggest that complexity science could be used to recast health professionals, organizations, and systems to be interactive, adaptive, and interdependent. Complexity science is a systems concept based on agility of thought, creativity, risk taking, and the value of diversity (Brainard and Hunter 2016). Current guidelines only report decisions and outcomes but do not report or measure context. This practice is problematic in a highly professionalized field such as healthcare and increases the risk of guidelines either becoming too mechanistic or allowing too much discretion. The science should inform the process as well as the decision. The logic of health systems informatics is based on a transformed clinical decision process and system design. Transformed organizations have been referred to as learning or knowledge organizations, are self-organizing, and are able to capture and use a range of knowledge, including tacit knowledge. They are purposeful systems in which individuals work together toward common goals to achieve extraordinary performance. Hierarchical structures and decision processes are as outmoded as the traditional role of professionals. Under the logic of an integrated clinical function, leaders must collaborate to redesign work processes. Physicians, nurses, and other clinicians must work in integrated teams and be jointly accountable for clinical outcomes. Working together requires flexible and adaptive systems that are “less bounded” but not “boundaryless” (Friedson 2001). Such systems operate with a degree of standardization—of both the clinical processes and the clinical outcomes. Standardization is an inherent condition of structured work processes and is a departure from the traditional independent or autonomous role of health professionals. Standardized measures and information systems do not mean mechanistic decision processes but rather information and decision support that transcend individual practitioners and institutions. Clinical decisions are structured around clinical work processes that include health professionals who have knowledge relevant to a treatment process. Leadership can be shared by professionals who have the dominant knowledge of the case at a given point in the clinical process and not be assumed by the traditional coterie of physicians. Such structured processes are evidence based, drawing on engineering, behavioral, and social sciences. Health professionals recognize the value of information and knowledge embedded within a CDSS and are increasingly structuring clinical decision making around its logic. They understand the power of IT and integrated teams, which require them to fundamentally redesign the structure of many clinical functions. Such a redesign requires new types of healthcare organizations and IT systems, as well as new perspectives on clinical decision making. It calls for a needed transformation of the organization itself (McGuinness 2014). Most healthcare facilities have implemented evidence-based decision making, but they have done so within the traditional role of clinicians and the traditional

53

54

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

structure of organizations. Small hospitals and clinics lag in this transformation because of shortage of funds, resistive cultures, and lack of time or capacity. Failing to change will put at risk their ability to position themselves strategically and even their very survival.

Effective Clinical Decision Support Systems Many studies that report positive outcomes from using CDSSs lack sufficient rigor because they do not recognize the nature of the clinical decision process. They demonstrate incremental improvements but seldom recognize the outcomes of the overall process and the essential tenets of a CDSS. These limitations are, in part, the result of the way these studies are conducted. Most studies are run by individuals who have considerable knowledge about clinical care but not about the science of decision making or the types of decisions and CDSSs being studied; they do not seem to understand that not all clinical decisions are equal. As a result, their findings tend to apply to specific clinical decisions and are not generalizable to other decision types, situations, or settings. In addition, they frequently refer to decision science concepts that they do not understand. For example, they use a term such as intuition but do not explain its intuitive decision-making basis or application. The lack of conceptual rigor in many studies limits the accumulated knowledge available for clinical decision making, which imposes a constraint on the design of effective CDSSs. Although such research provides evidence to inform clinical decisions, it does so without context. The healthcare field has thus not been able to draw on what is known about designing effective CDSSs or to contribute to the body of decision science. Healthcare leaders and system designers must be critical readers of published research on CDSSs and use the findings to design evidence-based clinical processes. Designing CDSSs is a complex process that requires time, skills in decision and systems science, and the freedom to test innovative solution strategies. They must be critical consumers and not assume that IT vendors or consultants serve their best interests. It is the function of researchers, IT companies, and policymakers to develop and test effective CDSS tools. The challenge is to gain an understanding of the types of clinical decisions being addressed as well as to design and test CDSSs that are consistent with those types of decisions, because different decision types require different mechanisms. Information system interfaces and information exchange are essential qualities of effective CDSSs.

The Science of Clinical Decision Making The Institute for Technology Assessment (2018) identified the need to apply decision science to the complex and competing functions in clinical practice

C h a p te r 3:   H e a l th Professions, Patients, and D ec isions

and public health. Decision science is the systematic analysis of the complexity and dynamic nature of decision making, applied here within clinical contexts. It recognizes that contextual factors must be explicitly examined so that decisionmaking processes enable a thorough but analytically appropriate method of achieving optimal outcomes. Decision science applied to healthcare includes quantitative analyses and supporting evidence for clinical decisions and clinical processes. Decision science in healthcare incorporates the clinical sciences as well as engineering, organization, and systems theory to achieve optimal clinical outcomes and system performance. Historically, clinical decision making was based primarily on the context of individual clinical decisions, but systemsbased clinical decision making broadens the focus of clinical decisions and adds the range of organizational, behavioral, and engineering sciences. This section focuses on clinical decision making and its context by type of clinical decision. (Chapter 4 focuses on the organization and system context.) Decision science applied to clinical decisions recognizes that different types of decisions follow different rules, each of which has a distinct logic and scientific basis (exhibit 3.1). The design and use of CDSSs are defined by and serve the inherent and essential properties of specific decision types and decision processes. The decision process and the type of evidence needed to support the decision are determined by the properties of the decision being made. Evidence is used to frame the decision process and inform the decision. The decision itself and its context are both dependent variables, making health informatics independent as a transformational science. Clinical decision making consists of several interactive components: (1) knowledge of a patient’s condition; (2) effective treatment options based on accumulated evidence; (3) the patient’s perspective; and (4) contextual factors, such as decision timeliness and urgency. Treatment options refer to the evidence supporting the decision and the associated composition and structure of the clinical team. Complete knowledge of a patient’s condition is never achieved or needed as a basis for making a diagnosis, primarily because continued search might compromise the timeliness, quality, and acceptability of the decision.

Type

Logic

Basis

Evidence-based Mechanistic decisions—Computational

Facts known

Evidence-based decisions—Judgment

Expert and probabilistic reasoning

Factual basis but sufficient facts not known

Intuitive

Subconscious

Holistic thought

Affective

Normative

Values

55

Decision science Systematic analysis of the complexity and dynamic nature of decision making

EXHIBIT 3.1 Decision Types and Logic

56

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Knowledge includes both descriptive information (e.g., patient preferences) and scientific evidence to support an accurate diagnosis and treatment. As various diagnostic iterations are completed, alternative treatment options—based on accumulated evidence—are considered. While the dimensions of the clinical problem are framed, based on iterations of the diagnostic process, a solution strategy is developed by searching a memory set to bring to bear the collected evidence and knowledge. Other factors, such as cost and the reimbursement mechanism, also influence clinical decision making but are considered clinical support factors, not deterministic, and are addressed throughout the book.

Types of Clinical Decisions Clinical knowledge is being generated at a fast rate but is difficult to access, synthesize, analyze, and report. IT has the power to locate, access, accumulate, analyze, and present information and knowledge in a usable form for clinical decision making. The manner in which this process is carried out depends on the type and context of the decision being considered; this is particularly true in decision making in professional domains such as healthcare. The design of a CDSS relies on decision type and context as well. Evidence-Based Decisions Evidence-based decisions are either mechanistic/computational in nature or probabilistic, based on judgment and expert reasoning. Computational decisions in clinical practice are constrained in health systems by limited access to large databases. In addition, most clinical decisions are inherently not mechanistic in nature and thus not appropriate for computational solutions. Probabilistic decisions have a factual basis, drawing on the best clinical evidence, but recognize that clinical judgment is required. This might be true because only limited clinical evidence is available or because of some exceptional properties of the case. The decision is enhanced by experiential knowledge and/or the inclusion of patient preferences. These multiple bases for decision making are represented by the shaded areas in exhibit 3.1, making the science of clinical decision making very complex. In general, most if not all clinical decisions are informed by facts, but the factual basis might not be the ultimate logic on which the decision is made. Computational Decisions Computational clinical decisions represent the highest level of evidence-based decision making and are mechanistic and factual in nature. They are based on a massive amount of data and knowledge as well as rapid processing speed, using technologies such as artificial intelligence (AI) or machine learning. These decisions are justified in that they are superior to decisions that involve human judgment. Broader applications of computational decisions in diagnosis

C h a p te r 3:   H e a l th Professions, Patients, and D ec isions

and treatment will increase as clinical process–outcome relationships become better known, including tailored interventions through precision medicine. (Computational decision technologies are further discussed in chapter 5.) More sophisticated decision support tools are being developed to inform clinicians, enabled by ever-increasing clinical repositories and processing speed. However, judgment in clinical decision making can never be substantially replaced by computational decision making because of the inherent requirement to incorporate other dimensions of decision making into the clinical decision process. Thus, such decisions cannot be appropriately made using AI, simply because they are not mechanistic in nature, and to force it would compromise the integrity of the decision. In addition to clinical decision support, smart systems will be increasingly deployed to carry out many operational functions in the health system, such as supply chain (supply robots) and other mechanistic functions. For example, new engineering technologies—such as medical drones—are being deployed to respond to alerts to access and rapidly transport urgently needed supplies, such as blood, to the emergency department (ED), reducing the transport of patients from rural hospitals to medical centers (Mayo Clinic 2017). These technologies are transformational, but the focus of this book is primarily on clinical decision making. Judgment-Based Decisions Judgment-based decisions by clinicians depend on accumulated clinical evidence as well as knowledge derived from training and experiences. Judgment is methodical, deliberative, and deployed when the evidence is overridden by experiential learning or when adequate decision support tools are not available at the point of decision making. Judgment-based decisions draw on the best available clinical evidence (exhibit 3.1). Experiential knowledge draws on individual judgment and heuristic reasoning. Heuristic reasoning is not fully consistent with mathematical representation and cannot be easily standardized. It is considered a structured way of thinking and includes introspection, conceptualization, and insight, which lead to problem solving that is interrupted and discontinuous. How and when, then, is heuristic reasoning applied, and how does it incorporate clinical evidence? Heuristics provides rigor to decisions that are probabilistic but not programmable. It reflects how clinicians make decisions and should be incorporated into the design of CDSSs. Evidence suggests that probabilistic decisions are best made by highly trained domain experts with clinical experience, using the science of expert reasoning. Judgment in decision making is important when time does not permit the decision maker to engage in a more deliberate process, including using available decision support tools, or when a specific clinical case includes conditions that a clinical professional believes justify overriding the expressed

57

58

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

decision support recommendation. Use of judgment cannot be justified by understaffing or the ineffectiveness or faultiness of a CDSS; such excuses are made in organizations that follow business rules or prioritize financial incentives instead of appropriately staffing to achieve optimal outcomes. Imposing operational restrictions or rules on clinicians is an inherent danger of organizational practice. Evidence suggests that a degree of uncertainty characterizes most clinical decisions and that this uncertainty cannot be eliminated for most types of clinical decisions, although it can be reduced by CDSSs. Judgment is an important element, but only when it builds on the best available evidence from a CDSS. In this way, judgment-based decisions are still evidence based and include probabilistic elements. Recognition of judgment in clinical decision making raises critical issues regarding monitoring decisions and imposing sanctions on deviations from evidence-based clinical guidelines. Judgment might be exercised in the absence of evidence and, in some cases, to override the available evidence. Healthcare organizations and systems must appropriately monitor and hold clinicians accountable for how evidence is used. Such information might provide a basis for guideline modification (changing the CDSS) or continuing education for the clinician (changing the clinician). Enabling the ability to retrospectively review cases related to the clinical decision is a value-added contribution of CDSSs. The use of guidelines and retrospective review (modified medical rounds) provide valuable information and serve the best interests of the patient. If negative sanctions are rigidly imposed when deviations from the stated guidelines are made, there is risk that clinicians will be incentivized to comply with the guidelines but will not exercise their best, informed clinical judgment on behalf of the patient. Such actions violate the oath of professionals and the moral obligations of the organization. Judgment in decision making draws on the traditional role of health professionals to always act in the best interest of their patients. Clinical judgment can be defended only when it builds on the best scientific evidence presented by the case. Statements such as “I have always done it this way” or “This is how I was taught” are not defensible. The use of judgment in decision making in medicine, nursing, and other health professions supports the tradition of educating professionals in mental processing that characterizes these forms of decision making. Such conditions can be programmed into patient simulation using mannequins, which are found in all quality medical and nursing schools (Patel and Kannampallil 2015). The use of judgment also logically supports its linkage to more evidence-based simulation in professional and continuing education. Clinicians should welcome a retrospective review of decisions that have been tested against the evidence, because these can serve as an effective basis for continuing education for those who are relatively new in practice and those who have been in practice for decades (Harteis et al. 2012).

C h a p te r 3:   H e a l th Professions, Patients, and D ec isions

Intuitive Decisions Although intuition is a frequently misused concept in the clinical literature, it is a legitimate decision process dominated by the decision maker’s subconscious mind (Ruhe and Wang 2007). Labeling decision contexts without drawing on recognized principles limits learning and informed application. Intuitive decisions are based on innate qualities of the decision maker, which are acquired but do not constitute reasoning or learning. These decisions do not replace or negate the importance of evidence but may use intuition to augment or justify the decisions. The decision itself, however, is holistic—drawing from evidence, facts, and the subconscious—and cannot be justified in the absence of an evidence-based decision support tool. Intuitive decision making views the information being processed in parallel rather than in sequence; that is, the quality of the outcome is dependent on intuition, facts, and evidence working together. Clinical decisions based on intuition must demonstrate superior outcomes. Affective Decisions Affective decisions are based primarily on values—presumably of the patient, not the professional—including emotion, feeling, and mood. Affective decisions are assumed to be thoughtful and deliberate and to incorporate evidence about the disease and treatment options. They are most notably made in clinical areas such as palliative care and are becoming increasingly important as patients and families become more involved in chronic and acute care decisions (National Academies of Sciences, Engineering, and Medicine 2017). Proxy decision makers represent the patient’s values and desires (Rolland, Emanuel, and Torke 2017). Beliefs and values can override but do not replace the best available science. Value-based decision making is not unique to healthcare and is considered a strength, not a weakness, of a well-conceived CDSS. Such CDSSs might include the patient’s social, cultural, religious, and family characteristics as well as clinical findings to frame decisions that are value based. Social information in medical records does not replace patient involvement but has been found to increase the appreciation and empathy for the particular social conditions of patients (Kotay et al. 2016).

Context of Clinical Decisions Decision context serves as the basis for designing CDSSs. At the most basic level are single-encounter decisions made by individual health professionals meeting with a patient for an acute episodic illness. Although this might be a single encounter with the medical system, it includes a continuous process of health maintenance. This decision context—not necessarily the decision—is relatively simple: The physician gathers relevant information, draws on available clinical decision support tools, makes a diagnosis, and prescribes treatment for an individual patient. Current CDSSs have been primarily designed with

59

60

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

this basic context. As CDSSs developed, their design has freOpportunity for Interprofessional Education quently included a systematic An elderly patient is at risk of falling. Her condition is characterized by search of the evidence on the a nexus of problems—urinary incontinence, diabetes, musculoskeletal decision but not on the cliniweakness, depression, and osteoarthritis. Describe a clinical team of cal context and broader clinical “health professionals working in collaboration with others, transcendprocess. The level of evidence ing specialties, professions, departments, and . . . organizations” that and quality of the decision makhas been tasked to make decisions about the care for this patient. ing have improved, but the How would you devise a simulation of such a decision-making process decision context has remained with fellow students—in this class and in others? the same. The episodic clinical encounter is giving way to a continuous process of health maintenance and wellness for chronic diseases. Such a process is informed by multiple consultations, patient networks, sensors, and self-diagnostic algorithms, making the process more complex. The clinical decision process thus increasingly involves the patient and family, extending the logic of the decision-making process and thus the design and use of decision support tools. Current information systems that allow patients to make appointments or access information on test results and clinical progress are not regarded as patient involvement in decision making. In addition, the context of clinical decision making has changed in the way clinical teams are now conceptualized as decision-making units. Clinical decisions are increasingly made by health professionals working in collaboration and transcend specialties, professions, departments, and even organizations. A CDSS for this integrated clinical decision process requires a reconfiguration of organizational design, management strategy, finance, and policymaking. All components must be aligned and inextricably linked, drawing on engineering and social sciences that transcend but are essential for evidence-based medicine. An Institute of Medicine study has found considerable variation in clinical decision-making behavior and outcomes, including how scientific evidence informs the decision process (Kaplan and Frosch 2005). Examining what constitutes knowledge and how it is acquired, transmitted, and applied in decision making is an age-old philosophical pursuit. A growing body of research is available to inform the development of effective CDSSs, but more needs to be done to test a range of clinical decisions and their context. There is considerable evidence to inform individual clinical decisions but less to inform the structure of the decision process. What is lacking is an exploration of the different types of decision processes, tested within different clinical structures and using different combinations of professionals. Researchers have not adequately addressed the fundamental transformation of the health system enabled by the logic and power of health systems informatics. In this regard, it is not an overstatement to say we are still in the process of automating obsolescence. A

C h a p te r 3:   H e a l th Professions, Patients, and D ec isions

common response by clinicians and health leaders is that “it works better than what we had before.” This is not a high bar for the science of decision making.

Types of Decision Context Research and practice add to our understanding of different types of decision context exhibiting different decision support requirements. Clinical decision support tools must be tailored to the type of decision context, resulting in better science, less variation, and improved outcomes. There will continue to be variation, requiring professional judgment, but it can be reduced where appropriate so that quality is improved. These differences can serve as a basis for developing more effective CDSSs, which will be better received by clinicians because, even though they result in change, they are based on the logic of clinical decision processes and not the structure of organizations or IT systems. The logic of IT systems and organizational design must serve the transformed clinical process. CDSSs can stimulate and guide changes in the structure of clinical decision processes based on evidence and the type of decision. Team Decision Context Clinical knowledge is structured around clinical specialties, resulting in “domain-specific guidelines, checklists and protocols to improve decision making” (Weinberger et al. 2015), which are essentially silos of decision making. Domain-specific CDSSs are considered inadequate for responding to most clinical cases, particularly complex cases. Improving guidelines using traditional logic will not unleash the potential of CDSSs because the logic, not the science, is what needs to be changed. The design of domain-specific guidelines assumes the logic of cognitive informatics, whereas health systems informatics guides the decision process and thus the logic. More and more clinical decisions are made by teams, an important development that promotes the redesign of the traditional, individual approach into a systems model. Team CDSSs, particularly in specialized areas such as the intensive care unit (ICU), have been analyzed. One study found that traditional CDSSs in ICUs have caused uncertainty, vagueness, and inconsistent standards of care, threatening patient safety and quality of care (Hawryluck, Bouali, and Meth 2011). Such findings do not suggest that decision support recommendations were not followed but rather that they were not structured to manage complex cases by teams of caregivers. CDSSs designed and tested for multiprofessional decision making were found to produce better clinical decisions, better coordination of care, and improved outcomes (Weinberger et al. 2015). The complexity of clinical cases in hospitals and the greater focus on the patient and wellness, which transcends institutions, suggest that CDSSs of the future must enable teams of health professionals to engage in shared decision making.

61

62

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Shared access to an electronic medical record (EMR) system is not sufficient as team decision support. Decision support tools that are designed around teams might first appear to be a relatively simple information exchange but in reality are built on evidence-based, shared decision-making models. The entire clinical milieu provides the context in which the decision-making process takes place. Actor–network theory includes the entire clinical setting in developing clinical guidelines and protocols, including tacit knowledge generated through group interaction (McDougall et al. 2016; McMurtry, Rohse, and Kilgore 2016). This theory posits that, instead of a defined team leader and structure, the combination of professionals (actors) is the power behind superior outcomes. Team composition and interaction, culture, professional language, decision support, and the care setting all affect decision making and contribute to improved outcomes. Clinical teams themselves need to develop guidelines based on the best available evidence and the team’s collective knowledge and experience. Teamdesigned guidelines can improve the decision process by using the best science applied to high-performing teams. Educational institutions are paying more attention to team-based competencies in training health professionals (Rajamani et al. 2015), introducing interventions such as collaborative curriculum, team competency practicum, collaborative team environment, and team-based CDSS development and testing. Evidence indicates that team CDSSs are effective in a range of clinical areas, but more research is needed to guide the design, testing, and use of team decision models and CDSSs. Such change will require new organizational designs. Healthcare organizations have not traditionally been oriented toward clinical teams but instead are characterized by functional hierarchies, personnel functions based on individual jobs and rewards, and dysfunctional financial incentives (Sullivan et al. 2016). The singular focus on improving CDSSs based on individual decision making is suboptimal and lacks the essential systems perspective. Patient Involvement Context IT enables caregiving teams to redefine the concept of patient-centered care, including patients’ communication with providers, online access to a breadth of information, content communities, and wearable sensors and self-diagnostics. More sophisticated systems allow patients to track their medical history and monitor their health. For example, the University of Missouri Health Care (2018) has developed a smartphone application that lets patients with depression enter their moods and symptoms into a log and share data with their psychiatrists. The Internet has opened a new level of patient interaction and access to clinical information, and greater change will occur in the decision process, depending on the demographics and cultural experience of patients

C h a p te r 3:   H e a l th Professions, Patients, and D ec isions

and the involvement of clinical teams. The quality of electronic information is enhanced by health portals where both human and nonhuman actors are important (Foroutani, Iahad, and Rahman 2013). The interaction of teams involved in the decision process with patients draws on actor–network theory as well. Changes in the clinical decision process go beyond the clinical encounter to include the entire clinical continuum, which requires transformation of the health system itself. The phrase “the patient will see you now” reflects the likely direction of the health system of the future (Topol 2012). The question is how patient involvement will change, enabled by the type and use of CDSSs. For many years, clinicians and hospitals have marketed themselves as patient centered. However, only in the past few years have most hospitals and clinics enabled patients to access the electronic health record, allowing them to make or change appointments and review certain clinical information, such as laboratory results and ranges as well as medication lists and dosages. This level of access and involvement raises issues of privacy and security of private and sensitive information (chapter 16). Giving patients access to test outcomes and other medical data has resulted in a more informed patient but has not substantially altered the decision process or used more than a fraction of the power of IT. One could argue that this type of access is not really patient centered. Giving patients access to the electronic health record system is based on the traditional model of the doctor–patient relationship, where the patient plays a passive role. Health systems informatics can transform how patients access and use health information and how they collaborate as co-producers with health professionals in the decision process (Butcher 2013). Patient participation in decision making introduces a complex set of dynamics, including the patient’s and family’s values and beliefs. As discussed earlier, values-based decision making is known as affective decision making. Affective decisions by patients do not negate the essential nature of evidence-based decision support tools, which patients also access, but patient values might override the science in some instances. This situation is frequently experienced in palliative care, where probabilistic evidence is available and accessible but ultimately defers to personal feelings, emotions, and values. Information systems that support patient participation provide access to not only evidence-based solutions but also relevant information posted on social media or websites through the patient health record. Knowledgeable and involved consumers are active online, creating multimedia content, coordinating support groups, running chat sessions, and maintaining disease-specific or health-related blogs. Providing access to informed consumer support alters the traditional clinician–patient relationship, enabling patients to ask questions, compare treatments, learn, and share advice or experiences (chapter 8) from and with people other than their clinicians. The means by which patients gain information itself generates controversy. For example, defenders of television

63

64

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

advertising of prescription drugs—a $5 billion industry allowed in only two countries in the world (the United States and New Zealand)—argue that these commercials provide patients with information about illnesses and their effective treatment. However, studies consistently report that most of these advertisements are misleading or false, although the ads may be of benefit to some patients (Faerber and Kreling 2014). One could argue that the pharmaceutical industry is not in the business of patient education but financial return and, along with insurance companies that reimburse for these medications, likely contributes to the fact that the United States has the highest-cost health system in the world. At a minimum, prescription drug advertising is not an effective means of communicating information to patients. The danger in planning for patient participation in the clinical decision process is making generalizations about populations and their beliefs, practices, and values. There is a danger that the health system will develop rigid guidelines for affective decisions as if they were mechanistic, factual ones. For example, it is easy to assume that all patients desire to be fully informed and to have an active role in considering options and selecting a course of treatment. However, many patients prefer their physicians to make the clinical decisions in certain situations. On the other hand, some patient populations do share certain values and beliefs even though each individual in those populations is unique. Older patients and those with severe illnesses, for example, have been found to place greater relative value on clinical judgment than on clinical evidence (Mira et al. 2014). This is not to say that these patients should not be informed about medication interactions, possible complications, and risks as part of the decision process. In addition, how patients and families are informed is important. Research indicates that patients involved in an evidence-based process have more confidence in and are more satisfied with the final decision. The ability to recognize values-related issues when they arise in practice is an essential competence for clinicians (Scherer et al. 2015). Urgent Decision Context Urgent situations that characterize many decisions are not unique to healthcare, but the urgency in healthcare is different from that in other industries. Judgment-based decisions might be the main or initial source of knowledge in certain decision contexts, such as emergency care. When the urgency of the situation allows only limited clinical decision support—because time is lacking or the case or setting is complex—professional judgment by the clinician is required to make a rapid assessment and arrive at a course of action as soon as possible. For emergency care, whether in an ambulance or the ED, the appropriateness of decisions might depend in part on the speed at which assessment and treatment decisions are made, possibly limiting the use of CDSSs

C h a p te r 3:   H e a l th Professions, Patients, and D ec isions

(Hall 2002). Training in simulated settings, along with experience, sharpens clinical judgment, an essential knowledge base, and is part of the role of the health profession. Even in emergency care, patient-monitoring data should be quickly integrated with the patient record and appropriate decision support tools. The ability to accomplish this electronically in an emergency environment demonstrates the obsolescence of paper records or institution-specific EMRs. What role might CDSSs play in urgent care? There are models for how IT can be engaged immediately to collect, monitor, and report vital information for urgent patients, but there is limited assessment of how quickly and effectively such information can be analyzed and presented to improve clinical decision making. The use of CDSSs has considerable potential given the advanced communication systems in ambulances, local EDs, and tertiary centers. Staff should be able to immediately access an electronic patient record, transcending institutional and regional delivery systems, providing important patient information, and further informing urgent care decisions. The limited ability to rapidly access patient information across disparate organizations and systems highlights a system failure perpetuated by the faulty design of EMRs and CDSSs. The timeliness and progression of integrated information and knowledge are essential for decisions to be knowledge based. Decision models in the emergency or urgent care context have the potential to outperform individual or group decisions. In an inpatient setting, these models can be employed for repetitive and complex decisions. Decisions to call a rapid-response team in hospitals, for example, are based on multiple and interrelated indicators of change in a patient’s condition and typically are made by experienced nurses and hospitalists. Learning occurs as experience increases. Repetitive events, however, generate considerable data on multiple and complex clinical factors exhibited by a patient. Research has found that nurses—using a computational decision model (chapter 5), which collects vital information and calculates complex interrelationships using large databases— activated the rapid-response team earlier and with fewer false positives than was possible with judgment-based or intuitive decisions. The results were fewer deaths and greater efficiency (Parker 2014). Such decision models can rapidly gather and process immense amounts of data, sometimes looking for obscure pattern relationships, and provide timely alerts. The volume of accessible data becomes a friend in urgent care and the basis for improving clinical decisions. The ability to access a large volume of data, to rapidly process it, and to generate and transmit knowledge-based guidelines, even in remote settings, will characterize urgent care in the future. The system will require integrated EHRs and HIE systems supported by machine learning to guide urgent care decision processes.

65

66

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Conclusion Much of the complexity of health systems management is in structuring a system that allows authorized clinicians universal access to individual patient records, presents the latest and best available evidence to inform clinical decision making, and gives individuals the freedom and flexibility (when appropriate) to make informed clinical decisions. The logic of such a system defines how hospitals and clinics and their information systems are structured and managed and what their strategic vision will be. Healthcare services finance will be aligned to reward quality and efficiency, not the number of patients seen or procedures performed. Bundled payment is a start, but it is still procedure oriented and just one step on a long journey. Health systems informatics can be the basis for evaluating health system design and for establishing institutional and national policies. Policy decisions will be informed by collected evidence, available resources, and values. The logic of IT will define the structure and function of clinical services, hospitals and clinics, regional systems, financing, and health policies. Healthcare has paid little attention to the study of organizational and system design, financing, and policymaking based on the assumptions of a clinical function enabled by advanced IT. The clinical process will be fundamentally altered when IT is used as an organizing principle and is tailored to each clinical encounter. The concept of mass customization is used in marketing to denote a high level of flexibility and personalization of custom-made products that are produced in a large, highvolume system that is knowledge driven. The health system needs to extend this concept by drawing on collected knowledge to tailor services to individuals. Health systems informatics provides the architecture for the fundamental redesign of the clinical decision-making process, as well as of organizational structure and strategy, financing, and policymaking. Desired change can occur only with the leadership of health professionals, educators, state licensing boards, financing agencies, policymakers, and executives throughout the US health system. Without such transformational change, we will continue to use IT to automate obsolescence.

Chapter Discussion Questions 1. How has the role of health professionals changed but not been diminished in an information-driven health system? 2. How does the assumption of evidence-based clinical decision making change the role of the health professions? 3. Give examples of the difference between a clinical decision and the decision process.

C h a p te r 3:   H e a l th Professions, Patients, and D ec isions

4. Should intuitive decisions include evidence from a clinical trial? 5. Make a case for and against public advertising of prescription drugs as a patient information strategy.

Case Study  Redesigning Futures: The First-Ever Engineering-Driven College of Medicine Phyllis M. Wise One of society’s greatest challenges today is to improve healthcare for more people at lower cost. To meet this challenge, physicians cannot continue to be trained in the same way they have been trained in the past. Instead, biomedical scientists, engineers, physical scientists, innovators, and physicians—nontraditional partners—must come together to develop dramatically different models of medical education. This collaboration will lead to the modernization of medical education and to higher-quality education, which in turn will lead to better healthcare. During the 1980s and 1990s, few new medical schools were established in the United States. In 2006, growing concern over physician shortages led the American Medical Association to issue a statement calling for a 30 percent increase in medical school enrollment. Since then, more than 20 new medical schools have been founded. These new medical schools have addressed the need for more physicians, but none of them has considered the proposition that a different kind of physician will be required to solve the challenges of our global healthcare needs. The University of Illinois at Urbana–Champaign, in partnership with the Carle Health System, has addressed that set of needs by instituting a distinctive, groundbreaking, engineering-driven college of medicine that redesigns the education of doctors and thus the delivery of healthcare. The new college brings together seemingly different disciplines and partners in ways that have not been possible in a traditional medical school environment. The bold and innovative vision of the new Carle-Illinois College of Medicine is to establish the first research-intensive college of medicine to focus expressly on the convergence of engineering, technology, Big Data, and healthcare. Because this was not a retrofit or add-on to an existing enterprise, it integrated these elements from the beginning. Both the university and the system were at critical junctures in their strategic plans that made the creation of a new college appropriate at this time in their respective histories. The university is a member of the (continued)

67

68

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

prestigious Association of American Universities, a society of public and private research universities. One-third of the faculty in the university’s internationally respected College of Engineering were already involved in biomedical and systems engineering research. Many bioengineers were not actually in the department of bioengineering but spread out in different departments in different colleges across the campus. The system—comprising the Carle Foundation Hospital, Carle Physician Group, and Health Alliance Medical Plan—is a vertically integrated healthcare delivery system. It is in the middle of a major recruitment campaign, hiring 150 physicians and physician-scientists, many of whom are involved in the Carle-Illinois College of Medicine. The hospital is ranked in the top 10 percent of hospitals in Illinois. In 2016, its combined outpatient and inpatient visits ranked among the highest volumes in the state. The university and the system are the largest and second-largest employers, respectively, in the micro-urban town of Urbana–Champaign. The establishment of Carle-Illinois College of Medicine was the result of a comprehensive and collaborative process. It was a thorough, extensive, consultative, and thoughtful planning process that involved hundreds of university faculty members, system physicians, and external consultants as well as many community leaders in Urbana and Champaign. The first goal is to transform the education of physicians by providing graduates with a distinctive degree in medicine. An inaugural class has been admitted for fall 2018. The university and the system had the opportunity to design a new paradigm for medical education, research, and the delivery of care from the ground up. Faculty and physicians designed a totally new medical curriculum. Each course was developed and will be taught by a team of faculty that includes basic scientists, physicians, and engineers. The new curriculum infuses principles of engineering, technology, Big Data, and innovation into all courses as a road map to the future of medicine. The curriculum emphasizes analytical thinking and problem solving and will include clinical immersion in all four years. This is being done, in part, through interdisciplinary, team-based, innovative approaches to achieving improved healthcare outcomes. One of the most important aspects of the curriculum is that students will be introduced to the clinical environment from the beginning. Principles of engineering will be introduced into all cases and the treatment of disease. The future physician-discoverers, physician-innovators, and physician-engineers who will train in the college of medicine will be encouraged to cocreate new devices, medications, technologies, and more with faculty mentors while they are students. Thus, this collaboration will create a new culture in medical education and practice.

C h a p te r 3:   H e a l th Professions, Patients, and D ec isions

The future professionals will advance care delivery far beyond what can currently be imagined. Not only will their patients benefit, but the impact they make will also extend to countless others who will be healthier as a result of their medical discoveries. The second goal is to redesign healthcare, taking full advantage of current and future discoveries in engineering, technology, and Big Data. Advances in these areas are driving medical breakthroughs. The future of precision medicine lies in discovering new sensors, materials, and devices as well as new uses for robotics, miniaturized imaging, Big Data, and remote monitoring and then applying these innovations to medicine and healthcare. The complex system for delivering healthcare services will draw on medical systems engineering. Graduates will advance patient care and develop dramatic technologies to deliver higher-quality healthcare to more people at lower cost.

Case Study Discussion Questions 1. Research the Flexner Report. How did this report transform medical school in the early 1900s, and how did existing medical schools respond? How did the change in curriculum requirements affect the profession of medicine? 2. Why did the American Medical Association not call for a new type of medical school and curriculum in 2006, instead of just a 20 percent increase in numbers? 3. Discuss the conflicts that might arise if a change in curriculum design were proposed in an existing and well-established medical school. 4. Consider a range of existing medical schools and assess their relative strategic positions if engineering is added as a new translational science. 5. What arguments can be made for preparing physicians with competencies in engineering? Against? 6. The curriculum at Carle-Illinois College of Medicine is designed around “interdisciplinary, team-based, innovative approaches to achieving improved healthcare outcomes.” How are nurses and other health professionals engaged in this new model?

Additional Resources Hogarth, R. 2010. “Intuition: A Challenge for Psychological Research on Decision Making.” Psychological Inquiry 21 (4): 338–53.

69

70

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Mattick, J. S., M. A. Dziadek, B. N. Terrill, W. Kaplan, A. D. Spigelman, F. G. Bowling, and M. E. Dinger. 2014. “The Impact of Genomics on the Future of Medicine and Health.” Medical Journal of Australia 201 (1): 17–21. Mickleborough, T. 2015. “Intuition in Medical Practice: A Reflection on Donald Schön’s Reflective Practitioner.” Medical Teacher 37 (10): 889–91.

References Alonso-Coello, P., A. Irfan, I. Solà, I. Gich, M. Delgado-Noguera, and D. Rigau. 2010. “The Quality of Clinical Practice Guidelines over the Last Two Decades: A Systematic Review of Guideline Appraisal Studies.” Quality & Safety in Health Care 19 (6): 1–7. American Nurses Association. 2012. “Florence Nightingale Pledge.” Accessed September 20, 2017. www.nursingworld.org/FunctionalMenuCategories/AboutANA/ WhereWeComeFrom/FlorenceNightingalePledge.aspx. Bates, D. W., G. J. Kuperman, and S. Wang. 2003. “Ten Commandments for Effective Clinical Decision Support: Making the Practice of Evidence-Based Medicine a Reality.” Journal of the American Medical Informatics Association 10 (6): 523–30. Brainard, J., and P. R. Hunter. 2016. “Do Complexity-Informed Health Interventions Work? A Scoping Review.” Implementation Science 20 (11): 1–11. Brenner, S. K., R. Kaushal, Z. Grinspan, C. Joyce, K. Inho, R. J. Allard, D. Delgado, E. L. Abramson, and I. Kim. 2016. “Effects of Health Information Technology on Patient Outcomes: A Systematic Review.” Journal of the American Medical Informatics Association 23 (5): 1016–36. Butcher, L. 2013. “Shared Decision-Making: Giving the Patient a Say. No, Really.” Hospitals & Health Networks 87 (5): 26–31. Faerber, A. E., and D. H. Kreling. 2014. “Content Analysis of False and Misleading Claims in Television Advertising for Prescription and Nonprescription Drugs.” Journal of General Internal Medicine 29 (1): 110–18. Foroutani, S., N. A. Iahad, and A. A. Rahman. 2013. “An Initial Framework for Interactive Health Portals: Using Actor Network Theory.” In Research and Innovation in Information Systems (ICRIIS): 2013 International Conference on Research and Innovation in Information Systems, 475–80. Piscataway, NJ: IEEE. Friedson, E. 2001. The Third Logic. Chicago: University of Chicago Press. ———. 1994. Professionalism Reborn: Theory, Prophecy, and Policy. Chicago: University of Chicago Press. Hafferty, W., and D. Levinson. 2008. “Moving Beyond Nostalgia and Motives: Towards a Complexity Science View of Medical Professionalism.” Perspectives in Biology and Medicine 51 (4): 599–615. Hall, K. H. 2002. “Reviewing Intuitive Decision-Making and Uncertainty: The Implications for Medical Education.” Medical Education (36): 216–24.

C h a p te r 3:   H e a l th Professions, Patients, and D ec isions

Harteis, C., B. Morgenthaler, C. Kugler, K.-P. Ittner, G. Roth, and B. Graf. 2012. “Professional Competence and Intuitive Decision Making: A Simulation Study in the Domain of Emergency Medicine.” Vocations and Learning 5 (2): 119–36. Hawryluck, L., R. Bouali, and N. D. Meth. 2011. “Multi-professional Recommendations for Access and Utilization of Critical Care Services: Towards Consistency in Practice and Ethical Decision-Making Processes.” Journal of Law, Medicine & Ethics 39 (2): 254–62. Hoomans, T., A. Ament, S. Evers, and J. L. Severens. 2011. “Implementing Guidelines into Clinical Practice: What Is the Value?” Journal of Evaluation in Clinical Practice 17 (4): 606–14. Institute for Technology Assessment. 2018. “What Is Decision Science?” Accessed May 9. www.mgh-ita.org/Frequently-Asked-Questions/decision-science.html. Kaplan, R. M., and D. L. Frosch. 2005. “Decision Making in Medicine and Health Care.” Annual Review of Clinical Psychology 1: 252–56. Khayal, I. S., and A. M. Farid. 2017. “The Need for Systems Tools in the Practice of Clinical Medicine Systems Engineering.” Systems Engineering 20 (1): 3–20. Kotay, A., J. L. Huang, W. B. Jordan, and E. Korin. 2016. “Exploring Family and Social Context Through the Electronic Health Record: Physicians’ Experiences.” Families, Systems & Health 34 (2): 92–103. Latoszek-Berendsen, A., H. Tange, H. J. van den Herik, and A. Hasman. 2010. “From Clinical Practice Guidelines to Computer-Interpretable Guidelines.” Methods of Information in Medicine 49 (6): 550–70. Mayo Clinic. 2017. “Medical Drones Poised to Take Off.” Accessed December 7. www.mayoclinic.org/medical-professionals/clinical-updates/trauma/ medical-drones-poised-to-take-off. McDougall, A., M. Goldszmidt, E. A. Kinsella, S. Smith, and L. Lingard. 2016. “Collaboration and Entanglement: An Actor-Network Theory Analysis of TeamBased Intra-professional Care for Patients with Advanced Heart Failure.” Social Science & Medicine 164: 108–18. McGuinness, K. M. 2014. “Institutional Decision Making: Empowering of Health System and Economic Transformation.” American Psychologist 69 (8): 817–27. McMurtry, A., S. Rohse, and K. N. Kilgore. 2016. “Socio-material Perspectives on Inter-professional Team and Collaborative Learning.” Medical Education 50 (2): 169–80. Mira, J. J., M. Guilabert, V. Pérez-Jover, and S. Lorenzo. 2014. “Health Barriers for an Effective Communication Around Clinical Decision Making: An Analysis of the Gaps Between Doctors’ and Patients’ Point of View.” Expectations 17 (6): 826–39. Muzio, D., and I. Kirkpatrick. 2011. “Professions and Organizations: A Conceptual Framework.” Current Sociology 59 (4): 389–405. National Academies of Sciences, Engineering, and Medicine. 2017. Integrating the Patient and Caregiver Voice into Serious Illness Care: Proceedings of a Workshop. Washington, DC: National Academies Press.

71

72

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Parker, C. C. 2014. “Decision-Making Models Used by Medical-Surgical Nurses to Activate Rapid Response Teams.” MEDSURG Nursing 23 (3): 159–64. Patel, V. L., and T. G. Kannampallil. 2015. “Cognitive Informatics in Biomedicine and Healthcare.” Journal of Biomedical Informatics 53: 3–14. Rajamani, S., B. L. Westra, K. A. Monsen, M. LaVenture, and L. C. Gatewood. 2015. “Partnership to Promote Interprofessional Education and Practice for Population and Public Health Informatics: A Case Study.” Journal of Interprofessional Care 29 (6): 555–61. Rolland, J. S., L. L. Emanuel, and A. M. Torke. 2017. “Applying a Family Systems Lens to Proxy Decision Making in Clinical Practice and Research.” Families, Systems & Health 35 (1): 7–18. Ruhe, G., and Y. Wang. 2007. “The Cognitive Process of Decision Making.” International Journal of Cognitive Informatics and Natural Intelligence 1 (2): 73–85. Scherer, L. D., M. de V. Tilburg, B. J. Zikmund, H. O. Witteman, and A. Fagerlin. 2015. “Trust in Deliberation: The Consequences of Deliberative Decision Strategies for Medical Decisions.” Health Psychology 34 (11): 1090–99. Sullivan, E. E., Z. Ibrahim, A. L. Ellner, and J. L Giesen. 2016. “Management Lessons for High-Functioning Primary Care Teams.” Journal of Healthcare Management 61 (6): 449–67. Topol, E. 2012. The Creative Destruction of Medicine: How the Digital Revolution Will Create Better Health Care. New York: Basic Books. Tricoci, P., J. M. Allen, J. M. Kramer, R. M. Califf, and S. C. Smith Jr. 2009. “Scientific Evidence Underlying the ACC/AHA Clinical Practice Guidelines.” Journal of the American Medical Association 301 (8): 831–41. Tyson, P. 2001. “Hippocratic Oath Today.” NOVA. Posted March 27. www.pbs.org/ wgbh/nova/body/hippocratic-oath-today.html. University of Missouri Health Care. 2018. “App Aids Patients with Depression.” Accessed May 9. www.muhealth.org/our-stories/app-aids-patients-depression. Weinberger, H., J. Cohen, B. Tadmor, and P. Singer. 2015. “Towards a Framework for Untangling Complexity: The Interprofessional Decision-Making Model for the Complex Patient.” Journal of the Association for Information Science and Technology 66 (2): 392–407.

CHAPTER

THE COMING OF THE CORPORATION: TRANSFORMING CLINICAL WORK PROCESSES

4

Gordon D. Brown

Learning Objectives After reading this chapter, you should be able to do the following: • Assess how the traditional structure of healthcare organizations defines the design of the information system and supports the traditional clinical function. • Apply the concepts of standardization of work to clinical decision making. • Conceptualize the design of a health system that achieves optimal clinical outcomes tailored to individual patients.

Key Concepts • Traditional, hierarchical corporate structures of healthcare organizations and clinical processes • Mintzberg’s model of coordinating work processes • Standardization of clinical work processes and outcomes • Medical homes, accountable care organizations, and communities of practice

Introduction “The Coming of the Corporation” is the title of the last chapter in Paul Starr’s Pulitzer Prize–winning book, The Social Transformation of American Medicine. It explores the increasing size and power of organizations in the US health system. Starr (1984) concludes that the growth in corporate presence resulted 73

74

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

from the creation of larger hospitals; the advent of integrated health systems; the market influence of suppliers, including pharmaceutical companies; and the likelihood that more and more physicians would move from independent to employed status. An even more profound change is occurring today: Organizations have greater responsibility for evidence-based decision support, making them more directly accountable for clinical outcomes and thus clinical processes (Muzio and Kirkpatrick 2011). This change has little to do with the employment status of physicians because it focuses on the structure of clinical processes and outcomes. Will organizations be regional or local systems? Will they be financed by public or private sources? Will they be held accountable by governmental regulations? There is no right or wrong answer because we are asking the wrong question. The answer may be “all of the above” if the organizations assume a systems perspective and are structured around the logic of information technology (IT) (Es-Sajjade and Wilkins 2017; Kornberger 2017; Lewis 2016). The structure of healthcare organizations should be nimble, highly efficient, tailored to communities and patients (with compatible information systems to enable information exchange), and evidence based. Once the parameters of services are defined, the structure of the system can be crafted to enable their delivery. Although health systems and commercial systems share some similarities in automating and integrating complex processes, healthcare and other social services have a different value set. Unlike most consumption goods and services, which are driven by the market, healthcare services are social goods, creating both private and public benefit, that are financed and provided through the private, nonprofit, and public sectors. Healthcare services do not cleanly fit the assumptions of either the private or the public sector, but each of these sectors brings values that are inherent in system design. The values, flexibility, and dynamics of the private and public sectors fit well with the assumptions of the third sector—the nonprofit, or plural, sector (Mintzberg 2015). At the consumer level, healthcare services have been episodic, fragmented, costly, and not sufficiently tailored to individuals’ cultural, religious, economic, and personal values. At the organizational level, services are provided and financed through multiple independent local and regional institutions. Private insurance companies primarily maintain high-cost, volume-based plans, which are fairly monopolistic and increasingly publicly funded. Pharmaceuticals are also provided through large, monopolistic private corporations. Social services are provided through independent local and state agencies, and states control licensure laws. The private, public, and plural sectors have crafted a flexible and dynamic but complex, uncoordinated, and costly system that is supported by a fragmented information system. At the policy level, most countries require that health and social services be available to all and either provided directly or financed by the government.

Ch a p t er 4 :  T h e C o mi n g o f th e C o r p o ra ti o n: Transfor m ing C linic al Wor k Proc esses

There is always a trade-off in national health systems between low costs and the availability of nonurgent and elective services. In the United States, the policy question is how a decentralized and disjointed private–public system that increasingly uses public funds can provide integrated, high-quality, and accountable healthcare services. The choice is not just between public and private systems, centralized and decentralized structures, or corporate and clinical values because policy must include elements of all of the above. The architecture for such a health system must be built around an integrated information system that is knowledge based, enables coordination of services, is tailored to individual patients, holds a range of disparate players accountable, and is self-organizing. The traditional policy direction in many countries has been to integrate services by forming large, centralized, bureaucratic, and largely public systems that define how the services are structured and consumed. This model is not considered workable in the United States, but our health system might be compelled to move in that direction if it cannot fix itself. The US health system need not pursue a national health system strategy or a highly regulated private sector, but it can pursue a system that is based on a systems logic with IT as the system architecture. The Health Information Technology for Economic and Clinical Health (HITECH) Act of 2009 mandated that healthcare institutions purchase computerized information systems and demonstrate their meaningful use. Meaningful use refers to the level at which healthcare organizations acquire and use IT for clinical decision making. Healthcare is one of the few sectors in which federal mandates and direct financial incentives are necessary to motivate organizations to invest in basic IT. Notably, the HITECH mandate is not necessarily to use computers in decision support or to transform the clinical process, but merely to buy and demonstrate meaningful use of IT—a very low bar indeed. Although healthcare organizations have historically undervalued IT and failed to make it a priority as a matter of policy, this chapter examines the transformative power of IT in healthcare. (See chapter 15 for a discussion of how healthcare financing has affected the clinical process.)

Traditional Corporate Structures as the Logic for Clinical Information Systems The internal design of and interorganizational relationships among hospitals, nursing homes, and clinics are based on two logical premises. First, the clinical structure was the dominant function based on the historical role of the professions. Second, the official structure was based on a corporate or management function. This presented a unique relationship, with the clinical function being external to the organization but the dominant operational and strategic logic.

75

76

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

As corporations, hospitals traditionally did not assume direct responsibility for clinical processes or outcomes. It was only in 1965, in Darling v. Charleston Community Memorial Hospital, that hospitals became legally liable for clinical errors in their facilities, and even then their legal responsibility was limited to ensuring their staff had the appropriate credentials and supervision and were receiving the required hours of continuing education—areas of corporate responsibility that focused on factor inputs, not clinical processes or outcomes. The hospital avoided interfering with physicians’ clinical decision making; its primary role was to support the clinical function by acquiring capital and overseeing administrative functions such as supply management, personnel, billing and collections, marketing, and legal services. The medical function traditionally bypassed the administrative function, with the medical director reporting to a joint conference committee comprising the CEO and members of the board of directors. Most physicians were not officially members of the corporation and were accountable only for the medical services they provided (compare the Ross-Loos medical group discussion in chapter 1). The radiology, laboratory, and pharmacy departments were traditionally under dual authority—they functioned somewhat autonomously and answered to clinical directors but were located under the administrative structure, which was overseen by a senior administrator (exhibit 4.1). Even today, the core medical and nursing functions are accountable to the chief medical officer and chief nursing officer, respectively. These traditional, autonomous, and hierarchical administrative structures are incompatible with the function of health professionals. Implementing an electronic medical record (EMR) presents technical challenges, such as standardization of vocabularies, but more importantly, it transforms the orientation of the information system structure—and in turn, the structure EXHIBIT 4.1 Traditional Functional Design

Medical director Medicine Surgery Obstetrics Radiology Pediatrics Pathology Anesthesia

Board of trustees

Joint conference committee

Medical staff

Finance

Senior administrator

Administrator, hospital services

Human resources

Purchasing

Administrator, clinical services

Food services

Maintenance

Nursing

Admissions Housekeeping

Laboratory Pharmacy

Medical records

Radiology

Ch a p t er 4 :  T h e C o mi n g o f th e C o r p o ra ti o n: Transfor m ing C linic al Wor k Proc esses

of the clinical function—from individual clinicians to the institution. It must be emphasized that the design and function of the clinical information system are not structured under the traditional administrative function but brought the clinical function more closely into the organization. These transformations are complex in that they required, to some degree, the transformation of the once independent clinical function. The traditional corporate structure of hospitals has been hierarchical, with departments led by managers or directors who report to higher-level administrators. This organizational design—known as a vertical structure— has the advantage of providing stability and control and enabling the efficient use of specialized equipment and staff across departments to support rapidly expanding medical technology. In this structure, information, communications, decisions, resources, and plans flow vertically among governing board, senior managers, middle managers, and frontline workers. The business function is the structure’s primary logic and includes the administrative departments of finance, human resources, public relations, housekeeping, food service, and maintenance. Thus, the business function has historically been parallel to the clinical function and has exercised considerable influence because of the importance of the finance, human resources, and IT departments. The business function has been characterized as follows: 1. Most hospital and health system CEOs and chief operating officers (COOs) in the United States have a business—not clinical—education and orientation. 2. The principal responsibility of healthcare governing boards is to oversee finances, prevent and manage legal risk, and develop a marketing strategy (usually growth). 3. Accreditation agencies in the industry historically based performance on input measures (e.g., the percentage of medical staff who are board certified) rather than outcomes (e.g., recovery rates). As such, they tended to inhibit innovation. The business orientation of the management and governance functions draws on universal standards for accounting and finance while supporting the clinical function. Hierarchical structures offer advantages but are poorly suited for innovation, transformational change, coordination of highly integrated clinical work processes, and empowerment of frontline workers. A hierarchical mind-set tends to be strongly embedded in the organizational structure and culture and is therefore difficult to change. It is the antithesis of the clinical structure and function. Although hierarchical functional

77

78

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

structures still dominate healthcare, they are ineffective in information-driven environments and are changing (Tseng 2011). Traditional, autonomous, and hierarchical structures also served as the logic for the early design of clinical information systems in hospitals. Clinical information systems were frequently incompatible across departments because each department had the freedom either to purchase its preferred system or to develop its own unique one. Each of these systems included specialized clinical service features. This logic heavily influenced (and continues to influence) the design and application of medical and nursing informatics. The power of IT to scan large amounts of data, identify patterns and relationships, and report information and knowledge across time, functions, and space was negated by the structure of the organization and thus the IT system. In short, information system design was subordinated to traditional clinical and administrative functions and was not transformative. Health systems informatics enables the transformation of the clinical function, but to do so it must change the traditional structures of clinical decision making, healthcare organizations, and information systems (Mintzberg 2017). The values, concepts, and traditions of organizations are deeply embedded in existing structures, which are not easily changed. When change does occur, it is often reactive and incremental. This chapter explores the process of transforming complex and profound values, traditions, and structures (Morgan 2010). Such transformation not only poses technical challenges in conceptualizing a new system but also challenges healthcare leaders to envision and skillfully transform how their organizations are currently structured, financed, staffed, and evaluated. This transformation will be incremental because of the complexity of the process, so the vision should be oriented toward the long term.

Standardization of Clinical Work Processes Clinical decision making and clinical work processes evolved outside the direct purview of healthcare organizations. They were (and still are) based primarily on the logic of episodic clinical decisions rather than on the logic of integrated, continuous care. Further, the structure of the clinical encounter and process served as the logic for medical informatics and the overall design of clinical information systems, which conflicts with management information systems (chapter 1). Clinical IT has focused on informing decisions but has transformed them only marginally. The business function continues to be based largely on the traditional logic of operational and management structures, with the clinical function the domain of the professions and the two only loosely integrated. This is not an easy fix because integrated information systems assume a high degree of standardization whereas professional work has traditionally rejected

Ch a p t er 4 :  T h e C o mi n g o f th e C o r p o ra ti o n: Transfor m ing C linic al Wor k Proc esses

standardization of work processes and measures of work outcomes. Because IT in healthcare organizations evolved within two different logic structures— that of the business function and that of the clinical function—IT became an automating technology in both the business and the clinical area, rather than an integrated and transformative technology. The standard measures of work outcomes and the design of work processes are a well-developed science in business and industry, where they serve as the basis for structuring and coordinating work. They have been broadly used in manufacturing and most service industries because of the nature of the tasks involved and the IT that informs and coordinates those tasks. In his classic model for the coordination of work, Mintzberg (2017) conceptualized five basic approaches to coordinating work: (1) direct supervision, (2) mutual adjustment, (3) standardization of skills, (4) standardization of outcomes, and (5) standardization of work processes. The principles of standardization were broadly applied in manufacturing and other service industries and served to transform industrial production and consumer markets. Standardization has been considered the nemesis of the clinical function.

Mintzberg’s Model Healthcare organizations continue to see a high degree of autonomy of the clinical function as it has been traditionally defined and structured in medicine and, to a degree, in nursing and other health professions. The coordination of the clinical process has evolved, according to Mintzberg’s model, by direct supervision, mutual adjustment, and standardization of skills—all of which are compatible with traditional professional roles. One need only walk into a healthcare facility to observe that standardization of skills is everywhere. Titles, name badges, dress codes, and uniforms visually represent standard skills according to the roles of healthcare personnel. Academic degrees, licenses, certifications, credentials, skills, and specialized training are some of the standard requirements of the health professions. Rooms for informal discussion and socialization are frequently segregated by specialty, reinforcing a culture of individual professions, hierarchy, and power. Work processes in healthcare, however, remain unstandardized. The work of health professionals, drawing on Mintzberg’s model, has been coordinated primarily through direct supervision and mutual adjustment. Both approaches deal with direct interaction or structured communication between two or more professionals who have separate tasks but are linked to the same interdependent process. This communication, which traditionally has taken place mostly face to face, allows parties to exchange explicit and tacit knowledge and thus gain a mutual understanding of how the treatment regimen should be coordinated. Organizational theory, however, recognizes the inherent limitations of direct supervision and mutual adjustment: Each is limited if the work process (patient care) is carried out by workers with dissimilar skill sets and competencies or

79

80

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

in different locations and at different times, such as across shift changes and repeat visits. Clinical process improvement is difficult because the processes are not standardized, thus limiting learning. Studies have found that handoffs over time, space, and specialty are the most vulnerable points for errors (Kohn, Corrigan, and Donaldson 2000). This breakdown is inherent in any complex work process that is incremental in structure, task oriented, and informed by fragmented decision support. The traditional structure of the clinical function has defined the design and function of clinical information systems, reimbursement of hospitals and doctors, interorganization relations, and other structural elements of the health system. This structure represents the DNA of healthcare organizations, process design, and information systems architecture.

Standardized Clinical Outcomes Professional autonomy is based on the assumption that clinical outcomes and processes cannot be standardized. The development of the EMR was a defining moment for the health system because it required the development of a standard clinical vocabulary and data set, which are inherent to electronic information systems. This standardization, a major contribution of medical and nursing informatics, was an exceedingly complex task and provided the basis for the transformation of the health system. A standard clinical vocabulary for terms and semantic relationships (chapter 12) made standardizing outcome measures possible and expanded the application of Mintzberg’s model. Standardization of outcomes, in turn, made standard measures of quality possible and comparative studies feasible. Associated with the standardization of outcomes was increased standardization of clinical decisions based on outcomes. Standardization of decision–outcome relationships inherently leads to greater exploration of clinical process–outcome relationships and thus greater standardization of clinical processes. The standardization of clinical outcomes and processes does not causally lead to mechanistic decision making but does require an understanding of the structure of clinical decision making and the demands of the information system to inform the decision process (chapter 3). In the late 1990s, the Institute of Medicine (IOM) report To Err Is Human garnered much attention by focusing directly on the quality of clinical outcomes in hospitals as measured by unexpected deaths from clinical errors (Kohn, Corrigan, and Donaldson 2000). IOM wisely chose outcome measures that were measurable and generalizable. Public discussions about the report revolved around the estimated 100,000 people dying in hospitals every year as a result of medical errors. The report framed clinical quality and safety as a public health concern—an important policy shift—and called for increased accountability for the public monies being spent on healthcare services. The federal financing of healthcare motivated the government, the public, and

Ch a p t er 4 :  T h e C o mi n g o f th e C o r p o ra ti o n: Transfor m ing C linic al Wor k Proc esses

employers to take a greater interest in improving quality and safety for everyone. A subsequent IOM (2001) report, Crossing the Quality Chasm, recognized the legitimacy of measuring and comparing clinical outcomes and called for a systems approach that focuses on the relationships among clinical quality, clinical processes, and organizational design. Since the publication of this report, the use of quality improvement techniques well known in other industries, such as root cause analysis, has proliferated. Such techniques are powerful in guiding changes in operational processes, but they have been used more for process improvement in administrative processes than in clinical processes. The IOM reports not only acknowledged the power of information systems, including decision support systems, to improve the clinical function and increase accountability but also encouraged the investment in such systems. Substantial evidence shows that healthcare organizations can achieve superior outcomes and elevate their clinical performance with the help of IT (Pinsonneault et al. 2017; Sharma et al. 2016). IOM viewed IT primarily from the perspective of medical informatics, including the EMR, computerized physician-order entry, clinical decision support systems, and health information exchange (HIE). HIE was not conceptualized as a shared decision support or knowledge-generating tool but as a means to share medical data and information across disparate systems. This view of HIE characterized the vision and level of technology at the time, but it is now evolving into a knowledge system based on large, integrated institutional databases (chapter 2). Decision support was applied primarily to the structure of existing clinical processes; although it informed and changed clinical decisions, it did not fundamentally alter the structure of clinical processes. Information-driven decision support using clinical guidelines and protocols increased the number of decisions supported by accumulated clinical evidence. IOM’s recommendation to apply IT to healthcare gave rise to the federal mandate and incentives to adopt electronic information systems in healthcare operations. Thus, the IOM reports not only generated dialogue about quality and safety but also increased the visibility and adoption of health IT. When applied to the clinical function, IT created a dialectic grounded on clinical quality and safety. The concept of clinical quality and safety is one that everyone supports in the abstract, but implementation requires disruptive change. The IOM reports also expanded the discussion on standardizing measures of clinical outcomes and processes. As mentioned, the development and linkage of clinical vocabularies enabled the development of decision support tools for gathering and reporting clinical evidence and informing clinical decision making. Decision support tools such as clinical guidelines and protocols informed the clinical decision but did not fundamentally alter the structure of the clinical process, which was still built around the traditional role of health professionals. The delay in applying informatics as a transforming science was, in part, due to the complexity of

81

82

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

developing essential clinical vocabularies and data standards. Even more costly was the failure to conceptualize the power of IT as transformative, not just automating.

Structuring Clinical Work Processes Study of evidence-based clinical outcomes necessarily leads to the development and measurement of standardized clinical processes and process–outcome relationships, and they set the stage for transforming the system. The standardization of work processes is a powerful basis for the coordination of work and the measurement of process–outcome relationships, as well as for the transformation of the health system (Mintzberg 2017). The Industrial Revolution introduced the world to mass production, characterized by highly structured tasks and work processes that ultimately gave rise to assembly lines. Standardization of work processes was the key to efficiency and effectiveness (Mintzberg 2017). Work processes and measures of quality were standardized to ensure high-quality and consistent outcomes. Unlike the work in factories, however, most clinical care cannot be completely standardized and is performed by people, not machines. In healthcare, tasks are carried out by a range of health professionals who are interdependent and interact with each other (not just with patients). The challenge is standardizing these complex human work processes while still allowing professionals and patients an appropriate amount of decision-making autonomy. The logic of jobs and work processes is based on many factors, including the nature of the work and the technology that supports it. In healthcare, the nature of the work is becoming more consumer oriented, with patients as coproducers. Patients want electronic health records (EHRs) to be linked to personal health records (PHRs), to be universally accessible, and to connect them to health and wellness information, their DNA profile, clinical evidence, and social networks. The effective application of this information will outpace the development of systems that enable this access, not because of technical limitations but because of the failure of design. The challenge is in standardizing work processes while tailoring them to individual patient conditions and clinician judgment. Process improvement is no longer confined to traditionally structured independent practitioners, departments, and caregiving units but encompasses the structure of the clinical process, transcending the organization itself. The IOM reports correctly concluded that IT could substantially improve quality and safety through decision support and clinical process improvement. However, the reports understated the potential of IT to fundamentally change the structure of the clinical process, transcending individual clinicians, institutions, and systems. The reports created a dialectic that enlisted healthcare leaders to take advantage of available evidence from clinical research, engineering,

Ch a p t er 4 :  T h e C o mi n g o f th e C o r p o ra ti o n: Transfor m ing C linic al Wor k Proc esses

management, information science, and social science to transform systems of care (Reid et al. 2005). It is incumbent on healthcare leaders to design such systems and not wait for the public sector to mandate change through regulations and controls. Creating innovative, high-performing health systems through regulations and controls is the ultimate oxymoron.

Integrated Systems Perspectives Currently, the US health system combines poorly designed clinical processes with poorly structured organizations and financing mechanisms (chapter 15). Compounding this problem is the overlay of information systems designed according to the logic of dysfunctional and disparate structures. It is little wonder that conflict, stress, poor quality, lack of continuity, and high costs still characterize the health system. The good news is that we have both the knowledge and technology to transform it and that visionary leaders are making transformational changes. Maintaining and improving old structures and practices will simply not be adequate for current and future demands. Doing the wrong thing better is not an appropriate vision for the future. Applying clinical guidelines and protocols within hierarchical functional structures increases the risk that decision support systems will be based on strict rules and controls that require standard procedures to be carried out in a culture of blame. That is the nature of hierarchical organizations, which do not fit the information age. If clinical guidelines and protocols become “institutionalized” (i.e., absorbed into operations as rigid standards of practice), contextual information might be lost. Strictly enforced standards not only undermine professional autonomy but also compromise clinical quality. Professional staff will and should resist them, either directly or indirectly. Leadership must come from health professionals who understand the transformative power of IT and who recognize that organizations need to be structured to support transformation. Organizations and information systems that follow the logic of clinical guidelines and protocols shift the structure of the clinical process from individual practitioners and departments to multiprofessional teams. Teamwork has always been regarded by health professionals as important, but more a behavioral function than the basis for work process design. In recent years, more attention has been given to teams, typically motivated and enabled by IT. Team structures in a hospital or clinic have merit but lack the full potential to transcend organizational and professional boundaries. Developing clinical teams with the logic of an integrated clinical process is an important precursor to restructuring organizations themselves (Leasure et al. 2013; Muzio et al. 2011; Wennerstrom et al. 2015).

83

84

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Team function can serve as an organizing principle and the basis for transforming clinical and organizational designs, and IT enables innovation in caregiving teams. Integrated teams of health professionals have applied clinical guidelines and protocols as guides in postsurgical and other inpatient cases and consistently demonstrated improved clinical outcomes (Buchert and Butler 2016; Macias et al. 2017; O’Malley et al. 2015; Tessier et al. 2016). These efforts are commendable given that it is difficult to use standardized work processes, such as clinical guidelines and protocols, in traditional institutions. Clinical guidelines and protocols were initially developed and used in areas such as surgery, a complex but highly procedural clinical service in a contained space. Medically complex diseases, such as type 2 diabetes, are less procedural and rely on multiple health professionals, services, specialties, and clinical pathways that one institution alone cannot normally provide. Here, standardization offers the opportunity for work coordination and evidence-based clinical decisions. Advanced IT is applied as the basis for informing and transforming clinical processes. (The term advanced does not refer to the level of technology per se but to the innovativeness with which it is applied to enable the transformation of clinical work processes.) An integrated systems perspective will inherently move from an illness and treatment model to a wellness model. That is the nature of systems thinking, which is based on improved outcomes, increased satisfaction, and reduced costs. (Some may argue that a healthy population will live longer, get complex diseases, and thus increase overall costs, but these are not considered to be negative attributes.) The structure and financing of the medical care system has historically not included wellness and prevention, except in limited cases such as capitation-based payment schemes (chapter 15). A health system that operates primarily under an illness model but strives for high quality and efficiency is the classic case of doing the wrong thing better. To be sure, transitioning to a population health model will be extremely complex because current structures, financing models, professional competencies, and government programs are based on a different fundamental assumption. (Population health is considered in depth in chapter 10. Several models designed to include population health are included in the discussion.)

Medical Homes Medical home Model of care in which the services provided by a team of health professionals are coordinated by a primary care physician and involve the patient

Patient-centered medical homes are based on the concepts of wellness, primary care, and care coordination. These settings are well suited for team structures—specifically, the work of many health professionals who function within corporate boundaries, are led by primary care physicians, and are committed to the coordination of care. Medical homes are promoted and sponsored by the American Academy of Family Physicians (AAFP), American Academy of Pediatrics, American College of Physicians, and American Osteopathic Association (AAFP et al. 2011). They have been adopted by US Public Health Service

Ch a p t er 4 :  T h e C o mi n g o f th e C o r p o ra ti o n: Transfor m ing C linic al Wor k Proc esses

community health centers and recognized by the National Center for Quality Assurance (NCQA 2018) as “a model of care that puts patients at the forefront of care.” Research shows that medical homes “improve quality, the patient experience, and staff satisfaction while reducing healthcare costs” (NCQA 2018). They also stimulate relationship building among team members and between clinicians and patients, and they address physician reimbursement issues. Because the organizing principles of both teams and patient centeredness are positive, medical homes have demonstrated an ability to provide better care coordination, greater patient satisfaction, and lower cost (Gittell et al. 2015; Shi et al. 2017). A focus on teamwork and the effective use of staff as team members are fundamental in clinical areas (Leasure et al. 2013; Solimeo, Stewart, and Rosenthal 2016) and are a first step toward system transformation. Assessments of medical homes have mostly focused on the services provided within the medical home, not those that extend beyond. Their use of IT as an enabling technology has been inconsistent, reflecting their inclination to be independent from rather than integrated with each other (Morton et al. 2015). They have followed the traditional organizational strategy of focusing on a clinical specialty area, resulting in poor integration, a lack of service continuity with other specialized care providers, and a focus on the medical model. For example, breakdowns often occur with critical care patients who need acute, nutritional, chronic, or palliative care services (Boucher, White, and Keith 2016; Kaslow et al. 2015). Primary care medical homes typically serve insured populations and have shown mixed results among underserved populations (Nwando et al. 2017). The closed-system concept characterizes the relationship between primary care medical homes and specialty providers. Many specialists form their own medical homes to coordinate care within their specialty, such as oncology medical homes (Baum 2014; Waters et al. 2015). A community could thus have a number of medical homes that have excellent internal coordination but lack coordination with other medical homes. A system of independent and specialized medical homes is not a system. Still, the medical home model has introduced important organizational concepts into a primarily clinical structure. This is progress and reflects improved integration of clinical and organizational functions. However, medical homes have adopted many approaches of traditional healthcare organizations and are not structured or conceptualized as complex, adaptive systems. The increasing desire to accredit medical homes (AAFP 2011) is a further step toward the institutionalization of clinical care. Accreditation is considered essential to the health system, but accrediting agencies also have been identified as inhibitors of innovation. Professional accreditation has many positive attributes, including assurance that minimum standards are met, but its focus on minimum standards is not innovation. The difficulty of integrating primary care with medical specialties reflects this limitation.

85

86

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

The financing of medical homes is based on historical reimbursement models and is a significant constraining factor. Most models are fee-for-service schemes supplemented by payments for nonprocedural services, such as care coordination, which are excluded from fee-for-service models. Some medical homes use care coordination financing to supplement the salaries of clinicians, who are likely not the care coordinators. (Financing is discussed in more detail in chapter 15.) Medical homes demonstrate innovation in the coordination of care, use of teams, and, to a degree, wellness services. They must continue to work on system transformation by coordinating care across systems, placing greater emphasis on population wellness, and substantially changing their reimbursement models. IT (including EHRs, clinical decision support, PHRs, and disease registries) is key to improving access to and sharing of patient information within a care coordination team.

Accountable Care Organizations Accountable care organization (ACO) “Group of doctors, hospitals, and other health care providers, who come together voluntarily to give coordinated high quality care to their Medicare patients” (Centers for Medicare & Medicaid Services 2018)

Recognizing the fragmented nature of healthcare delivery, the Affordable Care Act of 2010 authorized Medicare to support the development of accountable care organizations (ACOs) and offer financial incentives to improve population health, increase care coordination, reduce costs, and enhance quality. The purpose of an ACO is to improve the integration of care for a defined population, bringing together and coordinating disparate healthcare organizations such as providers of inpatient, primary, specialist, pharmacy, home health, wellness, and other services. If this goal can be achieved, it would be a major step toward aligning organizations and payment models with the core values of the health professions (Ganguli and Ferris 2018). ACOs move financing away from the traditional cost-based, fee-for-service system that rewards volume and intensity toward a value-based system that is accountable for overall quality and cost of care. ACOs are thus designed to align financing incentives with quality, efficiency, and continuity. Progress, to be sure, but many clinical specialties continue to be paid on a fee-for-service basis, which conflicts with the ACO concept (chapter 15). There is work yet to be done. The ACO is an innovative model for structuring a local, integrated system composed of providers and payers with a shared interest in wellness, quality, continuity, and efficiency. This integration of independent and diverse organizations motivated by the same goals is consistent with the assumptions of systems theory and the values and political structures of the United States. In addition, the ACO model provides a system context for deploying an integrated EHR to enable the operation of the enterprise (Huber, Shortell, and Rodriguez 2017; Wu et al. 2016). As previously noted, developing an integrated information system is a challenge for ACOs because member organizations have their own information systems and rely on HIE (a fragmented technology) to coordinate clinical services. However, this HIE capability, which allows access to clinical

Ch a p t er 4 :  T h e C o mi n g o f th e C o r p o ra ti o n: Transfor m ing C linic al Wor k Proc esses

information across the system, is an essential quality of ACOs and provides the basis for developing shared knowledge-based systems (Brown et al. 2016). Drawing firm conclusions about the impact of ACOs is impossible given their newness and their range of models and settings. Comprehensive analyses have only recently started to be reported and highlight the structural deficiencies in health systems, including the incompatibility of EHRs between institutions (Huber, Shortell, and Rodriguez 2017). However, evidence suggests that ACOs have demonstrated improvements in health maintenance and in care continuity, quality, and efficiency. ACOs have reduced costs, improved clinical outcomes, and achieved better integration of clinical processes, increasing physician satisfaction (Phipps-Taylor and Shortell 2016). Many ACOs have also demonstrated that the health of the population they serve has improved. This raises the question of whether they serve the enrolled population or the population at large. There is some evidence that best practices in an ACO are carried over into the general population, making ACOs an effective agent of change (Casalino, Erb, and Shortell 2015). Of course, improvement in a system that is inefficient and almost devoid of integration is not a high standard, but it is an important milestone and provides evidence on which future policy directions may be crafted. Another caution is that the semiautonomous and diverse organizations forming integrated ACO models today could develop into large, formal, regional corporations with considerable market power, regional monopolies, and higher costs (Scheffler 2015). Such a trend would return the US health system to its previous corporate structure, except with larger corporations. This is a real danger because corporations have considerable economic, technical, and political power. ACOs might provide an effective basis for linking health information systems, giving patients greater capacity to seek consultations when traveling outside a given ACO. The same constraints are encountered in ACOs as in current systems—namely, greater emphasis on medical care than on population health, financial motivation to keep the patient in the system, and system-specific IT systems limiting access. IT problems can be resolved by the development of integrated ACOs—loosely configured systems integrated by an IT architecture. These are complex systems with many variables and interdependencies. Traditional approaches to organizational design will not apply and, if pursued, will end up doing the wrong thing better. ACOs require disruptive innovation that will transform structures, strategies, and clinical practice (see end-of-chapter case study). Structuring and managing change in ACOs is extremely complex because ACOs are innovative while simultaneously maintaining and growing their legacy practice (Ganguli and Ferris 2018; Hwang and Christensen 2007). Their future success will be determined in large part by how well they are able to manage this complexity and by the potential benefits of a new delivery model.

87

88

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

ACOs are a major advance over the design of regional medical care systems because they are invested in wellness and patient-centered, integrated services and they are enabled by IT—all of which are important qualities of future organizations and systems. The greatest prospect is that ACOs will continue to evolve into flexible, tailored, dynamic, and innovative communities held together by a strong dedication to a shared purpose and a value-based payment system. These are the qualities of future organizations in an information-driven world. One danger, mentioned earlier, is that ACOs will develop into large, formal, and regional corporate entities with hierarchical structures, central planning, dominant financial function, and a human resources office. These are the traditional structures of industrial-era corporations. Ganguli and Ferris (2018) warn system leaders to resist the pressure to “build larger and more prestigious institutions, increase market share, buy expensive technology, and evaluate physicians based on the financial bottom line as they too are evaluated by their financial success.” In contrast, future health systems must be flexible and innovative, and they must be evaluated and rewarded on the basis of population health, integration of services, quality, and efficiency.

Communities of Practice and Future Designs

Community of practice Clinical team that is both selforganizing and structured and rewarded on the basis of the needs of the patient, thus transcending individual clinicians, organizations, or systems

Flexible communities of practice are evolving in the commercial world and are ideally suited to healthcare given that they involve highly professionalized staff, tailored services, and the principle of patients as coproducers (which puts healthcare on the cutting edge of organizational innovation). The community of practice concept originated in open systems theory related to learning communities but was applied to new, dynamic organizations tailored to the information- and knowledge-driven world. They have been aptly described by Wenger and Snyder (2000) as “Communities of Practice: The Organizational Frontier.” How might they be tailored to a patient’s condition and be transformed as the patient’s condition changes? Communities would not be bound by dedicated clinical specialties and a defined range of services but would be adapted to patients’ needs; they would not be limited spatially but would include consultations with specialists across systems. In addition, communities would not be highly structured but would be drawn together by a common purpose— improving clinical quality, efficiency, and continuity as well as maintaining a healthy status—and motivated by a strong commitment to that purpose. These are the strong, traditional motivations of health professionals. They are also conditions that cannot be achieved at a high level in bureaucratic structures. A few minor details have to be worked out in communities of practice, such as how value is assessed and rewarded—functions that are made more complex as teams change when a patient’s condition changes. On the upside, value-based reimbursement allows flexibility because it not necessarily based on specialty, formal position, or corporate negotiation but on the value

Ch a p t er 4 :  T h e C o mi n g o f th e C o r p o ra ti o n: Transfor m ing C linic al Wor k Proc esses

added. Such flexibility has been demonstrated in the commercial industry for joint venture arrangements. The challenge is assessing the value added in a dynamic environment. The approach would be based on the value of the knowledge-based contribution to effective and efficient care, not on fixed fees among professionals or corporations. Professionals could work as independent practitioners or in medical groups, ACOs, or medical centers. Physicians could return to independent practice but participate in communities of practice and be rewarded for their value-added contribution to the community, not for the volume of services provided. Value added might be an evidence-based consultation via electronic technology, including dialogue but without direct contact with the patient. Professionals driven by personal economic gain or professional dominance would not be invited to the community. Professionals would be independent but accountable to the ACO for their demonstrable value-added contribution. Physicians would participate in a number of communities, depending on their specialty and quality and value contribution. Once the valuation challenges are worked out, the community of practice would be tailored to individual patients and enabled by an integrated information system. The IT exists to support such a function, but the organizational, financial, and professional details can only be conceptualized. When such a system is developed and tested, no hierarchical, bureaucratic, and formal organization will be able to compete with it or match its ability to adapt to rapid change. Emerging models might serve as a guide for future designs. One innovative model is Cityblock (www.cityblock.com/about), which focuses on integrating primary care, behavioral health, and social services with community health services. Cityblock specializes in underserved communities (based on a community development model), engaging members in those communities to address both their health and their social service needs. At its core is an integrated digital platform that links health and social services within Cityblock and to external community medical centers. This innovative model did not come from the traditional health system but is the creation of Google’s parent company, Alphabet. Transformational change might be initiated, in large part, from outside the health system. The slow, evolutionary development process of the current health system might itself be replaced by a seismic change from an external industry that bypasses the traditional values, structures, and strategies. Open systems theory operates under the assumption that the health system is itself part of a larger social system and must function and be accountable to a higher purpose. Let the games begin!

Summary of Integrated Systems For healthcare organizations to support integrated clinical functions, they must undergo fundamental change. The change process must involve a leadership team comprising members who have a mastery of clinical science, organizational and systems theory, and IT. Each member of the team needs not only to

89

90

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

share in the strategic vision but also to bring individual expertise to enable that vision. Team members are more than domain experts; they are conceptualizers and transformative leaders. The current trend to create and fill positions such as chief clinical (or medical) information officer is encouraging but too often focuses on people with primarily technical or clinical skills or, worse, people who have only resume credentials and lack vision or knowledge. Such officers must possess a vision of the future and an understanding of the complexity of getting there, as well as how their own area of technical expertise can enable the journey. All training and curricula for health professionals should include systems theory and health systems informatics (not just IT), and graduate programs in health administration should include courses in these competencies, made available to midcareer health professionals as well. Graduate programs should expect mastery of and high-level competency in these disciplines, which might require restructuring existing curricula at many universities.

Conclusion The dominant organizational structure continues to be based on the logic of the traditional, hierarchical business function, although some organizations have begun the complex work of restructuring. Healthcare differs from commercial industries and, to some degree, from other service sectors. Optimal system performance is achieved by knowledge from disciplines such as medicine, engineering, finance, economics, organizational science, ethics, and sociology. The highly individual nature of health and illness has historically assigned medical care responsibility to health professionals, a model ideally suited for managing variation inherent in personalized care. These factors remain important considerations in future systems. The power of IT offers healthcare organizations an opportunity to expand this caregiving model to a systems-based structure in which the business, clinical, and IT functions are integrated and aligned to enable truly evidence-based practice in a highly efficient and dynamic but tailored system.

Chapter Discussion Questions 1. Describe how US hospitals have been structured to accommodate the roles and decision-making autonomy of physicians. 2. Describe the traditional structure and function of IT in hospitals in the context of its state of development and the structure of hospitals. 3. What is a community of practice, and how has it been used in other industries? How might value be measured and rewarded in a community of practice?

Ch a p t er 4 :  T h e C o mi n g o f th e C o r p o ra ti o n: Transfor m ing C linic al Wor k Proc esses

Case Study  Not All Innovation Is Created Equal in the Transition to Value-Based Care Ryan Marling A recent essay published in JAMA recounts the competing pulls physicians must reconcile while practicing with one foot in the fee-for-service world, and the other in the value-based, accountable care world. The piece, by Ishani Ganguli, MD, MPH, and Timothy G. Ferris, MD, MPH [2018], does a great job of conveying the conflicting financial incentives and seemingly contradictory processes physicians attempting to bridge the gap between both of these worlds regularly face. They point out that value-based care models (which intend to reward better patient outcomes and lower spending) are beginning to proliferate the healthcare landscape. Yet fee-for-service purchasing (in which each episode of a patient visit, surgical procedure, hospital stay, etc. has its own individual fee) still remains dominant. In this way, the new game is beginning before the old game has ended. This reality commonly leads to temptation among providers and managers to merely incrementally tweak the existing business in efforts to accommodate the change in payment. But, by attempting to bridge both worlds, are providers actually creating one-size-fits-none models that leave no party better off?

When Change from Within Isn’t Enough Attempting to innovate within the existing business model is only appropriate in cases when sustaining and/or incremental innovations, like a process improvement or upgrade in equipment, are the intended result. In cases of more comprehensive and fundamental changes initiatives, like the shift to value-based care, such a strategy is grossly ineffective. It’s left caregivers to pick up the slack and bridge the gap between the diverging models. It’s also resulted in lower margins for practices, along with greater complexity and increasingly burdensome administrative tasks referred to in the article by Dr. Ganguli and Dr. Ferris. Instead, fundamental change initiatives like this one require an autonomous business unit—separate from the influence of the existing business—to foster growth and success, as outlined in The Innovator’s Solution by Clayton Christensen and Michael Raynor [2013]. Autonomous units are needed when the organization’s current processes aren’t appropriate for the new project, and/or the organization’s values are not likely to adequately prioritize the new initiative. (continued)

91

92

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

The necessity for this distinction stems from a general law of organizational nature. The business’ existing management will either starve the new model of necessary resources or force the new model to conform with the existing processes and values—resulting in mere incremental change and/or failure.

The Pull of the Existing Business Model An example of this strong force to conform, as described in The Innovator’s Solution, is F.W. Woolworth, then one of the world’s leading retailers, and their establishment of Woolco in 1962. Woolco was Woolworth’s response to the rise of discount retail stores, in essence, a discount department store arm of their own. It was, in fact, originally established as a free-standing, autonomously managed business unit, and like other discount retailers of the time, averaged gross margins of 23 percent and turned inventory five times a year, compared to the 35 percent margins and 3.4 turns per year Woolworth averaged. But, in 1971, Woolworth corporate executives decided to integrate Woolco back into Woolworth in attempts to leverage the fixed costs of management, buying, and logistics functions across both businesses. What came of this integration? Within the span of a year, Woolco’s margins were pushed up to 34 percent, and their inventory turns declined to four times annually—mirroring the profit model of F.W. Woolworth, and resulting in Woolco’s ultimate closure. For Woolco to make it as a discount retailer, they needed to remain a separate business unit, as their values and processes were too different from those of Woolworth. Likewise, this pull of processes and values within the existing business has resulted in the failure of new growth ventures in many industries, being the scourge of innovation efforts from companies the likes of Bank One, Charles Schwab, Merrill Lynch, and IBM.

What Does This Mean for Providers? For many providers, the switch to delivering care within a value-based system requires a fundamental change in processes and values. As such, the change initiative should ideally be housed in a separate and autonomous business unit, just as Woolworth originally did with Woolco. Such a prescription is easier said than done. Smaller practices and specialist practices may be particularly reluctant to go “all in” on value-based care with an autonomous unit, as they have historically been dependent upon fee-for-service arrangements. More so than others, they may be tempted to hold onto their fee-for-service model and do the bare minimum to accommodate for value-based care within their existing model. Careful assessment of their situation is likely to reveal that their transition to value-based care will entail more drastic change and warrant an autonomous business unit.

Ch a p t er 4 :  T h e C o mi n g o f th e C o r p o ra ti o n: Transfor m ing C linic al Wor k Proc esses

If not assessed and acted upon properly, reinvention efforts, necessary for future success, will only result in further frustration and one-size-fits-none solutions. Source: This case study is reprinted with permission from the Clayton Christensen Institute. www.christenseninstitute.org/blog/not-all-innovation-is-created-equal/?_sft_topics=businessmodel-innovation,interoperability. Copyright 2017.

Case Study Discussion Questions 1. When one speaks of innovation, what are the relevant factors to consider for hospitals and clinics under a business model, and what types of innovations might be implemented in such a model? 2. What is the range of forces and values in a fee-for-service business model that might diminish or conflict with a value-based model? 3. How would you conceptualize a separate and autonomous business unit in an ambulatory care clinic? 4. How might the structure of a clinical process change under a valuebased reimbursement system? 5. What are the properties or qualities of an integrated information system that might differentiate a value-based model from a cost reimbursement model?

Additional Resources Affordable Care Act: www.healthcare.gov/glossary/affordable-care-act/. Cityblock: www.cityblock.com/about.

References American Academy of Family Physicians (AAFP). 2011. “Nation’s Top Primary Care Physician Organizations Release Guidelines for Patient-Centered Medical Home Recognition Programs.” Published March 8. www.aafp.org/media-center/ releases-statements/all/2011/pcmh-elements.html. American Academy of Family Physicians (AAFP), American Academy of Pediatrics, American College of Physicians, and American Osteopathic Association. 2011. “Guidelines for Patient-Centered Medical Home Recognition and Accreditation Programs.” Published February. www.aafp.org/dam/AAFP/documents/ practice_management/pcmh/initiatives/PCMHJoint2011.pdf. Baum, N. 2014. “Business of Urology: Patient Centered Medical Home: What May Be on the Horizon for Urologists.” Urology Practice 1 (2): 53–56.

93

94

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Boucher, N. A., S. White, and D. Keith. 2016. “A Framework for Improving Chronic Critical Illness Care: Adapting the Medical Home’s Central Tenets.” Medical Care 54 (1): 5–8. Brown, B. B., C. Patel, E. McInnes, N. Mays, J. Young, and M. Haines. 2016. “The Effectiveness of Clinical Networks in Improving Quality of Care and Patient Outcomes: A Systematic Review of Quantitative and Qualitative Studies.” BMC Health Services Research 16: 360–75. Buchert, A. R., and G. A. Butler. 2016. “Clinical Pathways: Driving High-Reliability and High-Value Care.” Quality of Care and Information Technology 63 (2): 317–28. Casalino, L. P., N. Erb, and S. Shortell. 2015. “Accountable Care Organizations and Population Health Organizations.” Journal of Health Politics, Policy and Law 40 (4): 821–37. Centers for Medicare & Medicaid Services. 2018. “Accountable Care Organizations.” Modified May 3. www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/ ACO/index.html. Christensen, C. M., and M. E. Raynor. 2013. The Innovator’s Solution: Creating and Sustaining Successful Growth. Boston: Harvard Business Review Press. Darling v. Charleston Community Memorial Hospital, 33 Ill. 2d 326, 211 N.E.2d 253, 1965 Ill. LEXIS 250, 14 A.L.R.3d 860 (Ill. 1965). Es-Sajjade, A., and T. Wilkins. 2017. “Design, Perception and Behavior in the Innovation Era: Revisiting the Concept of Interdependence.” Journal of Organization Design 6 (1): 1–12. Ganguli, I., and T. G. Ferris. 2018. “Accountable Care at the Frontlines of a Health System: Bridging Aspiration and Reality.” Journal of the American Medical Association 319 (7): 655–56. Gittell, J. H., J. Beswick, D. Goldmann, and S. S. Wallack. 2015. “Teamwork Methods for Accountable Care: Relational Coordination and TeamSTEPPS®.” Health Care Management Review 40 (2): 116–25. Huber, T. P., S. M. Shortell, and H. P. Rodriguez. 2017. “Improving Care Transitions Management: Examining the Role of Accountable Care Organization Participation and Expanded Electronic Health Record Functionality.” Health Services Research 52 (4): 1494–510. Hwang, J., and C. M. Christensen. 2007. “Disruptive Innovation in Health Care Delivery: A Framework for Business-Model Innovation.” Health Affairs 27 (5): 1329–35. Institute of Medicine (IOM). 2001. Crossing the Quality Chasm: A New Health System for the 21st Century. Washington, DC: National Academies Press. Kaslow, N., S. Kapoor, S. Dunn, and C. Graves. 2015. “Psychologists’ Contributions to Patient-Centered Medical Homes.” Journal of Clinical Psychology in Medical Settings 22 (4): 199–212. Kohn, L. T., J. M. Corrigan, and M. S. Donaldson. 2000. To Err Is Human: Building a Safer Health System. Washington, DC: National Academies Press.

Ch a p t er 4 :  T h e C o mi n g o f th e C o r p o ra ti o n: Transfor m ing C linic al Wor k Proc esses

Kornberger, M. 2017. “The Visible Hand and the Crowd: Analyzing Organization Design in Distributed Innovation Systems.” Strategic Organization 15 (2): 174–93. Leasure, E. L., R. R. Jones, L. B. Meade, M. I. Sanger, K. G. Thomas, V. P. Tilden, J. L. Bowen, and E. J. Warm. 2013. “There Is No ‘I’ in Teamwork in the PatientCentered Medical Home: Defining Teamwork Competencies for Academic Practice.” Academic Medicine 88 (5): 585–92. Lewis, N. 2016. “Independent Practices Seek Care Coordination Strategies: Physician Collaboration Is Difficult for Small Practices Under Today’s Payment Models, but Its Importance Is Growing.” Medical Economics 93 (9): 24–29. Macias, C. G., J. N. Loveless, A. N. Jackson, and S. Suresh. 2017. “Delivering Value Through Evidence-Based Practice.” Clinical Pediatric Emergency Medicine 18 (2): 89–97. Mintzberg, H. 2017. Managing the Myths of Health Care. Oakland, CA: BerrettKoehler Publishers. ———. 2015. “Time for the Plural Sector.” Published summer. https://ssir.org/ articles/entry/time_for_the_plural_sector. Morgan, M. I. 2010. Executing Your Business Transformation: How to Engage Sweeping Change Without Killing Yourself or Your Business. San Francisco: Jossey-Bass. Morton, S., S. C. Shih, C. H. Winther, A. Tinoco, R. S. Kessler, and S. H. Scholle. 2015. “Health IT–Enabled Care Coordination: A National Survey of Patient Centered Medical Home Clinicians.” Annals of Family Medicine 13 (3): 250–56. Muzio, D., D. Hodgson, J. Faulconbridge, J. Beverstock, and S. Hall. 2011. “Towards Corporate Professionalization: The Case of Project Management, Management Consultancy and Executive Search.” Current Sociology 59 (4): 443–64. Muzio, D., and I. Kirkpatrick. 2011. “Professions and Organizations—a Conceptual Framework.” Current Sociology 59 (4): 389–405. National Center for Quality Assurance (NCQA). 2018. “Patient-Centered Medical Home (PCMH) Recognition.” Accessed May 9. www.ncqa.org/programs/ recognition/practices/patient-centered-medical-home-pcmh. Nwando, O. J., S. Sheth, V. Mleczko, A. L. Choi, and A. E. Sharma. 2017. “The Impact of the Patient-Centered Medical Home on Health Disparities in Adults: A Systematic Review of the Evidence.” Journal of Health Disparities Research & Practice 10 (1): 68–96. O’Malley, A. S., R. Gourevitch, K. Draper, A. Bond, and M. A. Tirodkar. 2015. “Overcoming Challenges to Teamwork in Patient-Centered Medical Homes: A Qualitative Study.” Journal of General Internal Medicine 30 (2): 183–92. Phipps-Taylor, M., and S. M. Shortell. 2016. “More Than Money: Motivating Physician Behavior Change in Accountable Care Organizations.” Milbank Quarterly 94 (4): 832–62. Pinsonneault, A., S. Addas, C. Qian, V. Dakshinamoorthy, and R. Tamblyn. 2017. “Integrated Health Information Technology and the Quality of Patient Care:

95

96

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

A Natural Experiment.” Journal of Management Information Systems 34 (2): 457–86. Reid, P. P., W. D. Compton, J. H. Grossman, and G. F. Fanjiang. 2005. Building a Better Delivery System: A New Engineering/Health Care Partnership. Washington, DC: National Academies Press. Scheffler, R. M. 2015. “Accountable Care Organizations: Integrated Care Meets Market Power.” Journal of Health Politics, Policy & Law 40 (4): 639–45. Sharma, L. A., K. Chandrasekaran, K. Boyer, and C. M. McDermott. 2016. “The Impact of Health Information Technology Bundles on Hospital Performance: An Econometric Study.” Journal of Operations Management 41: 25–41. Shi, L., D. Lee, M. Chung, H. Liang, D. Lock, and A. Sripipatana. 2017. “PatientCentered Medical Home Recognition and Clinical Performance in U.S. Community Health Centers.” Health Services Research 52 (3): 984–1004. Solimeo, S. L, G. L. Stewart, and G. E. Rosenthal. 2016. “The Critical Role of Clerks in the Patient-Centered Medical Home.” Annals of Family Medicine 14 (4): 377–79. Starr, P. 1984. The Social Transformation of American Medicine. New York: Basic Books. Tessier, J. E., G. Rupp, J. T. Gera, M. L. DeHart, T. D. Kowalik, D. Tom, and P. J. Duwelius. 2016. “Health Policy and Economics: Physicians with Defined Clear Care Pathways Have Better Discharge Disposition and Lower Cost.” Journal of Arthroplasty 31 (9): 54–58. Tseng, S.-M. 2011. “The Effects of Hierarchical Culture on Knowledge Management Processes.” Management Research Review 34 (5): 595–608. Waters, T. M., J. A. Webster, L. A. Stevens, L. Tao, C. M. Kaplan, L. Graetz, and B. L. McAneny. 2015. “Community Oncology Medical Homes: Physician-Driven Change to Improve Patient Care and Reduce Costs.” Journal of Oncology Practice 1 (6): 462–67. Wenger, E. C., and W. M. Snyder. 2000. “Communities of Practice: The Organizational Frontier.” Harvard Business Review 78 (1): 139–45. Wennerstrom, A., L. Hargrove, S. Minor, A. L. Kirkland, and R. Shelton. 2015. “Integrating Community Health Workers into Primary Care to Support Behavioral Health Service Delivery: A Pilot Study.” Journal of Ambulatory Care Management 38 (3): 263–72. Wu, F. M., T. G. Rundall, S. M. Shortell, and J. R. Bloom. 2016. “Using Health Information Technology to Manage a Patient Population in Accountable Care Organizations.” Journal of Health Organization and Management 30 (4): 581–96.

CHAPTER

PREDICTIVE ANALYTICS IN KNOWLEDGE MANAGEMENT

5

Gordon D. Brown, Kalyan S. Pasupathy, and Mihail Popescu

Learning Objectives After reading this chapter, you should be able to do the following: • Critically assess the data-mining applications in healthcare organizations. • Classify data-mining applications according to the method and structure of databases. • Critique the applications of modeling to alternative problem scenarios. • Discuss the contribution of the electronic health record to a knowledgebased system.

Key Concepts • Knowledge-based systems • Stored data and mathematical algorithms • Prerelational, relational, object-oriented, resource descriptive format, and NoSQL databases • Classification, clustering, and association-rule data mining • Discrete event, agent-based, and systems dynamics modeling • Complexity leadership

Introduction The development of information technology (IT) applications to support clinical decision making has been significant and rapid, overcoming complex issues in establishing vocabularies and integrating systems. However, only a limited amount of real innovation and systems transformation, which are frequently enabled by advanced IT, has occurred. This chapter, which builds on the 97

98

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

concepts discussed in chapter 2, explores the concepts of data mining, analytics, and analytical modeling and their potential for generating sources of knowledge that can be used to improve health outcomes and increase system efficiency. Electronic health record (EHR) and health information exchange (HIE) systems generate large and valuable databases that have spawned decision support tools, dashboards with indicators, and other metrics, all claiming success in improving decision making and performance. However, organizations and clinicians have been reluctant or unable to apply higher-order analytics to these databases to add knowledge of complex relationships to guide clinical and managerial decision making. Scholars conclude that with the information age, most organizations have “increasing amounts of data but don’t analyze the information to inform their decision making” (Davenport, Harris, and Morrison 2010, 9). In this chapter, data mining and modeling are framed as higher-order analytical tools that have the potential to generate sources of knowledge from various databases. The EHR makes inherent assumptions about the transformational structure of clinical processes and their support within the system. Developing and testing innovative models requires a higher order of analysis of complex systems. The personal health record (PHR) can be used as the basis for conceptualizing innovative and transformational models of healthcare services delivery, such as the development of accountable care organizations (ACOs).

Data Mining and Analytics The EHR is a warehouse of accumulated information that includes relationships and patterns about a given patient and potentially a group of patients. IT applications such as barcoding have facilitated the use of technology for information integration, process improvement, and safety. In this respect, the EHR is still inert in itself and characteristic of many electronic databases in hospitals and clinics. Health informatics enables the collection, integration, and presentation of patient-specific information from the EHR at a more rapid rate, allowing more trend data to be presented than was possible with paper charts. However, limited data mining and modeling are possible with the patient-specific information drawn from the EHR, making such information systems function much as the paper charts did. Enabling clinical decision support systems (CDSSs), such as clinical guidelines and protocols, through the EHR substantially changes decision processes and clinical outcomes. These guidelines, however, are made up of general, population-based information and are not tailored to patient-specific populations. Large databases are a natural outgrowth of electronic information and are of considerable potential value to enterprises. Known as Big Data, these

C h a p te r 5 :   Pre d i c tive A naly tic s in Knowledge Managem ent

large databases in healthcare contain considerable information on behaviors, individuals, or small population sectors such as patients from a selected group of physicians. Here, the meaningful pattern is unknown going in, but analytics can identify and describe this pattern. For instance, how patients respond to a certain medication can be best gleaned through data mining. Individual clinicians are restricted to what they observe in their respective practices. Data can pool and reveal patterns across multiple physicians, regions, and patient cohorts. Big Data can define the pattern as the accumulated information for a single patient, for patients with similar diagnoses seen by a single clinician, or for patients with similar diagnoses seen by a community of practice. Business intelligence and analytics, including information on customers, have become valuable assets to organizations and have been identified by corporate leaders as priorities for business development (Davenport, Harris, and Morrison 2010; IBM 2011). Service and product industries consider business intelligence and analytics to be the competitive differentiator in many future markets. It is easy to see that the health system will be affected directly and indirectly by these new technologies that capture, analyze, and guide operational and strategic decisions. Chapter 2 examines the four types of knowledge-based IT systems that inform and transform clinical decision making—EMR (electronic medical record), HIE, EHR, and PHR. Each of these databases supports a level 1 analysis, including measurements, trends, comparisons, and tests of statistical significance. This chapter discusses a higher order, or level 2, of analytics. Data mining of EHRs and of financial, utilization, and other organizational and clinical files uses mathematical algorithms to convert the accumulated experiential knowledge embedded in data files into explicit knowledge. Data mining is the use of sophisticated search capabilities and analytical techniques on large databases to discover patterns, correlations, and trends that can be leveraged to produce knowledge. What constitutes a large database is relative and depends on the application and the types of analytics to be deployed. The size of the database is defined by the statistical tool and level of analysis being performed. For example, a database is typically considered small if it stores less than a few hundred thousand records, and it is considered very large if it has more than a billion records. This technique is commonly used to identify relationships within an existing decision-making structure, but it can be used as the basis for transformational change (“6” in exhibit 5.1). Data mining extracts accumulated knowledge embedded within an EHR of individual patients or populations of similar patients of a clinician, a group of affiliated physicians, or, increasingly, a group of affiliated institutions such as the Shared Health Research Informatics Network (SHRINE) discussed in chapter 2. Such accumulated knowledge forms a level of empirical evidence that complements the memory and recall (tacit knowledge) of clinicians. A clinician could choose to mine

99

Data mining Use of sophisticated search capabilities and analytical techniques on large databases to discover patterns, correlations, and trends that can be leveraged to produce knowledge

100

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

EXHIBIT 5.1 Data Mining and Analytics

System structure

Community of practice

Training A

1. EMR ACO

Internet, social media

Scientific evidence B 2. Decision support (EHR) e-trials

3. HIE e-trials

E

C

Experiential

Decision

Knowledge accumulation

5. Integrated PHR

6. Data mining, analytics

D 4. Health data vault

Outcome–decision relationship

Note: ACO = accountable care organization; EHR = electronic health record; EMR = electronic medical record; HIE = health information exchange; PHR = personal health record.

patient-specific information from an EHR to examine relationships that might confirm, refine, or refute population-based clinical guidelines contained in an EHR. The findings of the analysis of relationships in a patient’s electronic record allow a clinician to balance the continuously accumulated empirical evidence against experiential knowledge. These data are not high-level scientific evidence based on clinical relationships that can be generalized to an entire population, but they can complement decisions because they represent the experiences of a patient or groups of patients. By linking, and thus integrating, EHRs across institutions, it is possible to create much larger databases and to improve confidence in the analyses. Such data might be used to support clinical decisions that deviate from population-based guidelines. Patient preferences (related to appointments, referrals, or other matters), patterns of compliance with medications, wellness behaviors, and other information can be analyzed as well to tailor services to individual patients. As patients become more informed consumers, such information might be of considerable value when selecting and retaining a clinician. (Privacy concerns related to data mining are addressed in chapter 13.) Exploring the value-added potential of data mining requires an understanding of the context in which it is applied and the analytics that drive it. The degree to which clinical data can be accessed has practical limits, some of which are institutional restrictions while others are technical challenges. The EHR

C h a p te r 5 :   Pre d i c tive A naly tic s in Knowledge Managem ent

serves as the initial, primary data source because it is viewed as the property of the institution and because the knowledge it generates may add value to the organization and its clinicians. The EHR limits data mining because data are frequently structured in a manner that makes data mining difficult. It might be technically possible to extract information from health records in other institutions through an HIE (“3” in exhibit 5.1) if the disparate data share a common vocabulary and architecture, which would create a larger database. However, that possibility also is limited because issues of politics, legal restrictions, proprietary interests, and privacy and security are involved. Some institutions are not willing to allow valuable patient and clinical information (a market asset) to be extracted from their EHRs because such data may be used by competitors to gain a competitive advantage. Some form of collaborative agreement and mutual benefit would likely be needed to support such an HIE. These are all complex but not unsolvable problems. The point is that the problems of data mining through HIE are related not only to IT and clinical capability but also to market and competition. As discussed in earlier chapters, the interrelationship among the technical, clinical, and institutional (business) perspectives is inherent in health systems informatics. Data mining and analytics allow this interrelationship to be examined and help guide the strategy for innovation and system transformation. Extracting the value of knowledge gained from such analysis will likely foster changes in system structure and create new systems-based forms that integrate the technical, clinical, and business functions that previously competed with each other and functioned in isolation. Healthcare leaders must understand the potential value and technical challenges of data mining. Data mining has two main components: stored data and mathematical algorithms. Stored data are the various databases that are available, many of whose relationships have never been analyzed. Mathematical algorithms comprise a set of precise rules followed by the computer in calculating relationships in a database used for knowledge extraction. The underlying data storage technology has a great impact on the nature of the data-mining methods used.

Database Types and Their Impact on Data Mining The five main types of databases are prerelational, relational, object oriented, resource descriptive format (RDF), and nonstructured query language (NoSQL). • A prerelational database stores data in tree-like hierarchies. Some EHRs have a prerelational format, such as the EHR of the US Department of Veterans Affairs; the EHR of Epic Systems Corporation; and MUMPS,

101

102

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Data warehouse Data from a broad range of sources linked together and stored for easy retrieval, reporting, analysis, and decision making

the healthcare database originally developed for Massachusetts General Hospital in the 1960s. • The relational database is currently the most widespread database format. It is used in some EHRs, such as those produced by the Cerner Corporation, and keeps the data in tables that represent real objects and that are connected through relationships (hence the name “relational”). For example, a “patient” table has columns such as patient last name and patient age, whereas a “visit” table represents the relation between patients and physicians. This structure is powerful and adaptable as a CDSS, but it is somewhat inflexible. Inflexibility is a problem with IT architecture that was designed for a given purpose because, if the problem context changes, modifying the tables to reflect the new situation is not easy. For example, two different EHR databases may have different data dictionaries (e.g., the column for patient surnames may be called variously “patient last name” or “last name”), resulting in great difficulties in data exchange. Another, more subtle drawback is that relational databases are not conducive to data mining for several reasons. First, the relational format minimizes data redundancy and is suitable for multiple input– output operations. For example, in the “visit” table, the information about each patient (e.g., last name, age) at each visit is not repeated. Instead, an identifier that can be found in the patient table is used. As a result of this structure, searching many tables—maybe tens of tables in a typical EHR—is necessary to retrieve the desired data. For this reason, conducting data-mining experiments on an EHR relational database is generally not a good idea. To alleviate this problem, a variation of the relation framework—a data warehouse—might be created, either within a given institution or in a common warehouse shared by institutions involved in the HIE. A data warehouse is a relational database with a special format that is more conducive to analysis than to transaction processing. For example, if one wants to relate financial and clinical information, a data warehouse would facilitate the extraction and transformation of different data formats into a common analytical framework. The data warehouse is highly optimized for output (fast data retrieval) but not for input. Second, the table format of the typical EHR does not allow the discovery of relationships among data that were not already known when the database was built. For example, in an EHR that contains data from four generations of the same family, finding a link between greatgrandparents and their great-grandchildren is hard technically because the EHR designers did not anticipate generational relationships to be clinically relevant. In the era of genomic medicine, however, generational relationships are relevant.

C h a p te r 5 :   Pre d i c tive A naly tic s in Knowledge Managem ent

• An object-oriented database is more flexible than a relational database in several respects. First, it can deal easily with a variety of objects, such as image sequences that are abundant in medical practice. Second, it integrates data with code under the object-oriented paradigm, resulting in a more structured representation of the problem domain. Third, it introduces the concept of class hierarchies, which allows for an incremental refinement of the domain model. However, from the knowledge-discovery point of view, an object-oriented database is not fundamentally different from a relational database. Most of the current EHRs are either in relational or object-oriented format. This online analytical processing (OLAP) architecture is highly optimized for fast data input/output or for satisfying the demands of medical personnel during the care delivery process. However, this data representation does not lend itself to more complex questions that usually arise in mining a clinical database to determine relationships and patterns emerging from EHRs. For example, a search for patterns related to a given clinical diagnosis usually involves relationships among numerous tests and diagnostic areas. For this reason, producing an answer to a complex data-mining question might take weeks in a regular EHR. Instead, an OLAP architecture, such as a data warehouse, should be used for a fast answer to such problems. • The resource descriptive format (RDF) database addresses most of the problems inherent in the other database types. In an RDF database, data are stored in just one table that has three columns: subject, predicate, and object. All information is stored as triples—for example, “patient id 11100” (subject) “hasLastName” (predicate) of “Brown” (object or value). This format is extensible, so it is easy to interconnect and conducive to mining. RDF databases, however, still require (as relational databases do) a way to “chunk” the information into triples—through natural language processing (NLP), for example. In fact, NLP is increasingly employed in contemporary EHRs to enable searching of unstructured text fields, such as physician notes or discharge summaries. In a sense, NLP contributes to the transition from EHR systems (“2” in exhibit 5.1) to data mining and analytics (“6” in exhibit 5.1) knowledge management systems. • The NoSQL database does not store data in a set structure (tables), and it uses a key-value mechanism for data retrieval instead of an SQL query. NoSQL databases, such as MongoDB, are more scalable and more suitable to diverse content than relational databases are, making them the technology of choice for Internet applications. However, the lack of transactional support makes them less suitable for health data. These five types of databases follow the historical evolution of information systems, from file based (single desktop computer) to relational and

103

Online analytical processing (OLAP) architecture IT setup that enables the user to “slice and dice” the data in multiple dimensions to provide insights

Natural language processing (NLP) Computational approach to processing human language

104

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

The Impact of the Internet The Internet has significantly affected the traditional data-mining process along multiple dimensions. The first and most notable effect

object oriented (desktop computer and local network) to RDF (desktop computer and Internet) to NoSQL.

of the Internet is the removal of the middle operative from the mining process. Traditionally, the process involved three entities: database, analyst, and decision maker (clinician or manager). The main drawback

Data-Mining Methods

of this setup is the delay between the situation reflected by the data and the decision time, which is analogous to getting dressed today for the weather conditions of one month ago. The new data-mining process provides just-in-time knowledge for the decision maker. For example, when you go to Amazon.com to buy a camera, you instantly know what other cameras customers who bought this camera also viewed; thus, you do not need any other report about cameras and their features. In this case, the middleman has been replaced by a sophisticated, “intelligent” computer program, reducing the decision lag. In the clinical setting, the clinician would know at the time of each consult the complex interrelationships among clinical conditions and between conditions and treatment patterns. The second effect is the interconnected architecture of the underlying database and its intelligent algorithms. Extraction of clinical knowledge (EMR; “1” in exhibit 5.1) requires a statistically significant amount of data that can be acquired only from multiple institutions. The EHR interconnection provides the architecture necessary for this endeavor (HIE; “3” in exhibit 5.1). At present, the required data are obtained through mandatory reporting by individual institutions and are deposited in registries (e.g., cancer registry) or national data sets (e.g., Healthcare Cost and Utilization Project). EHR interconnectivity eliminates the need for reporting and reduces the decision lag. As a result, the data can be mined and reports can be generated by the decision makers in real time. Obviously, this powerful capability has its own security challenges. However, both the banking industry and online commerce have proved that society can find a balance between utility and security. The third effect is represented by the semantic orientation. Controlled vocabularies and ontologies are required for extracting knowledge from diverse and complex data (chapter 12). Currently, several data-mining frameworks coexist: • Framework 1. Data are extracted from the EHR or business database of a single organization and then processed by (continued)

From a research perspective, data-mining methods are distinguished from hypothesisdriven data analysis (Wickramasinghe et al. 2008) in the sense that data mining generates rather than verifies hypotheses. While this statement is true, the distinction is rather artificial. According to the datato-information-to-knowledgeto-wisdom paradigm (chapter 2), knowledge and data have a feedback relationship, and we need to know something about problems before starting data mining. In our view, data mining includes statistical analysis. The term data mining is restrictive because it suggests that the analysis part is performed after a large quantity of data has been accumulated. While this approach is part of data mining, it does not account for just-in-time knowledge generation (see Frameworks 2 and 4 in the sidebar). Moreover, in some instances, knowledgediscovery methods use current data that are acquired online while the application is running. Imagine, for example, an intelligent mammography

C h a p te r 5 :   Pre d i c tive A naly tic s in Knowledge Managem ent

system that assists the physician (continued from previous page) during the screening process. While the image is acquired, an the analyst who makes a report for the decision maker automated system suggests to (manager or clinician). This approach is typically employed the clinician a possible area of for institutional management and quality assurance. concern; the area is then further • Framework 2. Knowledge is automatically generated just in investigated by the acquisition time for the clinician and manager. In this case, the analyst of further images. The term is involved in method development but not in knowledge pattern recognition includes creation per se. If the decision process is knowledge this kind of decision support intensive, a data warehouse might be necessary. system and is therefore proba• Framework 3. Data from multiple institutions are obtained bly more suitable for describing by reporting, analyzed, and then published. This approach all knowledge-discovery methis appropriate for guideline production (EHR; “2” in exhibit ods. However, for the sake of 5.1) and for strategic healthcare management (“System consistency, we use “data minstructure” in exhibit 5.1). ing” throughout this chapter. • Framework 4. Knowledge is extracted from the entire A detailed discussion of datahealthcare organization just in time for the decision mining methods is beyond the process. Unlike Framework 3, which can generate general scope of this book but should clinical guidelines, Framework 4 is able to account for local be included in clinical, manaexperiential knowledge. For example, it might be interesting gerial, and informatics currito find out what clinical methodology is employed by some cula. Students are encouraged highly experienced physicians who consistently achieve to draw on any general referabove-average outcomes but who systematically draw on ence for gaining more knowlexperiential knowledge and do not strictly follow published edge about this important area clinical guidelines. of analysis (Witten and Frank 2005). In this section, we provide some context for thinking about data mining in healthcare organizations. There are three main types of data-mining methods: classification, clustering, and association-rule mining: 1. Classification is an algorithm that attempts to assign an unknown object to one of the available classes of known objects. For example, when a patient is chronically late for clinical appointments, we might want to know as soon as possible if he will or will not show up for the next scheduled visit. Two classes can be used in this situation: show and no show. The classification algorithms can be data driven or knowledge driven. In the data-driven algorithms, previously stored data are used together with the known class labels to develop classification models, such as neural networks, support vector machines, or simple Bayes ones (Theodoridis and Koutroumbas 2009). In the no-show case, a data set

105

106

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

is needed that includes examples of patients described by several relevant characteristics (e.g., age, complaint, insurance, distance from the clinic) and the related show/no show labels. After the classifier is trained, it can be used online (just in time) to find the status of a scheduled patient. In some applications, however, either the classification data are not available or the problem domain can be easily described by simple rules. In this case, rule-based classifiers—such as fuzzy rule (FR) systems, first-order logic (FOL) rules, or description logic (DL) rules—are used (Brahman and Levesque 2004). For example, to develop an automatic system for detecting heart attacks on the basis of a set of body sensors, sensors can be designed, but collecting sufficient data to construct the algorithms needed to provide the alert is difficult. This problem is related to the power calculation in clinical trials, and in the case of an innovative medical system, the analysis is not afforded that luxury. Instead, clinical rules (implemented as FR, FOL, or DL) provided by clinicians can be established. When the class labels are not known, such as show or no show, the available objects must be grouped to discover possible similarities or relationships. 2. Clustering is the quintessential data-mining problem, where the number of groups and the relationships between the objects are typically not known. For the show/no show example, before implementing the classifier, some questions can be asked: How many types of patients does the clinic have? What is the profile of the patient from each group? A large variety of clustering algorithms is available, such as k-means, hierarchical clustering, and self-organizing feature maps (Theodoridis and Koutroumbas 2009), each of which has its advantages and disadvantages. No perfect data-mining algorithm exists, but a range of techniques (each designed to fit a problem at hand) is available. The important part is to select the algorithm that fits the problem. 3. Association-rule mining is another typical data-mining algorithm that tries to find relationships among characteristics of objects stored in a database. The discovery is based on the frequency of association; that is, if two characteristics associate often, then their relation might be relevant to the problem being analyzed. In the show/no show example, a correlation may be discovered between distance from home to the clinic and no-show status. This correlation may then be used in a rule-based classifier.

Dynamic Systems Modeling Knowledge management is described in chapter 2 as a concept “that promotes an integrated approach to identifying, managing, and sharing all of

C h a p te r 5 :   Pre d i c tive A naly tic s in Knowledge Managem ent

an enterprise’s information needs” (Lee 2000). Here, enterprise refers to the system, which includes all units that bring knowledge, including organizational and financial, to the services being delivered. It extends beyond individual clinicians, the organization, and the health system into other sectors such as education and social services. The structure of the enterprise is defined by the nature of the problem being addressed and the collective knowledge needed to serve the consumers. Through this process and the interactions among all participants in the system, knowledge is gathered and integrated. In the health system, part of the enterprise is the community of practice. An ACO may be defined as a form of community of practice. The ongoing political debate in the United States raises caution against assigning a label to ACOs that could imply more traditional structures. Chapter 4 concludes with a discussion of community of practice, which requires a systems perspective and thus necessitates the transformation of the traditional role of health professionals (i.e., new values, behaviors, and skills) and the development of new structures for delivering and financing services, new legal and regulatory mandates, and new relationships with supplies and other social sectors (e.g., education, social services). The complexity inherent in such analysis cannot be understood through statistical analysis and data mining alone and use of traditional decision models; it requires a dynamic systems perspective. Applying knowledge management includes considering how the clinical and business enterprise might be structured to bring maximum knowledge to bear on a problem being considered. In a dynamic enterprise system, one can test the impact of alternative designs, such as the change in structural orientation from a static system that focuses on clinicians and the organization to a patientcentered system with a community of practice that focuses on the patient. The latter is characterized by considerable variability and complexity and thus lends itself to predictive modeling (“7” in exhibit 5.2). Modeling starts with building conceptual models that challenge current assumptions and introduce futuristic thinking. Analytical models can then be developed and applied to test assumptions, refine the conceptual model, and present alternative futures. The modeling of complex systems frames the need to redesign elements of the health system, and modeling is the process through which complex change is framed and analyzed. The analytical process is too complex to be carried out using traditional data and information-driven models. In exhibit 5.2, modeling transforms the system structure, making it the independent variable. Dynamic modeling allows the transformation to be accomplished through complex analytical models, changing financing, organizational design, and other system components until the system achieves an optimal state. Implementing the solutions is difficult but can chart a policy direction toward transformation. Organizational leaders who are engaged in the process of IT design bring with them, to some degree, traditional ways of thinking, professional

107

Modeling Method of studying, understanding, and then replicating the complexities of the real world in order to design, change, and improve systems

108

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

EXHIBIT 5.2 Modeling and Analytics

System structure Community of practice

Training A

Internet, social media

Scientific evidence B

7. Modeling 1. EMR System

2. Decision support (EHR) e-trials

3. E-trials (HIE)

5. Integrated PHR E

C

D Experiential knowledge

Decision

Outcome–decision relationship

Knowledge accumulation

6. Data mining, analytics

4. Networked HIE, health data vault

Note: EHR = electronic health record; EMR = electronic medical record; HIE = health information exchange; PHR = personal health record.

and institutional perspectives, and a desire for stability. All of these can impose demands that conflict with innovative IT models and change. For organizational leaders to effectively develop new models, they need tools that frame the issues and test alternative assumptions. These tools draw from several theories, including knowledge management, organizational learning, and complex systems. Building conceptual mental models is a process of finding a common framework. It draws on the perspectives of each participant but also challenges and extends these perspectives. An outside facilitator might be helpful in this exercise if the leadership team is new or if the institution is transitioning from a defender (relies on existing strategies) or reactor (reacts to others) to a prospector (innovates) (Miles and Snow 1978). The pursuit of coming together as a team to build knowledge-based IT has increased the amount of research on complexity leadership (Lichtenstein and Plowman 2009). Complexity leadership is characterized by leadership teams that interact within and across organizational domains and that emphasize both informal and positional leaders (Hanson and Ford 2010). Exhibit 5.2 adds modeling as a means of increasing knowledge for both clinicians and organizational leaders. As mentioned, data mining brings together disparate databases and identifies relationships that raise important questions about service delivery and serve as a framework for new delivery

C h a p te r 5 :   Pre d i c tive A naly tic s in Knowledge Managem ent

109

models. Modeling differs in concept from various data-mining techniques that provide information and knowledge extracted from various databases (Pasupathy 2010). Data mining increases knowledge of clinical and operational decisions and is useful for framing larger systems questions. However, to increase systems knowledge and engage in transformational decision making, IT should progress from an operational perspective to a strategic and systems perspective. Because today’s healthcare organizations are highly complex, an individual’s mental models or data mining alone cannot provide the necessary analytics and evidence to support learning and decision making. Mental models are restricted by personal assumptions, which an individual has learned or observed from life experiences. Modeling is a method of studying, understanding, and then replicating the complexities of the real world to design, change, and improve information systems. This process, in turn, introduces and increases learning. The system structure thus becomes dynamic (exhibit 5.2). Modeling also helps to validate the information generated by data-mining models. A simple modeling process is shown in exhibit 5.3. Here, information from the real world provides some understanding that forms the paradigm or mental model of the decision makers. The modeler uses this paradigm to build a model. The simulated model, along with additional information from the real world (e.g., through data mining), brings about a paradigm shift. This process is repeated until the model represents the real world sufficiently, and the decision maker gains knowledge or learns along the way. The model built can then be used to predict future states of the system and to evaluate what-if scenarios. Finally, the decision maker uses this information and knowledge to act in the real world. Models can be statistics based, optimization based, or simulation based (Crown et al. 2017; Marshall et al. 2015a; Marshall et al. 2015b; Marshall et al. 2016). These models can be used to study static and dynamic systems. Because healthcare organizations have a high degree of variability and are dynamic, our discussion focuses on simulation-based models. EXHIBIT 5.3 The Modeling Process

Real world

Model Information

Action Paradigm

110

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Models do not come without limitations. Models are all “wrong” from an idealistic standpoint because they cannot represent each and every aspect of reality. However, they can replicate the essential elements of a system and, in doing so, frame key questions and propose approaches to solutions (Sterman 2002). The validity of models is important to ensure that they succinctly and sufficiently capture the real-world system. Models are built by human beings and are restricted by the assumptions built into the paradigms (Silvert 2001). However, the modeling process in itself helps to challenge and break down mental models (or silo thinking) and prompts paradigm shifts. In short, models can evaluate existing systems and what-if scenarios and predict future states. Modeling has three basic types, which should be considered when testing different assumptions and interrelationships: discrete event, agent based, and system dynamics (Marshall et al. 2015a; Marshall et al. 2015b; Marshall et al. 2016).

Discrete Event Modeling Discrete event (DE) modeling Type of modeling used primarily to study processes, streamline them, and reduce bottlenecks through better resource allocation, capacity utilization or standardization, and mechanization of routine processes

Discrete event (DE) modeling is used to study processes, streamline them, and reduce bottlenecks through better resource allocation, capacity utilization or standardization, and mechanization of routine processes (Law and Kelton 1991). This type of modeling has been used in healthcare settings over the years (Brandeau 2004; Murray and Berwick 2003; Schaefer et al. 2004). Most of these applications have been used to improve processes within (but not to restructure) healthcare organizations and systems—that is, the operations in a system, termed decomposed systems, where the process has the greatest potential value. This use is consistent with the typical application of DE modeling. Examples of such applications include centralizing the information center in physician clinics to manage all nonmedical operations (Swisher et al. 2001) and studying the flow of medication orders in inpatient pharmacy settings (Buchanan 2003; Dean et al. 1999; Ghandforoush 1993; Shimshak, Gropp Damico, and Burden 1981; Zhang and Pasupathy 2009). The medical informatics and bioinformatics literature also describes the use of complex systems models to understand the clinical progression of specific diseases, which is a context similar to data mining (Cross et al. 2008). The application of DE models typically focuses on operations (including processes that transcend professional and institutional boundaries) within a fixed, stated, or assumed strategy (Fowler and Schömig 2003). Application of DE modeling to knowledge systems, however, should focus on operations that are considered within the dynamics of alternative strategies. Fowler and Schömig (2003) are correct in stating the following: It is clearly erroneous to delegate operations to the status of an afterthought. Rather it should, as the custodian of the value adding process, be promoted into a much

C h a p te r 5 :   Pre d i c tive A naly tic s in Knowledge Managem ent

111

more central role as a potentially rich source of understanding and creativity within the dynamic, systemic paradigm of strategic thinking.

Knowledge-based systems require insights into the dynamic workings of the total system and the total value-added potential of work processes and their interrelationships. Such an approach focuses on information systems to support operations within alternative strategies and not operations within a given structure. Within such a model, information can and frequently does itself become a strategic asset (Davenport and Glaser 2002). Because DE modeling concerns the flow of entities (e.g., patients, medication orders), understanding queuing theory is necessary. The skills required to perform DE modeling include flowchart or process mapping, data collection, fitting arrival and service distributions, model building in simulation software, and quantitative analysis for staffing or capacity planning. Software that performs DE modeling includes Arena (www.arenasolutions.com), ProModel (www.promodel.com), and Simio (www.simio.com).

Agent-Based Modeling Agent-based (AB) modeling is used to study the behavior of systems according to the interactions among agents or entities. Such an approach uses the behavior of individual agents under given circumstances to model the overall changes in the system over time. Departments or divisions, people, projects, products, and services have served as agents to be modeled in traditional applications in health systems. This approach is also related to DE modeling or the bottom-up approach, where the agent is typically an individual or a functional area of an organization. Even applications that seem to be systems based are driven by individuals (agents) and their behavior. For example, AB models are applied in public health informatics to study patterns of individual behavior and their relationships, such as the link between exercise and health status and the effect of food prices on dietary disparity as a result of low income (Maglio and Mabry 2011). Such analysis derives information from the PHR to inform the patient (consumer) and the clinical process, but it does not address the issues that alter the structure of the clinical process. Specifically, it provides the basis for examining the relationships among sectors, such as health, social services, and education, that have traditionally been structured and function in relative isolation. This examination can lead to system transformation, broadening the concept of community of practice. The strength of AB modeling is its interdisciplinary nature; that is, it can synthesize knowledge from different disciplines. The disciplines relevant to the discussion depend on the level of analysis within the system. Health systems informatics addresses the complexity of sectors and systems to enable leaders to develop the most complete understanding of the interactions among

Agent-based (AB) modeling Type of modeling used to study the behavior of systems on the basis of the interactions among agents or entities

112

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

disparate systems. Here, the question is not the relevance of the models applied to the problem but the problem that these models address. Some of the theories underlying AB modeling are game theory, artificial intelligence, and complexity science. The skills required to perform this approach include developing state charts, modeling interactions, collecting data, building models in simulation software, and performing scenario analysis. For instance, AB modeling can be used to study a wide range of systems from the micro to the macro levels. Examples at the individual level include smoking patterns (smokers can influence their friends to take up smoking) and sexual behavior (teenagers without proper sex education are likely to engage in unprotected sex). Organizational-level examples are interactions among agents (stakeholders) such as patients, clinicians, and hospital administrators regarding care provision and insurance coverage. Software for AB modeling includes AnyLogic (www. anylogic.com) and NetLogo (https://ccl.northwestern.edu/netlogo/).

System Dynamics Modeling System dynamics (SD) modeling Type of modeling used to model complex, nonlinear relationships between components and to study the dynamics of the system over time

System dynamics (SD) modeling is used to model complex, nonlinear relationships between components and to study the dynamics of the system over time. This framework operates under the premise that structure predicts behavior over time. The underlying principle behind SD modeling is complexity theory. Complexity theory views the organization as a learning system (Senge 1990) that uses knowledge to drive the organization’s strategies and structures. In addition, complexity theory challenges the school of thought that promotes prescriptive structures, plans, and strategies (Fowler and Schömig 2003). Leadership in such learning organizations is team based, includes informal leaders, and guides a process in which strategies and vision emerge as an evolving supra system (a larger unit that encompasses subsystems and components). The skills most relevant for SD modeling include systems thinking, cause-and-effect formulation, data collection, stock-and-flow modeling, differential and integral calculus, model building in simulation software, and analysis for decision making and policymaking. Several SD models have been developed over the years and are discussed in the literature (Lattimer et al. 2004; Sterman 2000; Taylor and Dangerfield 2005). Software used for SD modeling includes Vensim (https://vensim.com/ vensim-software) as well as iThink and STELLA (www.iseesystems.com). In the healthcare field, SD modeling has been applied primarily to freestanding institutions and population groups, but it has equal (if not greater) potential when applied to higher-order systems (Hirsch 1979). The traditional application is due not to the limitation of SD modeling but rather to the lack of vision of the leaders who use it. The value given to higher-order systems applications has been low because of the highly regulated and subsidized nature of the healthcare industry, the power held by health professionals, and healthcare

C h a p te r 5 :   Pre d i c tive A naly tic s in Knowledge Managem ent

Modeling Approach

Level of Abstraction

Time Basis Amount of for Simulation Detail Needed Uses

Discrete event (DE)

Middle to low

Discrete

Medium to high

Tactical, process

Agent based (AB)

Wide range (high, medium, and low)

Discrete

Varies depending on the model

Multiple

Mostly continuous

Less detail

Strategic decision, policy analysis

System High dynamics (SD)

leaders’ traditional orientation toward the business function. Current US health policy that seeks to regulate the health system to become high performing and integrated is a fallacy. Government initiatives and regulations can cause disruption that might lead to the health system transforming itself. The challenge to enterprise leaders has always been developing systems that are innovative and address the issues of cost, quality, safety, continuity of services, wellness, and patient centeredness. These issues are not new, but they require new solutions.

Comparison of the Three Modeling Approaches The three types of modeling differ in their level of abstraction and the amount of detail required for modeling (exhibit 5.4). DE modeling is at the middle to low level of abstraction, but it requires a medium to high amount of detail for modeling purposes. SD modeling is at the high abstraction level and is used for strategic decision making and policy analysis. AB modeling spans a wider range of abstraction, including high, middle, and low levels of abstraction. DE and AB modeling are predominantly discrete, whereas SD modeling is continuous. Because health systems are highly complex, one modeling approach does not suffice to capture all aspects (strategic, tactical, and operational) of a real-world system. A combination of modeling approaches may be used. For example, DE and SD modeling can be used together to study the nuances of operations and to learn the impact of strategic decisions. Regardless of the type applied, modeling helps in better understanding systems and increasing the knowledge and learning of workers (chapter 14).

Conclusion According to the literature on decision making, the first step in the process is to identify the correct problem. This chapter focuses on defining the problem,

113

EXHIBIT 5.4 Three Modeling Approaches

114

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

using complex health systems as the frame of reference. This is the correct perspective for strategies that aim to design a knowledge-based information system. Almost any analysis or use of evidence will add knowledge to a process, but it does not necessarily maximize the performance of the process. Such an approach might undervalue IT as a knowledge asset. An information system is optimized when it contains the collective knowledge of all agents, who engage in dialogue to define the problem and develop an optimal solution strategy. The EHR, HIE, and individual clinicians and institutions are all valuable sources of data and knowledge. The application of the best evidence derived from CDSSs has been shown to improve clinical outcomes and safety. However, such systems draw on only a fraction of the clinical data available within these records, some of which could be used to provide tailored solutions to individual clinical problems. Data mining represents a range of analytical techniques for accessing and analyzing such data and, in so doing, increases the volume of information and knowledge and the level of evidence that can be brought to bear in a given clinical decision. CDSSs exist within given structural contexts and do not in themselves provide the basis for analyzing and improving the performance of the overall system to optimize performance, including clinical outcomes. Networked HIEs and health data vaults provide an information system that could be the basis for analyzing system performance for individual treatment measured as clinical outcomes, including wellness and cost. Dynamic systems modeling can provide the basis for innovation and the transformation of health system structures. Systems change is complex and examines the structure of clinical processes and organizations; the role of the professions, financing, law, and regulations; and the potential contribution of other sectors such as education and social services. Innovation is needed in the health system and supported by analytical models for testing complex interrelationships based on alternative assumptions. Such models provide support for innovation by visionary leaders and innovation centers that build and test conceptual and physical models. The models in themselves will not cause change and must not be limited to research centers. Models must become the tools of systems leaders to chart a direction for change. That part will be easy; more difficult is initiating change. If innovation becomes the strategy of healthcare organizations, the models can direct an extremely complex process. A system that is truly knowledge based focuses on operations but is also innovative and transformational. This is the contribution of the science of health systems informatics.

Chapter Discussion Questions 1. Identify technical and clinical factors that limit the application of datamining tools in the EHR.

C h a p te r 5 :   Pre d i c tive A naly tic s in Knowledge Managem ent

2. What information might be included in an analytical model for identifying individuals in a given diagnostic category who have the most repeat visits and consume the most resources? Where would you go to access this information? 3. What impediments do healthcare leaders face in applying modeling to guide the transformation of the health system? 4. What additional evidence for improving clinical outcomes is introduced by analytical modeling that is not included in clinical guidelines that are based on a high level of evidence from clinical trials? 5. Differentiate the three types of modeling, and give examples of how each might be applied to the health system.

Case Study  Analytics for Disease Management and Wellness Central Medical is a multispecialty group practice that has embraced the community-of-practice concept of an ACO. The practice formed an interdisciplinary innovation team to identify how to improve clinical outcomes, maintain a healthy population, increase efficiency, and coordinate care. It comprises 60 physicians and four clinics. Two of the clinics deliver primary care and focus on family medicine, internal medicine, and obstetrics/ gynecology. One is a specialty clinic with a multidisciplinary staff devoted to metabolic disorders, and the other (the largest) offers a range of clinical specialties. All four clinics are centrally managed with an integrated EMR, which has the capacity to present clinical information in meaningful and actionable ways, including trend lines for patients, evidence-based clinical guidelines, and integrated treatment protocols. Central Medical has multispecialty teams whose composition is tailored to the needs of the particular illness. The team members include primary care physicians, chronic care nurses, therapists, nutritionists, and health educators. After the start-up period, these teams developed a level of comfort and respect that enables the team members to bring the best discipline-based evidence to support diagnosis and treatment, healthy patient lifestyles, and the total-person concept of patient care. Because of its emphasis on improving patient behavior and knowledge about healthy lifestyles, Central Medical decides to invite patients with specific diseases to join the multispecialty team focused on that illness. Two patients and their respective family caregivers are added to the committee organized to identify and treat high-risk, chronic-care patients. The innovation team identifies the investment in institutional capacity for processing information as a key strategic resource, where knowledge on wellness and health (continued)

115

116

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

maintenance can be embedded in treatment protocols. The team agrees that the planned strategy’s focus should be on quality, safety, efficiency, and alignment of the clinical and financial functions. This approach is consistent with the philosophy and values set forth by the ACO—to coordinate care and address the overall health of patients, not just treat their illness. The patient population is distributed, demographically diverse, and covered by a range of insurance companies with different eligibility criteria and benefit packages. Among the insurance providers is Health First, a large capitation-based plan with more than a million members, including 20,000 of Central Medical’s patients. People are generally satisfied with Health First, in part because it enrolls a significant number of Central Medical’s patients. However, its mission does not align with that of the practice. Central Medical has an incentive to negotiate higher capitation rates, which will enable the practice to realize a financial return by aggressively managing utilization. Health First is motivated to (1) set lower capitation rates because of the aggressive market, where businesses are willing to change insurers for small differences in cost, and (2) shift the risk to Central Medical. As a strategy, Central Medical enters negotiations with Health First, basing its utilization rates and costs on historical performance and taking a strong position on setting higher capitation rates. The practice believes patients will stay loyal to the clinics because of the high quality of staff and strong patient orientation. Both organizations accept their adversarial relationship, which they consider to be inherent in the healthcare and health insurance industries. Central Medical’s innovation team moves forward with developing a strategy for carrying out its mission. It focuses on disease management of complex chronic illnesses, such as type 2 diabetes, as well as accessing and using evidence-based clinical guidelines. The practice discovers from available literature that intensive interventions, such as life coaches, demonstrate a 20 percent reduction in glycosylated hemoglobin (HbA1c) within 6 months for some patients. However, the team cannot justify the cost of adding the staff needed for intensive health maintenance for its sizable type 2 diabetes patient population. This particular population is inherently high risk, although some patients with the same diagnosis use resources at a much higher rate than others. If the team could systematically identify diabetes patients who are at highest risk, it could better focus its intensive maintenance strategy, improve health, and increase efficiency. The challenge is identifying the characteristics of the patients with the highest risk in this high-risk population. These parameters are not revealed in the

C h a p te r 5 :   Pre d i c tive A naly tic s in Knowledge Managem ent

existing evidence from systematic reviews. Senior clinicians suggest relying on the clinical judgments of individual clinicians; although logical, this approach cannot accurately identify the highest-risk patients because there are complex interdependencies. The team decides to analyze the collective medical records of its physician panel, but the predictive models lack rigor because of low population samples. During a brainstorming session, the team proposes a collaboration with Health First to use Big Data to identify the highest-risk patients on the basis of their rate of resource utilization. From analyzing Health First’s population-based enrollment data, the team is able to demonstrate that 1 to 2 percent of all patients with diabetes account for up to 30 percent of the total costs for this diagnostic group. Variables used in the predictive models include total annual prescriptions, unique (disease-specific) annual prescriptions, physician visits, hospital utilization (including emergency services), comorbidities, age, gender, occupation, family composition, benefit coverage, and treatment history. The team concludes that, by deploying a predictive model and a focused strategy, the information system could identify patients from a high-risk population who have the highest risk as well as a range of targeted interventions, such as life coaches and other intensive treatments. One team member says, “Precision medicine from a managed care firm. Who would believe it!”

Case Study Discussion Questions 1. Is the decision to provide intensive therapy for this high-risk population evidence based? How does it relate to clinical judgment? 2. What are the advantages of the Health First database? What are some concerns about the reliability and predictive validity of the Health First data set? 3. Are insurance claims data valid measures of clinical diagnosis and treatment? 4. How many variables used in the data-mining analysis of Health First data are available in the EHR? 5. Is there sufficient data in Central Medical’s EHR, and in a format to be mined, to identify these high-risk patients? Are there enough data to apply the results of the analysis to manage the high-risk patients? 6. What are the implications of, and potential for, aligning the strategies and corporate interests of Central Medical and Health First for better serving patients?

117

118

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

References Brahman, R. J., and H. J. Levesque. 2004. Knowledge Representation and Reasoning. San Francisco: Elsevier. Brandeau, M. L. 2004. “Allocating Resources to Control Infectious Diseases.” In Operations Research and Health Care: A Handbook of Methods and Applications, edited by M. L. Brandeau, F. Sainfort, and W. P. Pierskalla, 443–64. Boston: Kluwer Academic Publishers. Buchanan, E. C. 2003. “Computer Simulation as a Basis for Pharmacy Reengineering.” Nursing Administration Quarterly 27 (1): 33–40. Cross, S. S., F. C. Hamby, J. R. Goepel, and R. F. Harrison. 2008. “Prostate Cancer: A Systems Approach Overview.” Diagnostic Histopathology 14 (3): 122–33. Crown, W., N. Buyukkaramikli, P. Thokala, A. Morton, M. Y. Sir, D. A. Marshall, J. Tosh, W. V. Padula, M. J. Ijzerman, P. K. Wong, and K. S. Pasupathy. 2017. “Constrained Optimization Methods in Health Services Research—an Introduction: Report 1 of the ISPOR Optimization Methods Emerging Good Practices Task Force.” Value in Health 20 (3): 310–19. Davenport, T. H., and J. Glaser. 2002. “Just-in-Time Delivery Comes to Knowledge Management.” Harvard Business Review 80 (7): 107–11. Davenport, T. H., J. G. Harris, and R. Morrison. 2010. Analytics at Work: Smarter Decisions, Better Results. Cambridge, MA: Harvard Business School Publishing. Dean, B., A. V. Ackere, S. Gallivan, and N. Barber. 1999. “When Should Pharmacists Visit Their Wards? An Application of Simulation to Planning Hospital Pharmacy Services.” Health Care Management Science 2 (1): 35–42. Fowler, J. W., and A. K. Schömig. 2003. “Simulation of Manufacturing Systems.” In Applied System Simulation, edited by M. S. Obaidat and G. I. Papadimitriou. Boston: Springer. Ghandforoush, P. 1993. “Optimal Allocation of Time in a Hospital Pharmacy Using Goal Programming.” European Journal of Operational Research 70 (2): 191–98. Hanson, W. R., and R. Ford. 2010. “Complexity Leadership in Healthcare: Leader Network Awareness.” Procedia—Social and Behavioral Sciences 2 (4): 6587–96. Hirsch, G. B. 1979. “System Dynamic Modeling in Health Care.” ACM SIGSIM Simulation Digest 10 (4): 38–42. IBM. 2011. Analytics: The New Path to Value. Armonk, NY: IBM Institute for Business Value. Lattimer, V., S. Brailsford, J. Turnbull, P. Tarnaras, H. Smith, S. George, and S. MaslinProthero. 2004. “Reviewing Emergency Care Systems I: Insights from System Dynamics Modeling.” Emergency Medicine Journal 21 (6): 685–91. Law, A. M., and W. D. Kelton. 1991. Simulation Modeling and Analysis. New York: McGraw-Hill. Lee, J. Sr. 2000. “Knowledge Management: The Intellectual Revolution.” HE Solutions 32 (10): 34–37.

C h a p te r 5 :   Pre d i c tive A naly tic s in Knowledge Managem ent

Lichtenstein, B. B., and D. A. Plowman. 2009. “The Leadership of Emergence: A Complex Systems Leadership Theory of Emergence at Successive Organizational Levels.” Leadership Quarterly 20 (4): 617–30. Maglio, P. P., and P. L. Mabry. 2011. “Agent-Based Models and Systems Science Approaches to Public Health.” American Journal of Preventive Medicine 40 (3): 392–94. Marshall, D. A., L. Burgos-Liz, M. J. Ijzerman, W. Crown, W. V. Padula, P. K. Wong, K. S. Pasupathy, M. K. Higashi, N. D. Osgood, and ISPOR Emerging Good Practices Task Force. 2015a. “Selecting a Dynamic Simulation Modeling Method for Health Care Delivery Research—Part 2: Report of the ISPOR Dynamic Simulation Modeling Emerging Good Practices Task Force.” Value in Health 18 (2): 147–60. Marshall, D. A., L. Burgos-Liz, M. J. Ijzerman, N. D. Osgood, W. V. Padula, M. K. Higashi, P. K. Wong, K. S. Pasupathy, and W. Crown. 2015b. “Applying Dynamic Simulation Modeling Methods in Health Care Delivery Research—the SIMULATE Checklist: Report of the ISPOR Simulation Modeling Emerging Good Practices Task Force.” Value in Health 18 (1): 5–16. Marshall, D. A., L. Burgos-Liz, K. S. Pasupathy, W. V. Padula, M. J. Ijzerman, P. K. Wong, M. K. Higashi, J. Engbers, S. Wiebe, W. Crown, and N. D. Osgood. 2016. “Transforming Healthcare Delivery: Integrating Dynamic Simulation Modeling and Big Data in Health Economics and Outcomes Research.” Pharmacoeconomics 34 (2): 115–26. Miles, R. E., and C. C. Snow. 1978. Organizational Strategy, Structure and Process. New York: McGraw Hill. Murray, M., and D. M. Berwick. 2003. “Advanced Access: Reducing Waiting and Delays in Primary Care.” Journal of the American Medical Association 289 (8): 1035–40. Pasupathy, K. 2010. “Transforming Healthcare: Leveraging the Complementarities of Health Informatics and Systems Engineering.” International Journal of Healthcare Delivery Reform Initiatives 2 (2): 34–54. Schaefer, A. J., M. D. Bailey, S. M. Shechter, and M. S. Roberts. 2004. “Medical Treatment Decisions Using Markov Decision Processes.” In Operations Research and Health Care: A Handbook of Methods and Applications, edited by M. L. Brandeau, F. Sainfort, and W. P. Pierskalla, 595–614. Boston: Kluwer Academic Publishers. Senge, P. M. 1990. The Fifth Discipline. New York: Currency Doubleday. Shimshak, D. G., D. Gropp Damico, and H. D. Burden. 1981. “A Priority Queuing Model of a Hospital Pharmacy Unit.” European Journal of Operational Research 7 (4): 350–54. Silvert, W. 2001. “Modeling as a Discipline.” International Journal of General Systems 30 (3): 261–82. Sterman, J. D. 2002. “All Models Are Wrong: Reflections on Becoming a Systems Scientist.” System Dynamics Review 18 (4): 501–31.

119

120

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

——— 2000. Business Dynamics: Systems Thinking and Modeling for a Complex World. New York: Irwin McGraw-Hill. Swisher, J. R., S. H. Jacobson, J. B. Jun, and O. Balci. 2001. “Modeling and Analyzing a Physician Clinic Environment Using Discrete-Event (Visual) Simulation.” Computers and Operations Research 28 (2): 105–25. Taylor, K., and B. Dangerfield. 2005. “Modeling the Feedback Effects of Reconfiguring Health Services.” Journal of the Operational Research Society 56 (6): 659–75. Theodoridis, S., and K. Koutroumbas. 2009. Pattern Recognition. London: Academic Press. Wickramasinghe, N., R. K. Bali, B. Lehaney, J. L. Schaffer, and C. M. Gibbons. 2008. Healthcare Knowledge Management Primer. London: Routledge; Taylor and Francis Group. Witten, I. H., and E. Frank. 2005. Data Mining: Practical Machine Learning Tools and Techniques. San Francisco: Morgan Kaufmann. Zhang, B., and K. Pasupathy. 2009. “Integration of Simulation Modeling into Hospital Pharmacy Delivery Network Planning.” INFORMS Conference Proceedings, San Diego.

CHAPTER

CLINICAL DECISION SUPPORT SYSTEMS IN MEDICINE

6

Pavithra I. Dissanayake and Karl M. Kochendorfer

Learning Objectives After reading this chapter, you should be able to do the following: • • • •

Define clinical decision support systems. Discuss the policies that drive clinical decision support systems. Explain the characteristics of effective clinical decision support systems. Understand the design and implementation of clinical decision support systems. • List barriers to adoption of clinical decision support systems.

Key Concepts • • • •

Meaningful use stage 3 requirements Artificial neural network and genetic algorithm Ten Commandments and the Five Rights Challenges and barriers to implementation

Introduction Decision support systems are information systems that facilitate the decisionmaking process by automatically performing some sets of tasks. Studies have shown that health information technology (HIT) has the potential to improve the quality of care provided to patients, increase provider performance and patient engagement, produce cost savings for individuals and institutions, prevent errors, reduce adverse events, and enhance patient health overall (Garg et al. 2005). To achieve these improvements, however, electronic health record (EHR) systems must incorporate and integrate tools that are generally known as clinical decision support systems (CDSSs) and include computerized physician 121

122

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

order entry (CPOE) (Berner and La Lande 2007; Buntin et al. 2011; Osheroff et al. 2012). Over the past few decades, CDSSs have evolved to become both strategic and tactical tools that offer strategic benefit to healthcare organizations by tactically assisting clinicians during the patient care process. They aid and augment but do not replace human intelligence (Brown, Patrick, and Pasupathy 2013). This chapter presents an overview of CDSSs in the United States, including the policies and regulations governing them, the characteristics that make them effective, the challenges and limitations of designing and implementing them, and some tools that have been used successfully in today’s clinical practice.

Definition Just as our knowledge of and experience with clinical decision support has improved over time, its definition has evolved to accommodate evolving healthcare needs. Several definitions are currently in use, and they vary from very targeted and precise to all-encompassing. Middleton, Sittig, and Wright (2016) offer the following: A process for enhancing health-related decisions and actions with pertinent, organized clinical knowledge and patient information to improve health and healthcare delivery.

Across all the available definitions, the overarching theme is that CDSSs provide clinicians, staff, patients, and other individuals with knowledge and personspecific information that is intelligently filtered and presented at the appropriate time to facilitate the delivery of care (Berner 2009; Osheroff et al. 2012).

History and National Policies In the 1950s, decision support in healthcare focused on the use of electronic systems to assist health professionals with the clinical diagnostic process. In 1959, Ledley and Lusted, in what is considered to be the first work in medical informatics, explored the reasoning foundations of medical diagnosis in terms of a mathematical model based on logic, probability, and value theory (Ledley and Lusted 1959). These initial systems were generally linear and Bayesian in nature. The late 1960s saw the beginnings of an integrated CDSS, with pioneering studies outlining the use of linear associations between clinical input data to suggest a diagnosis and/or therapy according to probability (Bleich 1969; Collen et al. 1964; Greenes 2006; Warner et al. 1961; Wright and Sittig 2008).

C h a p te r 6:   C l i n i c a l D ec ision Sup p or t Sy stem s in Medic ine

Information systems with basic functionality were first introduced in the hospital setting in the early 1990s. These early CDSSs, however, were only seen at academic institutions that developed their own systems (Middleton, Sittig, and Wright 2016). Early CDSSs were mostly concentrated on medication management, laboratory testing, and quality assurance. In late 1999, with the publication of To Err Is Human, the Institute of Medicine shifted the focus of healthcare to improving patient safety by preventing medical errors through use of HIT (Donaldson 2008; Kohn, Corrigan, and Donaldson 2000). Multiple advances were made following this publication, including HIT adoption by health systems and the development of government policies related to HIT and CDSS use. The development of standard data architecture, Health Level Seven International (HL7), and standard terminologies such as International Classification of Diseases (9th and 10th revisions) and Systematized Nomenclature of Medicine Clinical Terms (SNOMED CT) has allowed the development of better-integrated CDSSs (Ahmadian et al. 2011; Boxwala et al. 2004; El-Sappagh and El-Masri 2014; Mead 2006). Despite increasing evidence supporting the benefits of HIT and CDSSs, their adoption in healthcare was slow until the establishment of the National Quality Strategy under the Affordable Care Act (ACA) of 2010, which aimed to improve the quality of healthcare delivered in the United States. To further this goal, the government passed the American Recovery and Reinvestment Act (ARRA), which included the Health Information Technology for Economic and Clinical Health (HITECH) Act, in 2009. HITECH is intended to ensure that healthcare institutions implement EHR systems and use them in a meaningful manner—that is, in a way that improves the quality of care (Blumenthal and Tavenner 2010; US Department of Health and Human Services 2011; Lyman et al. 2010; Mardon et al. 2014; Reider 2016). The Centers for Medicare & Medicaid Services (CMS) was tasked with setting specific meaningful use goals (CMS 2016a), and the CDSS has been leveraged as a tool to achieve these goals. Stage 2 of meaningful use requires healthcare institutions to implement a minimum of five CDSS interventions associated with four or more clinical quality measures. In 2015, CMS published meaningful use stage 3 requirements, which further expand CDSS use by mandating the application of clinical decision support to drug–drug and drug–allergy interactions (CMS 2016c). Given the increased demands to deliver a higher quality of care and the potential utility of a CDSS as an effective tool for meeting these demands, numerous software companies have developed innovative decision support products that target different aspects of healthcare. The US Food and Drug Administration (FDA) is tasked with regulating CDSS software and devices. The 21st Century Cures Act, enacted in 2016, redefined device to help the agency in generating guidelines and determining what devices fit under FDA regulation (FDA 2016, 2018). The FDA is (at the time of this writing) in the process of drafting a regulatory framework and strategies. This light regulatory

123

Meaningful use Measure of the level of application of IT in clinical decision support systems

124

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

touch has created a vacuum that has spurred members of the software and healthcare industries to form the Clinical Decision Support Coalition, which sets self-regulatory standards and encourages the FDA to establish more defined guidelines (Karnik 2014; Thompson 2017). Multiple attempts have been made to describe the evolution of CDSSs. Bokhari (2017) explains this evolution in terms of the underlying design and technical complexity of CDSSs, dividing it into four phases: legacy decision support, smart decision support, intelligent decision support, and artificial intelligence and artificial neural networks (exhibit 6.1). Wright and Sittig (2008) also frame this evolution in four phases—stand-alone systems, integrated systems, standards-based systems, and service models—depending on the level of CDSS integration with the clinical workflow.

CDSS Types A wide range of CDSS applications are used, both synchronously and asynchronously. These CDSSs differ in the type of output, type of target audience, and type of information used as the knowledge base. Exhibit 6.2 lists the types that are currently used. EXHIBIT 6.1 Evolution of Clinical Decision Support Systems

Legacy Decision Support

Smart Decision Support

Intelligent Decision Support

Example from car industry

• Traditional cruise • Adaptive cruise control control

Description

• Relies on • Aids decision mak- • discrete data ing by presenting • Historically recommendations • evidence based • Primarily relies on • Simpler rules and discrete data alerts • Involves multiple algorithms •

Clinical example

• Drug–drug interactions • Dose range • Allergy alerts

• Lane departure • Autonomous driving • Blind spot warnings and assist Learning and • reasoning Presents precise decision and reasoning for an action • Real-time surveillance of discrete • and non-discrete data • Involves complex • algorithms

• Venous throm• boembolism prophylaxis • Radiology advisors • • Sepsis alert •

Source: Adapted with permission from Bokhari (2017).

Artificial Intelligence and Artificial Neural Network

Learning, reasoning, forecasting, and answering “what ifs” (run simulations) Implements decisions automatically No action needed from the decision maker Highly complex

Venous throm• Oncology boembolism • Radiology prophylaxis • Pathology Radiology advisors Sepsis alert

C h a p te r 6:   C l i n i c a l D ec ision Sup p or t Sy stem s in Medic ine

Type

Example

Alerts and reminders

• • • • • •

Drug–drug interactions Drug–allergy interactions Dose range checking Reminders for labs Cancer screenings Duplicate drugs and labs

Documentation • Initial assessment forms templates • New patient intake forms • Urgent care visit forms • Malignant protocols in pathology Order sets

• Disease-specific order sets • Admissions order sets • Discharge order sets

Knowledge and • Clinical guidelines data displays • Relevant policies • Point-of-care reference information • Reference websites (e.g., NIH, CDC) • Drug-related information from vendors Diagnostic deci- • Image analysis sion support • Electrocardiogram analysis tools • Differential diagnosis tools Disease • Care plans management • Apps to monitor blood pressure for hypertension and prevention management • Apps to monitor glucose levels for diabetes management Management and strategic assistance system

• • • •

Tracking length of hospital stay Tracking antibiotic usage Tracking hospital-acquired infectious diseases Assisting with meeting meaningful use goals

Note: CDC = Centers for Disease Control and Prevention; NIH = National Institutes of Health.

A number of attempts have been made to categorize CDSSs, and multiple taxonomy frameworks have been published. Metzger and colleagues categorize CDSSs by the time of support, whether the support is active or passive, and the ease of access by the user (Miliard 2015; Perreault and Metzger 1999). Berlin and colleagues also propose a multifaceted CDSS taxonomy with five categories: context, knowledge and data source, decision support, information delivery, and workflow (Sim and Berlin 2003). Traditionally, CDSSs have been classified as either knowledge based or non–knowledge based.

125

EXHIBIT 6.2 CDSS Types and Examples

126

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Knowledge-Based CDSSs Many early and current CDSSs are knowledge based. They present clinicians with information that assists with decision-making processes, rather than making the decisions for clinicians. Clinicians are expected to take into consideration the information provided and then make the final determination (Berner and La Lande 2007). A knowledge-based CDSS comprises three components: a knowledge base, an inference engine, and an interface or communication mechanism (Berner and La Lande 2007; Brown, Patrick, and Pasupathy 2013). Much of the electronically accessible scientific and medical information, as well as the pertinent individual patient information found in the EHR, must be connected for the CDSS to function properly. Moreover, the knowledge base must encompass the rules and associations of the compiled information. These rules and associations are usually in the form of an if-then scenario—that is, if a certain condition is true, then a specific defined action should be taken (Berner and La Lande 2007). Using structured logic or formulas, the inference engine combines the rules of the knowledge base with the patient information in the EHR. The interface provides a platform for collecting information, such as the patient’s condition, that must be entered (data input) into the EHR and for presenting the outcome or result (data output) to the user.

Non-Knowledge-Based CDSSs Non-knowledge-based CDSSs employ artificial intelligence algorithms along with machine learning. Machine learning allows the computer program to analyze the data, recognize patterns in the data, and learn and adapt with experience. This type of program is further divided into artificial neural networks (ANNs) and genetic algorithms (GAs). Artificial Neural Networks ANNs are computer programs that, once trained, imitate some of the high-level functions of the human brain (Rodvold et al. 2001). They are composed of nodes with weighted connections between them, much like a human brain with neurons and connections between the neurons (Kononenko 2001; Rodvold et al. 2001). Most ANNs today are organized into three layers: (1) an input layer, where the data are entered into the system; (2) a hidden layer, where the data are processed via weighted connections between the nodes; and (3) an output layer, where the final results from the system are displayed (Berner and La Lande 2007; Singh Gill 2017). The hidden layer may contain one to many sublayers, depending on the complexity of the system; many types of learning rules are available to train the hidden layer of the ANNs. ANNs can be further divided into feedforward and feedback networks, depending on the orientation of the weighted connections (Agatonovic-Kustrin and Beresford

C h a p te r 6:   C l i n i c a l D ec ision Sup p or t Sy stem s in Medic ine

127

2000; Singh Gill 2017) (exhibit 6.3). In feedforward ANNs, the data flow unidirectionally from the input layer through the hidden layer to the output layer. In contrast, in feedback networks, also known as recurrent networks, the information flow is bidirectional, allowing for feedback loops. ANNs have been applied in a variety of areas in healthcare, including diagnostics (e.g., cancer diagnosis), biochemical analysis (e.g., analysis of urine and blood samples, monitoring of glucose levels), analysis of medical images or scans (both in pathology and radiology), and drug development (Patel and Goyal 2007). Genetic Algorithms GAs are selection algorithms based on Darwin’s principles of natural selection and survival of the fittest (Berner and La Lande 2007). A random set of solutions is evaluated to identify the best suited on the basis of a “fitness function” (Aishwarya and Anto 2014; Ghaheri et al. 2015). The solutions with the highest fitness function are combined and altered to produce the next set of solutions for evaluation. This process is repeated until a proper solution is identified. GA allows for complex data analysis, and its application has been explored in many specialties in medicine (Ghaheri et al. 2015). EXHIBIT 6.3 Artificial Neural Network (ANN) Architecture Input

Output

Input layer

Hidden layer

Output layer

A. Feedforward ANN architecture

Input

Output

Input layer

Hidden layer

B. Feedback ANN architecture

Output layer

128

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Effective Characteristics CDSSs are used in almost all aspects of healthcare delivery, from the clinical decision-making process to medication and lab order monitoring to quality assurance and cost control. This broad application is partly the result of national policies and mandates, described earlier. However, studies have found that not all CDSSs are equally effective. In response, experts in the field have proposed the characteristics most likely to produce a more effective CDSS. The Ten Commandments and the Five Rights are two such guides.

The Ten Commandments The actual use of CDSSs by clinicians is less than optimal. Even though a CDSS provides guidance, clinicians often override the system, do not open it when prompted, or simply ignore it. Bates and colleagues (2003) published a set of guidelines—the Ten Commandments for Effective Clinical Decision Support—based on the lessons learned from CDSS implementation. They are listed here along with our explanations: 1. Remember that speed is imperative. Clinicians are inundated with many tasks that require significant time. Therefore, clinicians consider speed to be the most valuable feature (Lee et al. 1996). 2. Anticipate the clinician’s need and provide assistance at the time of need. Information should be provided at the time of need. Information provided well before or after the clinician makes a decision is not useful. 3. Fit the CDSS into the existing workflow. The CDSS’s designers should understand the existing clinical workflow and implement it seamlessly, without asking the clinicians to alter their workflow. 4. Know that usability is essential and nuances will deter utility. Design and layout factors that improve usability (e.g., color palette, font size, ability to see all the data without having to scroll down) encourage use, whereas things that clinicians consider a nuisance discourage use. Usability testing must be done before implementation. 5. Recognize that clinicians will strongly resist stopping an action. Generally, clinicians do not accept CDSSs that simply inform them to stop an action without providing alternatives. 6. Understand that clinicians are amenable to suggested alternatives. CDSSs that attempt to change the clinician’s behavior by suggesting a different direction have been found to be more successful. 7. Know that simple CDSSs that require little additional work from clinicians are the best. CDSSs that require clinicians to perform additional work, such as clicking to open another window or scrolling down, are less effective.

C h a p te r 6:   C l i n i c a l D ec ision Sup p or t Sy stem s in Medic ine

8. Ask for additional information only when absolutely necessary. Clinicians will accept taking the time to provide additional information for a CDSS only if they gain data they deem important. 9. Monitor the CDSS’s impact, get feedback, and respond. Measure how often the CDSS’s suggestions are followed, and get feedback from the clinicians regarding the system. When necessary, modify the CDSS for better usability. 10. Actively maintain the CDSS’s knowledge base. Medicine is a field in which knowledge is constantly changing, guidelines are regularly altered, and medications are steadily introduced or changed. As knowledge changes, the knowledge base of the CDSS should also be updated to ensure the system remains accurate and effective.

The Five Rights Osheroff and colleagues (2012) describe a framework, known as the Five Rights, for successfully developing and implementing a CDSS. To improve outcomes, the CDSS should provide the (1) right information to the (2) right person in the (3) right format via the (4) right channel at the (5) right time in the workflow. These rights address the five basic components of a CDSS—what, who, how, where, and when. The Five Rights guide has become the gold standard for designing a successful CDSS and has been recommended by CMS (2016a) as the best practice approach for effective CDSS-driven healthcare quality improvement.

Design and Implementation In 2005, following a request of the Office of the National Coordinator for Health Information Technology, the Roadmap Development Steering Committee was established by the American Medical Informatics Association to develop guidelines for the effective implementation of CDSSs. Subsequently, Osheroff and colleagues (2007) published a framework for implementation, known as the “three pillars for realizing the promise of clinical decision support,” that included “best knowledge available when needed, high adoption and effective use, and continuous improvement of knowledge and methodology.” Successful implementation of a CDSS requires many steps, as well as collaboration among many members of the healthcare team. This process is the same whether it is applied to a single intervention or a complete system. The article “Advancing Clinical Decision Support: Key Lessons in Clinical Decision Support Implementation” outlines these steps (Byrne et al. 2017): 1. Involve stakeholders. The implementation of a CDSS is an organizationwide project that involves a wide range of health professionals, other providers,

129

130

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

and management. The clinical processes affected by the CDSS should be mapped, and all stakeholders should be identified. The engagement of stakeholders should start with the initial planning phases. The project’s goals should be communicated to all stakeholders early on, and collaboration should continue throughout the project. 2. Assess readiness for implementation. Assessment of readiness, which evaluates the organization’s culture and the viewpoints of the stakeholders, is a crucial early step. It helps the organization understand the stakeholders’ opinions of the project and their willingness to change, and it identifies potential barriers or areas that may require attention during implementation. All stakeholders should be engaged in this assessment. If a lack of readiness is recognized, the implementation of the CDSS should be postponed, and steps should be taken to improve readiness. 3. Assemble the CDSS implementation team. A well-assembled team is critical to the success of the implementation. Defining the goals, scope, and clinical workflows involved, and determining the resources required, will allow the right team members to be identified. The role of each member should be defined. At minimum, managers, physicians, nurses, pharmacists, and information technologists whose input and support are necessary should be represented on the team (Osheroff et al. 2012). These team members will manage the implementation process and act as a bridge to other stakeholders. Once the CDSS has been implemented, the team may become part of the ongoing CDSS governance structure (see Case Study). 4. Select champions and leaders. Champions are individuals in the organization who believe in the project and usually hold a key position in the department they represent. They are well respected by their peers and are committed to the project’s success. Champions play a critical role in the success of any project and should be engaged early on. 5. Gain clinical buy-in and support. To be successful, a CDSS must be used by the clinicians. Although clinicians today may be accustomed to using technology in their workflows, they may still resist using a CDSS. Their reasons may range from concerns about losing their autonomy to mistrust of the knowledge base to the need for training (Berner 2009). Involving clinical leaders and champions, and keeping the channels of communication open, will address some of these concerns early and improve buy-in. 6. Integrate the CDSS into the clinical workflow. Ensuring that the design follows the guidelines of an effective CDSS, especially integration into the existing workflow, will help gain stakeholder support. A lessintrusive CDSS is easier for clinicians to accept. Conversely, a CDSS

C h a p te r 6:   C l i n i c a l D ec ision Sup p or t Sy stem s in Medic ine

that disrupts the existing workflow is viewed as a barrier to work and inhibits successful implementation. 7. Plan for the go-live. The go-live date should be determined and announced early on to give stakeholders time to prepare for it. The implementation team should develop a go-live plan and schedule and determine specific roles for team members. An assessment of readiness should be performed, and the CDSS should be tested before it goes live. 8. Train and support. Training is critical to successful user adoption. Providing the necessary training before the CDSS goes live will help minimize anxiety and increase acceptance. The training should be customized to each individual’s needs. The amount of support may vary depending on the individual’s prior experience with technology and the CDSS’s complexity. The go-live plan should build in extra training for the implementation team. Training and support should continue to be available for a while after go-live. 9. Monitor and evaluate the CDSS’s impact. How the CDSS utility will be monitored should be defined before go-live. Pose questions such as, Is the CDSS integrated well into the clinical workflow? Is the CDSS perceived as useful? Are there any areas of concern that should be studied? Monitoring of the CDSS should continue after the initial implementation has ended because the CDSS design may have to be adjusted as a result of a change in workflow or conflict with another CDSS or new technology. In addition, the CDSS may cease to activate, or it may activate inappropriately because of a software update (Wright et al. 2016). If the CDSS is not monitored on an ongoing basis, it may malfunction and become useless or a nuisance. 10. Perform knowledge management. A CDSS relies on a knowledge base that is continuously evolving. The policies and guidelines on which a CDSS is based may change with time, and the information in a CDSS may become irrelevant (e.g., if a medication is discontinued). The data architecture the CDSS uses to access patient information may become unavailable because of a change in software or clinical workflow. All of these situations underscore the need for knowledge management. In large institutions with extensive CDSSs, knowledge management can be the responsibility of a committee.

Challenges and Barriers Despite decades of use and a push for more widespread adoption, CDSSs still face many challenges to successful implementation. Middleton, Sittig, and

131

132

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Wright (2016) list six areas where barriers are encountered: data, knowledge, inference, technology, interoperability, and users.

Human Factors Human factors affect the success of a CDSS in several ways. Humans are involved in the development, implementation, and utilization phases of CDSSs (Brown, Patrick, and Pasupathy 2013). Developers must know both users’ needs and clinical workflows when designing CDSSs. Moreover, they should understand that not all users will believe a CDSS is needed. This potential problem may be exacerbated at larger institutions that have a large number of users. Addressing and managing these concerns will improve buy-in during CDSS implementation. Clinicians are trained to critically analyze the information on which they base their clinical decisions. One of their concerns about CDSSs is credibility—the clinicians’ ability to trust the information the CDSS provides (Berner 2009). Thus, the CDSS knowledge base should be actively updated as new information and new guidelines become available. It should draw from reputable sources and be transparent. Human–computer interaction can present major issues as well (Musen, Middleton, and Greenes 2014; Porat, Delaney, and Kostopoulou 2017; Thum, Kim, and Genes 2014; Wu, Davis, and Bell 2012). A CDSS should be designed so that it is easy to view, understand, and use. Poor usability—a lack of “userfriendliness”—may not only diminish user satisfaction but also negatively affect the user’s opinion of other CDSSs and lead to errors (Wu, Davis, and Bell 2012). Usability testing should be done before and after implementation. The CDSS committee must reach out to the users after implementation for feedback.

Interoperability The Healthcare Information and Management Systems Society (HIMSS 2013) defines healthcare interoperability as “the ability of different information technology systems and software applications to communicate, exchange data, and use the information that has been exchanged.” With the expansion of technology in healthcare, most organizations use multiple Interoperability Moment information systems to achieve HIMSS’s (2013) definition of interoperability is “the ability of differeffective care delivery. To funcent information technology systems and software applications to tion effectively, CDSSs must be communicate, exchange data, and use the information that has been able to communicate smoothly exchanged.” This definition does not include people as part of the with these various systems. The equation. Is the HIMSS definition sufficient? Consider interoperability need for interoperability is furas involving the reconciliation of various points of view and value ther emphasized by the rapid systems held by different clinicians and other healthcare providers. diffusion of many vendor-based CDSSs. Interoperability is one

C h a p te r 6:   C l i n i c a l D ec ision Sup p or t Sy stem s in Medic ine

of the major challenges of today’s CDSSs (Marco-Ruiz and Bellika 2015; Marco-Ruiz et al. 2016).

Free-Text Technology Free text is a method of entering information into the EHR in an unstructured, natural language format. Half of the patient information entered into EHRs by clinicians is done in free-text format (Sittig et al. 2008). Fields such as medical history, physical exam and progress notes, pathology and radiology reports, discharge summaries, and consultation notes are often entered as free text. Moreover, clinicians find structured fields that require searching for specific words or codes to enter information (e.g., “diagnosis list” or “procedure list”) to be time consuming (Porat, Delaney, and Kostopoulou 2017). They may also consider entering information in a structured format to be a duplicate effort because they already enter such information as free text in the history and physical note. This perception may decrease compliance of data entry, which in turn negatively affects the CDSS’s function. Nevertheless, most of today’s CDSSs are still designed using structured fields. Multiple new CDSSs are being developed that use natural language processing (NLP) and artificial intelligence technologies. These programs are able to extract relevant information from unstructured, free-text fields.

Alert Fatigue Despite the benefits, the increase in computerization and use of CDSSs has had some unanticipated negative consequences. Most CDSSs within EHRs give actions in the form of reminders and alerts. As a result, clinicians face an overwhelming number of alerts each day. A survey of primary care practitioners (PCPs) at the US Department of Veterans Affairs found that PCPs received an average of 63 alerts per day, and about 70 percent of PCPs reported more alerts than they considered manageable (Singh et al. 2013). A systematic review of literature on physicians’ response to drug safety alerts noted that physicians override 49 percent to 96 percent of all safety alerts (Kesselhem et al. 2011; van der Sijs et al. 2006). The review also noted that lower-level alerts are overridden more than high-level alerts (van der Sijs et al. 2006). A CDSS is simply a tool to aid clinicians in their decisions and, in the case of drug alerts, to prevent a potentially harmful situation. The final decision is made by the clinicians. If the clinicians are overwhelmed by a volume of alerts that they perceive as having little to no clinical significance, they may unintentionally also override clinically critical alerts, leading to preventable harm (Kuperman et al. 2007; van der Sijs et al. 2006). Possible solutions include reducing the number of alerts to include only those with high significance, personalizing the alerts to fire only in specific clinical settings, and categorizing alerts according to levels of importance depending on the clinical significance (Kesselhem et al. 2011; Kuperman et al. 2007; van der Sijs et al. 2006).

133

134

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Clinical Domain Examples Many CDSS tools (e.g., MYCIN, AI/RHEUM) have been built in various academic settings since the beginning of the field of informatics. However, few of these tools have had large commercial success or broad adoption, mostly because of the challenges of system integration. Only recently have CDSSs gained enough traction that they have become a large enough HIT sector to be cataloged or blogged about. In 2016, the National Committee on Vital and Health Statistics recommended to the US Department of Health and Human Services (HHS) that HL7 standards be adopted. Institutions must move to integrate and implement effective electronic tools to meet the meaningful use criterion for HL7 (Cimino, Jing, and Del Fiol 2012). These tools must be both useful and readily accessible and integrated in such a way that they are simple, timely, and contextually relevant (Cimino and Borovtsov 2008). KLAS Research is a company that provides information on HIT software and services used by providers and payers. By surveying users, the company generates in-depth reports that rank EHRs and CDSS tools among many other HIT products and services. In 2013, KLAS Research’s first report on CDSS tools revealed that about one-third of surveyed users reported that their EHR vendor plays a significant role in driving their organization’s CDSS technology, and more than half of respondents said their EHR vendor will continue to play this role in the future. Many of the most widely deployed CDSS tools today come from some of the largest EHR vendors, such as Allscripts, Cerner, and Epic. With the emergence of the Fast Healthcare Interoperability Resources HL7 specification, there is now the potential that more third-party (non-EHR) tools will become available that can be embedded or integrated into the EHR workflow, providing the “right” place for clinicians to get the message. The top six categories of focus for healthcare organizations are (in order) order sets, surveillance, care plans, point-of-care disease reference, infection control, and diagnostic support (Cherrington and Christensen 2016). Some CDSS tools that can be embedded in the EHR are discussed in this section.

AgileMD CMS’s value-based payment programs began in 2013, a few years after the ACA was passed; hospital-based reimbursements have been adjusted since then as a result of the programs. The passage of the Medicare Access and CHIP Reauthorization Act in 2015 increased the attention paid to the cost of care delivery, because that cost will soon become an important component of physician reimbursement. Controlling costs will require increased standardization, reduced variation, and adoption of best practices. AgileMD is a company that embeds protocols and clinical pathways into the EHR workflow to help standardize best practices. With EHR integration,

C h a p te r 6:   C l i n i c a l D ec ision Sup p or t Sy stem s in Medic ine

clinicians can easily place orders and create documentation that follows these best practices. This approach can lead to improved patient quality and safety, a reduction in unnecessary tests, decreased lengths of stay, fewer preventable admissions, and better adherence to guidelines.

MedSocket In 2016, total healthcare expenditures in the United States were estimated to be $3.4 trillion, and that spending is growing faster than the gross domestic product (CMS 2016b). Of the money spent on healthcare, 86 percent was associated with patients with one or more chronic conditions (Gerteis et al. 2014). Drug spending is also outpacing overall healthcare expenditures, with the most recent estimate (2015) being $325 billion (CMS 2016b). With such a rapid increase in costs but only modest improvement in quality, many companies are trying to bend the cost curve and improve quality through CDSS technology. MedSocket, founded by one of this chapter’s authors (K.M.K.), develops technology to reduce the cost of chronic disease care and improve treatment outcomes by providing clinicians with the information and on-demand decision support they need directly from the EHR. The company offers a suite of CDSS solutions to tackle these problems. The first, called 1-Click Decision Support (1-CDS), is a patented decision support solution for personalized chronic disease management. 1-CDS automatically populates calculators with data from the EHR and overlays patient data on top of national evidence-based guidelines. MedSocket has also built a patented search technology called 1-Search, which is embedded within the EHR to aggregate the best medical resources that an institution licenses. 1-Search presents results according to the type of clinician performing the search. 1-Search has received National Library of Medicine funding from Phase I and Phase II Small Business Innovation Research awards (Alafaireet et al. 2017). MedSocket Rx, another product, connects patients with discount medication programs in the EHR workflow for medications they are currently taking.

VisualDx For every three patient visits, physicians are asked two clinical questions about the care of their patients. Unfortunately, up to 70 percent of those questions go unanswered (Covell, Uman, and Manning 1985; Ely et al. 2005; Gorman and Helfand 1995). Given that more than a billion patient visits occur each year in the United States, this situation may lead to suboptimal care in more than 400 million patient encounters annually (National Center for Health Statistics 2012). CDSS tools can quickly provide answers to those unanswered questions at the point of care. Dermatologist Art Papier developed VisualDx, a CDSS at the point of care, to aid in diagnosis. The system has a library of more than 32,000 images to help match patient conditions to known cases. Based on a patient’s chief

135

136

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

complaint, VisualDx references differential diagnosis and performs image searches. It also provides a machine-learning platform, based on Apple’s machine-learning toolkit, to allow a user to take a photo of a skin ailment and have the system look for similar images.

Infobutton Infobutton is a clinical decision support tool that draws on health knowledge resources and provides context-sensitive links embedded and integrated into the EHR (Cimino 2006). Infobutton is free, open source, and adaptable, and it is available online (www.openinfobutton.org/home). It has been shown to be “effective at helping clinicians answer questions at the point of care and demonstrate a modest incremental change in the efficiency of information delivery for routine users” (Del Fiol et al. 2008). Infobuttons are generally located next to data points and provide useful information to users. The tool extracts data from the patient record, builds a search query, and passes on the formatted query to select resources. The EHR then contextualizes the query by noting the problem term(s), the patient’s demographics (e.g., gender, age, weight), and the current EHR task (e.g., problem tasks, medication orders, lab orders, lab results). Finally, it notes who is requesting the information because different users may want different information and have different access permissions. Data from the query results are then packaged and sent to the Infobutton Manager. This application evaluates the data, matches them to a list of potential resources, and returns links to the user. Within seconds, the user can see resources relevant to the current tasks.

Fast Healthcare Interoperability Resources Standard In the spirit of open and effective interoperability of healthcare information, HL7 created the Fast Healthcare Interoperability Resources (FHIR) standard. FHIR (pronounced “fire”) is the latest standard developed to enable the exchange of healthcare information electronically (www.hl7.org/fhir/ overview.html).

Conclusion CDSSs have been shown to improve physician performance and patient care, as well as to reduce healthcare cost. This is especially true when CDSSs are designed to provide the right information to the right person in the right format via the right channel at the right time. Significant interest in the area has encouraged many researchers and innovators to develop new technologies to improve CDSSs and expand their use. CDSSs related to diagnosis, for example, have made significant strides as a result of the application of image recognition

C h a p te r 6:   C l i n i c a l D ec ision Sup p or t Sy stem s in Medic ine

technology and NLP. NLP has been studied in other areas of the EHR, with the goal of creating a more integrated CDSS. Another source of excitement is the expansion of the CDSS to a patient-oriented one, or a patient decision support system (PDSS). Such advances will not only encourage patient autonomy but also increase patients’ involvement in their care. This is an exciting time for CDSSs, and many advances are yet to come.

Chapter Discussion Questions 1. What role has the federal government played in the expansion of CDSS technology? 2. Which types of CDSS do you think have the largest impact on patient care? 3. What makes for an effective CDSS deployment? 4. What are some of the biggest challenges to CDSS adoption, and how can they be overcome? 5. How do you see CDSSs evolving in the future?

Case Study  Effective CDSS Implementation The University of Illinois Hospital and Health Sciences System (UI Health) has been a leader in adopting health information technology. When the hospital was built in 1982, a light pen–based computerized physician-order entry (CPOE) system was deployed—decades before CPOE use was incentivized through the Centers for Medicare & Medicaid Services’ meaningful use program. UI Health was one of the first to deploy a modern commercial electronic health record (EHR) system in 1997, for which it won the Healthcare Information and Management Systems Society’s Nicholas E. Davies Award of Excellence. Work on decision support rules began shortly after the EHR was deployed and, over the ensuing decades, resulted in hundreds of custom-built decision support rules. Creation of so many rules can lead to challenges in maintenance requirements and unintended consequences if not managed properly. Over the years, a clinical decision support system (CDSS) governance committee structure was formed at UI Health (exhibit 6.4) to help ensure the creation of evidence-based, highly effective CDSS rules. Researchers at UI Health have published many articles on the effectiveness of such rules in facilitating problem identification, reducing contraindicated medications, (continued)

137

138

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

EXHIBIT 6.4 UI Health CDSS Committee Governance Structure

Medical Executive Committee (MSEC)

Pharmacy and Therapeutics Committee

Electronic Medical Record Committee

CDS Committee Co-Chairs: Physician Informatician and Clinical Informatics Team Member Medication System Review Committee

Direct Report Informational Report

Active Members: Hospital IT, Informatics Team, Nursing, Pathology, Radiology, Pharmacy, Physicians, Member from Quality Assurance

Documentation Committee

Visiting Members: CDS Requesting Department

Source: Reprinted with permission from the University of Illinois Hospital and Health Sciences System.

and improving venous thromboembolism prophylaxis and warfarin dosing, among other findings (Falck et al. 2013; Galanter, Didomenico, and Polikaitis 2002, 2004, 2005; Galanter, Liu, and Lambert 2010; Galanter, Thambi, et al. 2010; Galanter, Heir, et al. 2010 Nutescu et al. 2013). Despite all the research and expertise, not every rule gets implemented smoothly. For example, a study by Lui and colleagues (2016) pointed out the need to appropriately identify psychiatric patients who are at risk for metabolic syndrome, which can be exacerbated by second-generation antipsychotic medications. The psychiatry department requested a new CDSS rule around the use of antipsychotics and the potential development of metabolic syndrome. This rule was then published into the UI Health production environment. After this CDSS rule was created, a primary care provider (PCP) updated a patient’s medications list because the patient’s outside psychiatrist changed her antipsychotic medication dose. The patient was a 50-year-old woman who had long-standing diabetes, hypertension, obesity, and bipolar disorder. A CDSS rule fired, indicating that the patient was on an antipsychotic and qualified for metabolic syndrome because of her abnormal girth, her systolic blood pressure (which was greater than 130), her triglyceride level (which was greater than 150), her HDL cholesterol

C h a p te r 6:   C l i n i c a l D ec ision Sup p or t Sy stem s in Medic ine

level (which was less than 40), and her glucose level (which was greater than 110). Clinically, most consider metabolic syndrome to be a precursor to the diagnoses of hypertension, obesity, and diabetes, all of which the patient already had—and all were documented in her medical record. The rule suggested that the PCP add “metabolic syndrome” to the patient’s problem list, but no link to do so was present. Recognizing that the rule was not firing as intended, the CDSS governance committee performed a root-cause analysis. The following lessons learned and opportunities for improvement were identified: • The rule was firing every time a medication reconciliation (a required task for any care transition) was performed on patients who had an antipsychotic on their medication list and met the criteria for metabolic syndrome. As a result, an alert went out to providers even when they were just acknowledging a patient’s current medication regimen. • The rule was firing in all clinical settings—inpatient, outpatient, surgicenter, urgent care, and emergency department—and for providers who did not typically address conditions such as metabolic syndrome. • The rule was designed by a psychiatrist and the CDSS committee’s information systems staff and was not vetted by the committee’s other clinical members. The creation and maintenance of the rule was outsourced to the EHR vendor and staffed by technical specialists who had little to no clinical expertise. All of this contributed to the confusion when reports of problems with the rule were being submitted. • The rule was firing for all medications in the antipsychotic drug class, not just second-generation antipsychotics shown to significantly contribute to metabolic syndrome. • Despite a lot of testing in nonproduction domains, creating a CDSS rule takes a lot of time and requires an iterative agile process to adequately address problems and improve the rule. • Acknowledging that a patient may already have clinical conditions (e.g., diabetes, hypertension, obesity) documented in the problem list can help make a CDSS rule “smarter” and increase its level of credibility—and thus increase the receiving clinician’s trust in the rule and its purpose. • Attaching guidelines and literature to a CDSS rule request can help the development team ensure the rule is evidence based. (continued)

139

140

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

• A rule that recommends an action can be made more convenient if the action is available inside the alert. The information provided is then a convenience rather than a burden.

Case Study Discussion Questions 1. Identify the Five Rights in this case. 2. Is there an opportunity to improve at least one of the Five Rights in the rollout of this rule? 3. What is the importance of governance in developing and maintaining CDSS rules? 4. Based on the governance structure, what roles can help avoid the problems described in this case? 5. What are the pros and cons of having practicing clinicians build CDSS rules?

Note We acknowledge the contributions to this chapter by Bill Galanter, Katie Ferraro, Rob Marvin, Jayne Williams, and Thomas Kannampallil.

Additional Resources 21st Century Cures Act: www.fda.gov/RegulatoryInformation/LawsEnforcedbyFDA/ SignificantAmendmentstotheFDCAct/21stCenturyCuresAct/default.htm. Centers for Medicare & Medicaid Services (CMS) EHR Incentive Programs: www. cms.gov/Regulations-and-Guidance/Legislation/EHRIncentivePrograms/ index.html. Clinical Decision Support Coalition: http://cdscoalition.org/. HITECH: www.hhs.gov/hipaa/for-professionals/special-topics/hitech-act-enforcement-interim-final-rule/index.html. HL7: www.hl7.org/. National Quality Strategy: www.ahrq.gov/workingforquality/about/index.html.

References Agatonovic-Kustrin, S., and R. Beresford. 2000. “Basic Concepts of Artificial Neural Network (ANN) Modeling and Its Application in Pharmaceutical Research.” Journal of Pharmaceutical and Biomedical Analysis 22 (5): 717–27.

C h a p te r 6:   C l i n i c a l D ec ision Sup p or t Sy stem s in Medic ine

Ahmadian, L., M. V. Engen-Verheul, F. Bakhshi-Raiez, N. Peek, R. Cornet, and N. F. Keizer. 2011. “The Role of Standardized Data and Terminological Systems in Computerized Clinical Decision Support Systems: Literature Review and Survey.” International Journal of Medical Informatics 80 (2): 81–93. Aishwarya, S., and S. Anto. 2014. “A Medical Decision Support System Based on Genetic Algorithm and Least Square Support Vector Machine for Diabetes Disease Diagnosis.” International Journal of Engineering Sciences & Research Technology 3 (4): 4042–46. Alafaireet, P. E., J. L. Belden, M. Botkin, K. M. Kochendorfer, R. L. Kruse, D. Strecker, and J. A. Williams. 2017. “Embedding a Medical Search Engine Within an Electronic Health Record.” Missouri Medicine 114 (4): 316–20. Bates, D. W., G. J. Kuperman, S. Wang, T. Gandhi, A. Kittler, L. Volk, C. Spurr, R. Khorasani, M. Tanasijevic, and B. Middleton. 2003. “Ten Commandments for Effective Clinical Decision Support: Making the Practice of Evidence-Based Medicine a Reality.” Journal of the American Medical Informatics Association 10 (6): 523–30. Berner, E. S. 2009. “Clinical Decision Support Systems: State of the Art.” AHRQ Publication No. 09-0069-EF. Published June. https://healthit.ahrq.gov/sites/ default/files/docs/page/09-0069-EF_1.pdf. Berner, E. S., and T. J. La Lande. 2007. “Overview of Clinical Decision Support Systems.” In Clinical Decision Support Systems: Theory and Practice, 2nd ed., edited by E. S. Berner, 3–18. New York: Springer. Bleich, H. L. 1969. “Computer Evaluation of Acid-Base Disorders.” Journal of Clinical Investigation 48 (9): 1689–96. Blumenthal, D., and M. Tavenner. 2010. “The ‘Meaningful Use’ Regulation for Electronic Health Records.” New England Journal of Medicine 363 (6): 501–4. Bokhari, Q. 2017. “Integrated Intelligent Decision Support System (IDSS) Vision.” Presentation at AMDIS Annual Conference, June 27, Nottingham, England. Boxwala, A. A., M. Peleg, S. Tu, O. Ogunyemi, Q. T. Zeng, D. Wang, V. L. Patel, R. A. Greenes, and E. H. Shortliffe. 2004. “GLIF3: A Representation Format for Sharable Computer-Interpretable Clinical Practice Guidelines.” Journal of Biomedical Informatics 37 (3): 147–61. Brown, G. D., T. B. Patrick, and K. S. Pasupathy. 2013. Health Informatics: A Systems Perspective. Chicago: Health Administration Press. Buntin, M. B., M. F. Burke, M. C. Hoaglin, and D. Blumenthal. 2011. “The Benefits of Health Information Technology: A Review of the Recent Literature Shows Predominantly Positive Results.” Health Affairs 30 (3): 464–71. Byrne, C., D. Sherry, L. Mercincavage, D. Johnston, E. Pan, and G. Schiff. 2017. “Advancing Clinical Decision Support: Key Lessons in Clinical Decision Support Implementation.” Accessed September 23. www.healthit.gov/sites/default/ files/acds-lessons-in-cds-implementation-deliverablev2.pdf. Centers for Medicare & Medicaid Services (CMS). 2016a. “Clinical Decision Support: More Than Just ‘Alerts’ Tip Sheet.” Published January 8. www.cms.gov/

141

142

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Regulations-and-Guidance/Legislation/EHRIncentivePrograms/Downloads/ ClinicalDecisionSupport_Tipsheet-.pdf. ———. 2016b. “NHE Fact Sheet.” Updated June 14, 2017. www.cms.gov/researchstatistics-data-and-systems/statistics-trends-and-reports/nationalhealthexpend data/nhe-fact-sheet.html. ———. 2016c. “Stage 3 Program Requirements for Providers Attesting to Their State’s Medicaid EHR Incentive Program.” Updated December 1. www.cms.gov/ Regulations-and-Guidance/Legislation/EHRIncentivePrograms/Stage3Medi caid_Require.html. Cherrington, A., and J. Christensen. 2016. “KLAS CDS Performance Report.” Accessed September 27, 2017. https://klasresearch.com/industry-reports. Cimino, J. J. 2006. “Use, Usability, Usefulness, and Impact of an Infobutton Manager.” In AMIA Annual Symposium Proceedings, 151–55. Bethesda, MD: American Medical Informatics Association. Cimino, J., and D. V. Borovtsov. 2008. “Leading a Horse to Water: Using Automated Reminders to Increase Use of Online Decision Support.” In AMIA Annual Symposium Proceedings, 116–20. Bethesda, MD: American Medical Informatics Association. Cimino, J., X. Jing, and G. Del Fiol. 2012. “Meeting the Electronic Health Record ‘Meaningful Use’ Criterion for the HL7 Infobutton Standard Using Open Infobutton and the Librarian Infobutton Tailoring Environment (LITE).” In AMIA Annual Symposium Proceedings, 112–20. Bethesda, MD: American Medical Informatics Association. Collen, M. F., L. Rubin, J. Neyman, G. B. Dantzig, R. M. Baer, and A. B. Siegelaub. 1964. “Automated Multiphasic Screening and Diagnosis.” American Journal of Public Health and the Nation’s Health 54 (5): 741–50. Covell, D. G., G. C. Uman, and P. R. Manning. 1985. “Information Needs in Office Practice: Are They Being Met?” Annals of Internal Medicine 103 (4): 596–99. Del Fiol, G., P. J. Haug, J. J. Cimino, S. P. Narus, C. Norlin, and J. A. Mitchell. 2008. “Effectiveness of Topic-Specific Infobuttons: A Randomized Controlled Trial.” Journal of the American Medical Informatics Association 15 (6): 752–59. Donaldson, M. S. 2008. “An Overview of To Err Is Human: Re-emphasizing the Message of Patient Safety.” In Patient Safety and Quality: An Evidence-Based Handbook for Nurses, edited by R. G. Hughes. Published April. www.ncbi.nlm. nih.gov/books/NBK2673/. El-Sappagh, S. H., and S. El-Masri. 2014. “A Distributed Clinical Decision Support System Architecture.” Journal of King Saud University—Computer and Information Sciences 26 (1): 69–78. Ely, J. W., J. A. Osheroff, M. L. Chambliss, M. H. Ebell, and M. E. Rosenbaum. 2005. “Answering Physicians’ Clinical Questions: Obstacles and Potential Solutions.” Journal of the American Medical Informatics Association 12 (2): 217–24. Falck, S., S. Adimadhyam, D. O. Meltzer, S. M. Walton, and W. L. Galanter. 2013. “A Trial of Indication Based Prescribing of Antihypertensive Medications During

C h a p te r 6:   C l i n i c a l D ec ision Sup p or t Sy stem s in Medic ine

Computerized Order Entry to Improve Problem List Documentation.” International Journal of Medical Informatics 82 (10): 996–1003. Galanter, W. L., R. Didomenico, and A. Polikaitis. 2005. “A Trial of Automated Decision Support Alerts for Contraindicated Medications Using Computerized Physician Order Entry.” Journal of the American Medical Informatics Association 12 (3): 269–74. ———. 2004. “Effectiveness of Automated Clinical Decision Support Alerts for Inpatient Digoxin Use with Computerized Physician Order Entry.” Journal of the American Medical Informatics Association 11 (4): 270–77. ———. 2002. “Use of Expert System to Prevent Exacerbation of Adverse Drug Event.” Journal of Healthcare Information Management 16 (4): 44–49. Galanter, W. L., D. B. Hier, C. Jao, and D. Sarne. 2010. “Computerized Physician Order Entry (CPOE) of Medications and Clinical Decision Support (CDS) Can Improve Problem List Compliance.” International Journal of Medical Informatics 79 (5): 332–38. Galanter, W. L., X. Liu, and B. L. Lambert. 2010. “An Analysis of Computer Alerts Suggesting Oral Medication Use During Computerized Order Entry of Intravenous (IV) Medications.” American Journal of Health-System Pharmacy 67 (13): 1101–5. Galanter, W. L., M. Thambi, H. Rosencranz, B. Shah, S. Falck, F. J. Lin, E. Nutescu, and B. Lambert. 2010. “Effects of Clinical Decision Support on Venous Thromboembolism Risk Assessment, Prophylaxis, and Prevention at a University Teaching Hospital.” American Journal of Health-System Pharmacy 67 (15): 1265–73. Garg, A. X., N. K. Adhikari, H. McDonald, M. P. Rosas-Arellano, P. J. Deveraux, J. Beyene, J. Sam, and R. B. Haynes. 2005. “Effects of Computerized Clinical Decision Support Systems on Practitioner Performance and Patient Outcomes: A Systematic Review.” Journal of the American Medical Association 293 (10): 1223–38. Gerteis, J., D. Izrael, D. Deitz, L. LeRoy, R. Ricciardi, T. Miller, and J. Basu. 2014. “Multiple Chronic Conditions Chartbook: 2010 Medical Expenditure Panel Survey Data.” AHRQ Publications No. Q14-0038. Published April. www. ahrq.gov/sites/default/files/wysiwyg/professionals/prevention-chronic-care/ decision/mcc/mccchartbook.pdf. Ghaheri, A., S. Shoar, M. Naderan, and S. S. Hoseini. 2015. “The Applications of Genetic Algorithms in Medicine.” Oman Medical Journal 30 (6): 406–16. Gorman, P. N., and M. Helfand. 1995. “Information Seeking in Primary Care: How Physicians Choose Which Clinical Questions to Pursue and Which to Leave Unanswered.” Medical Decision Making 15 (2): 113–19. Greenes, R. A. 2006. Clinical Decision Support: The Road Ahead. Burlington, MA: Elsevier. Healthcare Information and Management Systems Society (HIMSS). 2013. “What Is Interoperability?” Accessed September 30, 2017. www.himss.org/library/ interoperability-standards/what-is-interoperability.

143

144

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Karnik, K. 2014. “FDA Regulation of Clinical Decision Support Software.” Journal of the Law and the Biosciences 1 (2): 202–8. Kesselhem, A. S., K. Cresswell, S. Phansalkar, D. W. Bates, and A. Sheikh. 2011. “Clinical Decision Support Systems Could Be Modified to Reduce ‘Alert Fatigue’ While Still Minimizing the Risk of Litigation.” Health Affairs 30 (12): 2310–17. Kohn, L. T., J. M. Corrigan, and M. S. Donaldson. 2000. To Err Is Human: Building a Safer Health System. Washington, DC: National Academies Press. Kononenko, I. 2001. “Machine Learning for Medical Diagnosis: History, State of the Art and Perspective.” Artificial Intelligence Medicine 23 (1): 89–109. Kuperman, G. J., A. Bobb, T. H. Payne, A. J. Avery, T. K. Gandhi, G. Burns, D. C. Classen, and D. W. Bates. 2007. “Medication-Related Clinical Decision Support in Computerized Provider Order Entry Systems: A Review.” Journal of the American Medical Informatics Association 14 (1): 29–40. Ledley, R. S., and L. B. Lusted. 1959. “Reasoning Foundations of Medical Diagnosis: Symbolic Logic, Probability, and Value Theory Aid Our Understanding of How Physicians Reason.” Science 130 (3366): 9–21. Lee, F., J. M. Teich, C. D. Spurr, and D. W. Bates. 1996. “Implementation of Physician Order Entry: User Satisfaction and Self-Reported Usage Patterns.” Journal of the American Medical Informatics Association 3 (1): 42–55. Lui, K., G. Randhawa, V. Totten, A. E. Smith, and J. Raese. 2016. “Is Metabolic Syndrome on the Radar? Improving Real-Time Detection of Metabolic Syndrome and Physician Response by Computerized Scan of the Electronic Medical Record.” Primary Care Companion for CNS Disorders 18 (1): 10.4088/ PCC.15m01849. Lyman, J. A., W. F. Cohn, M. Bloomrosen, and D. E. Detmer. 2010. “Clinical Decision Support: Progress and Opportunities.” Journal of the American Medical Informatics Association 17 (5): 487–92. Marco-Ruiz, L., and J. G. Bellika. 2015. “Semantic Interoperability in Clinical Decision Support Systems: A Systematic Review.” Studies in Health Technology and Informatics 216: 958. Marco-Ruiz, L., A. Budrionis, K. Y. Yigzaw, and J. Bellika. 2016. “Interoperability Mechanisms of Clinical Decision Support Systems: A Systematic Review.” Presentation at the 14th Scandinavian Health Informatics Conference, April 6–7, Gothenburg, Sweden. Mardon, R., L. Mercincavage, M. Johnson, S. Finley, E. Pan, and D. Arora. 2014. “Findings and Lessons from AHRQ’s Clinical Decision Support Demonstration Projects.” AHRQ Publication No. 14-0047-EF. Published June. https:// healthit.ahrq.gov/sites/default/files/docs/page/findings-and-lessons-fromclinical-decision-support-demonstration-projects.pdf. Mead, C. N. 2006. “Data Interchange Standards in Healthcare IT—Computable Semantic Interoperability: Now Possible but Still Difficult, Do We Really Need a Better Mousetrap?” Journal of Healthcare Information Management 20 (1): 71–78.

C h a p te r 6:   C l i n i c a l D ec ision Sup p or t Sy stem s in Medic ine

Middleton, B., D. F. Sittig, and A. Wright. 2016. “Clinical Decision Support: A 25 Year Retrospective and a 25 Year Vision.” Yearbook of Medical Informatics Suppl 1: S103–S116. Miliard, M. 2015. “Clinical Decision Support: No Longer Just a Nice-to-Have.” HealthcareIT News. Published July 14. www.healthcareitnews.com/news/ clinical-decision-support-no-longer-just-nice-have. Musen, M. A., B. Middleton, and R. A. Greenes. 2014. “Clinical Decision-Support Systems.” In Biomedical Informatics: Computer Applications in Health Care and Biomedicine, edited by E. H. Shortliffe and J. J. Cimino, 643–74. London: Springer. National Center for Health Statistics. 2012. Health, United States, 2011: With Special Feature on Socioeconomic Status and Health. Published May. www.cdc.gov/ nchs/data/hus/hus11.pdf. National Committee on Vital and Health Statistics. 2016. “Recommendations for the Electronic Health Care Attachment Standard.” Letter to the Secretary of the US Department of Health and Human Services, July 5. www.ncvhs.hhs.gov/ wp-content/uploads/2013/12/2016-Ltr-Attachments-July-1-Final-ChairCLEAN-for-Submission-Publication.pdf. Nutescu, E. A., K. Drozda, A. P. Bress, W. L. Galanter, J. Stevenson, T. D. Stamos, A. A. Desai, J. D. Duarte, V. Gordeuk, D. Peace, S. S. Kadkol, C. Dodge, S. Saraf, J. Garofalo, J. A. Krishnan, J. G. Garcia, and L. H. Cavallari. 2013. “Feasibility of Implementing a Comprehensive Warfarin Pharmacogenetics Service.” Pharmacotherapy 33 (11): 1156–64. Osheroff, J. A., J. M. Teich, D. L. Levick, L. Saldana, F. Velasco, D. Sittig, K. Rogers, and R. Jenders. 2012. Improving Outcomes with Clinical Decision Support: An Implementer’s Guide, 2nd ed. Chicago: Healthcare Information and Management Systems Society. Osheroff, J. A., J. M. Teich, B. Middleton, E. B. Steen, A. Wright, and D. E. Detmer. 2007. “A Roadmap for National Action on Clinical Decision Support.” Journal of the American Medical Informatics Association 14 (2): 141–45. Patel, J. L., and R. K. Goyal. 2007. “Applications of Artificial Neural Networks in Medical Science.” Current Clinical Pharmacology 2 (3): 217–26. Perreault, L. E., and J. B. Metzger. 1999. “A Pragmatic Framework for Understanding Clinical Decision Support.” Journal of the Healthcare Information and Management Systems Society 13 (2): 5–21. Porat, T., B. Delaney, and O. Kostopoulou. 2017. “The Impact of a Diagnostic Decision Support System on the Consultation: Perceptions of GPs and Patients.” BMC Medical Informatics and Decision Making 17: 79. Reider, J. M. 2016. “Impact of National Policies on the Use of Clinical Decision Support.” In Clinical Decision Support Systems: Theory and Practice, 3rd ed., edited by E. S. Berner, 111–32. New York: Springer. Rodvold, D. M., D. G. McLeod, J. M. Brandt, P. B. Snow, and G. P. Murphy. 2001. “Introduction to Artificial Neural Networks for Physicians: Taking the Lid Off the Black Box.” The Prostate 46 (1): 39–44.

145

146

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Sim, I., and A. Berlin. 2003. “A Framework for Classifying Decision Support Systems.” In AMIA Annual Symposium Proceedings, 599–603. Bethesda, MD: American Medical Informatics Association. Singh Gill, N. 2017. “Overview of Artificial Neural Networks and Its Applications.” XenonStack. Published May 5. www.xenonstack.com/blog/data-science/ overview-of-artificial-neural-networks-and-its-applications. Singh, H., C. Spitzmueller, N. J. Petersen, M. K. Sawhney, and D. F. Sittig. 2013. “Information Overload and Missed Test Results in Electronic Health Record– Based Settings.” JAMA Internal Medicine 173 (8): 702–4. Sittig, D. F., A. Wright, J. A. Osheroff, B. Middleton, J. M. Teich, J. S. Ash, E. Campbell, and D. W. Bates. 2008. “Grand Challenges in Clinical Decision Support.” Journal of Biomedical Information 41 (2): 387–92. Thompson, B. M. 2017. “CDS Coalition Calls for Comments on Draft Industry Guidelines by July 1, 2017.” Published April 27. http://cdscoalition.org/positions/. Thum, F., M. S. Kim, and N. Genes. 2014. “Usability Improvement of a Clinical Decision Support System.” In Design, User Experience, and Usability: User Experience Design for Everyday Life Applications and Services, edited by A. Marcus, 125–31. London: Springer. US Department of Health and Human Services. 2011. “2011 Report to Congress: National Strategy for Quality Improvement in Health Care.” Published March. www.ahrq.gov/workingforquality/reports/2011-annual-report.html. US Food and Drug Administration (FDA). 2018. “Digital Health.” Accessed May 9. www.fda.gov/MedicalDevices/DigitalHealth/default.htm. ———. 2016. “21st Century Cures Act.” Accessed May 9, 2018. www.fda.gov/ RegulatoryInformation/LawsEnforcedbyFDA/SignificantAmendmentstothe FDCAct/21stCenturyCuresAct/default.htm. van der Sijs, H., J. Aarts, A. Vulto, and M. Berg. 2006. “Overriding of Drug Safety Alerts in Computerized Physician Order Entry.” Journal of the American Medical Informatics Association 13 (2): 138–47. Warner, H. R., A. F. Toronto, L. G. Veasey, and R. Stephenson. 1961. “A Mathematical Approach to Medical Diagnosis: Application to Congenital Heart Disease.” Journal of the American Medical Association 177 (3): 177–83. Wright, A., and D. F. Sittig. 2008. “A Four-Phase Model of the Evolution of Clinical Decision Support Architectures.” International Journal of Medical Informatics 77 (10): 641–49. Wright, A., T. T. Hickman, D. McEvoy, S. Aaron, A. Ai, J. M. Andersen, S. Hussain, R. Ramoni, J. Fiskio, D. F. Sittig, and D. W. Bates. 2016. “Analysis of Clinical Decision Support System Malfunctions: A Case Series and Survey.” Journal of the American Medical Informatics Association 23 (6): 1068–76. Wu, H. W., P. K. Davis, and D. S. Bell. 2012. “Advancing Clinical Decision Support Using Lessons from Outside of Healthcare: An Interdisciplinary Systematic Review.” BMC Medical Informatics and Decision Making 12 (1): 90–99.

CHAPTER

NURSING INFORMATICS

7

Carol G. Klingbeil, Pei-Yun Tsai, and Timothy B. Patrick

Learning Objectives After reading this chapter, you should be able to do the following: • Recognize the role of nurses as knowledge workers in the healthcare arena. • Discuss the evolving roles and competencies for nurses in nursing informatics. • Identify clinical workflows in nursing that involve informatics applications. • Explore the integration of nursing informatics education and research in the advancement of nursing in healthcare.

Key Concepts • • • • •

Transformation of clinical care Nurse informaticists as leaders of healthcare transformation Clinical workflows in nursing Interprofessional collaboration Consumer engagement

Introduction The historical roots of nursing informatics go back to the nineteenth century and Florence Nightingale. She is recognized as the first nursing informatics specialist because she argued that nurses should systematically collect data in patient care and use statistics to analyze the data to improve care processes and patient outcomes (O’Connor and Robertson 2003; Saba 2001). During the 1860 International Statistical Congress, Nightingale argued for “the uniform collection of hospital statistics, so that outcomes could be compared by 147

148

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

hospital, region, and country” (McDonald 2001). In addition, Nightingale was a pioneer in data visualization (Thompson 2016): During the Crimean War . . . Nightingale became appalled at the squalid conditions of army hospitals and soldier barracks, which were mired with feces and vermin. She persuaded Queen Victoria to let her study the issue, and Nightingale teamed up with her friend William Farr, the country’s leading statistician, to analyze army mortality rates. They uncovered a stunning fact: Most of the soldiers in the Crimean War hadn’t died in combat. They’d died of “preventable diseases”—precisely the sort caused by terrible hygiene. Clean up the hygiene and you’d save lives. Nightingale adroitly realized that tables of numbers and text would be too hard to parse. They needed, she said, a data visualization—“to affect thro’ the Eyes what we fail to convey to the public through their word-proof ears.” Her invention was the elegant “polar area chart,” a new variant of the pie chart: Each slice of the pie showed deaths for one month of the war, growing larger if the deaths increased, and color-coded to show the causes of death. Fans called it the “rose diagram,” because it looked like a flower. [It was also known as a “coxcomb”; see https://understanding uncertainty.org/coxcombs for an example of such a diagram.] Nursing informatics “A specialty that integrates nursing science with multiple information management and analytical sciences to identify, define, manage and communicate data, information, knowledge and wisdom in nursing practice” (ANA 2015, 1)

Modern nursing informatics began in the 1980s and encompassed the use of computer technology, computer systems, and data processing in nursing (Saba 2001). Graves and Corcoran (1989, 227) defined nursing informatics as a combination of computer science, information science, and nursing science designed to assist in the management and processing of nursing data, information, and knowledge to support the practice of nursing and the delivery of nursing care.

Nursing informatics has evolved significantly since the 1980s and, driven by advances in health information and communication technologies, has been applied in all areas of nursing, including clinical practice, administration, education, and research. However, information technology (IT) is used not just to automate existing nursing care processes but also, more importantly, to transform nursing practice. IT has been used to improve processes, bridge nursing research to nursing practice, and enhance the quality of healthcare. In 1992, the American Nurses Association (ANA) recognized nursing informatics as a nursing specialty along with medical-surgical, pediatric, mental, and community health nursing. The definition of nursing informatics was established as follows (ANA 2015, 1; Bickford 2009): A specialty that integrates nursing science with multiple information management and analytical sciences to identify, define, manage and communicate data, information, knowledge and wisdom in nursing practice. Nursing Informatics supports nurses,

C hap ter 7 :   N ur sing Infor m atic s

consumers, patients, the inter-professional healthcare team, and other stakeholders in their decision-making in all roles and settings to achieve desired outcomes. This support is accomplished through the use of information structures, information processes, and information technology.

According to this definition, to provide safe, efficient, and effective care, nurses need to integrate concepts from other fields, including “computer science, cognitive science, the science of terminologies and taxonomies, information management, library sciences, heuristics, archival science, and mathematics” (ANA 2015, 1).

Informatics, Nursing, and the Transformation of Clinical Care Florence Nightingale’s use of information methods to study army mortality rates during the Crimean War to effect change in clinical care is an example of the transformative power of informatics in nursing and clinical practice. Nurses are key to efforts to use IT to transform nursing practice and clinical care. The crucial role of nurses was recognized by the Healthcare Information and Management Systems Society (HIMSS) board of directors in a position statement released in 2011: Nurses are key leaders in developing the infrastructure for effective and efficient health information technology that transforms the delivery of care. Nurse informaticists play a crucial role in advocating both for patients and fellow nurses who are often the key stakeholders and recipients of these evolving solutions. Nursing informatics professionals are the liaisons to successful interactions with technology in healthcare. As clinicians who focus on transforming information into knowledge, nurse informaticists cultivate a new time and place of care through their facilitation efforts to integrate technology with patient care. Technology will continue to be a fundamental enabler of future care delivery models and nursing informatics leaders will be essential to transforming nursing practice through technology.

The leadership roles that nurse informaticists can fill are crucial to the success of ongoing and much-needed clinical transformation efforts, according to Sensmeier (2011): Many healthcare organizations are embracing the concept of clinical transformation . . . but they still require the tools and capabilities to make data available in real time and reduce the burden on scarce resources. . . . The role of clinical informaticists, including nurse informaticists, continues to be a much valued and necessary position

149

150

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

in today’s healthcare organization because these experts are essential to the success of quality initiatives, participating in the executive clinical team that analyzes clinical data. Survey results indicate that these human resources are necessary to ensure that clinical transformation efforts benefit from appropriate access to clinical data that is derived from the EHR [electronic health record].

Of course, not all nurses involved in informatics in a healthcare organization are leaders in that organization, but all nurses—and the roles they play in healthcare—are increasingly influenced by informatics. In the next section, we describe a hierarchy of roles nurses play in the use and leadership of informatics methods and systems in transforming clinical care.

Roles of Nurses in Informatics The ANA definition of nursing informatics implies roles for nurses in the collection of data, the conversion of data into information, the conversion of information into knowledge, and the conversion of knowledge into wisdom to support patients, nurses, and other healthcare providers in clinical decision making.

As Users of Information Technology In their most basic informatics role, nurses are users of IT. Nurses need to have basic computer skills to access, enter, and collect clinical data. Data are discrete entities that are described objectively without interpretation (Graves and Corcoran 1989). An example of data nurses must collect and use is patient vital signs. Nurses collect and employ clinical data using IT, which includes computer hardware and software as well as communication and network technologies (Snyder-Halpern, Corcoran-Perry, and Narayan 2001).

As Creators of Information Information is data that are interpreted, organized, or structured to reveal meaning (Graves and Corcoran 1989). Nurses analyze, aggregate, and compare data with standard normal measures to create meaningful information. For example, data on a patient’s blood pressure might be graphically displayed over time in relation to medications the patient has received. (Nightingale’s rose diagrams, mentioned above, are an example of creating useful information from data.) Such information is used to guide patient care decisions.

As Creators of Knowledge After meaningful information is created from data, the information is then converted into knowledge. One example of nursing knowledge is knowledge

C hap ter 7 :   N ur sing Infor m atic s

of patient characteristics associated with a positive response to treatment (e.g., antihypertensive medication). As identified by Synder-Halpern, Corcoran-Perry, and Narayan (2001), nurses at this level of nursing informatics are knowledge users and acquirers and are able to provide convenient and efficient means of capturing and storing knowledge. Interrelationships among different sets of information are identified by nurses, and this knowledge is then used to evaluate the processes they have implemented in practice. Furthermore, it may facilitate evaluation of outcomes to improve patient care, administration, education, and research (Graves and Corcoran 1989; Snyder-Halpern, Corcoran-Perry, and Narayan 2001).

As Innovators Wisdom is “the appropriate use of knowledge in management or solving human problems” (Nelson and Staggers 2018, 23). Nurses as innovators apply knowledge to nursing practice, including patient care, appropriate use of compassion, and integration of ethics into practice. The innovator role is closely related to nursing research and quality management. Nurse innovators must have the capability to conduct informatics research and design innovations that lead to the development of new domain knowledge to improve clinical practice and patient outcomes (McGonigle and Mastrian 2018; Snyder-Halpern, CorcoranPerry, and Narayan 2001; Staggers, Gassert, and Curran 2001).

Nursing Work and Information System Applications Although this section provides only a snapshot of the multiple arenas in which nursing care is delivered on a daily basis, it discusses several clinical workflows to serve as examples of the nature and complexity of nursing care and the kinds of supporting and clinical informatics applications that capture the work, intensity, and expertise of nurses as care providers.

Clinical Workflows Clinical care is delivered in a highly organized and integrated fashion not only according to policies and procedures but also in accordance with clinical processes most often referred to as clinical workflows. McGonigle and Mastrian (2018, 68–69) define workflow as “the action or execution of a series of tasks in a prescribed sequence.” Nursing is involved in and has a large number of different workflows that drive clinical care for patients in multiple settings in healthcare. As informatics work is planned, implemented, and evaluated, nurses must be involved in multiple work teams with the goal of improving multiple aspects of healthcare. Healthcare administrators need to understand that viewing a clinical care problem through a single lens will never fully address that

151

152

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

problem and, in most cases, will only complicate clinical matters. Identifying all stakeholders in different workflows is crucial to a multidisciplinary approach to problem solving and developing innovative solutions to clinical care issues.

Documentation One of the primary focus areas for clinical informatics is documentation of important aspects of care and planning care delivery. Nurses provide care in multiple settings, including acute care, long-term care, homes and community centers, and schools, as well as remotely through telehealth services. Important workflows in nursing represent the nursing process and include initiating care with nursing assessments, diagnosing nursing problems, administering medical and nursing care such as medications and treatments, initiating care plans to direct the care provided over time, and evaluating patient outcomes. Transitions in care, such as admissions and transfers from unit to unit or facility to facility, involve detailed documentation and communication supported by clinical information systems. Quality and safety measures are built into multiple workflows that are guided by national regulatory bodies, clinical care guidelines, and financial and meaningful use incentives. Even legal aspects of care, such as patients Interoperability Moment signing consent forms or leavWhat is the difference between nursing care documentation and ing the clinical setting against medical care documentation? Are they fundamentally different? Do medical advice, are important they incorporate different sets of values? Is it necessary to support clinical workflows that nurses interoperability between nursing care documentation and medical must attend to and document care documentation? What problems do you foresee? through the electronic health record (EHR).

Care Planning Nursing care plans contain a list of nursing problems called nursing diagnoses. This list identifies goals and specific interventions that nurses carry out to meet those goals with the patient. For example, a patient with leukemia undergoing chemotherapy is at risk for infection. The care plan outlines the goal of keeping the patient free of infections and the multiple interventions that need to be carried out during a 24-hour day, such as frequent handwashing, limiting sick visitors, and monitoring temperature. Nurses document their care on flow sheets in the EHR and in relation to the care plan. As a problem is resolved, nurses document the resolution to update the care plan. Care plans in an EHR format are often somewhat generic, and many institutions have found this aspect of the EHR to be in need of customization—especially when specialty populations, such as pediatric, transplant, or oncology patients, are involved. However transformational and disruptive we may desire change in health systems to be, it must take place in the context of current reality. Thus, we have

C hap ter 7 :   N ur sing Infor m atic s

to recognize that care planning is a focus of regulatory agencies when they evaluate care for credentialing. Whether care plans are current and tailored to patients are common areas of concern and evaluation by the agencies. Organizations detail and monitor their policies and procedures for care planning to ensure they are ready for the detailed review of surveyors from organizations such as The Joint Commission, the Centers for Medicare & Medicaid Services, and public health departments and other state agencies.

Care Transitions Transition-in-care workflows supported by the EHR include admissions, transfers, and discharges. These workflows often integrate the entire healthcare team, with different components of the workflow delegated to the role that is most suited for implementing and directing that aspect of care. Admissions When patients are admitted to the hospital, they undergo numerous evaluations, including social, emotional, physical, safety, and functional assessments. Nurses carry out these evaluations by asking the patient questions and gathering physical data such as height, weight, body mass index, and other data specific to areas of the body. Once the patient admission is completed, nursing care plans, including patient education plans, are created either manually or automatically based on the results of the admission assessments. Patient-centered care involves tailoring a patient’s plan of care and sharing this information with the patient and family. Transfers Transferring patients within a system of care requires a number of workflows that are detailed and complex. Nurses are involved in both the electronic and physical transfer of the patient, generating verbal reports and a handoff report in the EHR when the patient is ready for the actual transfer. Interoperability issues among health systems, which are discussed in other chapters of this book, affect transfers between health systems. There is almost always a nurse-to-nurse verbal communication, which relies heavily on the EHR, to review the plan of care, including details of treatments, medications, assessments, social and family support, and resources, as well as notes about whether communication with the patient and family has been easy or difficult. Discharges Discharging patients involves many different disciplines to prepare the patient and family to return home to continue outpatient care, home care services, or end-of-life home hospice care. Nurses orchestrate a number of different aspects of care when planning and implementing a patient’s discharge. Education of

153

154

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

the patient is a central workflow that nurses must coordinate so that the patient and family can continue care safely; this includes instructions on changing dressings, giving medications, and knowing what to do when problems arise. Coordinating care with other disciplines, such as physical therapists and dietitians, to ensure that all aspects of care are provided is often complex and time consuming. The EHR provides the structure for nursing coordination and preparation for discharge by producing checklists and workflow documentation that show a real-time picture of the patient’s readiness to go home. Informatics can support a detailed and comprehensive summary for patients and families, including patient home care instructions, electronic prescriptions, home care orders, and follow-up appointments and tests. The stakes for healthcare organizations are high because certain readmissions after discharge are currently not reimbursed. Thus, the discharge process and workflow has been an area that consumes many hours of interprofessional collaboration.

Quality and Safety of Care Medication Delivery and Safety Medication delivery is one of the most complex, easily disrupted, and risk-laden aspects of nursing care. Nurses are the final stop in a complex process of medication ordering by physicians or nurses, delivery by the pharmacy, and acceptance by the patient. Many opportunities for error exist throughout the process, and human factors can cascade into a serious adverse event for the patient. The Institute of Medicine report To Err Is Human estimated that medication errors cause 1 of 131 outpatient deaths and 1 of 854 inpatient deaths (Kohn, Corrigan, and Donaldson 2000). The documentation record for medications is generically termed the medication administration record (MAR). The electronic form of the MAR is often referred as the eMAR. In many systems, it is integrated with clinical decision support tools, such as drug references, patient information handouts, and alerts about system issues or individual dosing irregularities. In combination with the eMAR, the use of barcode medication administration (BCMA) has significantly improved patient safety. The most recent evidence reveals a 50 percent error reduction in medication delivery owing largely to the use of computerized physician-order entry (CPOE), a pharmacy dispensing system, and BCMA, which scans medications and patients before medications are administered to confirm that multiple orders align and that the medications are correct and safe to administer to the patient (Hassink et al. 2013). BCMA is most effective when combined with CPOE and eMAR (Nelson and Staggers 2018). Smart-pump technology for administration of IV or enteral tube medications and unit-dose medication dispensing stations provide additional support in this high-risk area of healthcare delivery. As with most emerging

C hap ter 7 :   N ur sing Infor m atic s

technology, security issues and cybersecurity threats are a concern with these medication systems. Given the significant amount of time nurses spend on medication administration, as well as the growing complexity of technology and increased emphasis on making the process safer for all involved, the amount of time and money spent on education for a large nursing workforce and on necessary technology updates is on the rise.

Medication Reconciliation and Safety Medication reconciliation occurs during patient admission, discharge, and transfer. In medication reconciliation, the patient’s current medications (including prescription, nonprescription, and supplements such as herbal remedies and vitamins) are reviewed. Nurses, physicians, and pharmacists are often the key professionals responsible for this critical aspect of care, which is driven by strict policies and procedures and includes highly complex clinical workflows. Getting a full picture of the medications a patient is taking is essential to safe care but is a challenging task. Prescribed medications can interact with herbal remedies, which patients and healthcare professionals can easily overlook. Moreover, patients may not be able to name or describe their most current medications. Hospitals and health systems spend a lot of time refining and reworking the care and information processes for medication reconciliation because it is one of the most challenging workflows in the EHR. On discharge, communicating this workflow to community pharmacies and then teaching the medication plan to patients and families (and making sure they understand it) are also challenging. E-prescribing has been a tremendous quality and safety component of the EHR, yet challenges with the workflow persist in every aspect of the prescribing process. There are opportunities for interprofessional workgroups to streamline and make e-prescribing safer for patients and families.

Clinical Decision Support Systems Evidence suggests that a clinical decision support system (CDSS) can positively affect healthcare providers’ performance by presenting preventive care reminders, medication information, clinical guidelines and screening tools, and drug–drug interaction and other trigger-based alerts (Jaspers et al. 2011; Nelson and Staggers 2018). State immunization registries may be integrated into the clinical information system (CIS) so that nurses can easily access resources and patient data that support timely administration of immunizations in any clinical setting where nurses see patients. Public health surveillance of a patient’s immunization status can now be integrated with hospital care in new ways, largely as a result of clinical informatics applications. Clinical care guidelines integrated into the CIS provide alerts and additional assessments on the basis of patient data and diagnoses. This automatic triggering of information brings evidence-based practice to the bedside for consideration in a timely and efficient

155

156

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

manner as care is delivered. One example of such evidence-based practice is a screening tool in acute care areas and emergency departments (EDs) for patients who are at risk for sepsis, a life-threatening emergency requiring life-saving tests, antibiotic initiation, and monitoring all within minutes of admission. Protocols are initiated and triggered on the basis of diagnoses, and data are entered into the record by a number of different clinicians. Safety and quality care may be achieved by implementing standardized, evidence-based drug and clinical reference applications in the CIS for use by multiple disciplines across the entire organization.

Adverse Events Detection of adverse events in hospitals is critical. Quality and safety issues affect patients and families directly; they may also affect the healthcare organization because regulatory and credentialing bodies report measures to the public, and reimbursement penalties are costly when errors occur. Detection has not been an automated process in the past, and it continues to rely on human reporting and detection. Nurses are often involved in this work as members of highly specialized teams that are assigned to quality and safety roles and departments. Classen and colleagues (2011) reported that, despite the decades-long focus on improving patient safety in the United States, current rates of adverse events among inpatients at three leading hospitals are still high, at 33.2 percent of adult hospital admissions. Recent patient safety efforts have focused on working with vendors to integrate surveillance and reporting systems into the EHR, along with trigger alerts to key people and roles when certain data values indicate problems. Diagnoses such as deep vein thrombosis or the use of Narcan, which reverses narcotic overdose, are two examples where detection by the EHR activates resources to review patient situations or records to determine potential problem areas. Multiple approaches to adverse event detection and reporting have been promoted to address these critical issues in healthcare.

Immunization Surveillance and Prescription Monitoring Programs Examples of key health information technology (HIT) applications that nurses regularly access are state-sponsored immunization registries and prescription drug monitoring programs (PDMPs). These applications are often integrated through interface connections with an organization’s EHR. Public health goals and regulatory and health promotion policy efforts push easy access and integration of these applications into the clinical workflow of nurses so that identification of risk areas may be assessed at multiple access points in the health system. Immunization registries are state-sponsored surveillance programs that provide a centralized data source for the immunization status of individuals in a population. Immunization needs can be identified by multiple health

C hap ter 7 :   N ur sing Infor m atic s

professionals who have access to the system’s records, including nurses, physicians, school personnel, and public health officials. One goal for health systems is to prevent overimmunization by tracking immunizations that are often administered in a variety of settings for both children and adults. As population health goals become a priority for health systems, access to and review of this information at every patient encounter is becoming a primary focus. In many cases, depending on the state, consumers can have access to the registry to keep track of and provide data to employers, schools, and childcare centers. Responsibility for monitoring the levels of immunization in health systems is often a part of the nurse’s workflow. Assessment of status, patient education, and documentation are all central to this workflow and are areas of expertise for nurses who promote health and work to prevent illness. These registries are excellent resources, but they present interoperability challenges because they are only state based. They are extremely useful from a state health perspective because they provide access to resources from the Centers for Disease Control and Prevention, such as standardized consent forms that all patients must be given, as well as links for reporting to a vaccine adverse event reporting system when adverse reactions occur. State-sponsored PDMPs are now active in 49 of the 50 states. These surveillance programs monitor certain controlled substances that are prescribed for conditions such as pain and attention-deficit disorder, to name just a couple. Medications for these conditions are at high risk for abuse; they can cause addiction issues and have become a diversion because of their high street value. Providers are required to monitor their prescription-writing practices and assess their patients’ prescription-filling practices to detect possible problems. Nurses are often delegates who monitor on a provider’s behalf, and nurse practitioners independently write prescriptions for patients. State laws are dictating the need for monitoring before writing a prescription for a controlled substance as a strategy to combat the current opioid epidemic and to increase patient and public safety. The overall impact and effectiveness of these programs have been mixed; however, more studies are needed given the policy efforts now in place and the level of opioid addiction in the United States (Chakravarthy, Shah, and Lotfipour 2012; Moyo et al. 2017; Pardo 2016). Access to these programs from multiple healthcare settings across a state is commonly integrated into the EHR for ease of use during patient encounters.

Consumer Engagement A critical element in the transformation of clinical care is the engagement of consumers in the process of care. Nursing informaticists play a central role in

157

158

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

using information management and systems to develop patient-centered care (HIMSS 2011): Nursing Informatics . . . has evolved to be an integral part of healthcare delivery and a differentiating factor in the selection, implementation, and evaluation of health IT that supports safe, high quality, patient-centric care. . . . When patient-centric processes encourage patient engagement, nurses and other healthcare team members across care settings can work in

Opportunity for Interprofessional Education

partnership to enhance the wellbeing of consumers.

We said that “a critical element in the transformation of clinical care is the engagement of consumers in the process of care.” Devise an exercise with your fellow students in which you incorporate patients as participants in their care. How would that modification of clinical care be transformative?

In this section, we discuss the roles of nursing informatics in communicating with patients and incorporating them as participants in their care.

Patient Education, Patient Portals, and Self-Management

Patient portal “Secure online website that gives patients convenient, 24-hour access to personal health information from anywhere with an Internet connection” (Office of the National Coordinator for Health Information Technology 2017)

Nurses work in multiple settings in which they assess patient and family knowledge and skills for self-management of chronic and acute conditions. Assisting with patient portal use and reviewing discharge instructions are areas in which nurses interface with patients independently, though they may consult other members of the healthcare team as needed. Knowledge of self-management theory and concepts and health literacy principles is critical in these interactions. Health literacy is “the degree to which individuals have the capacity to obtain, process and understand basic health information and services needed to make appropriate health decisions” (Baur 2011, 63). Nurses’ interactions with patients can take place in remote-access and telenursing encounters or in face-to-face visits in the clinic or hospital. Patient education materials—for example, information from Krames Patient Education or MedlinePlus from the National Library of Medicine— are often built into patient portals or integrated with the EHR by third-party vendors. Nurses are central to helping patients and families access these portals and encouraging them to sign up for portal use during encounters. Many factors affect patients’ comfort level with and trust in portals, and higher health literacy is one factor associated with greater ease and confidence in using a portal (Mackert et al. 2016). Barriers to portal use include security concerns, lack of technical skills or interest, and preference for face-to-face communication (Tieu et al. 2015). Facilitators of portal use include convenience, the ability to monitor one’s health, and improved patient–provider communication (Tieu et al. 2015). More evidence is needed to understand the multiple effects of portal use on patients and healthcare providers and on the overall

C hap ter 7 :   N ur sing Infor m atic s

costs of healthcare delivery. In addition to the many components of portals, the secure-messaging interface—where patients can send a message to their healthcare team to ask questions, request referrals or prescription refills, and seek urgent care evaluations—frequently involves nurses. Nurses independently manage certain aspects of patients’ health, and they also collaborate with providers as the intermediary between patients and providers. Answering the patient’s or family member’s questions and clarifying their understanding of instructions, laboratory or imaging results, medications, and continuing care needs are common. Nurses often provide additional resources for patient education, including explanations of how patients can access their discharge instructions from hospital stays, ED visits, or clinic visits. With the emergence of patient portals, the workflow for telephone management has changed drastically.

Telenursing Telehealth has been in existence for a number of years but has recently expanded to encompass a broader range of services delivered through telecommunication tools, including computer, phone, and video capabilities. Many telehealth services today can monitor patients remotely (e.g., medication delivery for hospice patients, vital signs for patients with chronic lung disease) or support body functions through implanted devices (e.g., heart-assist devices). Data from monitors can be transmitted in an instant or at regular intervals to a nurse in a healthcare setting, who in turn provides adjustments or reassurance based on the patient’s status and concerns. Incentives to keep patients at home and out of the hospital are increasing. The need to provide access to care in rural communities is also growing. Nurses play a central role in day-to-day monitoring, patient education, and other telehealth support services. Telehealth can ease transportation issues and issues caused by time constraints or the limited resources of family members (Vinson et al. 2011). Even minor urgent care needs can be managed through portals via symptom review with nurses, who can follow protocols and use their clinical judgment to implement treatment for a prescribed set of problems that do not require a physical exam, such as sinus infections or urinary tract infections. As an increasingly large percentage of the population ages, services provided at home in the context of home health not only are preferred but are also more efficient and timely, improving the quality and safety of complex care that once could be provided only in hospital settings.

Virtual Monitoring Nurses are often the beneficiaries of remote-monitoring HIT, which can augment their resources and detect acute and critical patient changes. Equipment that alerts nurses that a patient is active in a room and at risk for a fall is just

159

160

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

one application of virtual monitoring, which can provide technology-based resources in low-resource units. Although remote-monitoring HIT offers a number of benefits, it also presents challenges when the technology is interrupted or not working properly.

Nursing Education and Research According to the American Association of Colleges of Nursing (2018, 12), advanced nursing education programs require a standard core that includes data analysis and health informatics. Clinical informatics is central to healthcare delivery and the transformation of clinical care, so nursing informatics must be integrated into courses and curricula (Hunter, McGonigle, and Hebda 2013). The complexity of clinical instruction is important given that students often rotate through multiple settings, all of which require some sort of access to the EHR. Many organizations give prelicensure nursing students “read only” access. Complex educational requirements and security, safety, and quality issues are only heightened as dependence on technology increases.

Simulation Simulation is increasingly used in nursing education as a way to expose students to clinical situations and to prepare them for clinical encounters with patients. Including the EHR in such education enables the accurate simulation of the care environment, and discussing the challenges of integrating the EHR in the care delivery model is essential for optimal learning. Integration of HIT in monitoring conditions and care delivery is also an important component of a simulation experience. With healthcare costs skyrocketing, many large hospital systems have included simulation in their continuing education initiatives. More consideration by healthcare administrators of simulation-based education and training is needed to support and encourage the development of interprofessional collaboration among members of the healthcare team.

Research and Practice As nursing research and quality improvement projects advance nursing practice, the impact of interventions must be measured. Nurse scientists are nurses with a doctoral degree and research expertise. More and more nurse scientists are hired by clinical organizations to conduct research and advance the science of nursing and healthcare to benefit patients and providers. Access to patient data from the EHR is necessary for these purposes; questions about nursing’s effect

C hap ter 7 :   N ur sing Infor m atic s

on patient outcomes and ways patient care and outcomes can be improved are pursued in research. Advanced practice nurses practice as nurse practitioners, clinical nurse specialists, and nurse anesthetists; they are also central to the delivery of high-quality and safe healthcare. Their focus is on providing evidence-based care and integrating the best available evidence with their expert judgment and patient input. Advancing the practice of nursing relies on access and evaluation of data for quality improvement projects and evidence-based practice implementation. Access to EHR data must be improved to identify both problem areas and progress. Including nursing informatics in master’sand doctoral-level curricula is essential to the development of nursing science and the practice of nursing. Education on how to conduct studies and access data from the EHR in complex clinical settings is an important aspect of both academic and clinical training.

Conclusion Many opportunities for innovation and advancement beckon nurses to this exciting and highly specialized area of nursing and healthcare. Nurses, as the largest workforce in healthcare, are the biggest users of EHRs and healthcare technologies. Innovation is needed to solve multiple problems in healthcare delivery and to manage the complex and ongoing need for high-touch and high-tech solutions to healthcare’s problems. Opportunities for interprofessional collaboration and education have increased, and teams of researchers and clinicians are collaborating now more than ever as the clinical care setting becomes more integrated. Recognition of the overall impact of informatics on the quality of care delivered to patients and families, and of the safety value of working together rather than in silos, is critical for all health professionals.

Chapter Discussion Questions 1. How does informatics change with different roles and different levels of expertise in nursing? 2. What workflows in nursing are most challenging and laden with quality and safety issues? 3. What opportunities for interprofessional education do the many aspects of nursing informatics offer? 4. What are some of the challenges in nursing education related to clinical practice and informatics, both in prelicensure and in clinical practice?

161

162

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Case Study  A Question of Evidence Timothy B. Patrick and Norma M. Lang Chris, an academic health informatics specialist; Parker, a registered nurse; and Alex, a hospital clinical IT manager and strong proponent of electronic health records (EHRs), are discussing the virtues of EHRs for managing nursing care data and information and the obstacles to EHR-based clinical decision support. Parker: To show you that the EHR can’t handle complex situations, listen to this case. Selina Jones is 80 years old. She is admitted to the hospital’s medical-surgical unit after a fall. It appears she has sustained an injury to her hip that requires a surgical intervention. She is in only moderate pain but has difficulty moving. She’s nauseated, anxious, and calling for her dead husband. She doesn’t understand what happened to her, let alone the diagnostic tests, surgical plan, and other treatments she is and will be receiving. Selina’s daughter has assigned herself the authority to make healthcare decisions for her mom. The daughter agrees to have the surgeon do a surgical pinning of Selina’s hip. The registered nurse is responsible for doing the patient assessment and making the decisions on how best to prepare the patient for surgery, and the nurse has to communicate these decisions to other members of the clinical team. The surgeon and registered nurse are also responsible for creating the postsurgical plans, which include visits to a physical therapist and a social worker. Surgery is now complete. Selina is sent home, but she continues to have unmanaged pain, confusion, and anxiety, and she’s not able to participate in her own treatment or therapies or to understand the risk for falls. She can’t sleep well, has pressure ulcers acquired during her hospital stay, and suffers from urinary incontinence. Chris: Her family situation and medical issues do sound complicated. Alex, you must admit that your EHR is no panacea for this case. Alex: Of course it’s no panacea—nothing is. The golden rule in IT is “good data in, good information out.” If you enter inadequate data, you can’t expect complete results. Chris: “Good data in, good information out” certainly requires more than entering the usual ICD* and CPT* codes into the EHR. But let’s back up and focus on decision support. At the university hospital, we designed a set of standard nursing practice recommendations for assessment and interventions related to fall risks. So how could we determine whether nursing practice in your hospital conforms to our practice recommendations? Alex: The data fields in our EHR database have to match the key concepts in your practice recommendations.

C hap ter 7 :   N ur sing Infor m atic s

Parker: A standard approach matches both sides to a reference vocabulary, such as SNOMED CT.* Chris: Suppose your EHR used SNOMED CT or some other standard vocabulary in the first place? Then you really could achieve good data in, good information out. Alex: Yes, but not everything’s perfect. Besides, operations has its own pressures. As Shakespeare wrote, “There are more things in heaven and earth, Horatio, than are dreamt of in your philosophy.” Chris: Is matching of data and concepts enough? It seems that your data have to show that the right assessments and interventions (according to our recommendations) were taken. For example, according to our recommendations, incontinence is a risk factor for falls, so if the patient assessment included that, a protocol for fall prevention is implemented. Alex: Good example! And you could find that in our data. Chris: Always? And why would you find that? Parker: Because that protocol for that assessment is common practice? Alex: Our clinicians are experts and follow good practice in their care plans. Chris: In our recommendations, we cite clinical studies that provide evidence for incontinence as a risk for falls. Alex: Isn’t it enough if our practice and your recommendations agree on what ought to be done even if the reasons are not strictly the same? Chris: I don’t think so. Maybe it would be sufficient in a simple or isolated case, but I’m not comfortable with that position in general—and certainly not in a complicated case like that of Selina Jones. * ICD = International Classification of Diseases; CPT = Current Procedural

Terminology; SNOMED CT = Systematized Nomenclature of Medicine— Clinical Terms

Case Study Discussion Questions 1. Can the case of Selina Jones be managed by dealing with each of her problems separately? If not, how can an EHR keep track of her case? 2. Alex’s statement that “operations has its own pressures” implies a conflict between informatics and operations. Do you agree? Why or why not? 3. Describe a procedure for matching the EHR data fields to the practice recommendation assessment and intervention concepts. 4. What is Chris’s concern at the end? Do you agree or disagree? Why?

163

164

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Additional Resources Coughlin, S. S., J. J. Prochaska, L. B. Williams, G. M. Besenyi, V. Heboyan, D. S. Goggans, and G. De Leo. 2017. “Patient Web Portals, Disease Management, and Primary Prevention.” Risk Management and Healthcare Policy 10: 33–40. MedLinePlus: https://medlineplus.gov/.

References American Association of Colleges of Nursing. 2018. “Draft AACN Vision for Nursing Education Position Statement.” Published March. www.aacnnursing.org/ Portals/42/Downloads/Draft-Vision-Report4-18-2018.pdf. American Nurses Association (ANA). 2015. Nursing Informatics: Scope and Standards of Practice, 2nd ed. Silver Spring, MD: American Nurses Association. Baur, C. 2011. “Calling the Nation to Act: Implementing the National Action Plan to Improve Health Literacy.” Nursing Outlook 59 (2): 63–69. Bickford, C. J. 2009. “Nursing Informatics: Scope and Standards of Practice.” Studies in Health Technology and Informatics 146: 855. Chakravarthy, B., S. Shah, and S. Lotfipour. 2012. “Prescription Drug Monitoring Programs and Other Interventions to Combat Prescription Opioid Abuse.” Western Journal of Emergency Medicine 13 (5): 422–25. Classen, D. C., R. Resar, F. Griffin, F. Federico, T. Frankel, N. Kimmel, J. C. Whittington, A. Frankel, A. Seger, and B. C. James. 2011. “‘Global Trigger Tool’ Shows That Adverse Events in Hospitals May Be Ten Times Greater Than Previously Measured.” Health Affairs 30 (4): 581–89. Graves, J. R., and S. Corcoran. 1989. “The Study of Nursing Informatics.” Image: Journal of Nursing Scholarship 21 (4): 227–31. Hassink, J. J., M. Duisenberg-van Essenberg, J. A. Roukema, and P. M. L. A. van den Bremt. 2013. “Effect of Bar-Code-Assisted Medication Administration on Medication Administration Errors.” American Journal of Health-System Pharmacy 70 (7): 572–73. Healthcare Information and Management Systems Society (HIMSS). 2011. “Transforming Nursing Practice Through Technology and Informatics.” Position statement. Approved June 17. www.himss.org/sites/himssorg/files/HIMSSorg/ handouts/HIMSSPositionStatementTransformingNursingPracticethrough TechnologyInformatics.pdf. Hunter, K., D. McGonigle, and T. Hebda. 2013. “The Integration of Informatics Content in Baccalaureate and Graduate Nursing Education: A Status Report.” Nurse Educator 38 (3): 110–13.

C hap ter 7 :   N ur sing Infor m atic s

Jaspers, M. W., M. Smeulers, H. Vermeulen, and L. W. Peute. 2011. “Effects of Clinical Decision-Support Systems on Practitioner Performance and Patient Outcomes: A Synthesis of High-Quality Systematic Review Findings.” Journal of the American Medical Informatics Association 18 (3): 327–34. Kohn, L. T., J. M. Corrigan, and M. S. Donaldson. 2000. To Err Is Human: Building a Safer Health System. Washington, DC: National Academies Press. Mackert, M., A. Mabry-Flynn, S. Champlin, E. E. Donovan, and K. Pounders. 2016. “Health Literacy and Health Information Technology Adoption: The Potential for a New Digital Divide.” Journal of Medical Internet Research 18 (10): e264. McDonald, L. 2001. “Florence Nightingale and the Early Origins of Evidence-Based Nursing.” Evidence-Based Nursing 4 (3): 68–69. McGonigle, D., and K. Mastrian (eds.). 2018. Nursing Informatics and the Foundation of Knowledge. Burlington, MA: Jones & Bartlett. Moyo, P., L. Simoni-Wastila, B. A. Griffin, W. Onukwugha, D. Harrington, G. C. Alexander, and F. Palumbo. 2017. “Impact of Prescription Drug Monitoring Programs (PDMPs) on Opioid Utilization Among Medicare Beneficiaries in 10 US States.” Addiction 112 (10): 1784–96. Nelson, R., and N. Staggers. 2018. Health Informatics: An Interprofessional Approach, 2nd ed. St. Louis, MO: Mosby. O’Connor, J. J., and E. F. Robertson. 2003. “Florence Nightingale.” MacTutor History of Mathematics Archive. Published October. www-history.mcs.st-and.ac.uk/ Biographies/Nightingale.html. Office of the National Coordinator for Health Information Technology (ONC). 2017. “What Is a Patient Portal?” Reviewed September 29. www.healthit.gov/faq/ what-patient-portal. Pardo, B. 2016. “Do More Robust Prescription Drug Monitoring Programs Reduce Prescription Opioid Overdose?” Addiction 112 (10): 1773–83. Saba, V. K. 2001. “Nursing Informatics: Yesterday, Today, and Tomorrow.” International Nursing Review 48 (3): 177–87. Sensmeier, J. 2011. “Clinical Transformation: Blending People, Process, and Technology.” Nursing Management 42 (10): 2–4. Snyder-Halpern, R., S. Corcoran-Perry, and S. Narayan. 2001. “Developing Clinical Practice Environments Supporting the Knowledge Work of Nurses.” Computers in Nursing 19 (1): 17–26. Staggers, N., C. Gassert, and C. Curran. 2001. “Informatics Competencies for Nurses at Four Levels of Practice.” Journal of Nursing Education 40 (7): 303–16. Thompson, C. 2016. “The Surprising History of the Infographic.” Smithsonian Magazine. Published July. www.smithsonianmag.com/history/surprisinghistory-infographic-180959563/. Tieu, L., U. Sarkar, D. Schillinger, J. D. Ralston, N. Ratanawongsa, R. Pasick, and C. R. Lyles. 2015. “Barriers and Facilitators to Online Portal Use Among Patients

165

166

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

and Caregivers in a Safety Net Health Care System: A Qualitative Study.” Journal of Medical Internet Research 17 (12): e275. Vinson, M. H., R. McCallum, D. K. Thornlow, and M. T. Champagne. 2011. “Design, Implementation, and Evaluation of Population-Specific Telehealth Nursing Services.” Nursing Economics 29 (5): 265–72.

CHAPTER

E-HEALTH AND CONSUMER HEALTH INFORMATICS

8

George Demiris and Blaine Reeder

Learning Objectives After reading this chapter, you should be able to do the following: • Identify and differentiate the different platforms that support e-health applications. • Explain consumer health informatics concepts and their role in the design of e-health systems. • Construct the essential elements of a health social network. • Critically assess the implications and barriers of health social networks. • Identify ethical and practical considerations pertaining to the use of consumer-facing technologies.

Key Concepts • • • • •

Personal health record Social media Health social network Smart homes and Internet of Things Patient empowerment

Introduction E-health encompasses the use of telecommunications platforms, mobile (and ubiquitous) hardware and software, and advanced information systems to support and facilitate healthcare delivery and education. E-health has triggered a fundamental redesign of healthcare processes, integrating electronic communication at all levels and affecting all stakeholders. E-health also supports patient

E-health Use of telecommunication platforms, mobile and ubiquitous hardware and software, and advanced information systems to support and facilitate healthcare delivery and education

167

168

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Consumer health informatics Area of health systems informatics that focuses on the implementation and evaluation of system design to ensure that it interacts directly with the consumer, with or without the involvement of healthcare providers

engagement and even patient empowerment—the transition from a passive role (where the patient is the recipient of care services) to an active role (where the patient is involved in and perhaps even leads the decision-making process). Feste and Anderson (1995) emphasize that the patient empowerment model introduces “self-awareness, personal responsibility, informed choices and quality of life.” Precision medicine—a new healthcare paradigm that uses vast amounts of data from multiple data sources to identify and classify disease processes (McGrath and Ghersi 2016)—is expected to create a new era of personalized medicine in which individuals’ biological, physiological, behavioral, social, and environmental parameters, as well as their values and preferences, will inform tailored disease prevention and treatment. In this context, e-health can play a significant role in collecting information about individual patients’ needs and preferences and monitoring their physiological and behavioral parameters, wherever they may be. E-health bridges the clinical and nonclinical sectors and includes both individual- and population health–oriented tools. It encompasses different platforms, including telehealth applications that can span geographic distances (e.g., videoconferencing), web portals and mobile apps, online support groups, social media, wearable devices, and passive monitoring sensors. In addition, e-health delivers healthcare information, diagnoses, treatment, and care in a nonlinear manner where traditional hierarchies are obsolete and patients may enter the system at an infinite number of points, each with his or her own pattern and frequency of utilization. Healthcare lawyers are challenged “to determine whether they are dealing with the sale of a product or the supply of a service [and] whether to apply strict products liability or professional negligence” (Terry 2000). Advances in telecommunication technologies and data networks have introduced innovative ways to enhance communication between health professionals and patients. The result has been a shift in focus for informatics researchers and system designers, who had primarily aimed to design information technology (IT) applications that met the needs of healthcare providers and institutions by using data models that included episodic patient encounters as one group of healthcare-related transactions. The emerging model instead centers on the life course of individual patients and aims to ensure continuity of care and the inclusion of other stakeholders, such as family members, in the decision-making process. New technologies and informatics approaches call for the development of informatics tools that support patients and their families as active consumers in the healthcare delivery system. This shift from institutioncentered to patient-centered information systems requires new approaches to design and evaluation that examine and maximize the system’s effectiveness. Consumer health informatics is the area of health systems informatics that focuses on the implementation and evaluation of system design to

C h a p te r 8 :   E - health and C onsum er H ealth Infor m atic s

ensure direct interaction with the consumer, with or without the involvement of healthcare providers. It is a fast-growing subdomain of biomedical and health systems informatics that emphasizes the potential of informatics tools to engage and empower patients and equip them with the means to explore choices (Demiris 2016). This domain also emphasizes the value of informatics tools not only for those who become patients when they are diagnosed with a condition and find themselves interacting with the health system but for all health consumers who wish to engage in decision making about their wellbeing, disease prevention, and self-management. The applications and systems described in this chapter all belong to the domain of consumer health informatics and aim to support individual patients’ healthcare needs and preferences as well as those of their families.

Review of Patient-Centered Systems Consumer health informatics applications reach out to patients in their homes or in clinical settings. They help patients access and manage their own health documentation or information; link them to friends, peers, and others; actively engage them in health-related decision making; and provide them with tools for managing their disease or maintaining wellness. This section reviews homebased e-health applications and social networks and discusses the barriers to, and facilitators of, the successful adoption of patient-centered systems.

Home-Based E-health Applications E-health applications offer a platform to support disease management for patients diagnosed with chronic conditions and receiving care at home. These applications address numerous diseases and conditions. The use of technologies to facilitate symptom management for patients in the home has been explored for many years, and the body of evidence has clearly evolved from a focus on feasibility studies with small sample sizes in the early 1990s to today’s focus on large, randomized clinical trials (RCTs), such as the IDEATEL study in New York (Shea et al. 2009), the National Health Service–funded telemonitoring study for chronic obstructive pulmonary disease in Scotland (Pinnock et al. 2013), and the Tele-ERA study in Minnesota (Takahashi et al. 2010), to name just a few. Wootton (2012) synthesized the evidence pertaining to the use of telehealth for chronic disease management by reviewing a total of 141 RCTs on a variety of telehealth interventions delivered to 37,695 patients. Most of these RCTs demonstrated positive effects for telehealth with no significant differences between the various chronic diseases. A limited number of costeffectiveness studies were included in this body of literature. Although this evidence base reinforces the value of telehealth in home-based disease and

169

170

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Smart home Personal living quarters with an automated network of devices and systems that operate together by sharing data

Internet of Things (IoT) “Interconnection via the Internet of computing devices [including smart home devices] embedded in everyday objects, enabling them to send and receive data” (Höller and Höller 2014)

symptom management overall, the evidence is not solid for all chronic conditions or populations, and in some cases the evidence is contradictory as to the effectiveness or performance of specific telehealth services compared with traditional care (Wootton 2012). Home-based monitoring can be classified as active monitoring (where a patient or family member is asked to initiate the use of software or hardware, often requiring user training) or passive monitoring (where technology embedded in the residential infrastructure enables the assessment of parameters without anyone initiating or operating the system). When passive monitoring features are integral components of the home, the setting is often referred to as a smart home. Smart home projects that aimed to support the independence of older adults or people with disabilities emerged initially as research demonstration projects, but since they grew in number and in the type of technologies and settings used, sensor-based solutions for home monitoring are now commercially available (Demiris and Thompson 2012; Reeder et al. 2013). A more recent trend that is expected to disrupt traditional home-based care is the Internet of Things (IoT), defined as “the interconnection via the Internet of computing devices embedded in everyday objects, enabling them to send and receive data (Höller and Höller 2014).” Commercially available examples include home automation (e.g., the control of lighting and heating) and the recording of movement through motion tracking. Users can control such systems via a web interface or by voice interaction, which is made possible by the development of artificial intelligence (AI)–based voice assistant smart speakers that may enhance the ease of use of IoT systems. IoT smart home technologies have a unique opportunity to support home care patients who wish to remain independent at home by identifying potential health patterns, detecting anomalous activities, and increasing quality of life. Research has shown that many patient groups are motivated to receive data from smart home technologies that provide insight into their health status and from sensor systems that have the potential to enhance their lives (Lee and Dey 2010). Such technologies support the detection of health-related trends (e.g., decrease in overall activity level, increase in sedentary behavior, reduced number of visitors) that may require intervention, thereby helping patients maintain an independent living style by connecting them with family members, support systems, or other caregivers, and ultimately improving their quality of life.

Personal Health Records E-health applications support not only the transmission of data from one’s home to a clinical setting but also tools that allow patients to store and manage their own health information as a personal health record (PHR). The National Alliance for Health Information Technology (2008) defines a PHR as “an individual’s electronic record of health-related information that conforms to nationally recognized interoperability standards and that can be drawn from

C h a p te r 8 :   E - health and C onsum er H ealth Infor m atic s

multiple sources while being managed, shared and controlled by the individual.” The PHR is a tool to use in “sharing health information, increasing health understanding and helping transform patients into better-educated consumers of health care” (Kahn, Aulakh, and Bosworth 2009). In recent years, several initiatives have explored the design and implementation of such PHR tools. The Veterans Health Administration launched a PHR system called MyHealtheVet, which documents and manages appointments and medication requests (US Department of Veterans Affairs 2012). It also assists veterans in selecting and accessing a variety of healthcare services. Epic, the electronic medical record (EMR) software vendor, developed a PHR application that is currently used by Kaiser Permanente, the Cambridge Health Alliance, and other healthcare organizations. These PHR systems are widely used by consumers because they offer important functionality that could lead to improved health (Mechanic 2008). Between 2008 and 2013, the number of PHR users in the United States increased by more than 23 million people to a total of 31 million users, and models show that PHR adoption could reach rates as high as 75 percent of the US patient population by 2020 (Ford, Hesse, and Huerta 2016). Intervention studies have shown that PHRs have the potential to increase patient activation (Solomon, Wagner, and Goes 2012), promote health behaviors (Chrischilles et al. 2014), and improve the patient–clinician relationship (Nagykaldi et al. 2012); larger RCTs, however, have not replicated these gains (Wagner et al. 2012). PHRs help patients manage their own health by enabling the sharing of information such as health finances (e.g., bills, insurance claims), diagnoses or conditions, allergies, immunizations, and medications (Hassol et al. 2004). In these systems, the patient—not a healthcare facility or provider—owns and manages the data. Traditional EMRs, in contrast, are owned by healthcare organizations and maintained by clinicians and other staff. The integration of EMR and PHR systems is a synergistic model, where PHR data can augment EMR data and allow a collaborative continuum of care. Several barriers present challenges to realizing this synergistic vision, however, including legal and regulatory issues as well as sociotechnical concerns (e.g., clinicians’ lack of trust in data owned and generated by patients, acceptance of the patient’s active role, and changes in clinical workflow).

Social Media and Consumer Health Informatics Social media and social networking have seen widespread adoption in the past decade. The sharing of personal details, including health information, has increased since the advent of social media. Networking has been the subject of social science research since the 1950s (Ackerson and Viswanath 2009; Berkman 1984; Berkman et al. 2000; Bott 1957; Burt and Schøtt 1985; Fowler

171

172

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Web 2.0 technology Technology that replaced the traditional, static World Wide Web by enabling more community-based input, interaction, content sharing, and collaboration

and Christakis 2010; Heckathorn 1979; Israel 1982; Milardo 2000). Our discussion in this section focuses on social networking made possible by social media technologies that connect consumers and their health information. Social media uses Web 2.0 technology and user-generated content (UGC) to allow information, both health related and personal, to be communicated in new ways between individuals and groups (in contrast to the older one-way flow of information in healthcare from care providers to patients). One of the challenges in discussing social media is the lack of universally accepted definitions for social media or Web 2.0, which creates confusion (Adams 2010; Doherty 2008). Doherty (2008) notes that Web 2.0 “is essentially a set of technologies and the range of affordances made possible by those technologies.” Kaplan and Haenlein (2010), meanwhile, define Web 2.0 as “a platform whereby content and applications are no longer created and published by individuals, but instead are continuously modified by all users in a participatory and collaborative fashion.” Kaplan and Haenlein (2010) define UGC as content that is publicly available online, shows creative effort, and is created outside of professional practices. Social media uses Web 2.0 applications and UGC that allow multiway communication, collaboration, and democratic content management (Orsini 2010). This technology also facilitates communication by breaking down language barriers through natural language processing and machine translations (McNab 2009). Social media exchanges are instantaneous; conversations rather than directives; active rather than passive; connective, linking people with similar conditions and concerns; and representative of how lessons from experiences can be shared. Health 2.0 is defined as “the affordances of Web 2.0 technologies for the healthcare community whilst recognizing that these affordances are manifest in a variety of ways” (Doherty 2008). As a means of sharing personal health information, social media is a substantial tool. More than 50 percent of the world’s population were Internet users in 2017, and that number continues to grow (Internet World Stats 2018). The question now about using social media for healthcare purposes is how control of personal health information will shift from government and healthcare organizations to patients supported by private service providers (Kidd 2008). A related question is what impact such a shift might have on patient–provider relationships, quality of care, and efforts to equalize health disparities (Bacigalupe 2011). According to Boyd and Ellison (2007), social networking consists of “web-based services that allow individuals to (1) construct a public or semipublic profile within a bounded system, (2) articulate a list of other users with whom they share a connection, and (3) view and traverse their list of connections and those made by others within the system.” Other definitions and characteristics of social networking include the following:

C h a p te r 8 :   E - health and C onsum er H ealth Infor m atic s

• “An online location where a user can create a profile and build a personal network that connects him or her to other users” (Lenhart and Pew Internet & American Life Project 2007). • “Involves the explicit modeling of connections between people, forming a complex network of relations, which in turn enables and facilitates collaboration and collaborative filtering processes” (Eysenbach 2008). • Enables “users to connect by creating personal information profiles, inviting friends and colleagues to have access to those profiles, and sending e-mails and instant messages between each other. These personal profiles can include any type of information, including photos, video, audio files, and blogs” (Kaplan and Haenlein 2010). • Makes “it possible for users to branch into different conversations and create special relationships” (Landro 2006). Health social networking supports the accessibility and exchange of health information for consumers and is therefore considered a consumer health informatics resource. With a health and wellness focus, a health social network is powered by Health 2.0 applications. Swan (2009) defines a health social network as “a website where consumers may be able to find health resources at a number of different levels.” The types of interactions or communications that a health social network supports are patient to patient, patient to provider, and provider to provider (Doherty 2008). Patient-to-patient and patient-toprovider communications are most relevant to a discussion of consumer health informatics. Understanding the health implications of social networking entails learning who uses these sites. The wants and needs of patients and other customers must be central to the design and features of a health social network. Civan and colleagues (2006) conducted a group study of participants interested in managing their personal health information and determined that a health social network should deliver three goals: “monitoring and assessing health, making health-related decisions, and planning preventive or treatment actions.” Meanwhile, Weiss and Lorenzi (2007) offer four considerations when designing and rolling out a pilot health social network for patients with cancer: (1) It should have a mechanism to inform patients whether information is actively sent or passively posted, (2) it should have clearly delineated spaces to avoid unintentional personal disclosures, (3) it should offer a spell-check feature, and (4) it should ensure that patients’ family members do not think site invitations are spam. For patients with breast cancer, Skeels and colleagues (2010) argue that a health social network’s features should support dissemination of caregiving information and management of help requests and offers, and these features should be implemented using the Facebook Connect platform.

173

174

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

According to Eysenbach (2007), people obtain health information in three ways: (1) intermediation, where the person receives information from a health expert or information gatekeeper; (2) disintermediation, where the person eliminates the information gatekeeper; and (3) apomediation, where the person receives guidance from network intermediaries. Use of a web portal with content vetted by health experts, such as WebMD (www.webmd.com), is an example of intermediation; use of a patient-initiated web search is an example of disintermediation; and use of a health social network is an example of apomediation (Eysenbach 2008). The value of health social networking lies in its capability to connect a person with others who have similar health conditions and with whom that person can share information (Swan 2009). As Ancker and colleagues (2009) note, people can obtain or share “advice, interpretation of medical language or events, and personal experience. Such patient-generated information is likely to be written in common terms, rather than in medical jargon, and it may be easier to understand by those with lower health literacy or numeracy.” A folksonomy (a combination of folk and taxonomy) is generated when people tag digital information with their own keywords and classifications for later retrieval or use; these tags are then found by other searchers (Dye 2006). On a health social network, folksonomies provide insight about how patients use clinical terminology and understand their conditions; for example, some patients discuss type 1 diabetes as a symptom and not a disease (Smith and Wicks 2008). Online health communities connect citizens who are faced with challenges of symptom management, dealing with behavioral or lifestyle changes, or in need of emotional support. Facebook, for example, has been found to be a platform where individuals share health-related information, give advice and opinions, and find value in others’ personal experiences (Thrul et al. 2017). In many cases, “secret” Facebook groups—groups that are not searchable to outsiders and whose content is accessible only to members—have been created to provide online communities where peers can find support for behavior changes (e.g., dieting, smoking cessation), discuss symptom and disease management, and access emotional distress resources. The information on health social networks is crowd sourced, however, and may therefore be unreliable or inaccurate. One study reports on the “very low quality user-contributed health information on three different sites. Half of all postings containing medical information were incomplete or contained errors. Of these, over 80% were potentially clinically significant . . . [and thus may] compromise patient safety as a distribution platform for persuasive, personally tailored, but harmful misinformation” (Tsai et al. 2007). Web 2.0 technology enables the creation of online health environments or communication channels for consumers, including those with alternative or nontraditional health beliefs, to engage in group discussions with others who share and validate their viewpoints, which may or may not be accurate (Wilson and Keelan 2009).

C h a p te r 8 :   E - health and C onsum er H ealth Infor m atic s

175

Challenges in E-health Applications Factors that are critical to the success and diffusion of e-health applications include privacy/confidentiality and accessibility.

Privacy and Confidentiality An important and ongoing issue in e-health and consumer health informatics is the privacy and confidentiality of personal health information. Privacy is a person’s right to be free from and to refuse interference, attention, observation, and other types of invasion. In the context of healthcare, privacy is the assurance that (1) one’s health information is collected, accessed, used, retained, and shared only when necessary and only to the extent necessary and (2) the information is protected throughout its life cycle using fair privacy practices and consistent with applicable laws and regulations as well as individual preferences. Confidentiality is the obligation of every HIPAA (Health Insurance Portability and Accountability Act)–covered entity to enact and enforce policies that protect patients’ privacy. HIPAA addressed the need for comprehensive, national safeguards from threats to privacy and confidentiality, mandating the implementation of standards to keep personal health information safe, secure, and private (US Department of Health and Human Services 2000). Compliance with these standards greatly affected the design and functions of e-health applications. Multimedia transactions (e.g., video and audio recording, transmission of still images) not only must be able to conceal identifiable data but also must be performed through a secure infrastructure or platform (e.g., mobile devices, satellite, Internet). The widespread use of the Internet and Web 2.0 has heightened HIPAA compliance issues. For example, web-based applications for disease management require organizations to consider issues of data ownership and access, such as who owns the information stored on a vendor’s or third-party’s server and who has the authority to use those data. Similarly, personal health applications that empower patients to collect, store, and maintain their own information must be examined, and their ownership, access and monitoring rights, and potential confidentiality violations must be defined. Privacy and Social Media High levels of personal health information disclosed or posted on health social networks create risks for privacy abuses. Using Facebook as a framework, Grimmelmann (2009) conducted an in-depth analysis of the social and psychological factors behind people’s use of social networks and their privacy expectations. Grimmelmann asks, “What motivates Facebook users? Why do they underestimate the privacy risks? When their privacy is violated, what went wrong?” He argues that failure of social network operators to ask similar questions of their users will result in privacy policies that do not work (Grimmelmann 2009).

Privacy In the healthcare information context, assurance that one’s health information is collected, accessed, used, retained, and shared only when necessary and only to the extent necessary and that the information is protected throughout its life cycle using fair privacy practices consistent with applicable laws and regulations and the preferences of the individual

176

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Privacy issues related to health social networks include the following: • Some people may not know the potential risks of giving up their anonymity (thus undermining their privacy) when they participate in social networking (Adams 2010). • Personal information disclosed on social networks may “reinforce existing stereotypes, making them more intractable” (Ellison, Lampe, and Steinfield 2009). • “Personal information may be misused by marketing agents or used for nefarious purposes such as stalking, bullying, and identity theft” (Ellison, Lampe, and Steinfield 2009). • Personal information requests are more likely to be sent to “friends” via a phishing scheme on social networks (Jagatic et al. 2007). • Social network users must find a balance between the ease and convenience of widely sharing their information with a network of people and the need to protect their identity and associated information by customizing their privacy settings, which limits the risk of privacy invasion (Pratt et al. 2006). According to a survey of 205 college students about privacy concerns and risk-taking attitudes, “almost 10% of the participants provided their phone number on their social network profile” (Fogel and Nehmad 2009). Most young consumers on Facebook have taken steps to manage their privacy settings to some degree, however (Boyd and Hargittai 2010). Regardless, providers and parents should understand social networking to help mitigate the potential risks and benefits for teenagers who use these sites (Moreno et al. 2009). Allied health professionals, consumer groups, and communication professionals should push for policies that would display warnings about risk and privacy before young consumers are permitted to create social network profiles (Fogel and Nehmad 2009). Favoring the positive, Swan (2009) points out that “just as patients are the only ones who can avoid HIPAA privacy regulations and open source their own data to the benefit of the greater community, patients can skirt the social taboos that other health care ecosystem members may encounter regarding economic issues. . . . Providers would be forced to develop consumer-presentable health service offerings and pricing.”

Accessibility A significant number of consumers requiring home care services or managing multiple chronic conditions are older adults and, in many cases, may be experiencing functional limitations because of their clinical condition or age. Such limitations are the result of a decrease in cognitive, motor, or sensory abilities associated with temporary or permanent injury or aging. The premise

C h a p te r 8 :   E - health and C onsum er H ealth Infor m atic s

of consumer health informatics applications is the empowerment of consumers to play an active role in managing and making decisions about their personal health information. In this context, consumers older than 50 years may be at a disadvantage given that hardware and software designers often fail to consider them as a potential user group. Addressing cognitive, functional, or sensory limitations and recognizing consumers’ diverse levels of experience and proficiency in using software and hardware are essential to designing systems that maximize the positive elements of the user experience (i.e., increase the system’s usability and accessibility to the largest possible number of consumers). To achieve this goal, designers of e-health applications need to develop system features that increase the functional accessibility of their products and subject these features to rigorous usability testing with a wide range of consumers who have various capabilities, interests, and motivations. Existing resources, such as design recommendations developed for web systems for older adults (Demiris, Finkelstein, and Speedie 2001), or considerations in implementing telehealth systems to accommodate the functional, cognitive, and other needs of consumers (Stronge, Rogers, and Fisk 2007), can guide designers of e-health systems to maximize the accessibility of their products. Another dimension of accessibility for e-health systems is the availability of the appropriate infrastructure for using these systems. Several e-health applications, for example, require broadband Internet, which is not available in every residence and geographic region. Other web-based programs require peripheral devices, such as cameras or microphones, for synchronous communication; but it cannot be assumed that everyone has such peripheral equipment. When planning a wide implementation effort, healthcare organizations and their IT teams need to consider the technological restrictions faced by their patient population and community at large, especially in rural areas, to avoid excluding them and further increasing the digital divide.

Success Factors for E-health Factors that determine the success and sustainability of e-health applications include outcomes, processes, cost, and acceptance by patients, family members, and providers.

Outcomes If e-health applications are to be adopted as part of standard care, their outcomes should be the same as or better than those achieved by traditional care. The effect of e-health on clinical outcomes has been investigated to some extent, but large RCTs are still needed to clearly demonstrate its impact (Johnston et al. 2000). For example, when the effect of home-based e-health on medication

177

178

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

compliance and self-care ability was examined in a quasi-experimental study with a control group (receiving traditional care) and an intervention group (receiving traditional care plus access to a remote videoconference system), the effect was found to be no different from that of traditional care (Jerant, Azari, and Nesbitt 2001). However, a one-year RCT that involved congestive heart failure patients equipped with a two-way videoconferencing device and an integrated electronic stethoscope showed that e-health can reduce hospital readmissions and emergency visits for this patient population. It is generally assumed that most technologies used in telemedicine allow early detection and intervention through more frequent and intensive physiological monitoring. In addition, they can monitor medication and treatment compliance and promote patient education. The time has come to test this hypothesis by measuring the technology in large RCTs rather than in smallscale feasibility studies.

Processes In a face-to-face visit, the patient–provider communication does not include addressing technical issues such as focusing the camera or adjusting the audio. But these issues can be common in, or even dominate, a virtual visit. It is not fully understood whether video-mediated or web-based communication alters the relationship between clinicians and patients or how virtual encounters may create barriers to care delivery processes. Use of technology during clinical encounters may intimidate patients, limiting their participation in their care and inhibiting their communication with providers, because face-to-face interactions are considered “more spontaneous and free-flowing” than technologyenabled contact (O’Conaill and Whittaker 1997). This diminished willingness or ability to participate and engage is a serious threat to caregiving because patients greatly value the opportunity to ask questions and voice concerns when interacting with their clinicians (Ende et al. 1989; Street 1992). Active patient participation contributes to greater satisfaction with care, adherence to treatment, and improved health outcomes (Kaplan, Greenfield, and Ware 1989; Lerman et al. 1990). Studying care delivery processes that use telemedicine is therefore important. One study reviewed 122 virtual visits and performed content analysis to determine the themes of interaction during these visits (Demiris, Speedie, and Finkelstein 2001a, 2001b). The research showed that visit time is spent on the following categories of communication: assessing the patient’s medical status, promoting medication and treatment compliance, addressing psychosocial issues, exchanging informal banter, educating the patient on health issues, discussing administrative and technical issues, evaluating patient satisfaction, and ensuring continued accessibility to the provider. Clearly, some of these topics do not need to be covered during a virtual visit, but overall the discussions indicate that e-health has the potential to enrich the care process. Further studies and

C h a p te r 8 :   E - health and C onsum er H ealth Infor m atic s

direct comparisons between face-to-face and virtual visits will provide insights into the process of virtual patient–clinician encounters—for example, whether e-health encourages or inhibits a patient telling his provider about his physical discomfort, medical symptoms, and emotional state and, conversely, whether it encourages or prevents a physician giving her patient treatment instructions or expressing empathy (Bashshur 1995).

Cost A comprehensive evaluation of e-health applications must include a cost analysis to compare the inputs and outputs of e-health with those of traditional healthcare. Telemedicine evaluation methods have improved in recent years, but more work needs to be done to ensure that the benefits and outcomes are balanced with cost and other inputs (Bergmo 2010). Here, the inputs include medical expertise, facilities, technology, service personnel, and client characteristics. During a cost analysis, the effects of known quantities of traditional healthcare (e.g., episodes of care, hospital stays) should be assessed. Cost savings from the use of e-health can be realized if the following outcomes are demonstrated: • Reduction of unnecessary visits to the emergency department • Reduction of unnecessary or unscheduled visits to the physician’s office • Early detection and intervention • Patient education that leads to improvement of lifestyle choices and medication adherence • Prevention of repeat hospitalizations or overall decrease in rehospitalization rates • Reduction of indirect costs and burnout by easing the burden on caregivers The number of face-to-face consultations could, in some cases, be reduced if some of the consultations are replaced by virtual or web-based visits, which in turn eliminate travel time and travel costs. The use of portable devices and e-health technologies allows the collection and interpretation of vital signs data several times a day, rather than only at scheduled visits. This permits early detection and intervention, which is especially important if signs of deterioration are missed or problems are misidentified. In addition, telemedicine technologies enable family members and other caregivers to participate in a collaborative care process, adding to the patient’s support network.

Acceptance by Patients and Family Members One unique aspect of e-health is that the required technology is installed in the patient’s home and operated by the patient or the patient’s surrogate. The success of this form of healthcare delivery hinges on the patient’s and family’s

179

180

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

acceptance of its use. What influences this acceptance (and its diffusion) is the understanding of the concept of e-health. Considering the patient’s possible functional limitations and inexperience with the e-health technology, this initial acceptance is essential. Few survey instruments that have been tested for reliability and validity can measure patients’ perception of or satisfaction with e-health applications. One instrument that can is the Telemedicine Perception Questionnaire (TMPQ), which was developed by the University of Minnesota to assess patients’ perceptions of the advantages and disadvantages of e-health. TMPQ was tested extensively and was found to show high levels of internal consistency and test-retest reliability. Its domains include the following (Demiris, Speedie, and Finkelstein 2000): • Perceived quality of and access to healthcare • Time and money (e.g., time savings for the patient and nurse, reduced costs for the patient and the healthcare agency) • Components of the virtual visit (e.g., ease of equipment use, equal acceptability of virtual and real visits, protection of privacy and confidentiality, lack of physical contact, reduced sense of intimacy, patient’s ability to explain her medical problems in a virtual environment) • General impression of home-based e-health and its role in the future

Acceptance by Providers The success of e-health applications that involve healthcare providers (e.g., videoconferencing) obviously depends on the acceptance of both patients and providers. Many e-health applications alter providers’ practice patterns and affect their workflow. Thus, they have to accept this alternate mode of care delivery and be comfortable using the required equipment to interact with their patients. As is the case with all technological innovations, organizational commitment is essential to optimal use of telemedicine. This dependency can be a challenge given that many complex, institution-centric information systems do not support (at least not currently) the e-health infrastructure or endorse a strategic agenda for e-health applications. Adoption of e-health implies a restructuring of the institution and a redefinition of its services.

Conclusion E-health and consumer health informatics have seen significant advances in recent years. New technologies and pilot implementations in a variety of settings

C h a p te r 8 :   E - health and C onsum er H ealth Infor m atic s

have emerged, showcasing a potential shift to patient-centered care. The landscape of modern healthcare has undergone a transformation as a result of new laws and policies, particularly in the United States. Precision medicine is introducing new roles and responsibilities for clinicians, consumers, and system developers. The viable models and long-term effects of e-health interventions have yet to be determined, but consumers, aided by technology and acting as collaborators of care providers, are destined to play a large role in their own health and wellness. To determine the most cost-effective cause-and-effect models of e-health, studies in many different areas must be conducted. On the consumer side, researchers can determine who is using new technologies, in what ways, for what purposes, and how those technologies allow people to be empowered and to manage their own health and related information. On the clinical side, studies should explore how technologies such as social media can be integrated into the information systems in clinical settings. In addition, researchers must find ways to improve the reliability of the information that new technologies make available to healthcare providers, because these technologies can change the role of everyone who has a stake in the health and wellness of an individual. Healthcare leaders and administrators, for their part, must engage and include providers in planning, selecting, designing, and implementing new technologies that change the dynamic of the patient–provider relationship. From an administrative perspective, all stakeholders must help determine how an integrated health system can keep pace with new technologies and how reimbursement models can be adjusted to make them sustainable. Privacy and confidentiality are paramount in e-health and consumer health informatics. The questions of who owns personal health information, who can access it, and for what purposes it may be used must all be answered, and better means of protecting such information must be developed. In addition, the benefits and risks of emergent technologies to populations and communities must be weighed, particularly in the context of improving community and population health and engaging traditionally marginalized or disenfranchised populations. Equally important is understanding how new technologies can be used to address and minimize healthcare gaps and disparities and how they widen the digital divide between people who have access to the Internet and technology and people who do not.

Chapter Discussion Questions 1. What are the different platforms that support e-health? Give examples of each.

181

182

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

2. Define consumer health informatics, and differentiate it from traditional clinical informatics approaches. 3. Discuss the various types of social media, and give examples of how social media can be applied to e-health. 4. How would you design a reimbursement system that recognizes the value of e-health applications? 5. Identify the success factors in e-health. 6. How can a lack of or inadequate access to communication technology (the digital divide) lead to care inequality in the emerging e-health model? Identify the social and ethical issues associated with modern health information technology. 7. What are some of the cultural challenges in shifting from a centralized health information technology controlled by the organization to a decentralized health information technology controlled by the patient?

Case Study  Blue River Home Care Blue River Home Care is a for-profit home care agency affiliated with Blue River Hospital, a private 60-bed hospital. The agency is an early adopter of telehealth services that are now integrated into the home care plans of patients with chronic diseases. The telehealth technology integrates data from portable monitoring devices, including spirometers, blood pressure cuffs, digital weight scales, and videophones. After eight years, the agency is beginning to reap the benefits of investing in the telehealth infrastructure, which both enhances the quality of delivered services and reduces costs. Specifically, regular patient monitoring enabled by technology has reduced the staff’s travel time and costs; has made scheduling home care visits more efficient; and, in many cases, has led to proactive interventions to prevent adverse events. So far, it is too early to tell if rehospitalization rates for home care patients with chronic conditions will be reduced, but given that the technology enables early detection of symptoms, such reductions seem likely. Currently, the agency’s administration is looking for ways to keep pace with emerging technologies, which would give it a competitive advantage. The leaders are examining the integration of a personal health record (PHR) and a health social network into the existing telehealth infrastructure. This social network would allow patients to access and manage their personal health information, enable them to find and connect with others with similar conditions for support and sharing, and let family members and caregivers

C h a p te r 8 :   E - health and C onsum er H ealth Infor m atic s

participate in virtual support groups. The agency has held informal meetings with its healthcare providers regarding this plan. Some clinicians like the communication and information opportunities that the plan will offer patients and their caregivers, but other clinicians have some concerns, including the following: • The health social network might propagate unreliable or wrong information. • PHR integration could pose privacy (and thus liability) risks. • Patients may not use the health social network. • Physician compensation could be negatively affected by the online patient–clinician interaction. • Older patients, and those who have no experience with technology, may find it difficult to enter, manage, and find health information (especially to improve health literacy). These reactions are indicative of the diverse views and attitudes of healthcare providers, administrators, patients, and family members toward consumer health applications.

Solutions and Considerations Following are factors that Blue River (and other healthcare organizations) should consider and discuss when proposing and implementing an integrated PHR and health social network: • Communication between patients and their formal and informal caregivers will greatly improve. This, in turn, may lead to better health outcomes because such a system would 1. form or strengthen the social support for patients, 2. serve as an early detection tool for new health or medical events, 3. facilitate care coordination across various caregivers and care networks, and 4. increase the patient’s feeling of independence. • The viewpoint of each stakeholder group should be solicited and taken into account. • The potential benefits of an integrated PHR and health social network may be offset by problems introduced by the new technology. (continued)

183

184

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

• Although more information may be available to patients, they may not know how to interpret it for decision-making purposes. Most patients lack medical knowledge, so they could misunderstand health research findings, notes, discussions, and other information. The same holds true for family members who have access to their loved one’s PHR and social network profile. • Different approaches to integrating the PHR and health social network should be tried. For example, before patients are allowed to own, manage, and control their data, the organization should first develop a web portal that lets patients access, annotate, and share their information. This may address the patients’ desire to play a more active role in their own healthcare. • The technology will introduce new privacy and information reliability issues. Specifically, family members’ access to their loved one’s information and the ability of the social network’s members to view each other’s profile or pages could breach a patient’s privacy and confidentiality. In addition, if the patient is in charge of the PHR, the reliability of information may be compromised and affect the provider’s clinical decision making. • The organization must assess the usability and compatibility of system interfaces, regardless of whether the system or software was purchased from a vendor or created for the organization’s use. The system cannot be difficult to operate or navigate, especially for frail or elderly people (who represent most of Blue River’s patients) or those who have little or no experience with technology. Userfriendliness must be maximized to enable and encourage access to the system. Initial and ongoing training and customer support must be offered whenever feasible. • Because the proposed integrated PHR and health social network is intended to facilitate frequent communication among members of the network and allow patients to get involved in the decisionmaking process, the organization should build an IT infrastructure that supports these functions. • To gain institutional support, leadership must demonstrate that the new system will yield concrete benefits to the organization and its staff (e.g., increased marketability and competitiveness, improved quality of care, enhanced efficiency) without placing an undue burden or additional tasks on the staff. Furthermore, leadership must show evidence that the proposed system has proven effective in other clinical settings or similar organizations.

C h a p te r 8 :   E - health and C onsum er H ealth Infor m atic s

Case Study Discussion Questions 1. Name some of the challenges in ensuring continuity of care for patients with chronic conditions. How can emerging technologies solve or at least ease these challenges? 2. List some of the benefits of integrating a PHR with a health social network for patients with chronic illness, their caregivers, and the organization. 3. Name the specific effects that adopting e-health and social media applications will have on the structure and strategies of a healthcare organization. 4. How might the adoption of an integrated PHR and health social network change the interaction and relationship between patients and health professionals?

References Ackerson, L. K., and K. Viswanath. 2009. “The Social Context of Interpersonal Communication and Health.” Journal of Health Communication 14 (Suppl. 1): 5–17. Adams, S. A. 2010. “Revisiting the Online Health Information Reliability Debate in the Wake of ‘Web 2.0’: An Inter-disciplinary Literature and Website Review.” International Journal of Medical Informatics 79 (6): 391–400. Ancker, J., K. Carpenter, P. Greene, R. Hoffman. R. Kukafka, L. Marlow, H. Prigerson, and J. Quillin. 2009. “Peer-to-Peer Communication, Cancer Prevention, and the Internet.” Journal of Health Communication 14 (1): 38–46. Bacigalupe, G. 2011. “Is There a Role for Social Technologies in Collaborative Healthcare?” Families, Systems, & Health 29 (1): 1–14. Bashshur, R. L. 1995. “On the Definition and Evaluation of Telemedicine.” Telemedicine Journal 1 (1): 19–30. Bergmo, T. S. 2010. “Economic Evaluation in Telemedicine—Still Room for Improvement.” Journal of Telemedicine and Telecare 16 (5): 229–31. Berkman, L. F. 1984. “Assessing the Physical Health Effects of Social Networks and Social Support.” Annual Review of Public Health 5: 413–32. Berkman, L. F., T. Glass, I. Brissette, and T. E. Seeman. 2000. “From Social Integration to Health: Durkheim in the New Millennium.” Social Science & Medicine 51 (6): 843–57. Bott, E. 1957. Family and Social Network: Roles, Norms, and External Relationships in Ordinary Urban Families. London: Tavistock Publications. Boyd, D. M., and N. B. Ellison. 2007. “Social Network Sites: Definition, History, and Scholarship.” Journal of Computer-Mediated Communication 13 (1): 210–30.

185

186

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Boyd, D. M., and E. Hargittai. 2010. “Facebook Privacy Settings: Who Cares?” First Monday. Published August 2. http://firstmonday.org/article/view/3086/2589. Burt, R. S., and T. Schøtt. 1985. “Relation Contents in Multiple Networks.” Social Science Research 14 (4): 287–308. Chrischilles, E. A., J. P. Hourcade, W. Doucette, D. Eichmann, B. Gryzlak, R. Lorentzen, K. Wright, E. Letuchy, M. Mueller, K. Farris, and B. Levy. 2014. “Personal Health Records: A Randomized Trial of Effects on Elder Medication Safety.” Journal of the American Medical Informatics Association 21 (4): 679–86. Civan, A., M. M. Skeels, A. Stolyar, and W. Pratt. 2006. “Personal Health Information Management: Consumers’ Perspectives.” In AMIA Annual Symposium Proceedings, 156–60. Bethesda, MD: American Medical Informatics Association. Demiris, G. 2016. “Consumer Health Informatics: Past, Present, and Future of a Rapidly Evolving Domain.” Yearbook of Medical Informatics Suppl 1: S42–S47. Demiris, G., S. M. Finkelstein, and S. M. Speedie. 2001. “Considerations for the Design of a Web-Based Clinical Monitoring and Educational System for Elderly Patients.” Journal of the American Medical Informatics Association 8 (5): 468–72. Demiris, G., S. M. Speedie, and S. Finkelstein. 2001a. “Change of Patients’ Perceptions of TeleHomeCare.” Telemedicine Journal and e-Health 7 (3): 241–48. ———. 2001b. “The Nature of Communication in Virtual Home Care Visits.” In AMIA Annual Symposium Proceedings, 135–38. Bethesda, MD: American Medical Informatics Association. ———. 2000. “A Questionnaire for the Assessment of Patients’ Impressions of the Risks and Benefits of Home Telecare.” Journal of Telemedicine and Telecare 6 (5): 278–84. Demiris, G., and H. J. Thompson. 2012. “Mobilizing Older Adults: Harnessing the Potential of Smart Home Technologies. Contribution of the IMIA Working Group on Smart Homes and Ambient Assisted Living.” Yearbook of Medical Informatics 7: 94–99. Doherty, I. 2008. “Web 2.0: A Movement Within the Health Community.” Health Care and Informatics Review Online 12 (2): 49–57. Dye, J. 2006. “Folksonomy: A Game of High-Tech (and High-Stakes) Tag.” E-Content 29 (3): 38–43. Ellison, N. B., C. Lampe, and C. Steinfield. 2009. “Social Network Sites and Society: Current Trends and Future Possibilities.” Interactions 16 (1): 6–9. Ende, J., L. Kazis, A. Ash, and M. A. Moskowitz. 1989. “Measuring Patients’ Desire for Autonomy: Decision Making and Information-Seeking Preferences Among Medical Patients.” Journal of General Internal Medicine 4 (1): 23–30. Eysenbach, G. 2008. “Medicine 2.0: Social Networking, Collaboration, Participation, Apomediation, and Openness.” Journal of Medical Internet Research 10 (3): e22. ———. 2007. “From Intermediation to Disintermediation and Apomediation: New Models for Consumers to Access and Assess the Credibility of Health

C h a p te r 8 :   E - health and C onsum er H ealth Infor m atic s

Information in the Age of Web 2.0.” Studies in Health Technology and Informatics 129 (Part 1): 162–66. Feste, C., and R. M. Anderson. 1995. “Empowerment: From Philosophy to Practice.” Patient Education and Counseling 26 (1–3): 139–44. Fogel, J., and E. Nehmad. 2009. “Internet Social Network Communities: Risk Taking, Trust, and Privacy Concerns.” Computers in Human Behavior 25 (1): 153–60. Ford, E. W., B. W. Hesse, and T. R. Huerta. 2016. “Personal Health Record Use in the United States: Forecasting Future Adoption Levels.” Journal of Medical Internet Research 18 (3): e73. Fowler, J. H., and N. A. Christakis. 2010. “Cooperative Behavior Cascades in Human Social Networks.” Proceedings of the National Academy of Sciences of the United States of America 107 (12): 5334–38. Grimmelmann, J. 2009. “Saving Facebook.” Iowa Law Review 94 (4): 1137–206. Hassol, A., J. M. Walker, D. Kidder, K. Rokita, D. Young, S. Pierdon, D. Deitz, S. Kuck, and E. Ortiz. 2004. “Patient Experiences and Attitudes About Access to a Patient Electronic Health Care Record and Linked Web Messaging.” Journal of the American Medical Informatics Association 11 (6): 505–13. Heckathorn, D. 1979. “The Anatomy of Social Network Linkages.” Social Science Research 8 (3): 222–52. Höller, J., and J. Höller. 2014. From Machine-to-Machine to the Internet of Things: Introduction to a New Age of Intelligence. Cambridge, MA: Academic Press. Internet World Stats. 2018. “World Internet Usage and Population Statistics.” Accessed January 20. www.internetworldstats.com/stats.htm. Israel, B. A. 1982. “Social Networks and Health Status: Linking Theory, Research, and Practice.” Patient Counseling and Health Education 4 (2): 65–79. Jagatic, T. N., N. A. Johnson, M. Jakobsson, and F. Menczer. 2007. “Social Phishing.” Communications of the ACM 50 (10): 94–100. Jerant, A. F., R. Azari, and T. S. Nesbitt. 2001. “Reducing the Cost of Frequent Hospital Admissions for Congestive Heart Failure: A Randomized Trial of a Home Telecare Intervention.” Medical Care 39 (11): 1234–45. Johnston, B., L. Wheeler, J. Deuser, and K. H. Sousa. 2000. “Outcomes of the Kaiser Permanente Tele-home Health Research Project.” Archives of Family Medicine 9 (1): 40–45. Kahn, J. S., V. Aulakh, and A. Bosworth. 2009. “What It Takes: Characteristics of the Ideal Personal Health Record.” Health Affairs 28 (2): 369–76. Kaplan, A. M., and M. Haenlein. 2010. “Users of the World, Unite! The Challenges and Opportunities of Social Media.” Business Horizons 53 (1): 59–68. Kaplan, S. H., S. Greenfield, and J. E. Ware. 1989. “Assessing the Effects of Physician– Patient Interactions on the Outcomes of Chronic Disease.” Medical Care 27 (3, Suppl): S110–S127. Kidd, M. R. 2008. “Personal Electronic Health Records: MySpace or HealthSpace?” British Medical Journal 336 (7652): 1029–30.

187

188

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Landro, L. 2006. “The Informed Patient: Social Networking Comes to Health Care: Online Tools Give Patients Better Access to Information and Help Build Communities.” Wall Street Journal, July 12, D1. Lee, M. L., and A. K. Dey. 2010. “Embedded Assessment of Aging Adults: A Concept Validation with Stakeholders.” IEEE Xplore Digital Library. Accessed May 9, 2018. https://ieeexplore.ieee.org/document/5482302/. Lenhart, A., and Pew Internet & American Life Project. 2007. “Social Networking Websites and Teens: An Overview.” Published January 7. www.pewinternet. org/~/media/Files/Reports/2007/PIP_SNS_Data_Memo_Jan_2007.pdf.pdf. Lerman, C. E., D. S. Brody, G. C. Caputo, D. G. Smith, C. G. Lazaro, and H. G. Wolfson. 1990. “Patients’ Perceived Involvement in Care Scale: Relationship to Attitudes About Illness and Medical Care.” Journal of General Internal Medicine 5 (1): 29–33. McGrath, S., and D. Ghersi. 2016. “Building Towards Precision Medicine: Empowering Medical Professionals for the Next Revolution.” BMC Medical Genomics 9 (1): 23. McNab, C. 2009. “What Social Media Offers to Health Professionals and Citizens.” Bulletin of the World Health Organization 87 (8): 566. Mechanic, D. 2008. “Rethinking Medical Professionalism: The Role of Information Technology and Practice Innovations.” Milbank Quarterly 86 (2): 327–58. Milardo, R. M. 2000. Review of Family and Social Network, by E. Bott. Journal of Marriage and Family 62 (3): 861–63. Moreno, M. A., M. R. Parks, F. J. Zimmerman, T. E. Brito, and D. A. Christakis. 2009. “Display of Health Risk Behaviors on MySpace by Adolescents: Prevalence and Associations.” Archives of Pediatrics & Adolescent Medicine 163 (1): 27–34. Nagykaldi, Z., C. B. Aspy, A. Chou, and J. W. Mold. 2012. “Impact of a Wellness Portal on the Delivery of Patient-Centered Preventive Care.” Journal of the American Board of Family Medicine 25 (2): 158–67. National Alliance for Health Information Technology. 2008. “Defining Key Health Information Technology Terms.” Published April 28. www.himss.org/ defining-key-health-information-technology-terms-onc-nahit. O’Conaill, B., and S. Whittaker. 1997. “Characterizing, Predicting and Measuring Video-Mediated Communication: A Conversational Approach.” In Video Mediated Communication, edited by K. E. Finn, A. J. Sellen, and S. Wilbur. Mahwah, NJ: Lawrence Erlbaum. Orsini, M. 2010. “Social Media: How Home Health Care Agencies Can Join the Chorus of Empowered Voices.” Home Health Care Management and Practice 22 (3): 213–17. Pinnock, H., J. Hanley, L. McCloughan, A. Todd, A. Krishan, S. Lewis, A. Stoddart, M. van der Pol, W. MacNee, A. Sheikh, C. Pagliari, and B. McKinstry. 2013. “Effectiveness of Telemonitoring Integrated into Existing Clinical Services on Hospital Admission for Exacerbation of Chronic Obstructive Pulmonary

C h a p te r 8 :   E - health and C onsum er H ealth Infor m atic s

Disease: Researcher Blind, Multicentre, Randomised Controlled Trial.” British Medical Journal 347: f6070. Pratt, W., K. Unruh, A. Civan, and M. Skeels. 2006. “Personal Health Information Management.” Communications of the ACM 49 (1): 51–55. Reeder, B., E. Meyer, A. Lazar, S. Chaudhuri, H. J. Thompson, and G. Demiris. 2013. “Framing the Evidence for Health Smart Homes and Home-Based Consumer Health Technologies as a Public Health Intervention for Independent Aging: A Systematic Review.” International Journal of Medical Informatics 82 (7): 565–79. Shea, S., R. S. Weinstock, J. A. Teresi, W. Palmas, J. Starren, J. J. Cimino, A. M. Lai, L. Field, P. C. Morin, R. Goland, R. E. Izquierdo, S. Ebner, S. Silver, E. Petkova, J. Kong, and J. P. Eimicke. 2009. “A Randomized Trial Comparing Telemedicine Case Management with Usual Care in Older, Ethnically Diverse, Medically Underserved Patients with Diabetes Mellitus: 5 Year Results of the IDEATel Study.” Journal of the American Medical Informatics Association 16 (4): 446–56. Skeels, M. M., K. T. Unruh, C. Powell, and W. Pratt. 2010. “Catalyzing Social Support for Breast Cancer Patients.” In Proceedings of the 28th International Conference on Human Factors in Computing Systems, 173–82. Atlanta, GA: ACM. Smith, C. A., and P. J. Wicks. 2008. “PatientsLikeMe: Consumer Health Vocabulary as a Folksonomy.” In AMIA Annual Symposium Proceedings, 682–86. Bethesda, MD: American Medical Informatics Association. Solomon, M., S. L. Wagner, and J. Goes. 2012. “Effects of a Web-Based Intervention for Adults with Chronic Conditions on Patient Activation: Online Randomized Controlled Trial.” Journal of Medical Internet Research 14 (1): e32. Street, R. L. 1992. “Communicative Styles and Adaptations in Physician–Parent Consultations.” Social Science & Medicine 34 (10): 1155–63. Stronge, A. J., W. A. Rogers, and A. D. Fisk. 2007. “Human Factors Considerations in Implementing Telemedicine Systems to Accommodate Older Adults.” Journal of Telemedicine and Telecare 13 (1): 1–3. Swan, M. 2009. “Emerging Patient-Driven Health Care Models: An Examination of Health Social Networks, Consumer Personalized Medicine and Quantified Self-Tracking.” International Journal of Environmental Research and Public Health 6 (2): 492–525. Takahashi, P. Y., G. J. Hanson, J. L. Pecina, R. J. Stroebel, R. Chaudhry, D. N. Shah, and J. M. Naessens. 2010. “A Randomized Controlled Trial of Telemonitoring in Older Adults with Multiple Chronic Conditions: The Tele-ERA Study.” BMC Health Services Research 10: 255. Terry, N. P. 2000. “Structural and Legal Implications of E-health.” Journal of Health Law 33 (4): 605–14. Thrul, J., A. Belohlavek, D. Hambrick, M. Kaur, and D. E. Ramo. 2017. “Conducting Online Focus Groups on Facebook to Inform Health Behavior Change

189

190

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Interventions: Two Case Studies and Lessons Learned.” Internet Interventions 9: 106–11. Tsai, C. C., S. H. Tsai, Q. Zeng-Treitler, and B. A. Liang. 2007. “Patient-Centered Consumer Health Social Network Websites: A Pilot Study of Quality of UserGenerated Health Information.” In AMIA Annual Symposium Proceedings, 1137. Bethesda, MD: American Medical Informatics Association. US Department of Health and Human Services. 2000. “Standards for Privacy of Individually Identifiable Health Information. Final Privacy Rule Preamble.” Published December 2. http://aspe.hhs.gov/admnsimp/final/PvcPre01.htm. US Department of Veterans Affairs. 2012. “My HealtheVet.” Accessed May 17. www. myhealth.va.gov. Wagner, P. J., J. Dias, S. Howard, K. W. Kintziger, M. F. Hudson, Y. H. Seol, and P. Sodomka. 2012. “Personal Health Records and Hypertension Control: A Randomized Trial.” Journal of the American Medical Informatics Association 19 (4): 626–34. Weiss, J. B., and N. M. Lorenzi. 2007. “Challenges of Social Networking Technologies for Cancer Care and Social Support.” In AMIA Annual Symposium Proceedings, 1151. Bethesda, MD: American Medical Informatics Association. Wilson, K., and J. Keelan. 2009. “Coping with Public Health 2.0.” Canadian Medical Association Journal 180 (10): 1080. Wootton, R. 2012. “Twenty Years of Telemedicine in Chronic Disease Management— an Evidence Synthesis.” Journal of Telemedicine and Telecare 18 (4): 211–20.

CHAPTER

PRECISION MEDICINE

9

Timothy B. Patrick and Aurash A. Mohaimani

Learning Objectives After reading this chapter, you should be able to do the following: • • • •

Explain the concept of precision medicine. Describe the characteristics and challenges of Big Data. Discuss the relevance of Big Data to precision medicine. Understand the risks associated with the Internet of Things as a source of data for precision medicine. • Explain what the crisis of reproducibility is, why it matters, and why it is important for precision medicine.

Key Concepts • • • • •

Precision medicine Exome GenBank Big Data Crisis of reproducibility

Introduction Precision medicine is a kind of evidence-based medicine that applies the results of scientific research to patient care by taking into account the variations among many patients and the unique characteristics of the individual patient whose health is the object of clinical care. Of course, disease prevention and treatment have always been intended to be as precise as possible—that is, tailored to the individual circumstances of the patient. For example, “a person who needs a blood transfusion is not given blood from a randomly selected donor; instead, the donor’s blood type is matched to the recipient to reduce the risk of

Precision medicine “Innovative approach to disease prevention and treatment that takes into account individual differences in people’s genes, environments, and lifestyles” (White House 2015)

191

192

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

complications” (National Library of Medicine 2018). What is radically new is the degree of precision possible. Instead of providing treatment for an average patient, treatment can be tailored to entire subpopulations and to individuals within those subpopulations. One study of Warfarin protocols, for example, reports a clinical trial simulation in which “a combination of age and genotype identified different optimal [Warfarin] protocols for various subpopulations” (Ravvaz et al. 2017).

Precision Medicine and Genomic Science A major factor in the development of precision medicine has been the advance of genomic science and its application to disease prevention, diagnosis, and treatment. One of the earliest and still one of the most astounding examples of the precision made possible through genomics is the case of Nicholas Volker (Johnson and Gallagher 2010): On a Saturday morning in June . . . pediatrician Alan Mayer composes the e-mail he hopes will persuade a colleague to try a costly new technology. He has been shaping the argument in his mind—the chance to take the first steps into the future of medicine and maybe save the life of a very sick little boy. “Dear Howard—I hope you are well,” he writes, addressing Howard Jacob, director of the Medical College of Wisconsin’s Human and Molecular Genetics Center. “I’m writing to get your thoughts on a patient of mine. . . .”

According to Okie (2016), at the age of two Volker had developed symptoms of a horrific illness that inflamed his intestines whenever he ate. They’d spring tiny holes that led to leakage, infections, malnutrition and wounds that refused to heal. Undersized and emaciated, nourished by tube and intravenous feeding, he pleaded for the food he couldn’t have.

Howard Jacob and his team at the Medical College of Wisconsin sequenced Volker’s exome—the protein-coding portion of the genome—and “traced the cause of his disease to a single error in the sequence of 3.2 billion chemical bases in his DNA. As a result of the diagnosis, Volker received an umbilical cord blood transplant that appears to have saved his life” (Johnson and Gallagher 2015). Before the work of Jacob and his team, “there was no published record of any doctors sequencing someone’s genes and using the information to pinpoint the cause of a disease” (Johnson and Gallagher 2017, 113). Since that initial case in 2009, the sequencing of exomes in clinical practice had increased dramatically by 2015 (Johnson and Gallagher 2015):

C hap ter 9 :   Prec ision Medic ine

In 2009, it cost some $75,000 to sequence Nic’s exome. . . . The cost of sequencing an exome has now dropped to about $7,000, and the practice has gained broader acceptance, according to David Adams, deputy director of clinical genomics at the National Human Genome Research Institute. “Over the past five years, the coverage of exome sequencing by insurance companies has been increasing,” Adams said, explaining that this reflects a recognition that sequencing can be a cost-effective alternative to conducting numerous tests of individual genes. Adams said that a straw poll of experts across the country indicates that doctors have sequenced the exomes of between 10,000 and 20,000 patients and relatives in the last 12 months. In roughly a quarter of all cases, sequencing leads to a diagnosis, he said.

Precision Medicine Initiatives The 2016 US federal budget included $215 million to launch the Precision Medicine Initiative, which aims to develop “patient-powered research” intended to support advances in precision medicine that are (White House 2015) leading to a transformation in the way we can treat diseases such as cancer. Patients with breast, lung, and colorectal cancers, as well as melanomas and leukemias, for instance, routinely undergo molecular testing as part of patient care, enabling physicians to select treatments that improve chances of survival and reduce exposure to adverse effects.

A key feature of the continuing Precision Medicine Initiative is the “All of Us” research program, which “seeks to extend precision medicine to all diseases by building a national research cohort of one million or more U.S. participants” (National Institutes of Health 2018). Other efforts toward the development of precision medicine include the Brazilian Initiative on Precision Medicine (BIPMed) and the Human Variome Project (HVP). BIPMed (2018) is based on a software platform, built following the guidelines and principles of the Global Alliance for Genomics and Health (GA4GH), and observing the responsible sharing of genomic and clinical data. This platform is the first of its kind in Latin America and aims to offer public access to genomic and phenotypic data. It is intended to be used by clinicians and scientists all over the world, to share and obtain information about various aspects of genomic medicine and human health, as well as to support dissemination and training in the areas of human molecular genetics, computational biology and others.

193

194

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

With aims similar to those of BIPMed, the HVP (2018) is an international non-governmental organization that is working to ensure that all information on genetic variation and its effect on human health can be collected, curated, interpreted and shared freely and openly.

The HVP (2018) maintains “operational relations with the United Nations Educational, Scientific and Cultural Organisation (UNESCO).” HVP and BIPMed are operationally related, as BIPMed is the Brazilian country node of the HVP (BIPMed 2018).

Precision Medicine and Popular Culture Precision medicine has begun to penetrate the US popular culture and imagination largely through cancer treatment advertisements, such as those by the Cancer Treatment Centers of America (CTCA). A recurrent theme in these advertisements is genomic-based precision medicine, as reflected in this statement in a recent CTCA (2018) commercial: Advanced genomic testing, now offered by Cancer Treatment Centers of America, reveals the genomic profile of the patient’s individual cancer, to help guide more precise cancer treatments.

The breadth of such popular advertising is worth noting (Vater et al. 2016): Cancer Treatment Centers of America . . . had the largest advertising expenditures [in 2014], accounting for 59% of total advertising spending by cancer centers. Cancer Treatment Centers of America spent $101.7 million, consisting of $58.7 million for national advertising, $24.2 million for local advertising, and $18.7 million for Internet advertising.

The packaging for many drugs includes labeling information about pharmacogenomic biomarkers. The US Food and Drug Administration (FDA) maintains a table of those drug labels and the relevant biomarkers on its website at www. fda.gov/Drugs/ScienceResearch/ucm572698.htm (FDA 2018).

Precision Medicine and Big Data One of the most remarkable aspects of the Volker case is the contribution of Elizabeth Worthey, a clinical bioinformatician and data expert. Worthey and her team of programmers developed a software program that was used in the Volker

C hap ter 9 :   Prec ision Medic ine

case to narrow the suspected gene mutations from 16,000 to 32 and finally to a list of 8 potential errors responsible for Volker’s condition (Okie 2016). Since its beginnings, genomics has been characterized by vast amounts and variety of data, such as those confronted by Worthey, as well as a plethora of databases and software resources for producing and analyzing the data. One example of the dramatic increase in genomic data available for research and analysis is the history of the public DNA sequence database GenBank. In 1979, Walter Goad at Los Alamos National Laboratory created the Los Alamos Sequence Database, which developed into the DNA sequence database GenBank (Cinkosky et al. 1992). In December 1982, GenBank’s release 3 (the third time GenBank released a list of submitted sequences for public view) contained 606 sequences. By November 1983, release 14 contained 2,427 sequences. The number of sequences in GenBank has increased exponentially since 1983, doubling every 18 months. The December 2017 release 223 included 206,293,625 sequences (National Library of Medicine 2017a). Since the mid-1980s, GenBank has been a partner, along with the DNA Data Bank of Japan (DDBJ) and the European Molecular Biology Laboratory’s European Bioinformatics Institute (EMBL-EBI), in the global International Nucleotide Sequence Database Collaboration (Karsch-Mizrachi, Nakamura, and Cochrane 2012). GenBank, DDBJ, and EMBL-EBI exchange sequence data on a daily basis (National Library of Medicine 2017b). In general, the growth in the variety and amount of data being produced and available for analysis has been alarming. An opinion piece published in 1990 in Science (Waldrop 1990) discussed this enormous collection of information, which is now called Big Data, and its implications. The article pointed out four fundamental challenges for Big Data, although this term was not used at the time: (1) A great variety of data from different scientific domains must be integrated to support decision making, including clinical decision making; (2) coordination of data resources scattered across the scientific and health communities is lacking; (3) comparability and consistency of the data collected by different researchers are needed; and (4) massive volumes of data must be analyzed. Barbara Mihalas, at that time the head of the National Center for Supercomputing Applications’s effort to build a digital library, explained (Waldrop 1990) that To get at the underlying causes of, say, liver cancer or diabetes, a researcher may need to integrate information from molecular biology, genetics, cell biology, pathology, patient records—“all the data from all experiments looking at the processes that are relevant.” . . . And yet the reality is that the databases are typically scattered through offices and institutions all over the country, if not all over the world. As if that were not enough of a problem, there is the sheer overwhelming volume of data involved, especially as mega-projects such as the Human Genome Project and the Hubble Space Telescope start pouring forth data by the gigabyte.

195

Genomics “A branch of biotechnology concerned with applying the techniques of genetics and molecular biology to the genetic mapping and DNA sequencing of sets of genes or the complete genomes of selected organisms, with organizing the results in databases, and with applications of the data (as in medicine or biology)” (Merriam-Webster 2018) GenBank “A comprehensive database that contains publicly available nucleotide sequences for almost 260,000 formally described species” (Benson et al. 2013) Big Data In healthcare, very large collections of data, typically represented in a variety of ways, on individual patients, populations of patients, animals, biological objects, and other healthrelated matters

196

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

The scale of data volume considered in the article is now outdated, but the general issues discussed by Mihalas and Waldrop continue to be relevant today, nearly 30 years later. The journal Nucleic Acids Research maintains the online Molecular Biology Database Collection, which contains 1,737 databases (Rigdin and Fernandez 2018). The number of software tools for analyzing the data in all those databases continues to grow. The journal’s 2017 web server issue describes 86 web-based and stand-alone analysis tools for molecular biology. The 2015 and 2016 web server publications described 97 and 115 web-based analysis tools, respectively. To underscore the systems approach and perspective that characterize the treatment of health informatics in this book, and to emphasize that technological advances in precision medicine should not be considered in isolation from the general advances in information technology (IT), we point out that the challenges of Big Data are not exclusive to molecular biology, genomics, or precision medicine but also apply to the US intelligence community. As Lieutenant General Robert Ashley, director of the Defense Intelligence Agency, reported at the 2018 US Senate Select Committee on Intelligence hearing on global threats and national security, near peer opponents of the United States—presumably both state and nonstate actors—are investing heavily in machine-learning technologies to help them “get to decision cycles faster, allow them to digest information in greater volumes, and have a better situational understanding in the battlespace, and in some cases just what’s happening in the strategic environment” (C-SPAN 2018, 42:39–42:50). At the same hearing, referring to his counterparts from other intelligence agencies who were present, Admiral Michael Rogers, director of the US National Security Agency, commander of the US Cyber Command, and chief of the Central Security Service, said (C-SPAN 2018, 42:51–43:45): Every organization at this table is faced with the same challenge—victims of our own success in some ways. The ability to access data at increased levels brings its own set of challenges. . . . I can remember five, ten years ago, looking at some data concentrations and thinking to myself, “this is so large and has such a disparate amount of information in it, boy it would be really difficult for an opponent to potentially generate insider knowledge from it.”. . . I don’t have those kinds of conversations anymore.

In addition, cybersecurity challenges related to the mass collection and storage of data in healthcare, such as those discussed in chapter 16, are also of concern to the intelligence community. When asked by Senator Richard Burr, chair of the intelligence committee, whether he expected an increasingly challenging environment with the emergence of new IT firms on an ongoing basis, Admiral Rogers said (C-SPAN 2018, 38:23–38:46):

C hap ter 9 :   Prec ision Medic ine

Quite frankly I wonder how bad does this have to get before we realize that we have to do some things fundamentally differently, and I would argue that if you look at the Internet of Things—you look at the levels of security within those components, folks, this [is] going to—orders of magnitude—if we think the problem is a challenge now—we just wait, it’s going to get much, much worse—exponentially from a security perspective.

As suggested in chapter 16, a source of Big Data—and potentially a valuable resource for precision medicine—is the Internet of Things (IoT). Yet, as valuable as the healthcare IoT may be as a resource for precision medicine, it is also—as Admiral Rogers pointed out regarding the IoT—potentially fraught with threats to information security and privacy.

Precision Medicine and Scientific Reproducibility This book discusses health informatics from a systems perspective, and when viewed from a systems perspective, the activity of precision medicine is remarkable. The health of each human and each patient is affected by larger interacting systems of activity and influence. For example, according to the One Health approach championed by the World Health Organization (2017), the health of humans at the individual level is affected by interacting forces of human populations, animal and plant health (including the food supply), and environmental conditions. Successfully winnowing through the massive amounts of data collected from these interacting systems and then carrying out the scientific research that provides the evidence to guide precision medicine require adherence to standards of good science. The effective pursuit of modern science is typically characterized by reproducibility. Reproducibility requires that meaningful scientific research be presented with transparency in mind, such that results of reported analyses may be replicated to exact—or nearly exact—effect. As Jonathan Russell (2013) has written, “Reproducibility separates science from mere anecdote.” Reproducibility in science is so paramount that its value surpasses that of immediate clinical results: “Not every paper needs to be medically relevant, but at the very least they should all be reproducible” (Russell 2013). Domain-specific attempts at improving study reproducibility have been made across a diverse range of fields, including paleolimnology, cardiology, psychology, and nutrition (Flint et al. 2000; Heiri, Lotter, and Lemcke 2001; Open Science Collaboration 2015; Yamashina et al. 2002). Some notorious scientific results have been rejected because they were not reproducible. The claimed discovery of N-rays by French scientist René

197

198

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Blondlot in 1903, which resulted in a large number of publications, ultimately could not be reproduced by other scientists and was rejected by 1904 (American Physical Society 2007). A more recent, and perhaps even more notorious, case is the claimed discovery by Martin Fleischmann and Stanley Pons at the University of Utah of a mechanism for producing cold fusion. The results announced by Fleischmann and Pons in the late 1980s were deemed not reproducible and were rejected, although research into fusion as a practical source of power continues today (MacRae 2014). Science—specifically the biomedical research that supports evidencebased medicine—is reportedly facing a “crisis of reproducibility” (Forbes 2017): In 2011, Glenn Begley, who ran the oncology division at Amgen, decided to try to reproduce 53 foundational papers in oncology. He was unable to reproduce 47 of them, which is 89%. Bayer, another pharmaceutical company, reported in the same year that it was unable to reproduce 65% of the papers in its sample of the biomedical literature.

Lack of adequate documentation of study methods may be one cause of this crisis. Problems with adequate documentation of studies in biomedical research are not new. For example, Patrick and colleagues (2004) studied meta-analyses in healthcare and found that methods of information retrieval, a crucial aspect of any meta-analysis, often were not documented in enough detail to allow the study to be replicated. Such deficiencies may be a result of the growth of Big Data itself, particularly in genomic science. In this, we may be “victims of our own success,” to borrow a comment of Admiral Rogers’s at the senate hearing. As mentioned, the available data in a number of genomic data sets have multiplied, and so have the available tools to analyze those data. These increases, particularly in the past 15 years, may be attributed to the introduction of nextgeneration sequencing (NGS) platforms (Morozova and Marra 2008). NGSs have enabled mass parallelization of both DNA- and RNA-sequencing experiments, such that whole organism genomes and transcriptomes may now be sequenced for subsequent analysis at costs that are many orders of magnitude less than the investment needed for the Human Genome Project in the 1990s. This precipitous drop in sequencing cost has far outpaced even Moore’s Law (National Human Genome Research Institute 2017). It has been a challenge to make the results of scientific studies associated with these data publicly available in a manner that supports reproducibility of experiments and studies. As an example, consider the Gene Expression Omnibus (GEO) of the National Center for Biomedical Information. GEO serves as a transparent data repository for all manner of microarray and high-throughput sequencing experiments (Barrett et al. 2013). To encourage submission of researchers’ work, GEO imposes minimal restrictions on user-defined metadata

C hap ter 9 :   Prec ision Medic ine

attributes and enforces no apparent standard on the results of sequencing of data sets. Although GEO’s flexibility has apparently contributed to its success, at least in volume, it has also resulted in poor documentation of studies as a data-collection point. For example, many studies reported in GEO fail to specify the names of the bioinformatics tools used, tool versions, usage parameters, and even hardware specifications for runtime. Such gaps in the documentation of these studies render them nearly impossible to replicate. Poor documentation of studies may not be the only cause of the crisis of reproducibility. A lack of proper understanding of the data analytics necessary for dealing with Big Data may also be contributing to the crisis (American Statistical Association 2015): [In] a Duke University cancer research project in 2006 . . . researchers published a paper claiming they had built an algorithm using genomic microarray data that predicted which cancer patients would respond to chemotherapy. A subsequent attempt to reproduce the results found a morass of poorly conducted data analyses with errors ranging from trivial and strange to devastating. The original study was retracted by Nature Medicine in 2011.

As a response to the gaps in knowledge among scientists regarding data analytics, the National Library of Medicine (2017c) recently issued a request for information (RFI) titled “Next-Generation Data Science Challenges in Health and Biomedicine.” One specific focus of the RFI is “Promising directions for new initiatives relating to open science and research reproducibility.” A combination of lack of documentation standards and lack of data analytical skills and knowledge is likely the major cause of the reproducibility crisis. As Collins and Tabak (2014) argue, the checks and balances that once ensured scientific fidelity have been hobbled. This has compromised the ability of today’s researchers to reproduce others’ findings. . . . Crucial experimental design elements that are all too frequently ignored include blinding, randomization, replication, sample-size calculation and the effect of sex differences. And some scientists reputedly use a “secret sauce” to make their experiments work—and withhold details from publication or describe them only vaguely to retain a competitive edge. What hope is there that other scientists will be able to build on such work to further biomedical progress?

Conclusion The Volker case represents the promise and the enduring fact of precision medicine. Yet many challenges remain in the development of precision medicine, such as Big Data, data security, and better standards for reporting studies and providing

199

200

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

a secure foundation for evidence-based medicine. Perhaps most important, the development and success of precision medicine depend on our adoption of a systems approach to evidence-based medicine and the informatics that supports it.

Chapter Discussion Questions 1. Given enough genetic information, a precise identification of a person can be made. How do you think data warehouses that integrate clinical information with genomic information can be used to accomplish meaningful research while protecting patient privacy? 2. What are some means by which clinicians can use information systems to manage and respond to advances in genomic medicine? 3. Why is the Internet of Things a risk for information security and privacy? 4. Do some research on the costs of sequencing a genome and sequencing an exome. As a hospital administrator, what policy or policies would you propose for choosing between the two approaches? Why? 5. Searching for and accessing information on the Internet was once described as “drinking water from a firehose when you don’t know where the water is coming from.” Explain that phrase in terms of Big Data, evidence-based medicine, and precision medicine.

Case Study  Whose Body? Timothy B. Patrick, Peter J. Tonellato, and Mark A. Hoffman Two health sciences graduate students, Sandy and Grace, are discussing the value of clinical uses of genetic and genomic patient information. Sandy: It’s always the same story—the supposed trade-off between the benefits to society and the sacrificed rights of the individual! Just remember the case of Henrietta Lacks. HELA cells—cancer cells taken from Henrietta before she died—have been invaluable to medical science; they led to the polio vaccine and other medical “miracles.” But Henrietta was never told what was going to be done with her cells; she never gave her permission. Nor did her close relatives and family know or give their permission. It’s a clear case of science overstepping its bounds to the detriment of the individual. Grace: Sandy, you know that scientific research’s benefit to society really means the medical care benefit to the individual. Don’t you remember the recent case that took place here in our own hospital—the case of Jean, a

C hap ter 9 :   Prec ision Medic ine

17-year-old who was visiting the home of a friend when she fell down, struck her head, and suffered serious injuries? She was raced to the emergency room, where she required emergency surgery, and neither her parents nor her relatives could be reached before the procedure. The mother of Jean’s friend provided the hospital with Jean’s name and home address, which allowed the ER personnel to associate Jean with her parents in the system. Using the hospital’s healthcare information system, the surgeon entered an order for the protocol that she was planning to use to treat Jean. Among the details included in the protocol was the use of halothane, a type of anesthesia. Jean had never been the subject of genetic testing, but her father had had a genetic test that found a mutation in the ryanodine receptor gene. When people with this mutation are exposed to halothane, they can experience malignant hyperthermia, an often-fatal reaction that can cause the body’s core temperature to reach 106 degrees Fahrenheit. The hospital’s information system used the demographic person– person relationship between the father and his daughter and embedded pharmacogenomics decision support capabilities to infer that Jean had a 50 percent risk of also possessing this rare mutation. The system flashed an interactive alert to the surgeon, who was unaware of this genetic association. The surgeon responded to the alert by activating an alternative surgical plan that did not include the use of halothane. It was only by taking advantage of the genetic information about Jean’s father that a potentially catastrophic clinical event was averted! Sandy: But you make my case for me. The potential for abuse of genetic data is magnified by the existence and use of sophisticated healthcare information systems. There’s no mystery about the potential for abuse. Jean’s father was the one who had the test, not Jean. Yet the information produced by the test was also about Jean. Sure, revealing that information happened to help Jean, but the principle is that the information was about Jean as much as it was about her father. And Jean never gave her permission for that information to be used or revealed! It’s her body and her genome, not her father’s, right? So it’s her right to privacy that was violated. Grace: It might be her body, Sandy, but given the genetic data and information, we are bound by our Hippocratic oath to do no harm—primum nil nocere in Latin.1 In practice and in effect, Jean’s life was ours to save. What other choice did we have? Sandy: What about consent and protecting her privacy? And what about Jean’s father? Did he give permission to release the information from his (continued)

201

202

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

genetic test to be used in ways other than for his diagnosis and treatment? How is that different from the Havasupai tribe’s lost-blood case? Grace: Remind me about that case. Sandy: Arizona State University researchers asked the Havasupai if they would provide blood for studies to discover clues about the tribe’s incredible rate of diabetes, presumably to help the Havasupai. But the researchers used the collected blood for other purposes. They used the extracted DNA for studies on mental illness. The initial diabetes studies seem to have led nowhere, but even if that effort helped save lives, it would have been lives saved without the Havasupai’s consent. Grace: Sandy, for goodness sake, it was only blood! Sandy: Not at all, Grace, not at all.

Note 1. The following Hippocratic oath is reprinted from North (2002): I swear by Apollo, the healer, Asclepius, Hygieia, and Panacea, and I take to witness all the gods, all the goddesses, to keep according to my ability and my judgment, the following Oath and agreement: To consider dear to me, as my parents, him who taught me this art; to live in common with him and, if necessary, to share my goods with him; To look upon his children as my own brothers, to teach them this art. I will prescribe regimens for the good of my patients according to my ability and my judgment and never do harm to anyone. I will not give a lethal drug to anyone if I am asked, nor will I advise such a plan; and similarly I will not give a woman a pessary to cause an abortion. But I will preserve the purity of my life and my arts. I will not cut for stone, even for patients in whom the disease is manifest; I will leave this operation to be performed by practitioners, specialists in this art. In every house where I come I will enter only for the good of my patients, keeping myself far from all intentional ill-doing and all seduction and especially from the pleasures of love with women or with men, be they free or slaves. All that may come to my knowledge in the exercise of my profession or in daily commerce with men, which ought not to be spread abroad, I will keep secret and will never reveal. If I keep this oath faithfully, may I enjoy my life and practice my art, respected by all men and in all times; but if I swerve from it or violate it, may the reverse be my lot.

Case Study Discussion Questions 1. Whose perspective do you agree with—Sandy’s or Grace’s? Why? 
 2. Do you think there are important differences between the cases of Henrietta Lacks, Jean, and the Havasupai? Explain your answer. 


C hap ter 9 :   Prec ision Medic ine

3. Are there cases of advances in medical knowledge that do not, at least potentially, threaten to violate the privacy of individual patients? 
 4. Does a patient have the right to use the genetic information that belongs to her direct-lineage family members? Extended family or relatives? Other patients with a similar condition? 
 5. What moral, ethical, and legal protocols can be considered in guiding clinicians in this case? 
 6. What moral, ethical, and legal protocols can be considered in guiding researchers in this case?

Additional Resources Brazilian Initiative on Precision Medicine: http://bipmed.org/. Gene Expression Omnibus: www.ncbi.nlm.nih.gov/geo/. Human Genome Project: www.genome.gov/10001772/all-about-the-human-genomeproject-hgp/. Human Variome Project: www.humanvariomeproject.org/. One Health: www.cdc.gov/onehealth/index.html.

References American Physical Society. 2007. “This Month in Physics History: September 1904: Robert Wood Debunks N-rays.” APS News. Published August/September. www.aps.org/publications/apsnews/200708/history.cfm. American Statistical Association. 2015. “Roadmap to Fight Reproducibility Crisis.” Science Daily. Published June 16. www.sciencedaily.com/releases/2015/ 06/150616123914.htm. Barrett, T., S. E. Wilhite, P. Ledoux, C. Evangelista, I. F. Kim, M. Tomashevsky, K. A. Marshall, K. H. Phillippy, P. M. Sherman, M. Holko, A. Yefanov, H. Lee, N. Zhang, C. L. Robertson, N. Serova, S. Davis, and A. Soboleva. 2013. “NCBI GEO: Archive for Functional Genomics Data Sets—Update.” Nucleic Acids Research 41: D991–D995. Benson, D. A., M. Cavanaugh, K. Clark, I. Karsch-Mizrachi, D. J. Lipman, J. Ostell, and E. W. Sayers. 2013. “GenBank.” Nucleic Acids Research 41: D36–D42. Brazilian Initiative on Precision Medicine (BIPMed). 2018. “The Project.” Accessed February 23. http://bipmed.org/the-project. Cancer Treatment Centers of America (CTCA). 2018. “Genomics Commercial.” Video. Accessed February 23. www.cancercenter.com/video/commercials/genomics.

203

204

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Cinkosky, M. J., J. W. Fickett, P. Gilna, and C. Burks. 1992. “Electronic Data Publishing in GenBank.” Los Alamos Science. Accessed February 15, 2018. http:// library.lanl.gov/cgi-bin/getfile?00326716.pdf. Collins, F. S., and L. A. Tabak. 2014. “NIH Plans to Enhance Reproducibility.” Nature 505 (7485): 612–13. C-SPAN. 2018. “Global Threats and National Security.” Video. Posted February 13. www.c-span.org/video/?440888-1/fbi-director-rob-porter-background-checkcompleted-july. Flint, A., A. Raben, J. E. Blundell, and A. Astrup. 2000. “Reproducibility, Power and Validity of Visual Analogue Scales in Assessment of Appetite Sensations in Single Test Meal Studies.” International Journal of Obesity 24 (1): 38–48. Forbes. 2017. “How the Reproducibility Crisis in Academia Is Affecting Scientific Research.” Forbes. Published February 9. www.forbes.com/sites/quora/ 2017/02/09/how-the-reproducibility-crisis-in-academia-is-affecting-scientificresearch/. Heiri, O., A. F. Lotter, and G. Lemcke. 2001. “Loss on Ignition as a Method for Estimating Organic and Carbonate Content in Sediments: Reproducibility and Comparability of Results.” Journal of Paleolimnology 25 (1): 101–10. Human Variome Project (HVP). 2018. “Legal Structure: Global Variome Limited.” Accessed February 23. www.humanvariomeproject.org/about/legal-structure. html. Johnson, M., and K. Gallagher. 2017. One in a Billion: The Story of Nic Volker and the Dawn of Genomic Medicine. New York: Simon & Schuster. ———. 2015. “Young Patient Faces New Struggles Years After DNA Sequencing.” Milwaukee Journal Sentinel. Published October 25. http://archive.jsonline.com/ news/health/young-patient-faces-new-struggles-years-after-dna-sequencingb99602505z1-336977681.html. ———. 2010. “A Baffling Illness.” Milwaukee Journal Sentinel. Published December 18. http://archive.jsonline.com/features/health/111641209.html. Karsch-Mizrachi, I., Y. Nakamura, and G. Cochrane. 2012. “The International Nucleotide Sequence Database Collaboration.” Nucleic Acids Research 40: D33–D37. MacRae, M. 2014. “Cold Fusion 25 Years Later.” American Society of Mechancial Engineers. Published May. www.asme.org/engineering-topics/articles/nuclear/ cold-fusion-25-years-later. Merriam-Webster. 2018. “Genomics.” Accessed June 5. www.merriam-webster.com/ dictionary/genomics. Morozova, O., and M. A. Marra. 2008. “Applications of Next-Generation Sequencing Technologies in Functional Genomics.” Genomics 92 (5): 255–64. National Human Genome Research Institute. 2017. “DNA Sequencing Costs: Data.” Updated April 25, 2018. www.genome.gov/sequencingcostsdata/.

C hap ter 9 :   Prec ision Medic ine

National Institutes of Health. 2018. “About the All of Us Research Program.” Accessed May 2. https://allofus.nih.gov/about/about-all-us-research-program. National Library of Medicine. 2018. “What Is Precision Medicine?” Genetics Home Reference. Published May 8. https://ghr.nlm.nih.gov/primer/precisionmedicine/ definition. ———. 2017a. “GenBank and WGS Statistics.” National Center for Biotechnology Information. Published December. www.ncbi.nlm.nih.gov/genbank/statistics/. ———. 2017b. “GenBank Overview.” National Center for Biotechnology Information. Published December. www.ncbi.nlm.nih.gov/genbank/. ———. 2017c. “Request for Information (RFI): Next-Generation Data Science Challenges in Health and Biomedicine.” Released September 26. https://grants. nih.gov/grants/guide/notice-files/NOT-LM-17-006.html. North, M. 2002. “The Hippocratic Oath.” Published September 16. www.nlm.nih. gov/hmd/greek/greek_oath.html. Nucleic Acids Research. 2017. “Editorial: The 15th Annual Nucleic Acids Research Web Server Issue 2017.” Nucleic Acids Research 45: W1–W5. Okie, S. 2016. “A Boy’s Mysterious Illness, a Bold Gamble and a Breakthrough in Genetic Medicine.” Washington Post. Published April 20. www.washingtonpost.com/ opinions/a-boys-mysterious-illness-a-bold-gamble-and-a-breakthrough-ingenetic-medicine/2016/04/20/13f20b16-e638-11e5-bc08-3e03a5b41910_ story.html. Open Science Collaboration. 2015. “Estimating the Reproducibility of Psychological Science.” Science 349 (6251): aac4716. Patrick, T. B., G. Demeris, L. C. Folk, D. E. Moxley, J. A. Mitchell, and D. Tao. 2004. “Evidence-Based Retrieval in Evidence-Based Medicine.” Journal of the Medical Library Association 92 (2): 196–99. Ravvaz, K., J. A. Wessert, C. T. Ruff, C. Chi, and P. J. Tonellato. 2017. “Personalized Anticoagulation: Optimizing Warfarin Management Using Genetics and Simulated Clinical Trials.” Circulation: Genomic and Precision Medicine 10 (6): e001804. Rigdin, D. J., and X. M. Fernandez. 2018. “The 2018 Nucleic Acids Research Database Issue and the Online Molecular Biology Database Collection.” Nucleic Acids Research 46: D1–D7. Russell J. F. 2013. “If a Job Is Worth Doing, It Is Worth Doing Twice.” Nature 496 (7443): 7. US Food and Drug Administration (FDA). 2018. “Table of Pharmacogenomic Biomarkers in Drug Labeling.” Updated February 8. www.fda.gov/Drugs/Science Research/ucm572698.htm. Vater, L. B., J. M. Donohue, S. Y. Park, and Y. Schenker. 2016. “Trends in CancerCenter Spending on Advertising in the United States, 2005 to 2014.” JAMA Internal Medicine 176 (8): 1214–16.

205

206

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Waldrop, M. M. 1990. “Learning to Drink from a Firehose.” Science 248 (4956): 674–75. White House. 2015. “Fact Sheet: President Obama’s Precision Medicine Initiative.” Published January 30. https://obamawhitehouse.archives.gov/the-pressoffice/2015/01/30/fact-sheet-president-obama-s-precision-medicine-initiative. World Health Organization. 2017. “One Health.” Published September. www.who.int/ features/qa/one-health/en/. Yamashina, A., H. Tomiyama, K. Takeda, H. Tsuda, T. Arai, K. Hirose, Y. Koji, S. Hori, and Y. Yamamoto. 2002. “Validity, Reproducibility, and Clinical Significance of Noninvasive Brachial-Ankle Pulse Wave Velocity Measurement.” Hypertension Research 25 (3): 359–64.

CHAPTER

INFORMATION SYSTEMS AS INTEGRATIVE TECHNOLOGY FOR POPULATION HEALTH

10

Julie M. Kapp

Learning Objectives After reading this chapter, you should be able to do the following: • • • •

Define and differentiate between public health and population health. Identify the essential components of population health management. Construct numerators and denominators of core metrics. Synthesize these ideas into a systems thinking approach to population health.

Key Concepts • • • • • • •

Triple Aim Registry Lifestyle, behavioral, and social determinants of health Open system Systems thinking Feedback loop Integration between healthcare and public health

Introduction This chapter explores population health from an open systems perspective. How does information technology (IT) assume and enable integration of medical care services with the health of individuals, families, and populations? What changes in structure, financing, and the role of patients are assumed when this open systems technology is applied?

207

208

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Status of Population Health in the United States A 2013 report, U.S. Health in International Perspective: Shorter Lives, Poorer Health, took an in-depth look at how the United States compares on a number of health outcome indicators in relation to 16 peer countries, including Australia, Austria, Canada, Denmark, Finland, France, Germany, Italy, Japan, Norway, Portugal, Spain, Sweden, Switzerland, the Netherlands, and the United Kingdom (Institute of Medicine 2013). The report described the United States as having a “health disadvantage.” In fact, the United States fares worse than average in the following nine health domains: 1. Adverse birth outcomes 2. Injuries and homicides 3. Adolescent pregnancy and sexually transmitted infections 4. HIV and AIDS 5. Drug-related mortality 6. Obesity and diabetes 7. Heart disease 8. Chronic lung disease 9. Disability Of the countries examined in the report, the United States has had the highest prevalence of diabetes and the highest prevalence of adult obesity since 2010. In addition, Americans have a shorter life expectancy than do residents in almost all of the United States’ peer countries, and the gap has been growing for the past three decades, especially among women. According to the report (Institute of Medicine 2013), deaths that occur before age 50 are responsible for about two-thirds of the difference in life expectancy between males in the United States and peer countries, and about one-third of the difference for females. And the problem has been worsening over time; since 1980, the United States has had the first or second lowest probability of surviving to age 50 among the 17 peer countries. Americans who do reach age 50 generally arrive at this age in poorer health than their counterparts in other high-income countries, and as older adults they face greater morbidity and mortality from chronic diseases that arise from risk factors (e.g., smoking, obesity, diabetes) that are often established earlier in life.

The United States is also an outlier in areas other than health outcomes. Its health spending per capita is higher compared with healthcare expenditures in other countries, but it falls below peer countries in life expectancy

Ch a p t er 10 :  I nf o r ma ti o n Sy ste ms a s In te g rative Tec hnology for Pop ulation H ealth

EXHIBIT 10.1 Life Expectancy at Birth and Health Spending per Capita, 2011

85

ISR KOR GRC SVN

Life expectancy in years

80

CHE ITA JPN ISL SUE FRA ESP AUS AUT NLD PRT GBR LUX NZL CAN DEU FIN IRL BEL DNK

NOR

USA

CHL CZE

POL EST

SVK

75

TUR CHN

HUN

MEX BRA

70 IDN

RUS R2 = 0.51

65

IND

0

1,000

2,000

3,000

4,000

5,000

6,000

7,000

8,000

9,000

Health spending per capita (US dollars) Sources: OECD Health Statistics 2013 for OECD countries (http://dx.doi.org/10.1787/health-dataen); World Bank for non-OECD countries.

(Organisation for Economic Co-operation and Development [OECD] 2013; see exhibit 10.1). Overspending with inferior results is not a sustainable model; the United States needs to reform the way it manages population health.

Population Health Provisions in the Affordable Care Act National Quality Strategy The Affordable Care Act (ACA) of 2010 recognizes the unsustainability of US population health outcomes and health expenses. Although the ACA is known for its focus on healthcare payment reform, it includes a number of provisions that attempt to address broader population health trends. These provisions have ushered in the National Quality Strategy, led by the Agency for Healthcare Research and Quality (2017), to improve healthcare quality. The National Quality Strategy has three overarching aims: • Better Care: Improve overall quality, by making health care more patientcentered, reliable, accessible, and safe.

209

210

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

• Healthy People/Healthy Communities: Improve the health of the U.S. population by supporting proven interventions to address behavioral, social, and environmental determinants of health in addition to delivering higher-quality care. • Affordable Care: Reduce the cost of quality health care for individuals, families, employers, and government.

A seminal article by Berwick, Nolan, and Whittington (2008) popularized the term Triple Aim. A framework developed by the Institute for Healthcare Improvement, the Triple Aim emphasizes the main goals of the National Quality Strategy: to improve the individual’s experience of care, to improve the health of populations, and to reduce the per capita costs of care for populations. Whether the ACA is repealed or replaced, the status of population health outcomes remains at a disadvantage.

National Prevention Strategy The ACA established the National Prevention, Health Promotion, and Public Health Council, or National Prevention Council (NPC). The council comprises 20 federal departments, agencies, and offices and is chaired by the US surgeon general. Through the ACA’s mandates, the NPC issued the National Prevention Strategy—the nation’s first such initiative—in 2011. The National Prevention Strategy is the key strategy and vision for improving the delivery of healthcare services, patient health outcomes, and population health. Its core values align with the objective of reducing the health disadvantage of the United States. The National Prevention Strategy “envisions a prevention-oriented society where all sectors recognize the value of health for individuals, families, and society and work together to achieve better health for Americans” (US Surgeon General 2011). The strategy’s overarching goal is to “increase the number of Americans who are healthy at every stage of life” by implementing four strategic directions and seven priorities (NPC 2011): Strategic Directions • Healthy and Safe Community Environments • Clinical and Community Preventive Services • Empowering People • Elimination of Health Disparities Priorities • Tobacco Free Living • Preventing Drug Abuse and Excessive Alcohol Use • Healthy Eating • Active Living

Ch a p t er 10 :  I nf o r ma ti o n Sy ste ms a s In te g rative Tec hnology for Pop ulation H ealth

• Injury and Violence Free Living • Reproductive and Sexual Health • Mental and Emotional Well Being

Difference Between Public Health and Population Health To better understand the context of these national strategic initiatives, one must understand the differences between public health and population health.

What Is Public Health? The World Health Organization (WHO) defines public health as “the science and art of preventing disease, prolonging life and promoting health through the organised efforts of society” (Acheson and Department of Health and Social Security Great Britain 1988; WHO 2017a). The science of public health has been around for a long time. Elements of public health can be found in many ancient cultures, such as the Greco-Roman and Egyptian. According to the Centers for Disease Control and Prevention (2016), all communities should implement the following ten activities when operationalizing public health: 1. Monitor health status to identify and solve community health problems 2. Diagnose and investigate health problems and health hazards in the community 3. Inform, educate, and empower people about health issues 4. Mobilize community partnerships and action to identify and solve health problems 5. Develop policies and plans that support individual and community health efforts 6. Enforce laws and regulations that protect health and ensure safety 7. Link people to needed personal health services and assure the provision of healthcare when otherwise unavailable 8. Assure a competent public and personal health care workforce 9. Evaluate effectiveness, accessibility, and quality of personal and populationbased health services 10. Research for new insights and innovative solutions to health problems

What Is Population Health? 
 Population health, as the term is used today, has many definitions, and these definitions vary (Stoto 2013). One of the most referenced definitions is that of Kindig and Stoddart (2003): “the health outcomes of a group of individuals, including the distribution of such outcomes within the group.” Population

211

212

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

health (1) identifies a defined population, which might be a geographic area, a service line, or another subgroup of focus, and (2) understands patterns of disease distribution and frequency, both overall and within population subgroups.

Population Health Management in the United States In today’s context, population health management is a nebulous, ill-defined term that generally refers to (1) identifying the health needs (numerator) of a healthcare service area or defined population (denominator); (2) measuring the percentage of those needs being met or unmet; (3) aligning the needs with targeted strategies to improve health outcomes; and, ideally, (4) tracking progress on improving those outcomes (Kapp et al. 2016). An example might be (1) identifying those with cardiovascular disease or those with high risk for cardiovascular disease among the residents of a particular county; (2) quantifying the level of disease or risk for disease through quantifiable metrics, such as the percentage of residents at high risk among the broader population; (3) identifying targeted strategies that have been proven to help reduce the burden of cardiovascular disease in that population or a similar one; and (4) tracking progress on these efforts. Basically, population health defines a specific subgroup (denominator) of the population, and population health management is the task of determining the health needs of that population and actively implementing efforts to improve the health of that subgroup. Because the focus of population health management is improving population health outcomes, it entails proactive examination of upstream social risk factors and lifestyle and behavioral factors over the course of life to systemically prevent disease. It also involves the alignment of evidence-based, effective policies with actions to improve identified disparities in health outcomes. According to Berwick, Nolan, and Whittington (2008), a population does not need to be defined by geography. What may best define a population is the concept of enrollment. (Note that enrollment as it is used here is different from enrollment as used in healthcare, where it denotes a financial transaction.) A registry is a component of population health management that tracks a defined group of people over time—a population—for the purposes of achieving the Triple Aim. Berwick, Nolan, and Whittington (2008) refer to population and registry in the context of defining a cohort—establishing inclusion and exclusion criteria and baseline measures for those who are a part of the cohort denominator—and then tracking and monitoring changes in that cohort’s health status over time (e.g., disease incidence) using an electronic database. A registry can feasibly be created through an electronic medical record (EMR), but doing so would reflect a denominator of patients enrolled in a health plan and not a broader, population-based enrollment. A registry could

Ch a p t er 10 :  I nf o r ma ti o n Sy ste ms a s In te g rative Tec hnology for Pop ulation H ealth

capture the broader population-based enrollment by recruiting from opt-in panels, and it might be accessed through a government or nonprofit website that is hosted independently of any healthcare organization. One way that participants might be recruited for opt-in panels is the Behavioral Risk Factor Surveillance System (BRFSS) surveys. The BRFSS is a nationwide surveillance system for collecting data on health-related risk behaviors, chronic conditions, and use of preventive services. It is currently implemented as a cross-sectional data collection effort in all 50 states, the District of Columbia, and three US territories. One advantage to hosting a registry externally is that it would reduce the impact of migration of patients in and out of the system as individuals change jobs. To be useful, this registry would need to be able to link to various EMRs. According to Dunn and Hayes (1999), population health management as an approach “focuses on interrelated conditions and factors that influence the health of populations over the life course, identifies systematic variations in their patterns of occurrence, and applies the resulting knowledge to develop and implement policies and actions to improve the health and well-being of those populations.” With a population-based registry, policies and actions would be beneficial to a broader enrollment base than to patients in a particular health system. Stoto (2013) notes the commonalities that encompass many population health perspectives: “Population health is seen as more than the sum of individual parts. . . . The population health perspective requires the consideration of a broader array of the determinants of health than is typical in either healthcare or public health.” It also includes health promotion and disease prevention (primary prevention of disease) and addresses upstream social factors that produce health disparities and inequities. Tracking and monitoring a population-based cohort through an external registry generates the incidence data that could provide evidence-based information on exposure risks, effects of lifestyle and modifiable behaviors, and programs or partners that have a positive (or negative) effect on outcomes. The monitoring entity could then disseminate this information to health systems or partners in the geographic proximity (or to related subgroups), which could then leverage this information to enhance their interventions.

Premature Mortality, Lifestyle and Behavioral Factors, and Social Determinants of Health Housing the registry outside of the healthcare organization may help to prevent the “medicalizing” (Magnan 2017) of a population health management approach. As mentioned earlier, life expectancy in the United States is shorter than in almost all of its peer countries. The National Prevention Strategy is founded on the idea that the shorter life expectancy in the United States is modifiable; it believes life expectancy can be extended by improving modifiable behaviors to reduce the number of premature deaths. McGinnis and Foege (1993)

213

214

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

suggested that about 50 percent of all deaths in the United Interoperability Moment States can be attributed to lifeThe WHO defines social determinants of health as “the conditions style and behavioral factors, in which people are born, grow, work, live, and age, and the wider such as tobacco use, poor diet set of forces and systems shaping the conditions of daily life.” Given and activity patterns, and alcothat those conditions may vary from region to region and from counhol and drug use. Mokdad and try to country, a health problem or disease associated with social colleagues (2004) confirmed determinants may not occur with uniform severity across different these findings a decade later. regions and countries. If this is so, how can data and information Medical care (e.g., hospitalizafrom different regions and countries regarding that health condition tions, treatments, prescription or disease be meaningfully aggregated or compared? How is this an medicines) accounts for only issue of interoperability? 10 percent to 15 percent of preventable mortality (Kindig, Asada, and Booske 2008; McGinnis, Williams-Russo, and Knickman 2002). Social factors are also powerful determinants of health (Braveman and Gottlieb 2014), an idea supported by a 2013 report: “Healthcare systems and the health services they deliver are not the only influences on population health. Lifestyles and behaviors, social and economic circumstances, environmental influences, and public policies can also play key roles in shaping individual and community health” (Institute of Medicine 2013). WHO (2017b) defines social determinants of health as “the conditions in which people are born, grow, work, live, and age, and the wider set of forces and systems shaping the conditions of daily life.” Common examples of these determinants include lack of transportation, unsafe or inadequate housing, and food insecurity (Fraze et al. 2016). Can we expect healthcare organizations to be effective in changing lifestyles and behaviors? If evidence-based information on effective interventions were identified in a local population-based registry, then we could expect the clinicians and other providers in a hospital or health system to integrate this knowledge into their patient care. The patient–provider relationship is, after all, a cornerstone of satisfaction, treatment adherence, and continuity of care (Rolfe et al. 2014; White et al. 2016). The provider may need to coach the patient on lifestyle or behavioral changes that are appropriate for the patient’s situation, or the provider may refer or connect the patient to an external program, such as a content community, that can better support the patient in his day-to-day life. In addition, the provider might find that the patient has basic social needs (e.g., transportation, housing, food) and arrange a connection with a community-based program that has been proven by the registry to be an effective resource or advocate.

Integration of Healthcare and Public Health Because lifestyle, behavioral, and social determinants of health play a larger role in population health than medical care does, a push for greater integration of

Ch a p t er 10 :  I nf o r ma ti o n Sy ste ms a s In te g rative Tec hnology for Pop ulation H ealth

215

healthcare and public health has become a component of population health management. Hardcastle and colleagues (2011) suggest that “public health and healthcare should be conceptualized as two interconnected parts of a single health system.” They further explain that a “fully integrated health system requires that all government policies reflect the ultimate goal of improving the health of the population, which necessitates the adoption of a Health in All Policies (HiAP) approach” (Hardcastle et al. 2011). In practice, healthcare organizations conceptualize their own definitions of the problems, outcomes, and the ultimate impact goal. The integration of healthcare and public health (nonmedical programs) has been stimulated by financial incentives such as bundled payments, accountable care organizations (ACOs), and managed care (Fraze et al. 2016). ACOs are beginning to address nonmedical needs given that lifestyle, behavior, and social determinants of health affect how patients engage with medical care (Fraze et al. 2016). For example, recognizing the need for stable housing for some of their patients, some ACOs have partnered with local housing authorities (Fraze et al. 2016). Similarly, an urban ACO is developing and funding a transportation service that patients in need can request via a mobile app (Fraze et al. 2016). Technology can help healthcare organizations address and fulfill population health needs. For example, medication can be delivered to a patient’s home via drone, minimizing the need to find and go to a pharmacy; a patient can summon a free ride to the doctor’s office as part of the doctor’s service; and a patient can request, schedule, and even receive a second opinion via a mobile app. Unfortunately, such technologies will continue to be considered cost-prohibitive until the healthcare industry starts functioning as an integrated system rather than as a collection of independent offices and units. Until we have standard IT platforms, medical record data file platforms, privacy and security agreements, outcome measures, and expectations for quality performance and reformed payment schemes, we cannot expect to see the successful development and deployment of these technologies. More important, the power of influence for systems-level change does not currently reside with the patient but with insurance companies, employers, and federal and state policies.

Population Health as a System Systems Thinking Systems thinking is an approach to a problem that considers how components within the larger structure operate and interact over the life cycle of the system, as well as how the design, implementation, and evaluation of that system can be optimized (Kapp et al. 2016). Systems thinking “can best be described as the application of systems concepts to frame our understanding of the world, and it is also about possible future action—what ought to be or could be”

Systems thinking Viewing parts or agents of a social or biological system as interdependent

216

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

(Rajagopalan and Midgley 2015). Systems thinking does not focus narrowly on a single technical solution but rather assesses a problem from a comprehensive, holistic view of the overall challenge.

What Is a System? Here, we expand on the ideas of systems first introduced in chapter 1. According to Russell Ackoff (1999), a leader in systems thinking, a system includes at least two elements that interact and are interdependent. Other systems leaders posit that the elements are integrated in ways that continually affect each other in feedback loops over time and that operate as a whole toward a common purpose (Kim 1999; Meadows 2008). Not all systems thinkers agree on what systems thinking is (Cabrera 2006), but for our purposes, the properties of systems are as follows: • • • •

They are composed of at least two elements. Elements are interdependent. Elements provide feedback loops to each other. Elements operate toward a common purpose.

As mentioned, systems thinking does not mean having a single, narrow focus but rather assessing a problem from a comprehensive, holistic view to see how the problem fits into the overall, bigger challenge (Ackoff 1981). Think of an organization as a system. An organization is purposeful, with a set of goals, objectives, and ideals, but an organization itself is just one part of a larger, purposeful system that has its own goals, objectives, and ideals (Ackoff 1974, 1981). The health system is, in turn, part of a greater community of human services.

What Is an Open System? Open system System with an environment that can affect its state and with which the system interacts

An open system is a system that has an environment, or context, that can affect its state and with which it interacts. A boundary separates a system and its environment and may be conceptual rather than tangible or defined. The system boundary may be arbitrary, subjective, and set by a particular stakeholder perspective. We might think of boundaries as “inside-the-box” thinking, where the boundary is an arbitrary line in the sand. For instance, a colleague has established her view with an arbitrary boundary if she says, “We can’t do that. We’ve always done it this way,” when, in fact, it can be done. Healthcare has many such examples of boundaries, such as its traditional focus on the health of individual patients and not on the upstream social determinants that lead to poor health conditions. Another example is expanding an emergency department (ED) because of increased patient volumes, when the sustainable solution may be to expand primary care to treat patients who are using the ED because they cannot easily make a doctor’s appointment.

Ch a p t er 10 :  I nf o r ma ti o n Sy ste ms a s In te g rative Tec hnology for Pop ulation H ealth

217

Suprasystems, or super systems, contain subsystems that interact with the environment. We might think of US population health as a suprasystem, containing many subsystems at federal, state, and local levels (Kapp et al. 2016).

Inflows and Outflows All systems structures have stock with inflows and outflows. An inflow might be likened to a faucet, and an outflow might be likened to a drain. In the case of the US population health system (exhibit 10.2), one can envision the stock as the status of US population health, as measured by disease rates and life expectancy. Some of the controllable inflows are modifiable individual behaviors, social service supports, medical care, and environmental factors. An outflow is premature mortality. Often, the solution to poor population health status is perceived to be a greater quantity of inflows (services or activities), which are bounded by resources (time and money). Instead, systems science (chapter 5) can be used to identify the inflows that are most influential at producing positive or negative effects on the stock and ultimately on the outflow (premature mortality). This may be done by using a registry, which follows US population health as a cohort and tracks the influences of the inflows of activities, services, and behaviors as well as the resulting outflow in a way that is interconnected and demonstrates efficacy. Feedback loops identify which activities and behaviors are working to reduce disease rates and increase life expectancy. The primary challenge is how to connect all of the system components into an operational, measurable, and integrated structure that is oriented toward improving preventable premature mortality. We can illustrate this in another way. A modification of Ackoff’s (1970) decision model includes an objective function and a set of constraints: P = f (C,U). Here, P is the performance of the system (population health status of premature mortality in the United States compared with peer countries), and f is some

Services, activities, behaviors, environmental factors

US population health status: Disease rates and life expectancy

EXHIBIT 10.2 US Population Health as a System Premature mortality

218

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

relationship between C (controlled variables) and U (uncontrolled variables). Controlled variables would include the 50 percent of premature mortality thought to be preventable. To further define the population health system under consideration, first define its five areas of focus: (1) context, (2) components, (3) connections, (4) infrastructure, and (5) scale (Eoyang and Berkas 1998). Context is the health and disease status of the population. Components are the measurable controlled constructs that contribute to either improving or diminishing the preventable premature mortality level. Connections are the interrelationships between these components and the strength of their connections, which include the interdependence of individual-level versus population-level change. Infrastructure refers to the organizations, stakeholders, programs, and policies that can either improve or diminish preventable premature mortality. Scale stands for the local, state, and federal levels.

Using Systems Thinking to Improve Population Health According to Meadows (2008), a “system generally goes on being itself, changing only slowly if at all, even with complete substitutions of its elements—as long as its interconnections and purposes remain intact.” To change the system, the relationships between the components in the system must be changed, much as Hardcastle and colleagues (2011) argued for the integration of healthcare and public health. As Meadows (2008) notes, “interconnections are . . . critically important. Changing relationships usually changes system behavior.” In addition, “one of the most powerful ways to influence the behavior of a system is through its purpose or goal. That’s because the goal is the direction-setter of the system, the definer of discrepancies that require action, the indicator of compliance, failure, or success toward which balancing feedback loops work” (Meadows 2008). Systems will produce what you ask them to produce, based on how you define the goal. “The only way to fix a system that is laid out poorly is to rebuild it, if you can” (Meadows 2008). Ackoff (1977) refers to such building as idealized design, which he describes as unattainable but progress toward which is possible. It is ideal-seeking in its purpose, technologically feasible, operationally viable, and capable of rapid learning and adaptation (Ackoff 1977, 1981, 1999). Chernichovsky and Leibowitz (2010) argue that a reformed health system would coordinate personal medical care with population health, provide health insurance coverage for all, and adopt new knowledge about medical and nonmedical determinants of health in a feedback loop. This kind of reformed system would change the focus from volume and profitability to health outcomes. Is there evidence that systems thinking can improve population health? Because the United States has not tried universal healthcare coverage (for

Ch a p t er 10 :  I nf o r ma ti o n Sy ste ms a s In te g rative Tec hnology for Pop ulation H ealth

example), we cannot know the answer. What we do know is that the practice of overspending on services but underperforming on outcomes is not working, and we have been talking about it for nearly a hundred years—at least as far back as the time of Winslow (1923): Compulsory health insurance is one panacea which has been strenuously advocated for the solution of this problem [the maintenance of physical health]. As introduced by Bismarck in Germany in 1885 and by Lloyd George in England in 1912 health insurance had two main objectives, the distribution of the financial burden of sickness and the provision of medical care not otherwise available. The first of these aims is a problem for the economist, not the sanitarian [i.e., public health office]; it is the second with which we are concerned at the present moment.

Winslow went on to state, “Preventive medicine must come, as a reality and not a pious phrase, through a fundamental change in the attitude of the physician and through a fundamental change in the attitude of the medical school where he is trained.” The ACA may have fostered the nation’s first formal prevention strategy, but medical and public health experts have been deliberating on the US population’s life expectancy and premature mortality since at least 1883 (Kapp 2013). Even more surprisingly, the nineteenth-century experts had a message that was resoundingly similar to that of today’s experts: “According to the census of 1880, 756,893 persons died during the census year. . . . One-half of these died from preventable diseases” (Pierson 1883). Moreover, the seven priorities of the National Prevention Strategy essentially mirror the 1912 recommendations of Dr. Henry S. Munro: One part of our [human] race is degenerating as the result of luxurious habits, such as lack of exercise, indiscretions in social life, eating, drinking and tobacco; while the modern industrial conditions have millions of working men, women and children ‘in the mill’ . . . We must teach the human animal how to live. Teach him (1) what to eat; (2) what to drink; (3) the kind of work and exercise adapted to his individual personality; (4) give him a sound philosophy of life, free from dogma, cant, or narrowness; (5) inform him concerning the value of the right kind of companionship and of self-education. These are the five life essentials, without which no life is complete—the absolute necessities of a life of health and happiness.

These examples suggest that, if we have been having the same conversations for at least a hundred years, it is time to transform the US health system. What complicates progress toward change is the inability to define, monitor, and improve on modifiable goals (e.g., premature mortality).

219

220

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Integrating Healthcare and Public Health Through Systems Design In the report For the Public’s Health, the Institute of Medicine (2012) recommends the following: The Secretary of the Department of Health and Human Services should adopt an interim explicit life expectancy target, establish data systems for a permanent healthadjusted life expectancy target, and establish a specific per capita health expenditure target to be achieved by 2030. Reaching these targets should engage all health system stakeholders in actions intended to achieve parity with averages among comparable nations on healthy life expectancy and per capita health expenditures.

Reaching these targets requires an integrated healthcare and public health system (perhaps through the population-based registry) working toward deliberate, shared goals—which is a powerful way to influence change in the system. Successfully integrating these two systems requires, among other things, explicitly defined goals; standard metrics and measurements shared by all stakeholders; and an IT platform that supports various stakeholders’ unique roles and responsibilities, enables all stakeholders to communicate, and tracks their contributions and progress toward the goals.

Applying Population Health Informatics to the Patient Population Efforts are under way to incorporate elements of a patient’s lifestyle and behaviors into the patient–provider encounter. IT assumes and enables the integration of medical care services with the health of individuals, families, and populations. Healthcare organizations are working to incorporate patient data not typically entered into the EMR to develop a more comprehensive picture of the patient, including the patient’s lifestyle, behaviors, and personal environment. Examples of such patient data include those from fitness trackers, lifestyle apps, electronic food diaries, and home medical devices such as blood pressure monitors, and much of it is contained in the personal health record. However, each of these data sources must be considered in the context of its limitations. For example, Hoch (2015) reports that 23 percent of users abandon an app after one use, and the average mobile app retention rate is 20 percent after 90 days (Perro 2017). Integration requires not only changes in technology, however, but also changes in incentives and financing (chapter 15).

Applying Population Health Informatics to the Community Population Successful integration of healthcare and public health requires a standardized, uniform IT platform. Imagine having to electronically sync the existing tailored knowledge management systems of each hospital, health system, ACO,

Ch a p t er 10 :  I nf o r ma ti o n Sy ste ms a s In te g rative Tec hnology for Pop ulation H ealth

nonprofit or charity organization, public health department, and federal, state, and local governments. At minimum, the IT platform would need to “plug in” to the myriad systems to create the population-based registry. Tracking individuals across disparate programs and systems would require a shared identification number for each person; in theory, this identifier would allow seamless coordination of care, gains in efficiencies, and reduction in duplication of services. All stakeholders would need to be mindful of a shared set of outcomes metrics—the health outcomes they are all trying to improve. These metrics would include quality and quantity of life, as well as the leading indicators of quality and quantity of life. A standardized IT platform would mean that patients receive consistent advice, regardless of where they are receiving care. For instance, if patient information is entered, the platform’s algorithm would create a recommendation for additional community-based services (e.g., transportation assistance) that support the direction of medical care. This recommendation would be based on eligibility and inclusion criteria for all users. Regardless of which medical provider the patient sees, the provider would be able to access a list of available, vetted, community-based services for which the patient is eligible. For the platform to generate these recommendations, information about modifiable lifestyle, behavioral, and social determinants of health would have to be entered. The platform would also require an up-to-date catalog of all available services and their requirements for enrollment, contact information, and metrics for measuring and tracking all stakeholders’ contributions to the larger system’s health outcomes. It would include a feedback loop to all stakeholders, creating a learning system that alerts providers and patients to which services are rated highly, which are reliable at producing outcomes, and so on. These evidence-based practices would be stored electronically so that historical knowledge is not lost.

Conclusion Despite having worse health outcomes, the United States spends more on healthcare than does any other country. In response, the ACA has fostered population health management initiatives as well as integrated healthcare and public health efforts to improve the health status of the US population. One of these initiatives is the National Prevention Strategy, which recognizes that medical care is less of a contributor to shortened life expectancy and premature mortality than are modifiable lifestyle, behavioral, and social factors. Successfully integrating healthcare and public health—two large and disparate systems—requires a systems thinking approach. This approach is characterized by explicitly defined goals, purposes, roles, and responsibilities; shared metrics; an understanding

221

222

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

of each organization’s role in improving the metrics; and a shared knowledge management system. By aligning healthcare organizations with population health initiatives that have a defined common goal, the US health system can be transformed with the help of an integrated knowledge management system.

Chapter Discussion Questions 1. Identify the essential components of population health management, and describe the role of each. 2. Apply the concepts of population health management to a healthcare organization. Who is the population? Provide specific definitions of who is and is not included in this population. 3. Identify three core metrics, and define how they are measured (including their numerators and denominators). How would you track progress on those metrics? 4. What are the benefits of integrating healthcare and public health? 5. Provide two examples of systems as they relate to healthcare or population health. Are they open systems? Define the environment and identify the boundaries.

Case Study  Pemiscot County The state of Missouri has 114 counties, plus the city of St. Louis. The Robert Wood Johnson Foundation’s County Health Rankings & Roadmaps website ranks Pemiscot County 115th in health outcomes (RWJF 2017), making it the least healthy county in Missouri. According to the website, Pemiscot County has a premature death rate of 16,000, compared to Missouri’s overall rate of 7,700 and the top US performers’ rate of 5,200 (measured by years of potential life lost before age 75 per 100,000 population, age-adjusted). Pemiscot County has 3.530 persons per primary care physician, whereas Missouri has 1.420 and the top US performers have 1.040. It has 56 percent of children in singleparent households, whereas Missouri has 34 percent and the top US performers have 21 percent. Residents of Pemiscot County, compared with Missouri citizens overall, are more likely to report poor or fair health, adult smoking, adult obesity, sexually transmitted infections, less access to exercise opportunities, significantly more teen births, significantly fewer dentists per capita, and significantly longer hospital stays. Excessive drinking does not appear to be worse in Pemiscot County, however. According to the Community Commons website, the county has only one hospital located in the middle of the county.

Ch a p t er 10 :  I nf o r ma ti o n Sy ste ms a s In te g rative Tec hnology for Pop ulation H ealth

The total population in Pemiscot County is 18,296, according to 2010 US Census Bureau data, and 2016 census estimates break down the population as follows: 70.8 percent white (compared with 76.9 percent in the United States), 26.6 percent black or African American (compared with 13.3 percent in the United States), and 2.5 percent Hispanic or Latino (compared with 17.8 percent in the United States). There are 6,975 households in the county, with 83.4 percent of the residents living in the same house they occupied the previous year. The median household income is $29,238 (compared with $53,889 in the United States), and 73.7 percent of residents have a high school degree (compared with 86.7 percent in the United States).

Case Study Discussion Questions 1. Describe several metrics that should be included in a population health registry to track the health status of residents in Pemiscot County. 2. Who is responsible for gathering those metrics? 3. Which stakeholders would be interested in accessing a population health registry? Who should have access to that registry? 4. What is the role of medical care in population health? The public health department? The healthcare industry? Community-based programs? Individuals? Employers? 5. What are some interventions that could help improve the health outcomes in Pemiscot County? How can the registry support these outcomes being evidence-based?

Additional Resources Community Commons: www.communitycommons.org. County Health Rankings & Roadmaps: www.countyhealthrankings.org.

References Acheson, D., and Department of Health and Social Security Great Britain. 1988. Public Health in England: The Report of the Committee of Inquiry into the Future Development of the Public Health Function. London: The Stationery Office. Ackoff, R. L. 1999. Ackoff’s Best: His Classic Writings on Management. New York: John Wiley & Sons. ———. 1981. Creating the Corporate Future: Plan or Be Planned For. New York: John Wiley & Sons.

223

224

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

———. 1977. “National Development Planning Revisited.” Operations Research 25 (2): 207–18. ———. 1974. Redesigning the Future. New York: John Wiley & Sons. ———. 1970. “A Concept of Corporate Planning.” Long Range Planning 3 (1): 2–8. Agency for Healthcare Research and Quality. 2017. “About the National Quality Strategy.” Reviewed March. www.ahrq.gov/workingforquality/about/index.html. Berwick, D. M., T. W. Nolan, and J. Whittington. 2008. “The Triple Aim: Care, Health, and Cost.” Health Affairs 27 (3): 759–69. Braveman, P., and L. Gottlieb. 2014. “The Social Determinants of Health: It’s Time to Consider the Causes of the Causes.” Public Health Reports 129 (Suppl 2): 19–31. Cabrera, D. A. 2006. “Systems Thinking.” PhD dissertation, Cornell University. Accessed January 1, 2017. https://pdfs.semanticscholar.org/bf64/6993a7fc 685dc6d84fc473fab68ea253a829.pdf. Centers for Disease Control and Prevention. 2016. “The Public Health System and the 10 Essential Public Health Services.” Updated September 20, 2017. www. cdc.gov/stltpublichealth/publichealthservices/essentialhealthservices.html. Chernichovsky, D., and A. A. Leibowitz. 2010. “Integrating Public Health and Personal Care in a Reformed US Health Care System.” American Journal of Public Health 100 (2): 205–11. Dunn, J. R., and M. V. Hayes. 1999. “Toward a Lexicon of Population Health.” Canadian Journal of Public Health 90 (Suppl 1): 7–10. Eoyang, G. H., and T. H. Berkas. 1998. “Evaluation in a Complex Adaptive System.” Published April 30. www.researchgate.net/publication/237571019_ Evaluation_in_a_Complex_Adaptive_System. Fraze, T., V. A. Lewis, H. P. Rodriguez, and E. S. Fisher. 2016. “Housing, Transportation, and Food: How ACOs Seek to Improve Population Health by Addressing Nonmedical Needs of Patients.” Health Affairs 35 (11): 2109–15. Hardcastle, L. E., K. L. Record, P. D. Jacobson, and L. O. Gostin. 2011. “Improving the Population’s Health: The Affordable Care Act and the Importance of Integration.” Journal of Law, Medicine & Ethics 39 (3): 317–27. Hoch, D. 2015. “App User Retention Improves in the U.S., but Declines Internationally.” Published June 12. http://info.localytics.com/blog/app-user-retentionimproves-in-the-us. Institute of Medicine. 2013. U.S. Health in International Perspective: Shorter Lives, Poorer Health. Washington, DC: National Academies Press. ———. 2012. For the Public’s Health: Investing in a Healthier Future. Washington, DC: National Academies Press. Kapp, J. M. 2013. “Evaluation of the National Prevention Strategy in Reference to Historical Population Health.” American Journal of Health Promotion 27 (5): 281–83. Kapp, J. M., E. J. Simoes, A. DeBiasi, and S. J. Kravet. 2016. “A Conceptual Framework for a Systems Thinking Approach to U.S. Population Health.” Systems

Ch a p t er 10 :  I nf o r ma ti o n Sy ste ms a s In te g rative Tec hnology for Pop ulation H ealth

Research and Behavioral Science. Accessed January 1, 2017. https://jhu.pure. elsevier.com/en/publications/a-conceptual-framework-for-a-systems-thinkingapproach-to-us-popu. Kim, D. H. 1999. Introduction to Systems Thinking. Waltham, MA: Pegasus Communications. Kindig, D. A., Y. Asada, and B. Booske. 2008. “A Population Health Framework for Setting National and State Health Goals.” Journal of the American Medical Association 299 (17): 2081–83. Kindig, D., and G. Stoddart. 2003. “What Is Population Health?” American Journal of Public Health 93 (3): 380–83. Magnan, S. 2017. “Social Determinants of Health 101 for Health Care: Five Plus Five.” Published October 9. https://nam.edu/social-determinants-of-health101-for-health-care-five-plus-five/. McGinnis, J. M., and W. H. Foege. 1993. “Actual Causes of Death in the United States.” Journal of the American Medical Association 270 (18): 2207–12. McGinnis, J. M., P. Williams-Russo, and J. R. Knickman. 2002. “The Case for More Active Policy Attention to Health Promotion.” Health Affairs 21 (2): 78–93. Meadows, D. H. 2008. Thinking in Systems: A Primer. White River Junction, VT: Chelsea Green. Mokdad, A. H., J. S. Marks, D. F. Stroup, and J. L. Gerberding. 2004. “Actual Causes of Death in the United States, 2000.” Journal of the American Medical Association 291 (10): 1238–45. Munro, H. S. 1912. “The Retraining of the Human Animal for the Restoration of Health.” Medical Herald 31 (2): 116–26. National Prevention, Health Promotion, and Public Health Council (NPC). 2011. “National Prevention Strategy: America’s Plan for Better Health and Wellness.” Published July 2011. www.surgeongeneral.gov/priorities/prevention/about/. Organisation for Economic Co-operation and Development (OECD). 2013. Health at a Glance 2013. Paris: Organisation for Economic Co-operation and Development. Perro, J. 2017. “Mobile Apps: What’s a Good Retention Rate?” Published March 2016. http://info.localytics.com/blog/mobile-apps-whats-a-good-retention-rate. Pierson, J. 1883. “The Relation of Sanitary Science to National Wealth.” The Sanitarian 12: 248–49. Rajagopalan, R., and G. Midgley. 2015. “Knowing Differently in Systemic Intervention.” Systems Research and Behavioral Science 32 (5): 546–61. Robert Wood Johnson Foundation (RWJF). 2017. “County Health Rankings & Roadmaps: Pemiscot, Missouri.” Accessed January 1. www.countyhealthrankings.org/ app/missouri/2017/rankings/pemiscot/county/outcomes/overall/snapshot. Rolfe, A., L. Cash-Gibson, J. Car, A. Sheikh, and B. McKinstry. 2014. “Interventions for Improving Patients’ Trust in Doctors and Groups of Doctors.” Cochrane Database Systematic Reviews (3): 004134.

225

226

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Stoto, M. A. 2013. “Population Health in the Affordable Care Act Era.” AcademyHealth. Published February 20. www.academyhealth.org/publications/2013-02/ population-health-affordable-care-act-era. US Surgeon General. 2011. “National Prevention Strategy.” Published June 16. www. surgeongeneral.gov/priorities/prevention/strategy/index.html. White, R. O., R. J. Chakkalakal, C. A. Presley, A. Bian, J. S. Schildcrout, K. A. Wallston, and R. Rothman. 2016. “Perceptions of Provider Communication Among Vulnerable Patients with Diabetes: Influences of Medical Mistrust and Health Literacy.” Journal of Health Communication 21 (Suppl 2): 127–34. Winslow, C. E. A. 1923. The Evolution and Significance of the Modern Public Health Campaign. New Haven, CT: Yale University Press. World Health Organization (WHO). 2017a. “Public Health Services.” Accessed January 1. www.euro.who.int/en/health-topics/Health-systems/public-health-services. ———. 2017b. “Social Determinants of Health.” Accessed January 1. www.who.int/ social_determinants/en/.

CHAPTER

GLOBAL HEALTH SYSTEMS INFORMATICS

11

Gordon D. Brown

Learning Objectives After reading this chapter, you should be able to do the following: • Understand different health information systems designs and how they use information technology as a decision support tool. • Discuss the advantages and disadvantages of centralized, distributed, functional, and integrated health information systems designs. • Critique various models of global corporate networks. • Explain how evidence-based knowledge can be applied globally. • Argue the premise that clinical protocols should be in the public domain and not proprietary. • Differentiate between publicly and corporately funded research findings.

Key Concepts • Centralized versus distributed data repositories • National health systems, national health insurance, and national information systems • Private–public partnerships • Global corporate networks • Generalizability and transferability of clinical evidence

Introduction This chapter has two distinct foci—the comparative analysis of health systems informatics and the development of global health systems. First, the chapter examines how different countries have designed and developed their health information technology (HIT) systems and what can be learned from these different approaches. The pace and breadth of applying clinical decision support 227

228

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

systems globally are not limited by the speed and power of HIT but by the economics, politics, culture, and information infrastructure of countries and regions. We examine the applications of HIT in countries, in particular the degree to which it has served as the basis for transforming health systems infrastructure (i.e., health systems informatics). Second, the chapter explores the development of global health systems harnessing the power of information technology (IT) to accumulate clinical knowledge and facilitate information exchange and patient communication. The design and testing of innovative models in various countries might guide how these technologies are applied in other countries, enabling the creation of integrated national and global systems. Health systems informatics embodies the complexity of developing an integrated health information infrastructure and is a strategic force in the design of the structure of health systems themselves.

Comparative Analysis of Health Systems Informatics: Design and Function Meaningful comparisons of the development and application of HIT across countries are complicated by differences in purpose, size, type, and structure of health systems in those countries, as well as the countries themselves. Countries can learn from each other in developing their own information system and understanding how HIT can serve as the logic and architecture for health system transformation. Most countries consider access to health services as a right and the responsibility of government, a perspective that differs from that in countries with private, locally based healthcare institutions such as the United States. Comparative analyses can extract conceptual and operational knowledge that is transferable. Each country follows a process for developing and designing its health information system, providing a good laboratory for examining alternative strategies and designs and their relationship to health system design and function.

Electronic Medical Records: Foundation of HIT Systems In all countries, the foundation of an information system is the development of the electronic medical record (EMR). Countries that have a national health system or national health insurance generally assume their public and private institutions and professionals will develop and use EMRs. This orientation influences the acceptance of IT and, in some countries, its design. This has not been the trajectory in health systems dominated by the private sector, such as in the United States. One measure of acceptance is the level of EMR use by physicians, particularly primary care physicians. Physicians in Sweden, Norway, Denmark, the Netherlands, and Australia, for example, were early adopters

C h a p te r 11:   Glob al H ealth Sy stem s Infor m atic s

and virtually 100 percent users of EMRs (Mossialos et al. 2016). Primary care physicians in New Zealand were among the world’s earliest and highest users of HIT, boasting almost 100 percent acceptance (Schoen et al. 2012). Several factors facilitate the rate of EMR use by physicians, including smaller, more homogenous populations as well as government IT policies and investments. But the most likely facilitator of adoption may be the cultural mind-set that healthcare is a social good and a national responsibility. A commonality among all countries is that a disparity exists between large and small, and between urban and rural, hospitals and clinics. This is a common system design failure because an information system that is accessible nationwide should be able to drive information and knowledge downward to remote, rural areas as well as upward to large, urban communities. It should also help patients remain in or return to their local communities. A conversion from paper records to an EMR, however, presents a great challenge (not to mention financial risk) to small institutions because of its cost and complexity. The time and expertise needed to install systems, provide training to users, and transform processes impose an additional administrative burden (Zhang and Zhang 2016). Thus, lack of such a system is a failure of design, investment, and deployment, not a limitation of HIT capability. Note that in countries with national or provincial health systems, automating the medical record is viewed by practitioners and institutions as a logical and accepted step in the development of HIT. However, the development of the EMR in many countries seems to lack a clear vision of the true power of HIT as a strategy for transforming the health system (chapter 4). It is not clear from the literature whether policymakers lack this vision or simply make a pragmatic decision as to what is achievable at a given time.

The Foundation Is Not a Design A common problem in designing an information system is the failure to envision what the final structure should look like. How will the disparate, separately developed parts of the system work together? Some pieces might collaborate and some may complement; there might be differences, but none should conflict by design. While Nielsen and Sæbø (2016) study this quandary in developing countries, it is also a problem in developed countries. This is dysfunction by design, and it typically results from different parent or sponsoring agencies investing in information systems, each with a silo architecture mind-set. IT architecture can be designed in a number of ways, with different capabilities and purposes as well as alternatives for achieving a given purpose— the equifinality principle. Information system design in the United States has generally followed a silo architecture (Ross 2003) that is primarily oriented to support traditional, individual clinical decision processes and organizational functions such as human and capital resources, logistics, and finance (Pascot,

229

230

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Information technology (IT) architecture Framework of information system infrastructure that aligns IT with enterprise strategy; could be hierarchical in a functional organization and focused on efficiency, or it could be strategic and form the basis for health system design

Bouslama, and Mellouli 2011). This is a dysfunctional information architecture, particularly for health systems in developing countries with fewer resources to spend on the initial investment and on the inevitable later redesign (Nielsen and Sæbø 2016). Nielsen and Sæbø (2016, 147–48) propose adopting a “connecting strategy” or systems perspective. More attention needs to be given to the overall IT architecture (chapter 15); often the completion of this structure is incremental and occurs through contributions from different agencies and functions. We are reminded again of the saying that “we cannot build a skyscraper by nailing together dog houses.” Another obstacle to designing IT architecture is that it does not involve enough individuals who have a strategic vision for an integrated, evidence-based, national information system. Developing an information system typically falls to a private EHR vendor and a few health system representatives, often IT technicians and clinical staff, who are operationally oriented and focused on technical aspects. The result is a design that does the wrong thing better. HIT systems in both developed and developing countries fail in the design phase when the process involves only operational staff and private vendors (Fragidis and Chatzoglou 2017). These individuals still conceptualize HIT as a technical project, not as a strategic resource. Another common issue is the security of accessible data. Granting patients, physicians, and health professionals access to medical records is a goal for many countries. Many national health systems are designed to allow controlled access to records by patients, hospitals, primary care centers, specialty clinics, laboratories, and policymakers. Each of these entities needs access to specific types of information in a secure environment. Secure access is a challenge, however, particularly for health systems with different ownership and incentive structures. Health systems in any country that do not conceptualize their HIT in an integrated, functional, and secure manner may find themselves underutilizing their investment in the long run and spending considerable resources redesigning their IT architecture.

Types of IT Architecture: Demise of “Form Follows Function” Historically, design was conceptualized as physical space, so that “form followed function.” We designed things rigidly based on function. In the digital world, however, design is about user interface; we design things to change. Most countries have focused on implementing institutional medical records linked to a national or regional repository for the purpose of accessing individual clinical information. The purpose of a national or regional health system orientation is to provide universal access to the EMR, which follows many different designs. These designs provide a laboratory for comparative analysis. National and regional access to medical records is not an insignificant

C h a p te r 11:   Glob al H ealth Sy stem s Infor m atic s

achievement, as witnessed by the slow progress in the United States. Patient information in many countries can be accessed regionally and nationally, but few systems have been developed to support clinical collaboration and a HITbased restructuring of clinical processes and institutions. A centralized design is one in which data from different institutional EHRs are gathered, duplicated, stored, and managed in a single repository. Canada, Australia, and Denmark have a centralized IT architecture, and each varies in the frequency of updates and the types of information included. This system requires considerable processing capacity and speed because all information exchange goes to and from a central repository. It is easy to manage and control but provides less flexibility to local and institutional healthcare providers. A distributed repository design is one in which all EHR data are stored locally in institutional databases. These data have reference points or locator codes so that they can be located, accessed, and sent in real time. This system needs less processing capacity but, because it is a distributed system, presents a greater challenge in ensuring the data are compatible, complete, and accessible. The Dutch Nationwide Electronic Health Record is a distributed repository (Fragidis and Chatzoglou 2017). In the United States, this design’s use of a national patient identifier raises security concerns. Another concern revolves around which health and human services would be included and who will manage the information exchange. These issues are more political than technical (Monica 2017). A semidistributed architecture characterizes the national data system in Greece. This architecture is suitable for countries with limited administration network bandwidth (Fragidis, Chatzoglou, and Aggelidis 2016). The national health system of Greece has seven health districts. Each district has a distributed repository of citizen health records that uses semantically interoperable standards. These distributed data sets include composites of local hospital and clinic EHRs. Access to patient information by institutions within a district is through the district data repository, called the Citizens Basic Health Data (CBHD). Access to patient information outside a given district is through the respective district CBHDs. Access is enabled through an encrypted “smart card” and personal identification number; the smart card also includes the location of the citizen’s CBHD. For security reasons, citizens’ access to the CBHD is limited. The system prides itself for being “citizen centered” as it allows considerable data to be captured, stored, and distributed, including data from social service agencies (which is a limitation in the centralized design). The semidistributed architecture is flexible, as most service coordination or exchange is done within communities and then within districts, but it is available nationally. This architecture supports health systems in countries with state or provincial systems. A blended architecture was developed around e-health and the logic of sensors and home health services (chapter 8). Ludwig and colleagues (2010)

231

232

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

developed and studied this architecture in Germany and characterized it as transinstitutional. This type of architecture is sensor enhanced and recognizes the growing importance of home health care and new sensor technologies in the population, as well as the large volume of information generated. The volume of information generated suggests a personalized “distributed system,” but processing data in the home would result in only aggregated information based on predefined metrics, which would be made available in external information systems. Home health IT architecture usually integrates at least “two legally independent persons: the senior herself/himself as a natural person and the health care provider as a legal person” (Ludwig et al. 2010, 212). The logic of the proposed architecture is based on the volume and value of knowledge generated from the system and its processing and storage capacity. Ludwig and colleagues (2010) recognize the value of integrated databases that enable secondary data analysis in different application scenarios. An information system architecture is transinstitutional if its components belong to at least two legally independent organizations, such as a hospital and clinic. The transinstitutional architecture would allow process sensor data to be externally saved in a hospital or a clinic or possibly in a local data repository. These data could be subjected to secondary data analysis. However, the volume of data might limit the data’s inclusion in home telehealth services under a semidistributed architecture. Home health and local social services have less utility in a national database oriented toward clinical care and consultations. The integration of sensor technologies into health information systems leads to architectural considerations that differ from the design of common architectures. Some countries include only medical services in their electronic databases, whereas others include a wide range of services, including social services. France, for example, is establishing an information exchange that includes social service professionals as well as primary and specialty providers. Australia’s system includes both health and disability support services and uses a unique national health number; this strategy adds not only considerable access but also coding and security complexity that are consistent with fully integrated patient-oriented systems of the future. Where do “medical services” stop and “health and human services” begin in such an integrated system? The answer is that they become one because it is a system that includes a range of health professionals such as dietitians, therapists, and social workers.

Restructuring Health Systems According to the Logic of Health Systems Informatics This book has demonstrated that the complexity of designing health information systems is exceeded by the difficulty of transforming the design of health

C h a p te r 11:   Glob al H ealth Sy stem s Infor m atic s

systems using the power of IT. This transformation is a long process, and the health systems in some countries may place greater value on the legacy pieces of the health system than on the measures of performance. If done well, transformation will take a lot of time and resources, but if not done well, it will result in suboptimal outcomes. This type of change will not have the speed and agility of a sports car; it will be more like turning a large oil tanker that needs a lot of space, time, and people working toward the same end.

The Australian Experience The logic of HIT can be used as the basis for conceptualizing health system architecture, which must be carried out within distributed institutional authority structures. This strategy is particularly messy in a private, decentralized health system such as that in the United States because the process will result in complex changes in system design that require analytical modeling (chapter 5). Establishing a national mandate to restructure services around an IT logic will encounter considerable disruption and will likely fail. Australia, for example, was among the first countries to envision an integrated health information system, starting in the early 1990s, and has considerable experience in implementing and evaluating various system designs and strategies (Garrety et al. 2016). Although IT can guide change, it has not proven to be an effective change agent when mandated. IT offers many ways to reach a desired destination, drawing on the concept of equifinality from systems theory (chapter 1). Australia initiated its first attempt at a national health information network in 2000, with a five- to ten-year development phase. Such a system would be unified, paving the way for the development and implementation of a national personal health record, which in turn would provide the template for changing how health services were delivered and financed across institutional, territorial, state, and federal boundaries and how patients accessed information and got involved in the care process. The design of the information system became the logic for determining changes in the system implemented by centralized planning and authority within a decentralized health system. Defining optimal outcomes includes engaging patients; health professionals; healthcare organizations; financial institutions; local, state, and federal government agencies; and possibly suppliers. All of these players have a degree of independence and are not inherently aligned to achieve optimal system performance, and one function does not dictate the structure of another. They can be aligned only by a common purpose and end point, and even then they seek their own pathway. The change strategy pursued in Australia was not sensitive to the relative decision-making autonomy of health professionals, the health policy mandates and responsibilities of different political units, and the autonomy and rights of patients. After six or seven years of top-down

233

234

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

reform, Australia abandoned the initiative because of widespread opposition by all elements in the system. Afterward, Australia rolled out a new initiative—the personally controlled electronic health record (PCEHR)—that accommodated patient values and made those being served the basis for system design. The PCEHR recognized the centrality of the patient and, among other things, allowed patients the ability to delete any medical findings and treatment protocols from their record that they did not want reported. Although the PCEHR was truly patient centered and patient controlled, it did not serve the essential requirements of health professionals because it omitted critical information needed for proper diagnosis and evidence-based treatment. It violated one of the essential tenets of health professionals—to base clinical decisions on the best information and evidence, including clinical judgment. Giving patients the prerogative to eliminate valuable medical information about themselves violates this tenet. The next iteration of the design—My Health Record or myHR—made the PCEHR optional. It allowed freedom of choice, but few professionals opted in and the national personal health record existed in concept only. Access, security, and privacy are important considerations for patients, but these considerations must be balanced with the other legitimate demands on the system (Nøhr et al. 2017). There is a distinction between a personal health record (PHR) and an EHR: The PHR is controlled by the patient, whereas the EHR has professional and system integrity and is therefore not subject to the patient’s restrictions. One important lesson from the Australian experience is that transformational change in healthcare is more complex than in other industries. Health systems are characterized by multiple levels of distributed responsibility and power, frequently conflicting but legitimate. An integrated information system for collecting and reporting data is easier to conceptualize in countries with universal health insurance or a national health system. In countries without a universal system, access to information on utilization, cost, and quality from health systems with different owners, reach, and types of services (including social services) first requires standardization.

Increased Private Sector Investment The trend in both developing and developed countries is toward increased diversification of health system ownership, and more and more public health systems, hospitals, and clinics are seeking funding from the private sector. Such involvement has the advantages of attracting needed capital to health systems (including capital for IT) and bringing in more creative methods for designing structures, strategies, and knowledge systems. Greater complexity is also a potential consequence, given private companies’ interests and motivations—particularly if they are investor owned. Private sector investment in

C h a p te r 11:   Glob al H ealth Sy stem s Infor m atic s

health systems in many countries consists of funding from both nonprofit and investor-owned firms. This public–private partnership is a strategy to increase access to services and to improve the quality of services by leveraging the capital, management capacity, and flexibility of private organizations (Reich 2002). However, diversity of ownership and public–private partnership may complicate the development of an integrated health information system. The challenge here is the coordination of data and services among these various entities. Countries such as Saudi Arabia, with traditional national and provincial health systems, are exploring the development of public–private partnerships, but they have concerns about the coordination of care across institutional and political boundaries (Bassi 2017). The organizing principle for system design is evidence of increased efficiency, quality, and integration, which may be difficult to achieve with public–private partnerships. As Rosenkötter and colleagues (2016) have said, the goal of having a fully integrated system is appropriate and offers considerable potential, but the difficulty of coordination remains a challenge. In the Netherlands, however, patient portals have achieved some success in integrating patient information across disparate regional health and information systems (Otte-Trojel et al. 2015). Integrated regional models have demonstrated greater patient involvement in wellness and medical care—particularly chronic care—with significantly improved health outcomes.

Health Systems Informatics and Health System Design In its biannual State of Health in the European Union report, the European Commission (2017) states, “The digital transformation of health and care has great potential for strengthening the effectiveness of health systems. . . . The healthcare workforce has to be prepared for technical innovation and patients should be at the centre of better health data for policy and practice.” This is what Australia attempted to do, in part, with its multiyear transformation efforts. The problem was that it focused too much on one group to the disadvantage of other groups. This same issue has been seen elsewhere. In Denmark, for example, physicians opposed the top-down rules for health record design imposed by the health ministry. They feared that such a system would be incomplete and inflexible and pose both a security risk and a financial burden on physicians. The Australian experience, which stems from the notion that the national government has the responsibility to design an IT architecture and to transform the national health system to accommodate that architecture, demonstrates the complexity of developing a national health information system (Garrety et al. 2016). The results confirm that such a wide-reaching change is a systems effort, requiring each function and entity in a system to change to reach an optimal solution. Accommodating dysfunctional components of the system is not the answer, but transforming all those parts is.

235

236

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Health systems informatics provides an innovative architecture that is transformational in nature and serves as the basis for conceptualizing and enabling innovations in health services delivery. A fundamental knowledge of systems theory and health systems informatics (chapter 1) can be acquired by clinicians, executives, suppliers, and policymakers and added to their own domain knowledge. Developing a health information infrastructure that is designed around the logic of traditional health systems structure and function is to automate obsolescence. Countries can learn from each other about advanced IT and essential correlates regarding system structure and function, as well as explore possible international collaborations. Caution should be taken when adopting seemingly new models, as often they are new only to the designers and are based on obsolete assumptions. Also, developed countries may not represent the best models of health information infrastructure or health systems design. Transforming health systems is difficult, and innovative models may originate in underdeveloped countries that need not make all the mistakes made by more advanced nations.

Knowledge-Based Health Systems Design A growing scientific literature supports innovation in health systems design, including information systems, and can inform policies and strategies in the global community. This does not mean the politics and strategies of countries should be unified; in fact, they will constantly change and evolve, creating different designs. Countries can learn from the diversity of design models tested in other countries. Improved health system performance is demonstrated by health systems designed according to the logic of health systems informatics. The aligning and motivating force is the commitment to achieve demonstrable superior outcomes; the strategy and structure for doing so differ from country to country and depend on the vision and mastery of industry leaders and policymakers.

IT-Enabled Innovations The Commonwealth Fund, in collaboration with the Institute for Healthcare Improvement, initiated an extensive study on what US healthcare providers can learn from innovations in other countries (Osborn and Goldmann 2017). Successful innovations were identified by an international panel of 200 experts, and the potential transferability and effect of these innovations (transformational change) were vetted by the Innovators Network. This network comprised 12 health systems representing different regions, population density (rural and urban), medical school affiliations, hospital design, and health plans.

C h a p te r 11:   Glob al H ealth Sy stem s Infor m atic s

237

Four areas were selected for experimentation in the various health systems (Osborn and Goldmann 2017): 1. Reducing Inappropriate Medication Use by Implementing Deprescribing Guidelines 2. Experience-Based Co-Design of Health Care Services 3. Postal Service “Call & Check Visits” for Isolated, Frail Elderly in the Community 4. Discharge to Assess: “Flipping” Discharge Assessment from Hospital to Home

All four of these pilot areas are enabled by IT—not as a supporting technology but as an architecture within which innovations in healthcare can be designed and tested. This study is important because it recognizes the potential of innovations in medical care delivery, not just in clinical science, to inform and guide change in other countries. The study’s most impressive aspect is its design, which is based on good science and not just expert judgment. More studies are needed on healthcare innovations tested in multinational settings and enabled by IT; such studies should not view IT as merely an automating technology but as a design concept.

Development of Global Health Systems: Collaborative Systems Our focus shifts now to global health. First, to what degree is clinical knowledge universal, and how can it be captured, assessed, transmitted, and applied in clinical decision support systems across countries? Second, to what degree can global corporate collaboratives be designed to enable the sharing of clinical and systems knowledge? A well-established and growing market exists in various countries for individuals who seek clinical services on the basis of costs and outcomes. At the most basic level are destination medical centers that provide niche specialty services to patients across international boundaries and, in doing so, gain institutional prestige and financial profit. At a higher order of complexity, global collaborative organizations gain new markets based on the value of the clinical knowledge embedded in their information systems and on their knowledge of transforming and managing information-based healthcare organizations.

Global Sources of Evidence for Clinical Decision Support Knowledge-based clinical decision support systems have been recognized for improving clinical decisions and outcomes. At the most basic level, this knowledge can be increasingly obtained from research articles available online.

Knowledge-based clinical decision support system Collection of scientific evidence and expert and experiential knowledge (including that generated through artificial intelligence) used to inform clinical decisions

238

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

The World Health Organization (WHO) launched the Health InterNetwork Access to Research Initiative (www.healthinternetwork.org) in 2002. This network is a partnership between leading biomedical publishers, academic institutions, and organizations in the United Nations. It provides local, nonprofit institutions, particularly in developing countries, with low-cost or free online access to 2,900 major journals in biomedical and social sciences. It has one of the world’s largest collections of biomedical and health literature. Currently, 1,400 institutions in 104 countries participate in the network. Systematic reviews of the literature on evidence-based clinical guidelines and protocols further inform the clinical decision process. The National Guideline Clearinghouse database (www.ahrq.gov/cpi/about/otherwebsites/ guideline.gov/index.html) includes valuable detail on the types and levels of evidence that support each guideline. Advances in the conducting and reporting of systematic reviews have increased the level of evidence available to support the guidelines and their acceptance by professionals in the country and globally. In addition, methods for weighing evidence and synthesizing results through systematic reviews have greatly improved (Clancy and Cronin 2005). The power of such databases is their access to the latest and best research on clinical diagnoses and treatment outcomes. A current limitation is the sample size of studies conducted in some nations, which do not represent the world population. Because of the differences in population characteristics and the science supporting the guideline, to what degree are the findings from these studies generalizable to practices in various countries? Information on sample size and integrity, confidence in the analysis, and the degree of acceptance by a country or health system are important considerations. Many policy questions are raised on clinical decision support, and the first is the degree to which guidelines are accessible across political boundaries. Technically, it is easy for information to be accessed, transmitted, and reported globally. Policy questions include who should access and evaluate (in terms of scientific merit) the decision support guidelines and how appropriate they are for the health conditions and cultural setting. Clinical decision support tools can be accessed by each country and evaluated and distributed on the basis of acceptable standards. As the number of clinical trials increases and evidence from practice accumulates throughout the world, the task of assessing the level of evidence may become more difficult and may be redundant. Governments, professional associations, and healthcare organizations continually sponsor the development and dissemination of clinical guidelines (e.g., through the journal BMJ Quality and Safety). Some countries look to international agencies, such as WHO, to gather and assess the level of evidence from accumulated clinical trials. As more countries become involved in conducting or sponsoring clinical trials, this task will become more complex but may yield greater benefit. The responsibility for generating the best knowledge, through valid and reliable guidelines available

C h a p te r 11:   Glob al H ealth Sy stem s Infor m atic s

globally, may be assumed by national and/or international clearinghouses, which will review the evidence to determine their scientific merit. The quality of the evidence depends on whether populations from an appropriate range of countries were included and whether the tests were based on sound scientific methods. The use of evidence depends on the values and policies of each country. The conducting and reporting of clinical trials must follow rigorous and consistent standards (i.e., the challenge of standardization again appears) (Bhatt and Mehta 2016). Research on clinical effectiveness, carried out by pharmaceutical companies, is increasingly multinational because their markets are global. Research carried out by private corporations enhances the quality of the research and expands the market and monopoly control of products. Again, health systems informatics promotes the alignment of all of the elements of the system along quality, cost, and continuity. Clancy and Cronin (2005) report that, in the 1990s, the scope of global clinical trials grew from including 28 to 79 countries. The objectivity of clinical trials and the level of evidence reported, including the nature of the population sampled, must be assessed to determine the generalizability of the study. Over the past ten years, there has been a dramatic rise in systematic reviews of clinical trials, increasing the reliability of evidence for use in healthcare decisions. Multinational studies have been conducted to assess the evidence and biases of reported clinical decision support. An evaluation of clinical guidelines, using the Appraisal of Guidelines for Research and Evaluation (AGREE), found an improvement in overall guideline quality but an unacceptable variation in the level of science that support those guidelines (AGREE Collaboration 2003; Armstrong et al. 2017). Certainly, the gathering and reporting of guidelines will require the studies supporting them to be reported and to meet high methodological standards. The proliferation of published guidelines in the United States and abroad requires the creation of internationally recognized standards for developing and reporting guidelines, as well as a reliable and trusted clearinghouse that compiles and makes them available (Grilli et al. 2000). Two factors that are important to investing in and valuing the clinical decision support infrastructure are the large and increasing volume of research being conducted worldwide and the rate of change in science that informs clinical decisions.

Destination Healthcare Destination healthcare (commonly known as medical tourism) is a major industry with a two-way flow of patients: (1) those going to other countries to undergo clinical procedures at a substantially lower cost, and (2) those coming into the country for services that are not available or of high quality in their own country. Medical services that appeal to low-cost markets are frequently bundled with tours, hotel accommodations, and other fringe benefits, like a minivacation. The cost savings might be considerable: A heart valve replacement

239

240

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

in the United States costs $170,000 but only $9,500 in India (Medical Tourism Association 2016). Medical tourism has been enabled by Internet marketing, but low-cost destinations are now using clinical evidence to support their claim that their clinical outcomes are equal or superior to those achieved in specialty centers in high-cost countries. The low-cost facilities’ ability to measure and compare clinical quality and outcomes could cause high-cost centers to reexamine their policies and practices in areas such as insurance plans, pharmaceuticals, specialty procedures, institutional profits, and personal incomes. IT has enabled medical markets to open, although they will likely remain niche or morph into more formal institutional networks. Destination healthcare, offering highly technical and complex procedures, presents a global market for a growing number of countries. In the United States, these services are provided by highly respected research medical centers with an established specialty, such as MD Anderson Cancer Center (www.mdanderson.org/patients-family/becoming-our-patient/international-center.html), the Cleveland Clinic (https://my.clevelandclinic.org/patients/international), and Mayo Clinic (https://dmc.mn/). Regional health systems also compete in this niche market, such as the University of Missouri Hospitals and Clinics, which offers its nationally recognized regenerative orthopedics services (www. muhealth.org/our-stories/programs-establish-mu-destination-health-care).

Global Corporate Systems Global corporate systems constitute legal arrangements between or among health institutions in two or more countries for the purpose of collaborating on the provision of clinical services; education; or the development of shared, evidence-based clinical decision support systems. There is a range of corporate models and of the health services they support.

Global institutional network Legal arrangement between two or more healthcare institutions in two or more countries for the purpose of collaborating on clinical care, education, or development of shared evidencebased clinical decision support systems

Collaborative Networks Global institutional networks are formal collaborations among clinical centers. In such networks, providers share clinical protocols, institutional knowledge, training and consultations, and other relevant information. One current model of this network is a collaboration that focuses on physician training and clinical consultations. MD Anderson Cancer Center, for example, hosts Project ECHO (www.mdanderson.org/education-training/global-outreach/project-echo. html), a teleconsultation and telementoring educational program for centers in rural and underserved communities. Another type of collaboration is a network of specialty institutions that share controlled access to clinical protocols, best practices, and training and joint research. Partners in this kind of network have controlled access to proprietary, evidence-based protocols through clinical consultations. They do not share interoperable EHRs across healthcare settings, which would permit joint clinical consultations (Clancy and Cronin 2005) because of their inability to

C h a p te r 11:   Glob al H ealth Sy stem s Infor m atic s

connect their disparate EHRs and to integrate their financing models. There is a sufficient level of shared corporate identity that the partners safeguard the network’s prestige and quality standards. This network could broaden to share common protocols and an integrated insurance model for specialty procedures. The IT problem can be addressed more easily than the joint insurance scheme, however. For example, the MD Anderson Cancer Center (2017) consists of national and international centers that provide services “ranging from quality assurance and best practice guidelines to full clinical integration.” Interoperable models have been extended to a global market but are limited by the current design of information systems and financial models. The failures of design include the functions of institution-centered information systems and financing, both of which can be overcome. These are impediments to global markets, but they are solvable and can bring value to patients and practitioners. The technical capacity is ahead of the business model. E-health Networks E-health networks are formed across national boundaries. E-health network partners have some form of interoperable information exchange and shared clinical protocols for providing services that transcend distances and geographic boundaries. Patients who need specialty service travel to the specialty hospital or clinic. Continuity of care is provided according to common protocols and information exchange. These networks have a high degree of complexity and require integration of clinical, information systems, and financing structures across national boundaries. They could boast being a true patient-oriented global health system. One example is the Pan-African e-Network, which connects more than 30 health centers across the African continent and is affiliated with 12 tertiary hospitals in India that provide medical teleconsultations (Duclos 2015). The network is structured around a “south to south” tradition of Indo-African trade and cooperation, with a specific emphasis on healthcare. It shows that healthcare does not have to be an isolated policy area in developing countries. The network has been described as an “integrated solution aimed at taking care of patients at a distance” (Duclos 2015, 156). Its success may be attributed to physicians valuing the relationship with other providers more than the access to information and protocols, an attitude that depends on changing behaviors, processes, and problem-solving models. Discussions of interoperable global systems generally assume that individual patients would come to specialty centers from developed and developing countries throughout the world. One might also conceptualize a market where a well-designed primary care system might form a collaboration with a specialized clinical center that could provide highly complex surgical and other procedural services. Local primary care centers would provide continuity of care after patients returned home, supported by common protocols and shared

241

242

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

EHR access. A system could have cost advantages and incentives, which are not currently available, to keep patients in their local health systems. Such a model is transformational and requires a new design and business plan, but innovative models should be considered. Forming international collaborations with new supportive structures may be easier than forming domestic networks. Joint Ventures and Other Corporate Health Systems The highest level of global collaboration is the corporate integration of hospitals and clinics. The reasons to pursue such formal structures include an integrated strategy, a common corporate culture, shared human resources, and access to capital. Like all formal partnerships, joint ventures are less nimble and more complex to manage. They might evolve from less formally structured systems. As with domestic integrated systems, their success depends more on a shared culture than on any business plan or IT integration. An example of a joint venture is the Johns Hopkins Aramco Healthcare Company in Dhahran, Saudi Arabia (www.hopkinsmedicine.org/international/ international_affiliations/middle_east/johns_hopkins_aramco_healthcare.html). Aramco is the Saudi petroleum corporation and one of the world’s largest companies in terms of size, complexity, and wealth. Historically, Aramco supported its own primary care system and a secondary or tertiary hospital in Dhahran. It has a defined population base of 350,000 employees and dependents, considerable capital, and an aggressive research and development culture. It is a world leader in exploration, production, refining, and distribution of petrochemicals, but its executives concluded they were in the oil business and not health. Integrating with regional centers in Saudi Arabia would have been the traditional strategy. The joint venture, however, was with Johns Hopkins, a global leader in clinical excellence, medical education, research, and evidence-based clinical protocols. An initial corporate strategy was to align the two organizations to use a common information system. Financial, cultural, and structural details also had to be worked out, but the architecture envisioned was based on IT. The partnership brought to Aramco an immediate identity and involvement at the highest levels of teaching, research, and clinical practice. Beyond the prestige and tradition, the partnership enabled the sharing of clinical knowledge and consultations. It is a complex global integration, and much can be learned from it. E-learning E-learning is one of the core functions of e-health. E-learning has evolved from standardized content and delivery based on global disease prevalence to targeted, customized, and evidence-based knowledge to inform and guide individual learners. Thus, it might be characterized as e-education. Learners are transformed from passive absorbers of information to active participants in knowledge acquisition. For example, the WHO Health Academy develops its own learning content, which enables the academy to select health areas it

C h a p te r 11:   Glob al H ealth Sy stem s Infor m atic s

deems important on the basis of global monitoring of health status. Health status is based on accumulated patient information augmented by national health surveys. Courses are customized to a health issue, as identified by good science, adopted to the local culture, and presented in the local language. These courses have an appropriate level of rigor and include hands-on practice, automated feedback, and immediate correction of errors. The limitation of closed systems such as the Health Academy is that they run on their own platforms and, although accessible online, are not discoverable through web portals. These are the trade-offs of closed systems. Developing and maintaining e-learning courses also require the sponsorship and support of organizations with a global reach and the capacity to understand and adapt courses to national and local conditions.

Global Health Policy and Population Health The recognition that infectious diseases and natural catastrophes in one country can have global consequences has emerged because people are becoming increasingly aware that they live in a global community and nations are better able to respond to such threats. Our interest comes from a humanitarian concern and from our instinct to protect ourselves and prevent a pandemic. As Freidman (2005) has said, “the world is flat,” meaning that we are increasingly interdependent, all living, playing, and competing on the same field and thus facing not only the same opportunities but also the same risks. Population health–related risks include global warming, climate change, foodborne and airborne illnesses, and mass migrations. These and other problems have greatly increased the interest in evidence-based solutions. A substantial body of research is available that assesses the effectiveness of population-based interventions to control global health risks (WHO 2015). Ongoing systematic reviews of the research literature by country ministries and linked to international organizations are continually needed, however. Policies and national or global strategies should be evidence based, drawing on the considerable research findings on health risks and effective interventions (WHO 2016). The existing and accumulating knowledge should be systematically collected and used to frame evidence-based policy decisions and intervention strategies. The urgency of developing evidence-based policy decisions is related, in part, to the growing interdependency among nations. National health policies affected by global forces recognize the risk of the spread of infectious diseases. This has been an age-old public health problem, but the level of risk is higher now because urban populations are more concentrated, travel has increased, more goods and services are exchanged between countries, and more populations are migrating across national boundaries. Countries are acutely aware that it is in their self-interest to participate in identifying and managing global

243

244

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

health challenges. Managing infectious diseases within countries as a global strategy has attracted a lot of support. Health systems informatics tools can help with the development of evidence-based policies and strategies. For example, WHO has initiated a Global Digital Health Index, an interactive digital platform built on national databases to measure and monitor digital capacity for tracking global health risks (Mechael and Kay 2016). Behavioral, environmental, occupational, and metabolic risk factors are determinants of national stability and worker productivity that can adversely affect economic growth. Instability is not in the best interest of an increasingly interdependent global economy. Evidence-based national health policies are applied in international markets, particularly in developing nations, whose population health (as well as trade and financial health) are important to nations with developed economies. Beyond disease and health risk factors are global risk factors, such as climate change and its effect on the environment and health status (National Academies of Sciences, Engineering, and Medicine 2017, 18): The effects of climate change on health will be felt in the form of malnutrition, drought, extreme temperatures, worsened air quality, and infectious disease spillover—and mitigation of these effects will require work well beyond the health sector, necessitating multidisciplinary collaboration and action.

Multidisciplinary and multisectoral policies and strategies require drawing on biological and systems science and developing policies that go beyond the healthcare sector and national boundaries. Such policies entail collaborating and sharing information with global organizations (such as WHO) and national health systems (both ministries and healthcare institutions). Accumulated evidence should be collected to guide these policies and strategies. Evidence may be drawn from national and international databases and predictive analyses. A systems perspective requires analytical models that assess global threats and develop global strategies. Evidence-based policies use predictive and systems dynamic modeling to guide economic and political decisions. Decades of cost–benefit analysis and disease modeling have produced analytical models that guide economic investments, trade relations, and development strategies. According to the World Economic Forum (2015), a $120 billion investment in noncommunicable diseases in low- and middle-income countries would result in a 10 percent ($377 billion) decrease in the cumulative cost of cardiovascular diseases. Such analyses are likely valid but not actionable because they lack an effective strategy for driving change in countries. Dynamic modeling links disease models to investment, trade, and growth strategies to develop alternative pathways. Scientific evidence does not always produce rational thought. One compelling argument is that changing the investment in health and social programs

C h a p te r 11:   Glob al H ealth Sy stem s Infor m atic s

requires changing the structure and dynamics of the economy itself (Korten 2015). Coordinating government policies and changing the behavior of large national and multinational corporations seems overwhelming and predictive of a dire global future. An alternative future, however, with an increased sense of global community and assisted by and in part derived from interactive models, informs us and potentially enlists our participation. One can make the case that life will be more chaotic if we do not work toward that future.

Conclusion The challenge in studying health informatics from a systems perspective is that systems are large, complex, dynamic, and frequently irrational. The path forward to better understanding their complexity and dynamics, and for developing better solutions, is by building evidence-based models that better inform the policies of governments and private corporations. Dynamic modeling in the health system has been based on historical data and projected alternative outcomes. These models apply to local health systems and might have greater utility for national and global health systems. The analytical models are not new, but their utility will increase as decision makers address higher-order problems from both individual and global perspectives, leading to changes in policy formation and system dynamics. Policy formation will constitute changes across health systems and countries. Changes in dynamics will include new initiatives, such as greater involvement of private corporations in the delivery of care, as well as greater coordination and accountability between service providers and funding agencies, including national health insurance. Systems will be more nimble, efficient, and accountable, which are not characteristics of large, complex, and bureaucratic corporations. New corporate models are needed that deploy the power of advanced IT. Achieving this state requires health systems to adopt an integrated local and global perspective. Like the commercial world, the healthcare delivery field is complex, interrelated, and constantly changing. In the process of change, we hope healthcare does not lose its humanity.

Chapter Discussion Questions 1. Debate the point that countries with different health systems structures all aspire to a similar goal: a system that is evidence based and nationally accessible. 2. How can an integrated regional health information system serve to support clinics and hospitals in rural and isolated communities?

245

246

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

3. Discuss why the design and selection of information systems in hospitals and health systems address technical issues but not strategic potential. 4. Debate the issue of giving patients the right to exclude sensitive information from their patient record; include health and risk factors for different age groups. 5. What are the implications of increased private funding and ownership of health facilities and services on the development of integrated information systems and coordinated care? 6. To what degree is clinical evidence generalizable across political boundaries? To what degree is it accessible? 7. Where do “medical services” stop and “health and human services” begin in an integrated system?

Case Study  Envisioning a Global Community Historically, health has been described as a community affair (National Commission on Community Health Services 1966), a perspective that still pervades cultures around the world. The transformation from a local, agrarian culture to a global, information-driven world characterizes most industries and services (Toefler 1980). For example, the construction and maintenance of roads in the United States were primarily the responsibility of townships, the most basic unit of government. Work to build and repair roads was carried out by local citizens who collaborated to complete the task. Most transportation was provided by horses and within the boundaries of a township. The next larger unit was the county, which was made up of townships with a city center called the county seat, which was physically accessible through a network of county roads maintained by county employees. The county seat was home to the county public health department, hospital, sheriff, jail, and courthouse. The county was managed by commissioners charged with collecting taxes and administering elections. Counties were generally configured such that a person could ride a horse-drawn buggy to the county seat, conduct business or attend events there, and then return home by nightfall. This township structure was logical, functional, and effective, based on horses being the primary mode of transportation. Fast-forward to modern times, when cars and trucks travel nonstop at high speeds on four- to six-lane interstate highways. The structures of townships and counties were not the feasible political units to enable this change. In 1956, the US Congress passed the National Interstate and Defense Highways Act (Public Law 84-627), placing much of the financial burden and responsibility for design and overall control of highways on the federal government. States strongly resisted the federal government’s involvement,

C h a p te r 11:   Glob al H ealth Sy stem s Infor m atic s

stalling the construction process. Only when national defense was added as a rationale for the interstate highway initiative did the project move along. President Dwight Eisenhower had encountered the design and utility of such roads in Germany during World War II, when the German government was building the autobahn to support the movement of military units. The reason for the opposition to the Highways Act was that it took away power and control over roads from townships, counties, and states, even though transportation technology and use had evolved beyond the horse and buggy. The same type of resistance to advances in technology and standards exists today—not only in the United States but throughout the world. The ability and availability of IT to enable greater global—or even national— information exchange are offset by the constraints of traditions, politics, economics, business models, and cultural values that were established and followed in a different era. The basic challenge globally is twofold: (1) Can existing technology based on proprietary IT systems, local and regional hospitals, cost-based insurance models, and other elements be adapted to provide a benefit to a global market? and (2) can new concepts and systems be designed around the logic and power of IT? The limitations of global health information systems are not primarily technical but economic, political, organizational, cultural, and historical. Imagine if the United States had not overcome the initial resistance to road building. We might be driving motor vehicles on narrow country roads that are disconnected from other routes. This is akin to what could happen without HIT transformation.

Case Study Discussion Questions 1. Review the forces at play that determined the nature and pace of transformation of the US highway system. Who were the change resisters? 2. What technology enabled the vision of a highway system that characterizes developed countries today? How was the technology related to the design and financing of the road system? 3. Why is an interstate highway system a means of moving large military units across the country, and why is it a defensive strategy? 4. Trace the development and application of HIT in the United States. How does it mirror the development of the interstate highway system? 5. Think of another industry—for example, the finance and banking industry—and its ability to inform and facilitate transactions that transcend space, politics, language, traditions, and culture. How has that industry outpaced the health system?

247

248

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Additional Resources Appraisal of Guidelines for Research and Evaluation (AGREE): www.agreetrust.org/. BMJ Quality and Safety: http://qualitysafety.bmj.com/. Buckley, G. J., J. E. Lange, and E. A. Peterson (eds.). 2014. Investing in Global Health Systems: Sustaining Gains, Transforming Lives. Washington, DC: National Academies Press. Health InterNetwork Access to Research Initiative: www.healthinternetwork.org. National Academies of Sciences, Engineering, and Medicine. 2017. Global Health and the Future Role of the United States. Washington, DC: National Academies Press. Partnerships for Enhanced Engagement in Research (PEER): www.usaid.gov/ what-we-do/GlobalDevLab/international-research-science-programs/peer. Sturchio, J. L., and A. Goel. 2012. “The Private-Sector Role in Public Health: Reflections on the New Global Architecture in Health.” https://csis-prod.s3.amazonaws. com/s3fs-public/legacy_files/files/publication/120131_Sturchio_PrivateSector Role_Web.pdf. World Economic Forum. 2006. “Global Health Initiative: Public–Private Partnership Case Example.” https://mosquitozone.com/sites/default/files/BHP%20 Billiton.pdf. World Health Organization Health Academy: www.who.int/healthacademy/courses/en/.

References AGREE Collaboration. 2003. “Development and Validation of an International Appraisal Instrument for Assessing the Quality of Clinical Practice Guidelines: The AGREE Project.” Quality and Safety Health Care Journal 12 (1): 18–23. Armstrong, J. J., A. M. Goldfarb, R. S. Instrum, and J. C. MacDermid. 2017. “Improvement Evident but Still Necessary in Clinical Practice Guideline Quality: A Systematic Review.” Journal of Clinical Epidemiology 81: 13–21. Bassi, J. 2017. “Vision 2030 and the Opportunities It Represents in Healthcare in Saudi Arabia.” Published January. www.tamimi.com/en/magazine/law-update/ section-14/dec-jan-2017/vision-2030-and-the-opportunities-it-representsin-healthcare-in-saudi-arabia.html. Bhatt, D. L., and C. Mehta. 2016. “Adaptive Designs for Clinical Trials.” New England Journal of Medicine 375 (1): 65–74. Clancy, C. M., and K. Cronin. 2005. “Evidence-Based Decision Making: Global Evidence, Local Decisions.” Health Affairs 24 (1): 151–62. Duclos, V. 2015. “Designing Spaces of Care in the Era of Global Connectivity.” Medicine Anthropology Theory 2 (1): 154–64. European Commission. 2017. “‘State of Health in the EU’ Report Recommends Digital Transformation of Health and Care.” Published November 27. https://ec.

C h a p te r 11:   Glob al H ealth Sy stem s Infor m atic s

europa.eu/digital-single-market/en/news/state-health-eu-report-recommendsdigital-transformation-health-and-care. Fragidis, L. L., and P. D. Chatzoglou. 2017. “Development of Nationwide Electronic Health Record (NEHR): An International Survey.” Health Policy and Technology 6 (2): 124–33. Fragidis, L. L., P. D. Chatzoglou, and V. P. Aggelidis. 2016. “Integrated Nationwide Electronic Health Records System: Semi-distributed Architecture Approach.” Technology & Health Care 24 (6): 827–42. Freidman, T. L. 2005. The World Is Flat: A Brief History of the Twenty-First Century. New York: Farrar, Straus and Giroux. Garrety, K., I. McLoughlin, A. Dalley, R. Wilson, and P. Yue. 2016. “National Electronic Health Record Systems as ‘Wicked Projects’: The Australian Experience.” Information Polity 21 (4): 367–81. Grilli, R. I., N. Magrini, A. Penna, G. Mura, and A. Liberati. 2000. “Practice Guidelines Developed by Specialty Societies: The Need for a Critical Appraisal.” The Lancet 355 (9198): 103–6. Korten, D. C. 2015. Change the Story, Change the Future: A Living Economy for a Living Earth. A Report of the Club of Rome. Oakland, CA: Berrett-Koehler Publisher. Ludwig, W., K. H. Wolf, C. Duwenkamp, N. Gusew, N. Hellrung, M. Marschollek, T. von Bargen, M. Wagner, and R. Haux. 2010. “Health Information Systems for Home Telehealth Services—a Nomenclature for Sensor-Enhanced Trans­ institutional Information System Architectures.” Informatics for Health and Social Care 35 (3–4): 211–25. MD Anderson Cancer Center. 2017. “MD Anderson Cancer Network.” Accessed November 10. www.mdanderson.org/about-md-anderson/our-locations/ md-anderson-cancer-network.html. Mechael, P., and M. Kay. 2016. “Unlocking the Potential of Digital Health.” Published June 10. Devex. www.devex.com/news/unlocking-the-potentialof-digital-health-88274. Medical Tourism Association. 2016. “Compare Prices.” Accessed May 5, 2018. http:// medicaltourism.com/Forms/price-comparison.aspx. Monica, K. 2017. “National Patient Identifier Gains Congressional Support.” Published May 11. https://ehrintelligence.com/news/national-patientidentifier-gains-congressional-support. Mossialos, E., M. Wenzl, R. Osborn, and D. Sarnak (eds.). 2016. 2015 International Profiles of Healthcare Systems. Published January. www.commonwealthfund. org/~/media/files/publications/fund-repor t/2016/jan/1857_ mossialos_intl_profiles_2015_v7.pdf. National Academies of Sciences, Engineering, and Medicine. 2017. Global Health and the Future Role of the United States. Washington, DC: National Academies Press. National Commission on Community Health Services. 1966. Health Is a Community Affair. Cambridge, MA: Harvard University Press.

249

250

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Nielsen, P., and J. I. Sæbø. 2016. “Three Strategies for Functional Architecting: Cases from the Health Systems of Developing Countries.” Information Technology for Development 22 (1): 134–51. Nøhr, C., L. Parv, P. Kink, E. Cummings, H. Almond, J. R. Nørgaard, and P. Turner. 2017. “Nationwide Citizen Access to Their Health Data: Analysing and Comparing Experiences in Denmark, Estonia and Australia.” BMC Health Services Research 17: 1–11. Osborn, R., and D. A. Goldmann. 2017. “Piloting Health Care Delivery Innovations from Abroad: A Systematic Approach.” Published November 27. www. commonwealthfund.org/publications/blog/2017/nov/pilotinghealth-delivery-innovations-from-abroad. Otte-Trojel, T., A. de Bont, M. Aspria, S. Adams, T. G. Rundall, J. van de Klundert, and M. de Mul. 2015. “Developing Patient Portals in a Fragmented Healthcare System.” International Journal of Medical Informatics 84 (10): 835–46. Pascot, D., F. Bouslama, and S. Mellouli. 2011. “Architecturing Large Integrated Complex Information Systems: An Application to Healthcare.” Knowledge and Information Systems 27 (1): 115–40. Reich, M. R. 2002. Public–Private Partnerships for Public Health. Harvard Series on Population and International Health. Cambridge, MA: Harvard University Press. Rosenkötter, N., P. W. Achterberg, M. J. H. van Bon-Martens, K. Michelsen, H. A. M. van Oers, and H. Brand. 2016. “Key Features of an EU Health Information System: A Concept Mapping Study.” European Journal of Public Health 26 (1): 65–70. Ross, J. W. 2003. “Creating a Strategic IT Architecture Competency: Learning in Stages.” Massachusetts Institute of Technology Sloan School of Management Working Paper No. 4314-03. Posted June 17. https://ssrn.com/ abstract=416180. Schoen, C., R. Osborn, D. Squires, M. Doty, P. Rasmussen, R. Pierson, and S. Applebaum. 2012. “A Survey of Primary Care Doctors in Ten Countries Shows Progress in Use of Health Information Technology, Less in Other Areas.” Health Affairs 31 (12): 2805–16. Toefler, A. 1980. The Third Wave. New York: William Morrow and Company Inc. World Economic Forum. 2015. “Maximizing Healthy Life Years: Investments That Pay Off.” Published January. www3.weforum.org/docs/WEF_Maximizing_ Healthy_Life_Years.pdf. World Health Organization (WHO). 2016. “Promoting Health While Mitigating Climate Change.” Technical Briefing for the World Health Organization Conference on Health and Climate, Geneva, Switzerland, August 27–29. www. who.int/phe/climate/conference_briefing_2_promotinghealth_27aug.pdf. ———. 2015. Operational Framework for Building Climate Resilient Health Systems. Geneva, Switzerland: World Health Organization. Zhang, X.-Y., and P. Zhang. 2016. “Recent Perspectives of Electronic Medical Record Systems.” Experimental & Therapeutic Medicine 11 (6): 2083–85.

CHAPTER

CONTROLLED TERMINOLOGY AND THE REPRESENTATION OF DATA AND INFORMATION

12

Timothy B. Patrick and Carmelo Gaudioso

Learning Objectives After reading this chapter, you should be able to do the following: • Understand the ways health informatics may be considered a branch of representational science. • Understand the concept of surrogate representation and its importance to health and biomedical information management. • Explain why concept-based controlled biomedical terminologies are important. • Describe the basic aspects of the terminology problem and its relation to interoperability of information. • Name the basic components and uses of controlled terminology. • Relate controlled terminology to metadata. • Explain the basic principles of terminology mapping. • Understand the relationship between controlled terminology and variation in healthcare.

Key Concepts • • • • • • •

Metadata Metadata schema Common data element Terminology mapping Interoperability Terminology problem Metathesaurus 251

252

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Introduction Controlled terminology Set of domainspecific multiword terms selected from natural language and organized by hierarchical and associative relationships

This chapter discusses the use of terminology standards (typically called controlled terminology) in representing data and information. Standard languages and terminologies are important in clinical care and research to enable the sharing and reuse of clinical information across providers, patients, organizations, and other stakeholders. Adherence to the same standard by different providers and organizations is not always possible, but even when the standard languages are not the same, the fact that they are standardized and controlled facilitates bridge-building among different users. A lack of standards in a healthcare setting may have consequences ranging from mere annoyances to poor quality of care and patient harm. Perhaps one of the more striking examples of the serious effect a lack of standard terminology can have is this account from the notorious Tuskegee syphilis study (Jones 1982): An official of the Centers for Disease Control (CDC) stated that he understood the term “bad blood” was a synonym for syphilis in the black community. Pollard [one of the subjects] replied, “That could be true. But I never heard no such thing. All I knew was that they just kept saying I had the bad blood—they never mentioned syphilis to me, not even once.”

If this account is to be believed, the lack of a common terminology led to the investigators’ failure to fully inform the study subjects about their disease and to obtain informed consent for their participation in the study. The thesis of this chapter is that shared controlled terminology standards are fundamental to high-quality, safe, and productive clinical care and research.

Health Informatics and Representational Science Ancient philosophers contemplated problems of perception and knowing. One issue that concerned them was the distinction between our human perception of objects and what was sometimes called the thing in itself—that is, the stuff of reality as it is apart from the way it is filtered through our sensory experience. A similar, although perhaps less mysterious, distinction is important for the management of health and biomedical data and information. There is the information itself, including its meaning and the facts and truths it carries, and then there is the way we represent that information. A patient’s discharge diagnoses may be represented in various ways, some perhaps better or worse than others. Similarly, a biomedical scientific research paper may be considered an original source of information (the thing in itself), while the manner in which

Cha p t er 12 :  Co n t ro l l e d Te r mi n o l o g y a n d th e Rep resentation of D ata and Infor m ation

that paper is indexed or described (including the language used to index or describe it) constitutes our representation of it. In this way, health informatics is a branch of representational science, the study of how data and information are represented. Clearly, health informatics as a science and profession is more than the study of how data and information are represented; analyses of data, and interventions based on those analyses that seek to improve the quality process and outcomes, are fundamental to health informatics (chapter 5). But at its core, health informatics is a representational science that enables fruitful analyses and interventions through the study of how data and information may be best represented.

Surrogate Representations and Controlled Terminology Components The ways in which data and information are represented are sometimes called surrogate representations—that is, what we initially encounter and make use of are our representations of the data and information; for example, when we search a repository of health information—whether patient records or biomedical research articles—what we are looking for are our representations of that information. This surrogacy applies more generally as well. When we search for and access information from collections or repositories, including the Internet, we do so through the mediating facade of representations in the form of indexes, bibliographic record databases, or catalogs. Such representations may be composed of words or phrases taken from the primary information objects themselves or of formally structured terms that do not occur in the information objects but that have been deliberately attached to those objects to represent their contents. These formally structured terms are components of a controlled terminology. The terminology is controlled in that the terms it contains have been carefully selected for a particular domain of information. A controlled terminology typically has the following components: • A thematically restricted set of terms (multiword noun phrases) that are selected purposely from natural language and that serve as precoordinated concept expressions for a corresponding set of concepts • A preferred expression for the concept where there are multiple expressions for the same concept • A set of semantic relationships that relate concepts and that are defined for the set of precoordinated concept expressions • A set of names for the relationships • A set of alphanumeric identifying codes for the concepts • A set of alphanumeric identifying codes for the terms

253

Representational science Branch of information science that studies how data and information are represented to support information storage and retrieval

254

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

• A set of alphanumeric identifying codes for the relationships • A set of rules for building postcoordinated concept expressions using the precoordinated concept expressions and relationships • Rules for the application of the terminology

Precoordinated Concept Expressions and Concepts

Precoordinated concept expressions Expressions that form the basis of controlled terminology; the building blocks of more complicated expressions of meaning

The basic unit of meaning in a controlled terminology is the concept. Concepts are unique and, although they may be replaced by other concepts, never change in meaning (Cimino 1998). A controlled terminology is typically focused on a subject domain or domain of practice. For example, a controlled terminology might be focused on health services and contain precoordinated concept expressions concerning healthcare services, providers, and programs, including “health professional,” “physician,” “primary care physician,” “hospital,” “outpatient clinic,” and “support group.” Precoordinated concept expressions form the basis of the terminology and are the building blocks of more complicated expressions of meaning. In a way, precoordinated concept expressions are like the atoms of the controlled terminology.

Preferred Concept Expressions Typically, several concept expressions are associated with a concept, with one expression selected as the preferred expression for the concept and the remaining expressions treated as synonyms for the preferred expression. For example, for the concept primary care physician, the preferred expression might be “primary care physician” and a synonym for the preferred expression might be “PCP.”

Semantic Relationships Semantic relationships may be hierarchical or nonhierarchical. Examples of a hierarchical relationship are is-a and is a subtype of. Examples of a nonhierarchical relationship are is clinically associated with and sponsors. Consider the hypothetical example of a controlled terminology for healthcare services. In such a terminology, the concept physician might stand in the is-a relationship to the concept health professional. Similarly, the concept hospital might stand in the sponsors relationship to the concept support group. Exhibit 12.1 depicts these types of relationships.

Alphanumeric Identifying Codes Typically, concepts, precoordinated concept expressions, and semantic (hierarchical or nonhierarchical) relationships are associated with alphanumeric codes of some sort. For example, the concept healthcare provider might be associated with the alphanumeric code C0001, and the concept expression “healthcare provider” might be associated with the alphanumeric code T4567. A key consideration in the design of the codes is whether the codes are pure

Cha p t er 12 :  Co n t ro l l e d Te r mi n o l o g y a n d th e Rep resentation of D ata and Infor m ation

Healthcare professional

Works in (associative) Hospital

255

EXHIBIT 12.1 Example of Semantic Relationships

Is-a (hierarchical)

Physician

identifiers of their associated concepts and expressions or whether they carry other information, such as information about hierarchical relationships among concepts. For example, if the concept physician stands in the is-a relationship to the concept healthcare provider, we might assign hierarchical codes to those concepts—C0001 to healthcare provider and C0001.1 to physician. Computer processing of codes in such a scheme has both advantages and considerable disadvantages, particularly regarding whether a concept may participate in multiple hierarchies (Cimino 1998).

Postcoordinated Concept Expressions As mentioned, the precoordinated concept expressions form the basis of the controlled terminology, serving as the building blocks of more complicated expressions of meaning. Precoordinated concept expressions are combined with semantic relationships to form postcoordinated concept expressions. For example, the precoordinated concept expression “obesity” might be combined with the precoordinated concept expression “heart disease” using the semantic relationship name “clinically associated with” to form the following postcoordinated concept expression: “Obesity clinically associated with heart disease.” Some controlled terminologies include, in addition to the rules for combining concept expressions and relations, specific grammar rules or syntax for creating those expressions. The controlled terminology SNOMED CT (Systematized Nomenclature of Medicine Clinical Terms), maintained by the International Health Terminology Standards Development Organisation (www.snomed.org) is an example of a terminology that has both semantic rules for combining concepts and relations and grammar rules for creating

Postcoordinated concept expressions Expressions that combine precoordinated concepts and semantic relationships

256

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

postcoordinated expressions of them. For example, clinical finding concepts may be combined with the attribute “due to” and an event concept to express a postcoordinated meaning. Thus, the clinical finding concept bruise of the head could be combined with due to and the event concept accident while engaged in sports activity. Specific grammar rules could then be used to form the following postcoordinated concept expression: 262528003|bruise of the head|:42752001|due to|=57701003| accident while engaged in sports activity|

Rules for Applying the Terminology Pragmatics, a subfield of linguistics, considers the aspects of language that take into account the context and goals of the language users. Think of the consideration of the rules for applying a terminology as a kind of pragmatics component of controlled terminology. Medical Subject Headings (MeSH) is the controlled terminology used by the National Library of Medicine (2016b, 2018b) to index biomedical literature in the PubMed system. One example of a rule for applying a terminology is the following annotation from MeSH for the concept “meta-analysis” (National Library of Medicine 2018a): This heading is used as a Publication Type; for original report of the conduct or results of a specific meta-analysis study; a different heading META-ANALYSIS AS TOPIC is used for general design, methodology, economics, etc. of meta-analyses.

International Classification of Diseases, 10th Revision, Clinical Modification (ICD-10-CM) is the clinical modification of ICD-10 used to classify morbidity (National Center for Health Statistics 2018; World Health Organization 2016). Following is another example of a rule for applying a terminology taken from the rules for using the ICD-10-CM code: I21 ST elevation (STEMI) and non-ST elevation (NSTEMI) myocardial infarction. The rules specify this: Use additional code, if applicable, to identify the following: • • • •

Exposure to environmental tobacco smoke (Z77.22) History of tobacco use (Z87.891) Occupational exposure to environmental tobacco smoke (Z57.31) Status post-administration of tPA (rtPA) in a different facility within the last 24 hours prior to admission to current facility (Z92.82) • Tobacco dependence (F17) • Tobacco use (Z72.0)

Cha p t er 12 :  Co n t ro l l e d Te r mi n o l o g y a n d th e Rep resentation of D ata and Infor m ation

257

Two Basic Uses of Controlled Terminology First, controlled terminology may be used to construct surrogate representations (classification or representation of existing information objects) of biomedical literature—for example, to represent the content of reports on clinical studies. As stated, concepts from the controlled terminology MeSH are used in PubMed to describe the subjects covered by biomedical journal articles. The article “Is sudden cardiac arrest the same as a heart attack? Johns Hopkins Med Lett Health After 50. 2011 Apr;23(2):8” is represented in PubMed by the record shown in exhibit 12.2. In exhibit 12.2, the fields in boldface are the MeSH terms (MH) used to indicate the subject or topic of the article. In this case, one of the subjects or topics of the article is described using the MeSH term “myocardial infarction.”

MID - 21523953 OWN - NLM STAT - MEDLINE DA - 20110329 DCOM - 20110524 IS - 1042-1882 (Print) IS - 1042-1882 (Linking) VI - 23 IP - 2 DP - 2011 Apr TI - Is sudden cardiac arrest the same as a heart attack? PG - 8 LA - eng PT - Journal Article PL - United States TA - Johns Hopkins Med Lett Health After 50 JT - The Johns Hopkins medical letter health after 50 JID - 9802902 SB - K MH - *Death, Sudden, Cardiac MH - *Health Knowledge, Attitudes, Practice MH - Humans MH - *Myocardial Infarction MH - Patient Education as Topic MH - Risk Factors EDAT - 2011/04/29 06:00 MHDA - 2011/05/25 06:00 CRDT - 2011/04/29 06:00 PST - ppublish SO - Johns Hopkins Med Lett Health After 50. 2011 Apr;23(2):8.

EXHIBIT 12.2 Sample PubMed Record

258

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Second, controlled terminology may be used to express, in a highly structured and controlled way, new information. For example, codes from the controlled terminology SNOMED CT may be used to express facts about a particular patient in a clinical information system. A physician might enter into the electronic medical record (EMR) a diagnosis for a patient using the SNOMED CT preferred concept description “myocardial infarction” using the SNOMED CT concept code 22298006. These two uses of a controlled terminology may work together. For example, corresponding to the case just described, another controlled terminology could be used to classify or otherwise represent (in a highly structured way) that new information. Thus, the diagnostic information entered into the EMR by the physician using the SNOMED CT concept “myocardial infarction” could be classified by the ICD-10-CM code “cardiac infarction.” In this case, the information recorded by the physician about the patient using the SNOMED CT concept could be considered a coding for primary use, whereas the classification or representation of that information using the ICD-10-CM code could be considered a coding for secondary use. The primary use could be clinical care or decision making, and the secondary use could be quality or safety measurement. Multiple controlled terminologies are used in healthcare because many were created for different purposes or for different contexts and cultures, and there was no compelling reason (or ability) to develop them as a single integrated system. The need to exchange information among healthcare providers and organizations justifies the development of an integrated terminology, because heterogeneity across terminologies may impede interoperability and health information exchange. Thus, bridging the semantic and lexical spaces among differing terminologies and for different purposes is an ongoing challenge in health informatics.

Metadata and Metadata Schemata

Metadata Data about data and information

The distinction between the two basic uses of controlled terminology is actually not sharp and clear. After all, from a certain point of view, data and information are just elements (e.g., patients, physicians, hospitals, diseases) that may be described and about which information may be recorded. Typically, however, metadata refers to data about data and information and, as such, has a natural affinity with the use of controlled terminology to construct surrogate representations of extant information objects. A controlled terminology may be used to record information about a patient or a journal article (primary information), whereas another controlled terminology may be used to describe facts about that information (secondary information or metadata), and further data (whether or

Cha p t er 12 :  Co n t ro l l e d Te r mi n o l o g y a n d th e Rep resentation of D ata and Infor m ation

not from a controlled terminology) may be used to describe facts about the metadata (metametadata). Such hierarchies of data and metadata are reminiscent of this amusing poem from the nineteenth century (Sharp 2011):

259

Opportunity for Interprofessional Education: Data Are in the Eye of the Beholder It might be true that “a rose is a rose is a rose,” but it’s not true that “a datum is a datum is a datum.” What constitutes valuable and important data is determined by conscious or unconscious judgment. Unilateral decisions as to what does or does not count as important and valuable data, rather than decisions about such matters made by an integrated team, may have significant unintended and negative

Big bugs have little bugs Upon their backs to bite them.

consequences. Many years ago, the school of medicine of a major Midwestern

Little bugs have littler bugs.

university built a center for health services and informatics research.

And so, ad infinitum.

The center maintained a large collection of books, journal articles, and other materials. The collection was represented by a catalog,

Data and metadata hierarchies, and various metadata were included in each catalog record, such as however, are typically less drathe item’s title, year of publication, and author. One of the fields in matic than the imagery of this the catalog record for each item named and described the research poem. project or study associated with the item—essentially, the reason The sample PubMed the item was included in the collection. Thus, for example, if the record in exhibit 12.2 demitem was a book on technology forecasting, this field in the catalog onstrates a modest hierarchy record might have explained that the item was part of a study on of data and metadata. The incorporating information technology into future medical education. journal article is the primary In effect, that field in the record contained the provenance of each object of interest; the MeSH item in the collection. term “myocardial infarction” Over time, the leadership of the center changed, as did is the metadata about that other personnel. In addition, as software and computer technology journal article; and the definiimproved, so did the information systems used by the center. At tion of the PubMed field name one point, the software for the catalog system was upgraded, which MH is the meta-metadata that required transferring the catalog records from the old system to the describes it and the use, in that new system. The technician in charge of this transfer made a fateful context, of the term “myocardecision—perhaps to save storage space in the new system—to dial infarction.” discard the field that described why each item was included in the Although metadata is collection. This field did not, in the technician’s mind, represent typically used to refer to data important and valuable data. With that action, an important record about data and information, of the intellectual history of the research center was needlessly and in common parlance metadata forever lost. are often considered to be data Do you think this decision could have been made by a team? about anything else—a patient, Why or why not? Who would be the members of this team? a book, a healthcare organization, or the like. However its use is restricted, metadata are usually organized in the form of a metadata

260

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Metadata schema Collection of fields or data elements, names for the fields, definitions for the fields, and specifications of what constitute permissible values for the fields; also known as a data dictionary

schema. A metadata schema is a collection of fields or data elements, names for the fields, definitions for the fields, and specifications of what constitute permissible values for the fields. In this sense, the fields included in a PubMed record (such as that in exhibit 12.2), together with specifications of the values allowed in those fields, constitute a metadata schema. The field “MH” and the specification that it has codes from MeSH as its value are part of that schema. Dublin Core Metadata Initiative (2018) is a widely used metadata schema. This general description of a metadata schema also applies, more or less, to the specification of fields in a structured patient record—for example, patient name, patient gender, patient age, and patient diagnosis. As mentioned earlier, the distinction between data and metadata (or between the primary use and secondary use of controlled terminology to record data) may blur somewhat. In the case of the schema for the structured patient record, the term information model is sometimes used rather than metadata schema. Two important web-based registries for healthcare data elements and their permissible values, which include fields such as those suggested for the structured patient record, are the United States Health Information Knowledgebase (USHIK) and the Cancer Data Standards Registry and Repository (caDSR) (Agency for Healthcare Research and Quality 2018; National Cancer Institute 2015). Both USHIK and caDSR are based on the international standard ISO 11179 (Information Technology—Metadata Registries [MDR]) (ISO/IEC 2015).

Common Data Elements and Repositories Common data element (CDE) Pair consisting of a variable and a set or domain of permissible values for the variable; used by different studies or healthcare settings

Common data element (CDE) repositories have garnered a lot of interest. A CDE is a pair consisting of (1) a data element name and definition, and (2) the specification of a permissible value domain for the data element. The permissible value domain, as its name suggests, is the set of values that are permitted for the data element. The permissible value domain for a data element may be of two types: (1) an enumerated value domain consisting of a list of values or (2) a nonenumerated value domain consisting of a range of values that are either discrete or continuous. The CDE shown in exhibit 12.3 is from the caDSR and has an enumerated value domain. CDE repositories are not new, as both USHIK and caDSR may be considered CDE repositories to some extent. However, there has been a push to standardize CDEs and to develop repositories for sharing CDEs among care providers and researchers. As the National Library of Medicine (2016a) explains, NIH encourages the use of common data elements (CDEs) in clinical research, patient registries, and other human subject research in order to improve data quality and opportunities for comparison and combination of data from multiple studies and with electronic health records.

Cha p t er 12 :  Co n t ro l l e d Te r mi n o l o g y a n d th e Rep resentation of D ata and Infor m ation

Data element: Tissue donor death status type Definition: Type of death the organ donor experienced Permissible values: Brain death (BD) Donation after cardiac death

Interoperability and the Terminology Problem As mentioned earlier, healthcare providers and organizations need to exchange information but heterogeneity across terminologies may impede interoperability. Thus, a significant terminology problem for health informatics is bridging the semantic and lexical spaces between differing terminologies to support interoperability. The terminology problem may be summarized as follows: 1. Different systems, individuals, and groups collect or produce data, information, and knowledge according to their needs and then represent and store the data according to their local preferences. 2. Other things being equal, the value of the data, information, and knowledge is increased to the extent that the data are available to and usable by others. 3. No given system, individual, or group is conversant with the local representation and storage preferences of every other system, individual, or group. 4. Therefore, some means for translating or mapping must be provided among different, locally preferred schemes for representing and storing data, information, and knowledge. The use of generally accepted and well-documented controlled terminologies by different systems, individuals, or groups may facilitate interoperability. Controlled terminologies have long been recognized as important to this end. For example, in 1963, the Groupe d’Etude sur l’Information Scientifique in Marseilles proposed a system of scientific information sharing based on the idea of an intermediate lexicon (Coates, Lloyd, and Simandl 1979). The basic idea was that a group of scientific study centers would share information (e.g., scientific reports) by using a standard switching language or intermediate lexicon. Each study center would use a controlled terminology, chosen according to local preference, to construct surrogate representations

261

EXHIBIT 12.3 A Sample Common Data Element

262

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

of its information. The information contained in a given document or report would be represented by terms in the controlled terminology. For example, if the document was about possible links between obesity and heart disease and the controlled terminology used by one study center was SNOMED CT, the terms used might be “obesity” and “heart disease.” A document with similar content maintained by another study center might use the controlled terminology ICD-10-CM to represent the document and the terms “obesity” and “heart disease, unspecified.” Interoperability between the two study centers would be achieved by using an intermediate language standard that would relate terms from SNOMED CT to terms in ICD-10-CM that are considered to express the same concepts. Using an intermediate switching language, a given study center can use its customary local terminology to search and acquire information from the collection of another study center, even though that information is represented by a different controlled terminology. The key to this scheme, aside from the intermediate lexicon or switching language itself, is that each study center uses a controlled terminology to represent the information in its collection.

Metathesaurus

Terminology mapping Process of matching a term from one controlled terminology (or natural language) to a term in another controlled terminology

A contemporary example of the switching language solution to the terminology problem is the Unified Medical Language System metathesaurus (National Library of Medicine 2009). The metathesaurus is a large database that relates terms from more than a hundred controlled vocabularies, terminologies, indexing languages, and coding systems, including ICD-10-CM, MeSH, and SNOMED CT. Collectively, these are referred to in the metathesaurus documentation as source vocabularies. The metathesaurus is organized by meaning, and terms from different source vocabularies are linked by metaconcepts. Two terms are assigned to the same metaconcept when those terms are considered to express the same concept. Thus, the metathesaurus provides a means of translating the representation of data and information based on one source vocabulary to a representation based on another source vocabulary. For example, the metathesaurus assigns both the SNOMED CT term “heart disease” and the ICD-10-CM term “heart disease, unspecified” to the metaconcept “C0018799.” Using the term “metaconcept assignments,” we can translate the ICD-10-CM term “heart disease, unspecified” to the SNOMED CT term “heart disease.”

Terminology Mapping Typically, terminology mapping, which supports interoperability of data and information, is a scheme for matching concept expressions among controlled terminologies by meaning. Simply put, it finds an expression in one terminology

Cha p t er 12 :  Co n t ro l l e d Te r mi n o l o g y a n d th e Rep resentation of D ata and Infor m ation

that has the same or nearly the same meaning as an expression in another terminology (Fung and Bodenreider 2005). However, it is possible and may be useful to have mappings between terminologies that are not intended to preserve meaning but are instead based on some sort of associative relationship. Types of knowledge for relating information from diverse information resources may include cross-domain integrative knowledge, such as the knowledge that the best way to interpret information on one subject is to relate it to information on some other subjects. This sort of knowledge may, for example, drive data mining using heterogeneous and cross-disciplinary data sources. For example, a case of cross-domain integrative knowledge might be the knowledge that, to understand prevention options for a particular human disease and to further prevention and wellness programs, it may be necessary to relate information about certain animal habitats to information about certain human habitats. This might necessitate retrieving information from different sources in a coordinated manner—for example, from the bibliographic databases PubMed and ABI/INFORM (ProQuest 2017). Consequently, this coordinated retrieval may require an associative mapping between the respective controlled terminologies—between the PubMed terminology MeSH and the ABI/INFORM terminology. For example, if we know that controlling mosquitoes may help prevent the spread of West Nile virus, that one method of mosquito management is recycling abandoned tires so that they do not provide mosquito breeding grounds, and that government regulations may promote appropriate tire recycling (policy issues covered by ABI/INFORM), then we may retrieve information from both PubMed and ABI/INFORM based on an associative mapping between MeSH and the ABI/INFORM terminology: mosquito control (MeSH) → tires AND recycling (ABI/INFORM).

Metadata Crosswalks A concept related to terminology mapping is the metadata crosswalk. A metadata crosswalk is essentially a mapping between the fields, or data elements, included in two different metadata schemata. A crosswalk is more than a mapping between the controlled terminologies that provide the name of the fields in the respective schemata; the permissible values for the fields and their meanings must also be considered. Numerous metadata crosswalks have been published (University of Texas Libraries 2017).

Quality Measurement, Variation, and Coding Accurate measurement of healthcare quality is a fundamental goal of science and evidence-based healthcare. Quality must be considered from the points of view of process and outcomes, both within a given healthcare institution and

263

264

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

at the system level across institutions. Quality measurement must take into account whether variation is acceptable or unacceptable. Furthermore, it must be determined whether variation found in data and information is indicative of variation in actual healthcare practice and delivery or the result of variation in the way that practice and delivery are represented. For example, if billing code data appear to indicate unwanted variation in the treatment of patients with the same diagnosis, it must be determined whether that variation is representative of healthcare practice or the result of variation in coding practice among billing coders in the institution. Thus, the study of coding variation is a core discipline of health informatics.

Conclusion Concept-based controlled terminologies are used to express biomedical and health information and to serve as building blocks of information. These terminologies are, in fact, key to effective information interoperability, information exchange, and the related goals of modern healthcare. Effective use of controlled terminology, however, requires recognition of the terminology problem and the development of standardized expressions of information. Only by addressing the need for controlled terminology, as well as the emergent heterogeneity it produces, can we ensure integration and interoperability of high-quality data and information across systems, people, and organizations.

Chapter Discussion Questions 1. Evidence-based medicine might be defined as the appropriate application of the best available evidence to determine diagnosis and treatment for patients. Should the controlled terminology used in knowledge representation be evidence based? What would that evidence be like? 2. Data, information, and knowledge resources are sometimes characterized by describing their inputs and outputs. Pick one resource with which you are familiar and describe its inputs and outputs. Describe the requirements for a controlled vocabulary to classify the inputs and outputs. 3. Data for public health work are often collected using structured forms such as the Acute and Communicable Disease Case Report form used by the Wisconsin Department of Health Services (www.dhs.wisconsin. gov/forms/F4/F44151.pdf). Would it be useful to use a controlled terminology for that form? Would MeSH be a suitable candidate?

Cha p t er 12 :  Co n t ro l l e d Te r mi n o l o g y a n d th e Rep resentation of D ata and Infor m ation

Case Study  A Problem of Display Codes During the course of an electronic medical record (EMR) implementation project at a large medical center, the question arises how best to design display screens for the system. In particular, the concern is what standard codes should be used for the information on the user screens. In the following e-mail, an information technology staff member involved in the EMR project asks a member of the health informatics group for advice: Hi Carol, I need your help with something, and it’s a pretty big something. I wonder if you or anyone on your staff would be interested in taking this on as a project? As you know, I’m working on the new EMR implementation. One of the things we’re stuck on right now is standard displays of information in the EMR. Because the record is integrated (which is the good news and the bad news), most of the tables and code sets are shared among multiple disciplines. We need to define what we will use for our displays. For example, the code set we’re working on right now is Units of Measure. Now, you wouldn’t think it would be too hard to decide what the display for something like “milligrams” would be—except that the choices are MG, Mg, mg, etc. And that’s one of the easy ones. According to the medical center’s standard abbreviations, all of the above are perfectly acceptable. The problem is, to build the EMR, we have to decide on just one. So far, since we began implementing the EMR, we’ve included a mishmash of displays depending on what department we were working with. Lo and behold, now we’re in a real pickle, and our database is a mess. In the truly integrated EMR, only one abbreviation can be displayed for “milligrams.” Again, this is just one example out of literally hundreds (if not thousands) of pieces of information we need to standardize. Do you have some time to help us figure out what standards are out there? We’d like to standardize on something that is nationally, if not internationally, accepted. It needs to include content for standards, definitions, and especially abbreviations for pharmacy, medicine, nursing, and purchasing units. We’d like to present some standards to the EMR Steering Committee and the Medical Records Committee for consideration. Do you have any suggestions? I’d be happy to meet with you and discuss further if you’d like. Thanks! Michael (continued)

265

266

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Case Study Discussion Questions 1. What is the problem that needs to be addressed? 2. What solution is being considered as viable? 3. How would you define (or not define) the problem differently, or define a viable solution differently? 4. What do you think is the correct solution?

Additional Resource Agency for Healthcare Research and Quality, National Guideline Clearinghouse: www. guideline.gov.

References Agency for Healthcare Research and Quality. 2018. “United States Health Information Knowledgebase.” Accessed January 23. http://ushik.ahrq.gov. Cimino, J. J. 1998. “Desiderata for Controlled Medical Vocabularies in the TwentyFirst Century.” Methods of Information in Medicine 37 (4–5): 394–403. Coates, E., G. Lloyd, and D. Simandl. 1979. The BSO Manual: The Development, Rationale, and Use of the Broad System of Ordering. The Hague, Netherlands: Fédération Internationale de Documentation. Dublin Core Metadata Initiative. 2018. “About the Dublin Core Metadata Initiative.” Accessed January 23. http://dublincore.org/about/. Fung, K. W., and O. Bodenreider. 2005. “Utilizing the UMLS for Semantic Mapping Between Terminologies.” In AMIA Annual Symposium Proceedings, 266–70. Bethesda, MD: American Medical Informatics Association. International Organization for Standardization and the International Electrotechnical Commission (ISO/IEC). 2015. “ISO/IEC 11179, Information Technology—Metadata Registries (MDR).” Updated November 5. http://metadatastandards.org/11179/. Jones, J. H. 1982. Bad Blood. New York: The Free Press. National Cancer Institute. 2015. “Metadata and Models.” Cancer Data Standards Registry and Repository (caDSR). Accessed January 23, 2018. https://cbiit.cancer. gov/ncip/biomedical-informatics-resources/interoperability-and-semantics/ metadata-and-models#caDSR. National Center for Health Statistics 2018. “International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM).” Accessed January 23. www.cdc.gov/nchs/icd/icd10cm.htm.

Cha p t er 12 :  Co n t ro l l e d Te r mi n o l o g y a n d th e Rep resentation of D ata and Infor m ation

National Library of Medicine. 2018a. “Medical Subject Heading.” Accessed January 23. https://meshb.nlm.nih.gov/search. ———. 2018b. “PubMed.” Accessed January 23. www.ncbi.nlm.nih.gov/pubmed/. ———. 2016a. “Common Data Element (CDE) Resource Portal.” Accessed January 23, 2018. www.nlm.nih.gov/cde/. ———. 2016b. “Introduction to MeSH.” Accessed January 23, 2018. www.nlm.nih. gov/mesh/introduction.html. ———. 2009. UMLS Reference Manual. Published September. www.ncbi.nlm.nih. gov/books/NBK9676/. ProQuest. 2017. “ABI/INFORM Collection.” Accessed January 1. www.proquest. com/products-services/abi_inform_complete.html. Sharp, J. W. 2011. “Desert Food Chain: The Insects, Part 12.” Accessed January 23, 2018. www.desertusa.com/mag06/jan/food12.html. University of Texas Libraries. 2017. “Metadata Basics: Crosswalks.” Updated March 16. https://guides.lib.utexas.edu/metadata-basics/crosswalks. World Health Organization. 2016. “ICD-10 Version:2016.” Accessed January 23, 2018. http://apps.who.int/classifications/icd10/browse/2016/en.

267

CHAPTER

INFORMATION MANAGEMENT STRATEGY

13

James D. Buntrock

Learning Objectives After reading this chapter, you should be able to do the following: • Name the key features of an information management strategy. • Describe the types of data in a healthcare organization. • Understand the relationship between operating model and information strategy. • Explain how information management strategy can guide investments and decisions.

Key Concepts • • • • • •

Data principles Information facets Reference data architecture Data governance Master data and reference data Binding

Introduction An information management strategy is an organization’s plan to acquire, manage, use, and deliver information through products and services to internal and external customers. Information management is a poorly recognized topic in many healthcare organizations, both big and small. When asked about an information management strategy, healthcare leaders may offer responses such as “We are using Cerner for our electronic medical record,” “We use a single medical record number for all our patients,” and “We use Oracle databases to manage our data.” Although these responses reflect aspects of an information

Information management strategy Plan to acquire, manage, use, and deliver information through products or services to internal or external customers

269

270

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

management strategy, they represent an incomplete and narrow understanding of this strategy. An information management strategy is often considered a component of an information technology (IT) strategy—an organization’s broad, long-term plan for its IT system and other use of technology. Healthcare administrators and senior leaders are increasingly confronted with health information technology (HIT) investment choices that provide new capabilities for care delivery, business (e.g., revenue cycle), and regulatory compliance and reporting (e.g., from the Centers for Medicare & Medicaid Services [CMS]) processes. They also face myriad IT-related decisions, such as replacing an antiquated system and purchasing a new electronic solution for a clinical service. Often, they think of an IT investment in terms of its functionality or ability to perform a task, but rarely do they consider its value or the value of data or information being collected. This chapter explores the many aspects of information management strategy to help healthcare leaders appreciate the value of their IT investments from a data and information perspective. Many healthcare organizations have already upgraded to an electronic health record (EHR) system, and its use is at an all-time high thanks to the mandates and incentives in the Health Information Technology for Economic and Clinical Health Act and the Federal Health IT Strategic Plan (Office of the National Coordinator for Health Information Technology 2015; US Department of Health and Human Services 2017). As organizations become more efficient and proficient with EHRs, they will naturally turn their attention to other data and information needs. Without a formal information management strategy, organizations could make IT investment decisions that do not further their goals or that lead to costly, mistake-riddled implementation. An information management strategy specifies the organization’s key data assets and their value, as well as the organization’s data usage.

Data as Assets Shouldn’t healthcare organizations focus on healthcare delivery rather than HIT? Why do organizations need to think about data? Why does an information management strategy matter? Healthcare delivery is the main focus of healthcare organizations, but data and information are their assets. Whether we realize it or not, healthcare is data intensive. At the most basic level, physicians and other health professionals do two things: perform clinical procedures and manage clinical information. The amount of information that needs to be collected and stored—from medical condition, history, and maintenance to procedures, billing, payment, test results, and care plans—is immense and continues to grow. In addition, the

C h a p te r 13:   Infor m ation Managem ent Strategy

amount of medical knowledge is expected to double every 73 days (Densen 2011). Clearly, health professionals use and process information in healthcare delivery. The first step in developing an information management strategy, therefore, is to acknowledge that data are an asset and must be treated as such. Assets have value, are maintained, and are kept operational. They need to be replaced and upgraded as their useful life is met. Recognizing this fundamental tenet is critical to an effective information management strategy.

Quantifying Value The value of data may be hard to quantify. The patient record has value, specifically from a security perspective (e.g., the value of a person’s identity on the black market). Fines and mitigations are imposed in the event of Health Insurance Portability and Accountability Act (HIPAA) privacy violation for improper disclosure of protected health information (PHI). In addition, managing or collecting data cannot just be altruistic (i.e., keeping data for the greater good of human kind). The value of data is more direct. Having the right data to perform a function, using data to improve a clinical function or process, or combining data to gain new insights are all part of the value proposition. The value of data increases when linked or combined with other data. Often, data may be used in a primary function—for example, maintaining the patient’s current residential address for correspondence and billing purposes. However, the broader use of this information may not be recognized by the primary function but is incredibly important for secondary uses. Maintaining an address history is valuable in epidemiology for understanding the geographic aspects of disease or continuity-of-care patterns based on where patients have lived. The demand for data is both a blessing and a curse. As healthcare consumerism grows, patients expect and want to share more information about their health status or medical condition. They also want information by which to measure their own dietary intake, exercise routines, vitals, sleep patterns, and other health-related “performance,” which enables a state of being known as the quantified self. Countless consumer-grade devices and wearables are available for monitoring one’s functioning and guiding health behaviors and performance. Direct-to-consumer genetic testing kits provide not only ancestry information but also potential disease risk data, which often cause patients to turn to their physicians for interpretation and explanation. On the patient care side, more data exist—whether documented in electronic or paper records and flowsheets, supplied by patients, or automatically collected through physiological monitoring and diagnostic procedures. All of these data increase the expectation that physicians and other care providers are using them to inform and improve their decision making.

271

272

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Understanding Informatics and Information Informatics is a simple word that is fraught with many meanings. In its basic form, it means people plus process plus technology. This definition is informative but lacks any aspect of information or data. Thus, informatics could be viewed as people plus process plus technology plus information, recognizing the role of data in health services delivery. Often, technology (the shiny object) distracts us from the people, process, and information aspects, but remember that technology is ever-changing and so is its value, whereas people, process, and information were important in the paper records era and will continue to be important in the digital era and beyond. Information is universal, unlimited, timeless, and even omnipotent. How it is used by people in processes (e.g., clinical workflows) and decision making (e.g., medication orders) is made possible by informatics. Informatics (technology) enables people to use the information or data in the right context (process) at the right time.

Determining the Data Facets Identifying the current state of data, the maturity of data, and other data facets is a key component of an information management strategy (and is discussed later). Organizations use data in different ways. For example, a physician practice that recently adopted an electronic medical record (EMR) may use the system to generate patient bills, whereas a hospital may use it to integrate patient and business data from its inpatient and outpatient units. The predominant use for an EMR is to support day-to-day patient care activities. Understanding today’s data environment is a related consideration. As the healthcare industry becomes more digitized, second-generation software and HIT companies are now flooding the market. Choosing among these products is not just about technology but also about financial value. Which solutions and vendors can provide the capability and functionality that the organization needs? The answer must be determined before money and other resources are invested. (See the sidebar for an example.) An information management strategy encourages a disciplined approach to understanding the current data state, data purpose, and data solutions and vendors and then fitting the details together to serve the organization for both the short and long term.

Steps in Strategy Development An organization’s information management strategy must be developed in the context of its business and IT strategies. An inventory of the major IT systems used is also important to include as input to understanding the current state. Next, formalizing information constructs and principles helps advise the data

C h a p te r 13:   Infor m ation Managem ent Strategy

273

architecture. Exhibit 13.1 shows the process for developing an information management strategy.

Understand the Current State During this step, the most important task is to determine the organization’s business strategy. Over the years, healthcare has continued to evolve through new types of healthcare delivery, mergers and acquisitions, and extended healthcare models for home health care. The size of healthcare organizations ranges from small group practices to mid-sized specialty clinics to large academic medical centers. Several models can be used to define the current state. Which model to apply is not important, but the process is necesPredicting Volumes in the Emergency Department sary to structure and define the Lakes Central Regional Hospital, located in a large metropolitan area, business. provides emergency medicine services. The emergency department The operating model (ED) administrator has noted wide fluctuations in patient demand, (how an organization delivresulting in long ED wait times and low patient satisfaction. Currently, ers value to its customers) can the ED department uses a software package for managing patients be measured against the level through the workflow from admission to discharge; the hospital itself of standardization or integrauses an EHR. In conversations with peers at a recent conference, the tion in the organization. As administrator learns about a new software that predicts ED demand integration and standardizaand begins to read more about it, including its pilot results. The tion increase, business and IT administrator’s proposal to buy the software is approved, but during decisions will become more contract negotiations with the vendor, the administrator learns that coordinated and centralized. the software is offered only as a software-as-a-service (SaaS). This Products, services, and cusarrangement requires the ED department to provide data several tomers will begin to share comtimes per day to the external system. mon data and use common systems. The organization should Questions determine whether it will be 1. What are some questions related to ED data and information changing its operating model that the administrator should discuss with the vendor? and adapting new strategies, 2. What would be the best way to evaluate the (nonfinancial) with the understanding that return on investment from this software? no one model is better than 3. What terms or conditions in the license agreement or another. The current model software contract should the administrator consider? often reflects the organization’s primary goals. NationDiscussion ally, with greater incentives for Many vendors want to sell the solution first—in this case, the SaaS— achieving meaningful use and and then figure out what it can do later. Thus, the ED administrastandards and incentives for tor, with guidance from the vendor, must determine what data are sharing data electronically, the required to make the SaaS perform optimally. Are the data available US healthcare industry is natuelectronically to be delivered to the SaaS? If not, the ED will likely rally increasing its integration (continued) and standardization. This trend

274

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

(continued from previous page)

need to implement new workflows or data capture methods, all of which require changes to clinical practice and clinical processes. Other questions the administrator should consider include the following: How does the SaaS protect the data submitted? Does the vendor provide any security attestation for the protection of the data? Does the vendor have secondary or derivative rights to use the data for other purposes? (This question may seem like a minor detail, but data have intrinsic value and software vendors may derive new value through improved products or new product capabilities.) How does the ED access the software output? Are the prediction results integrated into the current operations? How do users authenticate access to the software? (Again, these are simple but important ques-

is also reflected in the adoption of data standards for coding, exchange, and transactions that enable better data sharing and portability across organizations. Even though a national patient identifier is not used, a National Provider Identifier, mandated implementation of new diagnosis codes (e.g., International Statistical Classification of Diseases, 10th Revision), and other data representation standards have been adopted.

tions that are often overlooked. Not understanding the basic data requirements and data usage leads to integration or usability issues in the clinical workflow.) Does any of the SaaS functionality already

Determine Information Facets

Fundamental information management facets are often product road map that may include this capability? (These last two described in the strategy. Even questions touch on the redundancy of the SaaS functionality and its when operations are distribcost to implement.) Systems and data interactions become complex uted and governance is federquickly, but an information management strategy creates some order ated, these facets still apply and and transparency for these interactions. influence the type of work that has to be performed. Master data are common business data intended to be shared across the organization. In healthcare, several types of master data exist, including patient, patient care provider (employee), location, and service. Master data drive the linking of data across the organization. Without master data, sharing exist in the current ED software? Does the current software have a

EXHIBIT 13.1 Process for Developing an Information Management Strategy

Determine information facets Business strategy IT strategy IT systems Inputs

Create reference data architecture

Understand current state Determine data principles

Information management strategy Outputs

C h a p te r 13:   Infor m ation Managem ent Strategy

any kind of data becomes difficult. The master data’s maturity, quality, and role in clinical operations are basic requirements to be described in the current state. Does your organization use a common medical record number/master patient index? Where are clinical services delivered (city, address, building, or floor)? What clinical services are performed? Addressing these questions is the start of the information inventory. Systems of record describe the organization’s information systems used for collecting core data—data that support the different operations necessary to run the business, such as financial accounting, human resources, supply chain, EMR, and revenue cycle. Systems of record are often referred to as the gold source of data—where data are created, maintained, and managed—and organizations usually have more than one such system. Small organizations may have one or two key systems of record, whereas large organizations may have multiple systems of record or multiple instances (two or more systems running the same software) of systems of record for a geographic region or site. Operational data are used to run the organization. Security logs, access logs, and audit logs are all examples of this kind of information. Operational data may be required for regulatory purposes or used for improving operations. They can be voluminous and may require organizations to develop data retention policies if not already stated by regulations. They also can easily grow to many times the size of core data. Reference data provide a set of permissible values that can be used in business operations. Although reference data are different from master data, the two are often confused with each other. Both types have value, and their use is important. Common types of reference data include billing code descriptions, laboratory test catalog codes and descriptions, and clinical problems and diagnoses. Reference data used consistently throughout the organization allow core data to be shared and more easily integrated. Third-party data are used or obtained from outside the organization. This type of data may be related to payer information, socioeconomic status, or pharmacy benefit management. Third-party data can be in the public domain (see http://catalog.data.gov) or licensed from another organization, and they are often used to enrich and gain additional insights from core data. These data need to be managed carefully because data use agreements can be easily breached if broader access is given to users when it is not allowed. Benchmark data are used for comparison purposes and are often provided in aggregate or by key measures. They may be obtained by organizations through participation in a quality program. The American College of Surgeons (2018) National Surgical Quality Improvement Program is a well-known program that provides data collection guidelines, measures, and benchmarking for participating organizations. Exhibit 13.2 depicts a sample map of all these data types.

275

276

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

EXHIBIT 13.2 Sample Data Map of an Organization

EMR

Lab

Patient

Provider

Thirdparty data

Benchmark data

ERP

Cardiology

Location

Services

Social data

Survey data

Revenue cycle

Radiology

Critical care

Supply chain

Finance

Other

Master data Lab catalog

FDB

ICD10

SNOMEDCT

External data Audit Logs

Access Logs

Operational data

Reference data

Systems of record

Note: EMR = electronic medical record; ERP = enterprise resource planning; FDB = First Databank; ICD10 = International Classification of Diseases, 10th Revision; SNOMED CT = Systematized Nomenclature of Medicine Clinical Terms.

The type of data to be included in the information management strategy depends on the goals of the organization. For example, if an organization has a goal to expand its patient base and integrate into a larger health system through mergers and acquisitions, then understanding master data and core data is paramount to enable the integration and sharing of data across organizations. If a specialty care–based organization is developing a network with primary care clinics for new referrals or more complex care, then the master data about patients and providers may be the most useful. As organizations increase the integration and standardization of business processes for clinical care and operations, the importance of shared data also increases. Shared data require common approaches to master data, reference data, and often systems of record (core data). Understanding the degree of operational convergence and where it is heading influences the type of information management strategy required to make the best use of data.

Determine Data Principles Establishing data principles is an important part of developing an information management strategy. Treating data as an asset is the right frame of reference when developing data principles. As a first principle, all data have value. Value must exceed the relevant level of investment to be worthwhile. Data collected and archived but never again accessed have limited value. Putting data into action for decision making increases the value, and so does linking data with other data. This principle is important to understand not only in one’s own organization but also in the broader healthcare ecosystem.

C h a p te r 13:   Infor m ation Managem ent Strategy

Integrating third-party data or aggregating shared data is increasing where the health ecosystem goes beyond the traditional healthcare delivery system. No matter its size, no single organization has enough data to further its value and advance its decision making. Awareness of this fact has grown and is reflected in the existence of national disease registries, multisite clinical trials, and quality measures and reporting across organizations. Data principles generally reflect fundamental truths. Organizations do not need to develop more than 10 or 20 data principles. Data policies or data standards further refine and operationalize data principles and will be greater in number. The following represent common data principles that inform the data architecture, policy, and standards of the organization: • Data are an asset. As an asset, data have a tangible value. Data need to be inventoried, managed, and maintained. Because assets have value, they need to be protected. Usage of data must be appropriate and in the best interest of the organization. • Data are governed. Data governance implies active ownership, management, and importance. Active ownership may be at the institution or department level and can be contentious. As part of data governance, defining the level or role of ownership is critical. Data ownership at the institution level provides broader flexibility in managing data. • Data are fluid. Data fluidity emphasizes movement and accessibility to data, for both primary and secondary uses. Accessibility should be provided through multiple, easy-to-use methods. Methods reflect the technical ways to access data for both transactional and analytical uses. • Data are mobile. Mobility is the ability to support the use, acquisition, and distribution of data regardless of where the data were generated. Data come from many different places, and more and more are provided directly by patients (e.g., from wearable devices or remote monitoring). Patients expect data related to their care to be accessible through multiple channels. • Data are current. Currency defines the timeliness of data and varies depending on the purpose for use. Direct patient care needs real-time data for clinical decision making. Clinical analytics requires timely access to data to support better business decisions. Monthly financial statements require weekly or monthly accessibility. • Data are categorized and inventoried. Data inventory and classification enable the organizational assets to be visible. A robust inventory greatly improves this visibility and decision making because it reduces the chances of redundancy and duplication of data assets. Categorization

277

278

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

helps manage the different types of assets and shows the types of security controls that may be required. For example, reference data have different data access controls than do core data (PHI). • Data are interoperable. Interoperability provides a predefined level of syntax and semantics that improves consistency and comparability of data from different systems. Often, interoperability is accomplished through the use of external data standards (e.g., Fast Healthcare Interoperability Resources) or internal standards (e.g., data interface). Interoperability can be improved by using reference data in the form of coding standards. For example, use LOINC (Logical Observation Identifiers Names and Codes) for mapping laboratory observations or SNOMED CT (Systematized Nomenclature of Medicine Clinical Terms) for mapping clinical problems. • Data are secure. Security is the provision of the right level of access to the right person at the right time. Data security requirements have changed dramatically over the past ten years because of regulations and laws (e.g., HIPAA Privacy Rules) and external security threats (e.g., ransomware). Often, least-privilege access is expected, whereby the minimum appropriate access to data is given to enable the performance of a job function. Organizations need to categorize data as the basis of security classification (Open Group 2018) if data are considered public, protected, and special access. This categorization serves as the type of controls required to adequately protect data. • Data are designed. Data design fundamentally formalizes the data architecture. Data architecture is the design of the data and data relationships across the organization.

Create Reference Data Architecture Understanding how to construct and manage data is the next step. Many types of architecture are used in IT design. The Open Group Architecture Framework (TOGAF) is modeled on business architecture, application architecture, data architecture, or technology architecture (Open Group 2018). An enterprise architecture / business architecture can provide enormous value in formalizing an information management strategy and data architecture. Business architecture defines the business strategy and business capabilities of the organization, improving the recognition and connection to data assets. A reference architecture provides a blueprint that can describe both a current state and a future state. An easy way to think about reference data architecture is to imagine building a city from an empty piece of land (think SimCity). Having forethought of the city’s design allows you to make important decisions about how you will build, organize, and grow the city. This metaphor of city planning can use zoning and other ordinances that govern construction and

C h a p te r 13:   Infor m ation Managem ent Strategy

land use. Imagine erecting a large commercial building in a residential zone. Imagine constructing a city park in an industrial zone. These design concepts apply to architecture and can be extended to how corporate data can be positioned and managed. Going back to the data principles, managing systems of record is different from managing master data. Secondary uses of data for analytics—compared with direct patient care— require different levels of fidelity and data quality. In addition, in building a city, roadways to connect zones, a communication infrastructure, utilities, and transportation and other public services are needed. Similarly, in data architecture building, how data are acquired, moved, managed, and delivered must be defined. For certain business purposes, daily batch processing is sufficient, whereas other uses of data may require more immediate access. Some areas of the data architecture allow more fungible data or data from a third party. It may be best to stage these data in a certain area of the data architecture. In city planning, zoning and ordinances change, and the same can be said in data architecture building. Areas will need to evolve, be rebuilt, or be rezoned to accommodate new directions of the business strategy. Without a data architecture, organizations tend to invest in a specific capability or function but may not know if such capability already exists or if similar information is already being captured in a separate system. Once the new capability is implemented, the function may fall short of expectations by requiring further work—for example, to integrate data from other systems. This can lead to unnecessary duplication of data and increases the risk that data will become out of sync with the systems of record. With proper design, data replication is a valid approach to supporting IT systems but needs to be controlled and managed to maintain integrity and limit costs to the support system. Integration still remains a key challenge for organizations. Countless enterprise data warehousing initiatives have been launched only to fail because of the challenges of data integration. Data standardization requires an organization to establish strong data governance roles and responsibilities. Without such governance, creating common definitions and standards remains almost impossible—not technically but from a business acceptance perspective. Without common definitions, there will be variability in mapping or translating data, which often results in incomplete or incorrectly mapped data. Exhibit 13.3 depicts a sample reference architecture that provides a view of systems of record, data acquisition and movement, and data management supporting different data delivery methods. Technology Aspects Beyond corporate data principles, a reference data architecture needs to reflect and support the data requirements of the organization. Technologies continually change and improve, so focusing on data requirements helps maintain consistency across technology changes. Too often, decisions are based on

279

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

EXHIBIT 13.3 Sample Reference Data Architecture

Streams

Messaging

Operational data stores

APIs

Enterprise data warehouse

Views, marts, cubes

Data marts

SQL

Enterprise data lake

Analytics and reporting

Data management

Data access/ delivery

Internal data

External data

Load

Extract

Replication Transform

280

Quality control Systems of record

Data acquisition

Note: APIs = application programming interfaces; SQL = Structured Query Language.

technology and not data requirements, which may not advance or support the organization’s information management goals. For example, changing from one database to another may be an improvement in technology but may not advance the organization’s data capabilities.

Data Movement Three of the data principles described earlier relate to data movement—currency, fluidity, and mobility. An aspect of data architecture is data movement—how data flow between systems—which is like the technical plumbing that moves data through different stages of data collection, acquisition, management, and delivery. Data movement influences the design of the data architecture. Systems of record generally support real-time access to data. In many organizations, there may be several systems of record, some of which require data from one another to be effective. The need for data movement continues to grow in an environment teeming with biomedical devices, sensors, and applications. Several data technologies support real-time capture and movement of data between systems, such as interface engine, enterprise service bus, and database replication. Each of these technologies supports different types of data movement and provides options for receiving systems to best obtain the data. Three requirements have evolved since the first data warehouses. First, the most current data are now needed. No longer is it sufficient to operate on last week’s or even yesterday’s data. Data must be fluid to support real-time analytics and uses. Second, patient health and population health have merged. Much of direct patient care is based on one patient, but with population health

C h a p te r 13:   Infor m ation Managem ent Strategy

methods, understanding how that patient compares with the population is equally important. This has raised expectations that transactional processing and analytics processing need to happen together. Anticipating this duality of data processing influences the design of the data architecture and use of technology. Third, patient data are no longer generated by just one organization. External data, such as external medical records, personal data, and other direct-to-consumer information, are growing. The integration of external data requires new designs to accommodate this new type of information. EMR-to-EMR data are not new, but patients wanting to evaluate other types of data is a growing trend.

Data Representation Data representation goes beyond the data structures and syntax of data to a more molecular level that includes semantics. Semantics attaches meaning to data and often is based on a standard that provides a common definition for sharing. The reference data architecture needs to anticipate additional semantics as part of data processing and data enrichment. Big Data technology is a fundamentally new type of technology for storing and managing data. Big Data provides new capabilities of processing larger volumes, varieties, and velocities of data, which were previously limited with traditional database technologies. One of the features of Big Data is the ability to acquire enormous amounts of data. Data acquisition is important, but just piling data together may not be wise. Acquiring and storing data is like putting things in boxes, storing them in your garage, and waiting to sort them out later. Boxes accumulate in the garage, making it difficult for you to find anything when you need to. Without disciplined data acquisition, there is a risk of limiting the data’s future use. One of the promoted concepts in Big Data is late binding of data. Late binding is the ability to generate a data model or scheme on demand. One of the most time-consuming steps in data warehousing is modeling and transforming data per corporate and model definitions. A value of late binding is that it defers this step to when data are actually needed or used. The disadvantage is that for the data to be useful, the data still have to be appropriately structured and formatted. Performing modeling and integration up-front for all corporate data would never be feasible and requires so much time that the value of the data may never be realized. Performing no modeling or integration is also unwise. This forces every data consumer to assemble the data before they can be used. Thus, a balanced approach is important. For secondary uses of data through an enterprise data warehouse or data lake, it is important to determine which corporate data are deemed high value. High value is a subjective term, but getting input from the business strategy and business areas can prioritize and focus the data for further structuring and

281

282

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

standardization. This approach focuses the benefit on highInteroperability Moment value data being consistent and When people say their data are “in the database,” they often mean comparable. A well-intended their data are in a spreadsheet like Excel. Nothing is wrong with using data initiative may provide easy spreadsheets to store data. Spreadsheets give us a pleasing, useraccess to data but still require friendly view of our data. However, by being stored originally in a considerable time for finding spreadsheet, the data are “bound” to that pleasingly human view. and preparing the data. PerHow can this be a problem? Consider the approach to data by two forming an analytics function different workgroups. can still be time-consuming; Group A has collected population data by county and year analysts spend 80 percent of for a given state—say, Wisconsin. The data are nicely arranged in a their time pulling together, spreadsheet as below: preparing, and normalizing Population by County and Year, Wisconsin data but less than 20 percent of their time doing analysis. Population Applying well-structured methCounty 1990 2000 2010 ods to high-value data will pay Randolph enormous dividends and allow Milwaukee analysts to focus on analytics Racine and not on data preparation. For some enterprise data warehousing, choosing or Group B has collected data on the number of full-time developing an enterprise data physicians in various clinical specialties for each county in Wisconsin model can be a difficult exercise for the year 2000. The data are arranged in a spreadsheet as below: that requires multiple business Full-Time Equivalent Physicians in 2000, areas to commit to and agree on by County and Specialty definitions. The implementation Specialty Randolph Milwaukee Racine of business definitions requires Cardiology data models to represent the Dermatology definitions. There will be debates and disagreements on Endocrinology which model is better. In healthFamily practice care, a number of data models already exist from standards or After reviewing these spreadsheets, you may ask, “How initiatives such as Health Level many family practitioners were there in 2000 in the county with Seven International (HL7), the lowest population?” How would you find the answer to that Fast Healthcare Interoperabilquestion? Suppose you had 100 such questions. Do you foresee any ity Resources (FHIR), Patientinteroperability problems? Can you express your answer using the Centered Outcomes Research concept of late binding? Network (PCORNet) Common Data Model, and Informatics for Integrating Biology and the Bedside (I2B2). Late binding recognizes that deciding which model to use is not as important as knowing when to apply

C h a p te r 13:   Infor m ation Managem ent Strategy

the model. All data models have issues, but committing to one model and one approach to model creation is necessary for successful data representation.

Data Accessibility Data accessibility is a critical part of an information management strategy. All the planning and designing undertaken to acquire, move, and manage data do not guarantee the data’s consumption. The critical drivers of data consumption influence how data are delivered. The types of data consumers in the organization, the preferred methods for data delivery, and the critical functions that require data all must be examined. Based on this analysis, design considerations for reporting functions, data access methods, and self-service tools may emerge. Self-service tools and methods for reporting and analytics encourage greater diffusion in the organization. Understanding the people skills and preferred tools influences how to provide data access. Potential power analysts or clinical operations analysts may influence the data architecture to allow users to have access to data without relying on IT to retrieve and deliver the data. For larger organizations, there may be several ways to access data. Current EMRs possess a wealth of reporting and analytics capabilities, allowing much of the work to be performed by clinical users. For many organizations, analytics and reporting are diffused throughout different groups. The value of the data is often realized by those who know the operations and use data to make changes. For organizations that desire to improve their use of analytics, the concept of center of excellence is important. A center of excellence is a concentration of expertise that shares and develops best practices and furthers the maturity of analytics across the organization.

Other Considerations The considerations in this section are not exhaustive but represent additional perspectives that may make the development of an information management strategy more successful.

Data Governance Program Data governance should be seriously considered as part of an information management strategy. The corporate world has steadily recognized the value of having a chief data officer. A chief data officer is an emerging executive role that recognizes the value of data and provides strategic direction for the acquisition, use, governance, and quality of corporate data assets (Morgan 2016). However, an organization does not need to hire a chief data officer to start a data governance program. Data governance can be formalized in

283

284

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

the organization in several ways, and entire books are dedicated to the topic. The important point is that a data governance program emphasizes corporate accountability, shared leadership, and organizational responsibility. Without this support, many aspects of the information management strategy may fail. Too often, a data project is considered an IT project, and data governance decisions are relegated to IT. Data decisions are then executed to complete the project, which engenders distrust in the data or data solution itself.

Data Quality Program “Quality is job one” was a slogan of the Ford Motor Company in the 1980s. Data quality is important as well, but how is this quality applied? Batini and Scannapieca (2006) describe data quality according to dimensions of accuracy, completeness, currency, consistency, and accessibility. Executing a data quality program requires effort and attention. Proactively maintaining data quality at the source improves the downstream use of that data. Many elements of the information management strategy and data architecture provide a solid foundation to understand what can be done to support data quality. Master data are often a starting focal point for implementing a data quality program. For healthcare, this program would focus on patients, providers, addresses, and locations. Data quality methods would start at the source of data collection. Standard operating procedures dictate how to capture information about a patient, such as whether the patient is existing or new. Edit checks can prevent or limit data issues, such as wrong age or address and missing fields. Data quality can be improved through human review and edit checks, but given the amount of data being generated, employing automated data monitoring is also needed. Data processing technology today permits continuous monitoring of data for any change that is not expected and notifies the appropriate data steward or operational area.

Data Value Protection Because data have inherent value, they should be protected. Data protection largely focuses on information security and privacy. Aspects of data inventory, classification, and access are essential components of many information security standards and controls. However, securing access to data does not necessarily protect the data’s intellectual property value. Many analytics companies are eager to access clinical data to create new products or to mine the data for new insights that can be commercialized. Data can be enriched, derived, or combined with other data to create new value, a fact that holds tremendous potential for some companies. Current security and privacy regulations, however, give technology companies and nonproviders limited access to data. Data rights and sharing are further complicated by the increasing presence of software-as-a-service, whereby the system and capability are hosted externally by a vendor. The information management strategy needs to address

C h a p te r 13:   Infor m ation Managem ent Strategy

how software contracts are written, clarifying access and usage rights by third parties and allowances for derivative data. Predictive models and other data mining can be performed on data and, once complete, can be leveraged in other products and services without having to use PHI. Considerations for understanding data value should be viewed through these dimensions. Ironically, contracts can be created that allow a wide range of secondary uses, which organizations unwittingly do not recognize or understand as the potential value of their data. The following are measures by which to judge the data access allowable in a data use agreement with a vendor or partner: • • • •

Amount or volume of data being shared (number of records) Breadth of data being shared (number of topics or kinds of data) Duration of data use (time limited or perpetual) Clinical density of data (simple structured data or detailed clinical notes and reports) • Data rights and usage for secondary use or derivative works (direct and limited or open and unrestricted)

Perils and Pitfalls Developing an information management strategy presents some common pitfalls. The cautions in this section may not be relevant for every organization but provide some counterpoints to forces that may influence the strategy. One peril to watch out for is truth and beauty in data. As data projects push through implementation, leaders or physicians may expect, unrealistically, that the data will be pristine and correct. The nature of healthcare data is imperfect and, some would say, messy. Throughout the healthcare delivery process, assessments, differential diagnoses, tests, treatments, and further assessments go through a natural evolution. Often, the EMR or EHR has conflicting and discordant data, which reflect the natural process of healthcare. Data may be collected as unstructured text with an expectation that the data will be consistent and error free. Humans have an amazing ability when processing data to fill in missing information and automatically correct problems, whereas a computer can only use what is provided. Setting the expectation that healthcare data are incomplete, noisy, and conflicting will help prevent data investments from collapsing; perfection is the enemy of good enough. Another pitfall is creating the “mother of all databases” for the organization. This pitfall may be exacerbated by Big Data technology, which promises that an enormous scale of data can be managed. You may hear, “If we just had access to data, or if we could put all of our data in one location, it would solve all of our problems.” This type of thinking leads down one of two paths—the creation of a large data warehouse or the purchase of a software solution for data aggregation. Neither path is wrong, but both reflect magical thinking—that if we did this, our data problems will vanish. Often overlooked is the relative maturity of the

285

286

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

organization and its ability to systemize and standardize master data or core data to bring them together. Even in a mature organization, these types of projects require effort. Integrating data and bringing value back to the organization are not easy and require commitment and patience that must be built in from the start. These kinds of investments imply an organizational commitment to some level of data governance. Data governance enforces the decision making required for a successful implementation. Active engagement from the clinical practice or business areas is vital to ensure their commitment and ownership of the finished product. Beware of the marketing and hype behind data analytics that suggests organizations are achieving efficiencies and improved outcomes as well as gaining insights for competitive advantage. Purchasing an analytics solution to address a business problem should be approached with caution. The fuel in analytics is data. Not understanding the quality, completeness, and level of integration of data could quickly sink a well-intended analytics investment. Starting with an information management strategy and reference data architecture will lead to better-informed decisions about analytics solutions.

Data as a Competitive Advantage An information management strategy executed with appropriate design and architecture enables organizations to be more agile in their ability to integrate, deliver, and share information. Better investment decisions can reduce overall IT expenses. Changes to the organization will be more easily executed with a cohesive strategy and design. Medicine is a continuous learning system. Practice guidelines and best practices are often evidence based and developed over time. An organizational culture that relies on data for decision making builds on evidence and begins to increase the opportunities for continuous improvement and measurement. Organizations often have difficulty improving or managing what they cannot measure. Some improvements can be applied from the top down, but a datadriven culture needs to be enabled on the front line, where the right tools and data should be provided to enable better decision making. In many organizations, commercial EMRs are maturing, along with improvements in fulfilling meaningful use and population health requirements. Today’s EMRs are embedded with a wealth of reporting and analytics tools. Because many organizations have access to these tools, the competitive advantage is held by those organizations that use data regularly to make actual improvements in care delivery.

Conclusion Healthcare is a data-intensive industry. An information management strategy is a key component of the larger IT strategy that supports the business strategy.

C h a p te r 13:   Infor m ation Managem ent Strategy

Formalizing an information management strategy increases an organization’s data intelligence and awareness of data assets and informs business decisions that may have an effect on data. The three main components of an information management strategy are information facets (types of data), data principles, and reference data architecture. The reference data architecture acts as the data blueprint, which informs technology choices. Defining the role of data governance in the organization supports the development of an information management strategy. Understanding and implementing the right level of decision rights and governance will improve the successful execution of the strategy. Lastly, the data business is a messy business. Raising leadership’s awareness of the challenges of data integration and use will help set and manage their expectations and guide the execution of IT projects and investments.

Chapter Discussion Questions 1. What are the basic components of an information management strategy? Is one component more important than the others? If so, which one and why? 2. How can an information management strategy reduce the costs of IT systems and investments? 3. If given a choice to implement an information management strategy or an analytics strategy, which would you perform first? Why? 4. What is IT’s role in the context of an information management strategy? 5. Why is taking an inventory of information systems and their data important for an organization?

Case Study  Guiding a Merger Lakes Central Regional Hospital recently entered discussions to acquire and merge with St. Vincent’s Hospital, a hospital with 60 beds, and with Lake City Physician Practice, an affiliated outpatient medical practice of 30 physicians. This merger will capture more of the regional market, keeping care within the proposed merged facilities. St. Vincent’s and Lake City use the same electronic medical record (EMR) software, but it is deployed as two separate systems—one for outpatients and one for inpatients. Both organizations have a small internal information technology (IT) department and staff who administer and maintain the software. Lakes Central uses a different EMR, and its leadership is interested in merging the EMRs as part (continued)

287

288

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

of the acquisition, knowing the benefit of sharing data between inpatient and outpatient care as well as the cost savings from reducing the number of software licenses and amount of IT maintenance. St. Vincent and Lake City use their EMR for care delivery, and the EMR has replaced most of their paper records. They have not explored the EMR’s additional functionality or advanced analytics because of lack of time, lack of physician interest, and cost. Lakes Central was an early adopter of the EMR and electronic systems and has actively developed key performance indicators for the practice. Its use of the EMR, not just for patient care but for improving operational efficiency and patient outcomes, has earned the hospital high marks and enabled it to maintain a healthy operating margin. As part of the planned merger, Lakes Central leadership has consulted its IT department about the effort involved in bringing together the EMRs.

Case Study Discussion Questions 1. Given that St. Vincent’s and Lake City have different backgrounds and different levels of EMR adoption, how would an information management strategy help guide the system merger? 2. Which area of emphasis should be considered first? Why? 3. What are some of the risks of this type of merger, and how would these risks influence the information management strategy?

Additional Resources National Institute of Standards and Technology. 2004. “Standards for Security Categorization of Federal Information and Information Systems.” http://nvlpubs. nist.gov/nistpubs/FIPS/NIST.FIPS.199.pdf. National Provider Identifier: www.cms.gov/Regulations-and-Guidance/AdministrativeSimplification/NationalProvIdentStand/.

References American College of Surgeons. 2018. “ACS National Surgical Quality Improvement Program.” Accessed January 24. www.facs.org/quality-programs/acs-nsqip. Batini, C., and M. Scannapieca. 2006. Data Quality: Concepts, Methodologies and Techniques. Accessed January 24, 2018. http://dc-pubs.dbs.uni-leipzig.de/ files/DQBook_part1.pdf. Densen, P. 2011. “Challenges and Opportunities Facing Medical Education.” Transactions of the American Clinical and Climatological Association 122: 48–58.

C h a p te r 13:   Infor m ation Managem ent Strategy

Morgan, L. 2016. “Rise and Fall of the Chief Data Officer.” Information Week. Published February 16. www.informationweek.com/strategic-cio/it-strategy/ rise-and-fall-of-the-chief-data-officer/a/d-id/1324280. Office of the National Coordinator for Health Information Technology. 2015. “Federal Health IT Strategic Plan: 2015–2020.” Accessed January 24, 2018. www. healthit.gov/sites/default/files/9-5-federalhealthitstratplanfinal_0.pdf. Open Group. 2018. “The TOGAF Standard—Version 9.2.” Accessed January 24. www.opengroup.org/subjectareas/enterprise/togaf. US Department of Health and Human Services. 2017. “HITECH Act Enforcement Interim Final Rule.” Reviewed June 16. www.hhs.gov/hipaa/for-professionals/ special-topics/HITECH-act-enforcement-interim-final-rule/index.html.

289

CHAPTER

THE ROLE OF PEOPLE AND INFORMATION IN DELIVERING PATIENT-CENTERED CARE

14

Naresh Khatri

Learning Objectives After reading this chapter, you should be able to do the following: • Understand the implications of service orientation and knowledge intensiveness in healthcare organizations. • Appreciate the fundamental role of human resources management and health information technology functions and capabilities in healthcare organizations. • Describe the complementary relationship between human resources management and health information technology functions and capabilities.

Key Concepts • • • •

Three dimensions of human resources management capabilities Three dimensions of health information technology capabilities Learning-oriented human resources management systems Information technology paradox

Introduction Human resources management (HRM) and health information technology (HIT) are fundamental rather than support functions in healthcare. Healthcare organizations can enhance their clinical outcomes significantly by managing their HRM and HIT functions more effectively. Unfortunately, many healthcare organizations have not understood the critical role that HRM plays in furthering their performance (Buchan 2004; Dussault and Dubois 2003; Khatri 2006a;

291

292

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Khatri, Gupta, and Varma 2016). Similarly, their investments in HIT typically have not produced the desired results (Curry and Knowles 2005; Grossman et al. 2007; Khatri and Gupta 2016; O’Malley et al. 2009). These shortcomings can be attributed to a lack of internal HRM and HIT capabilities. Healthcare organizations are not factories or fast-food restaurants. They are highly service-oriented and knowledge-based entities that demand proactive work behaviors from their employees. Various dimensions of HRM and HIT capabilities are discussed in this chapter, along with the critical role they play in the health services delivery process.

Major Features of the Health Services Delivery Process Service Orientation Professional services such as healthcare are different from manufacturing organizations in several ways (exhibit 14.1). Unlike manufacturing firms, customers of service firms typically interact with the production process. In doing so, customers inject a high degree of variability and customization into the service production process. To meet this challenge of variability, service organizations need employees who are empowered and are proficient at diagnosing problems, thinking creatively, and developing novel solutions (Herzenberg, Alic, and Wial 1998; Korczynski 2002). Understanding the customer is fundamental to service work (Korczynski 2002), and to be responsive to customer needs, service workers need to be given more discretion. Consequently, service organizations need to manage the worker–customer dyad instead of the management–worker dyad in the traditional industrial model; the more flexible, interpretive model of decision making is more appropriate than the fixed and structured engineering model (Herzenberg, Alic, and Wial 1998). The human element takes on a central role in effective service operations in that it produces the service product, and EXHIBIT 14.1 How Service Organizations Differ from Manufacturers

1. Service organizations show greater variability than manufacturing organizations and thus are less amenable to standardization. 2. Service organizations require contact between the employees and the customers. 3. Service organizations need to give greater discretion to the employees in overcoming issues that arise out of variability and in meeting customer needs. 4. Customer satisfaction in services is mediated by employee satisfaction. 5. The employee–customer dyad is the key to managing services effectively, whereas the manager–worker dyad is the key to managing manufacturing processes.

Cha p t er 1 4 :  T h e Ro l e o f Pe o p l e a n d In f o r ma t ion in D eliver ing Patient- C entered C are

the interactions of frontline employees with customers form key attributes of the service product itself. Using industrial models to manage service-based organizations makes as little sense as using farm models to run factories (Davis 1983). Research evidence suggests that service firms often have difficulty improving performance by using organizational practices devised for manufacturing firms (Gera and Gu 2004; Karreman and Alvesson 2004). Indeed, translational shortcomings in introducing best practices from manufacturing into healthcare organizations abound. Healthcare organizations that build on the industrial model become trapped in bureaucratic structures that do little to meet customer needs but meet outdated ideas of internal efficiency (Herzenberg, Alic, and Wial 1998). To deliver patient-centered care, healthcare organizations need to adopt a “service production and delivery logic” rather than a “manufacturing goods logic” (Schneider and White 2004).

Knowledge Intensiveness Knowledge management is a relatively new management tool that organizations are trying to incorporate into their management systems. The capacity to learn is considered critical to organizational adaptation and long-term survival: “Ultimately an organization’s only sustainable competitive advantage lies in what its employees know and how they apply that knowledge to business problems” (Pearlson 2001, 193). Despite the increasing realization that knowledge management and organizational learning are sources of sustainable competitive advantage, current knowledge management practices in most healthcare organizations are elementary. About 42 percent of organizational knowledge is housed exclusively in the minds of employees (Hansen and Thompson 2002). Healthcare organizations can exploit this knowledge by designing learning-oriented HRM systems that (1) promote positive learning attitudes and (2) nurture an organizational climate of self-renewal (Jaw and Liu 2003). Doing so, however, requires human resources (HR) professionals to play a pivotal role in designing, shaping, and reinventing the work context—that is, implementing organizational change and development strategies. Similarly, HIT is essential to managing information and knowledge in healthcare organizations in that it provides the underlying infrastructure for organizational change and renewal.

Proactive Work Behaviors in the Health Services Delivery Process To be able to deliver exceptional service, professional service firms, such as healthcare organizations, require their employees to have a proactive rather

293

294

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

than standard behavior (Hyde, Harris, and Boaden 2013; Korczynski 2002; McClean and Collins 2011; Schneider and White 2004). Proactive work behavior consists of initiative and flexibility: Employees do something without being told or without their role explicitly requiring it (Crant 2000; Frese and Fay 2001), and employees are able to adapt to changing situations. Self-starting and flexible work behaviors are valuable organizational resources for two reasons (Beltran-Martin and Roca-Puig 2013). First, they spare the firm the costs of people not adjusting to changed situations; because self-starting and flexible workers easily adjust to new situations, losses associated with lack of change are minimized. Second, they facilitate change implementation processes by imparting necessary organizational agility. To provide exemplary patient care, healthcare employees must have proactive work behaviors hardwired in them (Korczynski 2002; Schneider and White 2004). Health services delivery involves high levels of task interdependence, task complexity, and uncertainty, and it often depends on the spontaneous actions of employees as they coproduce services with their patients (Hyde, Harris, and Boaden 2013). Thus, instead of changing a particular set of HR and information technology (IT) practices and systems to produce standard employee behaviors, healthcare organizations must encourage proactive employee actions and behaviors. This could be accomplished by implementing HRM and HIT practices and systems that align with unique organizational strategies, cultures, and histories; a one-size-fits-all set of HRM and HIT practices does not exist (Chadwick et al. 2013; Guest 2011; Tremblay et al. 2010; Woodrow and Guest 2014). In short, from the perspective of HRM and HIT capabilities, eliciting proactive work behaviors—not a prescribed set of HRM and HIT practices, which can vary from one organization to another—is what is important. Proactive employee behavior seems a prerequisite in service organizations given that customer perceptions and buying behaviors are greatly influenced by employee–customer interactions (Baluch, Salge, and Piening 2013; Guest and Conway 2011; Korczynski 2002; Towler, Lezotte, and Burke 2011). The notion of proactive behavior recognizes individuals as cognitive and emotional beings who possess free will (Wright, Dunford, and Snell 2001). In fact, MacDuffie (1995) stresses the importance of discretionary behavior and argues that a firm can achieve competitive advantage only if the members of its human capital individually and collectively choose to engage in behavior that benefits the firm. In short, the effect of HRM and HIT on the distal outcomes of organizational performance—that is, the quality of patient care in healthcare—is likely to be mediated through the proximal outcomes of proactive work behaviors (Baluch, Salge, and Piening 2013; Hyde, Harris, and Boaden 2013; Ketkar and Sett 2009).

Cha p t er 1 4 :  T h e Ro l e o f Pe o p l e a n d In f o r ma t ion in D eliver ing Patient- C entered C are

HRM Capabilities HRM is more complex in healthcare than in many other industries because of healthcare’s labor intensiveness; its well-established separation of professions and occupations, each with its own locus of practice and control; and the sheer scale of operations (Dussault and Dubois 2003; Khatri, Gupta, and Varma 2016; Townsend, Lawrence, and Wilkinson 2013). The logic underlying resource-based theory and dynamic capabilities suggests that HRM may be a more potent source of competitive advantage in healthcare than in other industries (Buchan 2004; Everhart et al. 2013; Townsend, Lawrence, and Wilkinson 2013). Few healthcare organizations are able to manage their HR function well because doing so is far too complex for them (McBride and Mustchin 2013). Consequently, those rare healthcare organizations that are able to harness their HR function effectively have a significant competitive edge. In fact, most healthcare organizations have not yet realized that HRM can be a source of competitive advantage because of clinical culture or clinical myopia (Khatri, Baveja, et al. 2006, 124); thus, HRM remains a hidden value (Khatri, Gupta, and Varma 2016). The importance of HRM in healthcare is evident from the fact that salary and wages constitute about 65 to 80 percent of a typical healthcare organization’s total operating budget—the single largest input in health services delivery (Buchan 2004; Dussault and Dubois 2003). It is only logical that healthcare organizations use their HR optimally if they want to improve their clinical outcomes. Several scholars suggest that although HRM is important in a healthcare context because of its service and knowledge intensiveness, it remains an outdated and overlooked function in most organizations (e.g., Baluch, Salge, and Piening 2013; Buchan 2004; Dussault and Dubois 2003; Kabene et al. 2006; Khatri, Wells, et al. 2006; McBride and Mustchin 2013; Townsend and Wilkinson 2010). Townsend and Wilkinson (2010) note that healthcare reforms during the past 30 years have focused largely on structural change and cost containment; HRM has been overlooked even when HR might dictate and constrain the introduction and rollout of these other initiatives. Similarly, Leggat, Bartram, and Stanton (2011) observe that healthcare organizations have not been effective in ensuring that basic aspects of HRM are in place. They found that a hospital’s organizational structure and hierarchy reinforce parallel care processes that fragment HRM practices and systems. Khatri and colleagues (Khatri, Gupta, and Varma 2016; Khatri, Wells, et al. 2006) argue that current HRM systems and practices in healthcare are premised on the old industrial model of management and thus are highly inadequate in managing knowledge-based and service-intensive healthcare entities. Healthcare organizations are not able to make the leap from traditional HRM practices to strategic HRM systems because they may lack the necessary HR capabilities to do

295

296

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

so (Boudreau and Lawler 2014; Khatri, Wells, et al. 2006; Lawler and Mohrman 2003; Townsend and Wilkinson 2010). Anecdotal evidence suggests that they have, in fact, been heading in the opposite direction in the past few years as they pursue lean management principles borrowed from manufacturing (McBride and Mustchin 2013). Many HRM tasks have landed in the lap of nurses and department administrators, who lack the time and professional expertise in managing the HR function (Khatri, Baveja, et al. 2006; McBride and Mustchin 2013). Two research developments signal the importance of building more effective organizational capabilities in HRM. First, the proper implementation of HRM practices and systems may be a more important determinant of organizational outcomes than the mere presence of a set of practices such as high-performance work systems (Baluch, Salge, and Piening 2013; Guest and Conway 2011; Tremblay et al. 2010). In turn, the effective implementation of HRM practices and systems may depend largely on whether an organization possesses sufficient HRM expertise and capabilities (Chow 2012; Khatri, Gupta, and Varma 2016; Lawler and Mohrman 2003). Many organizations in all industries have poorly managed HRM because they lack the necessary emphasis, support, and capabilities (Boudreau and Lawler 2014; Lawler and Mohrman 2003). HRM capabilities are necessary to create a “strong HR system” (Bowen and Ostroff 2004) and to reduce the gap between intended and implemented HRM practices and systems (Woodrow and Guest 2014). HRM capabilities in an organization are mechanisms to acquire, develop, renew, reconfigure, and deploy its HR function so that the organization is able to adapt to and remain in alignment with changing strategic business needs and the external environment (Khatri, Gupta, and Varma 2016).

Three Dimensions of HRM Capabilities According to Khatri, Gupta, and Varma (2016), HRM capabilities in healthcare consist of three dimensions (exhibit 14.2): 1. The CEO is enlightened about and supportive of the critical role of an effective HRM in boosting organizational outcomes (Brandl and Pohler 2010; Chadwick, Super, and Kwon 2013; Stanton et al. 2010). 2. The chief HR officer or head of the HR function is a visionary and understands the fundamentals of the health services delivery process, professional and leadership competence, relationship building with other department heads or units, and developing a well-articulated HR strategy (Boselie and Paauwe 2005; Murphy and Southey 2003). 3. The HR staff have the competencies and overall proficiency of an HR department. They have a thorough expertise in HRM activities and a good understanding of the psychological and social behaviors of

Cha p t er 1 4 :  T h e Ro l e o f Pe o p l e a n d In f o r ma t ion in D eliver ing Patient- C entered C are

CEO’s Enlightened View and Support of HR 1. The head of HR interacts with the CEO frequently and understands the priorities and strategic initiatives of the CEO very well. 2. The head of HR is an important member of the top management team. 3. HR plays a critical role in implementing major strategic initiatives of the hospital. 4. The hospital CEO realizes fully the important role that HR can play in health care delivery processes. 5. The CEO provides wholehearted support and resources to HR activities and programs. 6. The CEO views HR more as an administrative function than a strategic/ transformative function. HR Head’s Vision and Competence 1. The head of the HR department in my hospital has a compelling vision of how to use HR to enhance hospital performance. 2. The head of the HR department in my hospital spends much time and effort in building relationships with senior managers of all the departments in the hospital. 3. The head of HR department in my hospital understands the fundamentals of the health care delivery process (e.g., patient needs and concerns, hospital operations, etc.). 4. The head of HR in my hospital understands the unique operating characteristics of all the departments in the hospital. 5. The head of HR in my hospital possesses a thorough knowledge of the health care industry. 6. The head of HR in my hospital has developed a well-understood HR strategy for the hospital. Professionalism of HR Staff and HR Department 1. HR employees have a thorough expertise in HR (e.g., knowledge of HR activities, such as recruitment, training, compensation, performance management, etc.). 2. HR employees have a solid understanding of psychological and social behaviors of the hospital staff. 3. HR employees are very knowledgeable in change management. 4. HR employees understand overall strategy, culture, and operations of the hospital very well. 5. HR employees have developed an excellent rapport with other departmental heads and employees in the hospital. 6. HR department has developed effective in-house tools (e.g., valid recruitment methods, effective training programs, fair and effective reward systems, etc.) 7. HR department is very proactive and quick in adapting to employee problems and concerns. (continued)

297

EXHIBIT 14.2 Dimensions of HRM Capabilities in Healthcare Organizations

298

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

EXHIBIT 14.2 Dimensions of HRM Capabilities in Healthcare Organizations (continued)

8. HR department is staffed with highly professional and courteous employees. 9. The policies and procedures coming from the HR department help administrators, clinicians, and other employees perform their jobs efficiently. 10. The HR policies and practices enable my hospital [to] achieve its mission and strategic objectives. 11. The HR department of my hospital is very responsive to the needs of the unit/departmental managers, clinicians, and other hospital staff. 12. The HR department provides useful and timely information and expertise on HR issues to the unit managers and employees. Source: Reprinted with permission from Khatri, Gupta, and Varma (2016).

organizational employees to implement appropriate HRM practices (Han et al. 2006; Quinn and Brockbank 2006). They are proactive and quick to address the concerns of other employees and develop excellent rapport with other department heads and employees. Previous HRM research suggests that the support of the CEO for HRM is a critical element. According to Khatri, Wells, and colleagues (2006), one major factor contributing to poor HRM in healthcare organizations is that their CEOs have yet to fully grasp the significance of HRM in health services delivery. Even if the HR department is trying to do things that may impart more dynamism to the function, it may not succeed if the CEO’s support is absent (Chadwick, Super, and Kwon 2013). For example, Stanton and colleagues (2010) found that the CEO is crucial in providing HRM legitimacy, leadership, and resources that create a distinctive function and in nurturing internal agreement and consensus on the role of HR among the senior executive team. Thus, to achieve dynamic HRM capabilities, the CEO’s support of and enlightened view about HRM are essential. The HRM literature describes the significant role of the head of the HR function in imparting dynamic HRM capabilities. Ulrich (1996) makes a persuasive case in favor of HR champions who can transform the traditional HR function into one that is strategic and thus contributes to organizational performance. The HR function’s influence on employee behaviors and clinical outcomes may depend crucially on the competencies and professionalism of its chief officer. The HR department and HR employees with deep knowledge of their function and professional domain are indispensable in a highly knowledge-based and service-oriented healthcare organization. An HR function with cuttingedge HR knowledge and competent HR employees is well situated to modify,

Cha p t er 1 4 :  T h e Ro l e o f Pe o p l e a n d In f o r ma t ion in D eliver ing Patient- C entered C are

CEO support

HR head

Proactive work behaviors

HR capabilities

Quality of patient care

HR department

Source: Adapted with permission from Khatri, Gupta, and Varma (2016).

reconfigure, and renew HRM practices according to the strategic needs and prevalent culture of the organization. In the absence of these three dimensions of HRM capabilities, healthcare organizations may struggle to influence the actions and behaviors of their employees and to build a workforce capable of delivering exceptional healthcare services.

Relationship Between HRM Capabilities and Patient-Centered Care Khatri, Gupta, and Varma (2016) investigated the relationship between HRM capabilities and the quality of patient care in a national sample of US hospitals. They report that HRM capabilities have both a direct and an indirect positive relationship with the quality of patient care. The indirect relationship is mediated by proactive employee behaviors (exhibit 14.3).

HIT Capabilities HIT consists of a diverse set of technologies for transmitting and managing health information for use by consumers, providers, payers, insurers, and all other groups that have an interest in healthcare (Blumenthal and Glaser 2007). Healthcare providers have been implementing HITs, such as electronic medical record (EMR) and computerized physician-order entry systems, in response to the Health Information Technology for Economic and Clinical Health (HITECH) Act of the American Recovery and Reinvestment Act of 2009, which has set aside up to $30 billion in incentive payments to support the adoption and meaningful use of electronic health records (EHRs) and other types of HIT (Blumenthal 2011; Kaushal and Blumenthal 2014). The number of certified HIT vendors in the United States mushroomed from 60 in 2008 to more than 1,000 in 2012 (Sittig and Singh 2012). Many experts,

299

EXHIBIT 14.3 Relationship Between HRM Capabilities and Quality of Patient Care

300

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

however, have expressed major concerns that the easy availability of federal funds to promote meaningful use may result in rushed or poor implementation of HITs without comprehensive functionality and processes in place, thus causing substantial and unexpected risks in health services delivery (Black et al. 2011; Furukawa, Raghu, and Shao 2010; Klauer 2013; Nanji et al. 2011; Pines 2013). For example, one study reported that the increase in staffing and decline in patient safety associated with the introduction of EMRs in their study might have been due to the poor implementation of EMRs, among other factors (Furukawa, Raghu, and Shao 2010). It is plausible that the meaningful use provision in HITECH may merely prop up the HIT market without improving quality (Klauer 2013). Even worse, this provision may end up saddling healthcare providers with dysfunctional HITs that will be hard to change or replace later (Pines 2013). Optimal computerization is likely to improve quality, but whether the HIT systems currently deployed in most hospitals achieve such improvement remains unclear (Himmelstein, Wright, and Woolhandler 2010). For example, many HIT systems in US hospitals are poorly designed for the kind of proactive, team-based, patient-centered care that the patient-centered medical home and other models are calling for (Fernandopuule and Patel 2010); the core structure of current EHRs is at odds with continuous, seamless patient care, and many HIT systems are primarily driven by the imperative to allow physicians to document, code, and bill visits at a more intensive—and thus higher-cost—level. Although these features increase practice revenue in a fee-for-service setting, they do nothing to improve care. Undoubtedly, HITs are a powerful tool and can potentially transform healthcare (Blumenthal and Glaser 2007). The pertinent question is how to get them right (Feld and Stoddard 2004). For example, by leveraging IT investments during the 1990s, banks saw a 25 percent reduction in branches and a 20 percent reduction in full-time employees (Blount, Castleman, and Swatman 2005). In healthcare, the capacity of HITs to realize the transformational vision envisaged in healthcare reform depends largely on whether the systems installed are designed to produce the information required to achieve the desired quality and cost levels (Blumenthal and Glaser 2007). Unfortunately, current HITs are not configured properly and do not sufficiently support aspects of care delivery that are vital to improving care and controlling costs (Jones et al. 2012). This situation has happened, in part, because healthcare organizations have not built the necessary in-house IT expertise, which is a must to getting HIT right (Khatri and Gupta 2016). IT investments produce operational improvements only when they are accompanied by effective IT capabilities and expertise (Lu and Ramamurthy 2011; Mithas, Ramasubbu, and Sambamurthy 2011; Yeh, Lee, and Pai 2012). An organization’s IT capabilities enable it to acquire, deploy, and adapt IT-based

Cha p t er 1 4 :  T h e Ro l e o f Pe o p l e a n d In f o r ma t ion in D eliver ing Patient- C entered C are

resources to improve organizational processes and performance (Yeh, Lee, and Pai 2012). An enterprise is successful not because it uses any particular leadingedge IT software but because it has the capability to apply IT to ever-changing business opportunities (Ross, Beath, and Goodhue 1996). In the absence of internal IT capabilities, healthcare entities tend to return to old paper-based processes or use partial information recorded in both paper and EHR systems (Curry and Knowles 2005), resulting in a loss of information, more time spent retrieving pertinent information and data, and the need for more IT staff. The IT function should be organized to support transformed processes rather than the other way around (McAfee and Brynjolfsson 2008), and HITs are far more complex to implement than are typical technologies, such as a new medical device or a new medical procedure (Tyagi et al. 2013). Unfortunately, in healthcare, IT initiatives seem to be more concerned with technical aspects, ignoring in the process the vital contextual factors that make or mar IT projects. This happens because typical healthcare providers lack sufficient IT expertise and, as a result, approach IT projects from a technical perspective and not as a transformational tool (Kellermann and Jones 2013; Kivinen and Lammintakanen 2013; Mandl and Kohane 2012). When IT spending is not properly channeled into improved organizational capability, greater IT spending has a negative effect on agility, suggesting that IT capability is critical in realizing organizational agility (Lu and Ramamurthy 2011). Huge, imprudent IT investments are therefore not necessarily beneficial to organizational agility in responding to market changes. The reason may be the wrong infrastructure or incompatible systems, delayed or rushed implementations, or islands of automation meeting local needs without integration across the enterprise.

Three Dimensions of HIT Capabilities HIT capabilities consist of three dimensions: (1) HIT infrastructure, (2) professionalism of HIT staff, and (3) vision and competence of the chief information officer (CIO) (Bassellier and Benbasat 2004; Khatri 2006b; Mithas, Ramasubbu, and Sambamurthy 2011; Peppard 2010). The first dimension is the underlying infrastructure. HIT infrastructure helps healthcare organizations identify and develop key applications for improving business processes rapidly; share information across services, locations, and specialties; and facilitate transaction processing and supply chain management (Lu and Ramamurthy 2011; Mithas, Ramasubbu, and Sambamurthy 2011). The professionalism of HIT staff (who have necessary technical, behavioral, and business skills) is the second dimension (Fink and Neumann 2007). Here, professionalism refers to the quality of employees in the IT department. To implement initiatives effectively, IT professionals (and thus IT departments) must possess technical and interpersonal skills, communicate effectively

301

302

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

EXHIBIT 14.4 Dimensions of HIT Capabilities in Healthcare Organizations

CIO’s Vision and Competence 1. The head of IT interacts with the CEO frequently and understands his or her priorities/strategic initiatives very well. The head of IT is an important member of the top management team. 2. The CIO interacts with the CEO frequently and understands the priorities and strategic initiatives of the CEO very well. 3. The CIO is an important member of the top management team. 4. The CIO of my hospital has a compelling vision of how to use IT to enhance hospital performance. 5. The CIO of my hospital spends much time and effort in building relationships with senior managers of all the departments in the hospital. 6. The CIO of my hospital understands the fundamentals of the health care delivery process (e.g., patient needs and concerns, hospital operations, etc.). 7. The CIO of my hospital understands the unique operating characteristics of all the departments in the hospital. 8. The CIO of my hospital possesses a thorough knowledge of the health care industry. 9. The CIO of my hospital has developed a well-understood IT strategy for the hospital. IT Infrastructure 1. IT infrastructure greatly helps my organization in identifying and developing key applications of IT for improving business processes rapidly. 2. The IT infrastructure enables my organization to share information across services, locations, and specialties. 3. The IT infrastructure of my organization facilitates implementation of transaction processing (e.g., billing) and supply chain management (e.g., ordering of medical supplies) across the organization. 4. The IT infrastructure of my organization creates synergies across specialties, services, and locations. Professionalism of IT Staff 1. IT people in my organization possess excellent technical IT skills. 2. IT people in my organization have great interpersonal and social skills. 3. IT professionals in my organization communicate and work effectively with employees and managers in other departments. 4. Employees in IT department of my organization understand how to facilitate organizational change using IT. 5. IT people in my organization have abilities to conceive and develop costeffective applications of IT that support business needs of the organization faster than peer organizations. Source: Reprinted with permission from Khatri and Gupta (2016).

with employees and managers in other units, be conversant about change management, and have the requisite expertise for conceiving and developing cost-effective applications to support the clinical and business needs of the organization (Bassellier and Benbasat 2004).

Cha p t er 1 4 :  T h e Ro l e o f Pe o p l e a n d In f o r ma t ion in D eliver ing Patient- C entered C are

303

The third dimension is the CIO’s professional competence and vision for HIT. CIO savviness is believed to be pivotal to the realization of HIT value in healthcare organizations (Broadbent and Kitzis 2005; Burke et al. 2008; Peppard 2010). Thus, the CIO is an important member of the senior management team and must have excellent technical expertise and a compelling IT vision for the organization. In addition, the CIO must develop and maintain a great rapport with the heads of other units or departments, as well as possess the necessary leadership skills and business acumen. The CIO’s vision and competence are likely to be the main driver of the first two dimensions. The reason is that the CIO makes decisions regarding the overall information management strategy, the HITs to acquire or adopt, and the IT professionals to hire. The three dimensions of HIT capabilities are summarized in exhibit 14.4.

Relationship Between HIT Capabilities and Quality of Patient Care Previous research has failed to find a robust and consistent relationship between IT investments and organizational performance, which has led IT experts to coin the term IT paradox. One plausible explanation for the apparent IT paradox is that even though HIT capabilities may not have a direct relationship with the quality of patient care (Khatri 2015; Khatri and Gupta 2016), the relationship is mediated by other factors. In a national sample of US hospitals, Khatri and Gupta (2016) found that the relationship of HIT capabilities with quality of patient care is mediated by proactive employee behaviors (exhibit 14.5). Specifically, the CIO’s IT competency and vision have a positive correlation with IT infrastructure and IT staff professionalism, and professionalism has a significant positive relationship with proactive employee behaviors, which in turn have a positive relationship with the quality of patient care. Khatri (2015) found that the negative influence of IT staff professionalism and IT infrastructure on mortality rate EXHIBIT 14.5 Relationship Between HIT Capabilities and Quality of Patient Care

Professionalism of IT staff

Proactive employee behavior

Competence and vision of CIO

IT infrastructure

Outsourcing of HIT system

Source: Adapted with permission from Khatri and Gupta (2016).

Quality of patient care

304

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

from heart attacks is mediated by employee effort and flexibility. In other words, IT staff professionalism and IT infrastructure enable greater employee effort and flexibility, which in turn lead to a lower mortality rate from heart attacks.

Is a Homegrown System Better Than an Outsourced System? Healthcare providers take different approaches to implementing their HIT systems. Some have rolled out internally developed systems, whereas others have adopted vendor-developed systems, such as Allscripts, Cerner, Epic, and GE. Still others (although only a few) have yet to introduce an EHR, an EMR, or any HIT at all. Research that compares the relative effectiveness of internally developed systems with that of outsourced systems is scarce. The complexity of health services delivery coupled with the extent to which HITs provide the infrastructure for the entire care process makes implementation of HITs a formidable task. Implementation is also influenced by various organizational factors, especially the availability of IT expertise (Kaushal and Blumenthal 2014; Khatri 2006b). It comes as no surprise, then, that most healthcare organizations struggle with this process. Outsourced HIT systems have a tendency to “airdrop” HITs without concomitant changes in organizational cultures and management systems (Kivinen and Lammintakanen 2013). However, computerization entails more than just buying computer hardware; it also involves investing in a broad collection of complementary products and innovations, some of which take years to implement (Brynjolfsson and Hitt 2003; Jones et al. 2012). For example, buying computer hardware and software but not having skilled IT employees with knowledge of how the system will transform the clinical function would result in ineffectual HITs. Consequently, the benefits of HIT will not be realized fully until HITs are properly configured and embedded in the health services delivery process (Kellermann and Jones 2013; Nanji et al. 2011). McCormick and colleagues (2012) observe that off-the-shelf commercial systems are often chosen for billing reasons and are more closely aligned with the needs of administrators than those of clinicians. Furthermore, outsourced systems are unlike highly customized systems developed by on-site IT experts, who are integrated with the clinical staff. Thus, internally developed HITs tend to be better configured and embedded in the clinical process than outsourced HITs are. The quality of patient care is likely higher for healthcare providers that deploy an internally developed HIT system than for those that use an outsourced one. Khatri and Gupta (2016) compared the quality of patient care as reported by patients and found that the quality of patient care is significantly higher in hospitals with in-house HIT than in hospitals that rely on externally built and managed systems.

Cha p t er 1 4 :  T h e Ro l e o f Pe o p l e a n d In f o r ma t ion in D eliver ing Patient- C entered C are

Conclusion HRM and HIT are crucial to achieving high-quality patient care, but they are not particularly well managed in healthcare organizations. This shortcoming may occur because healthcare organizations have failed to build their HRM and HIT capabilities. Nonetheless, in the United States, more and more attention is being paid to developing HIT capabilities in healthcare organizations. Unfortunately, despite the centrality of people in the health services delivery process, HRM continues to receive inadequate attention or strategic focus from many US healthcare executives. Just as individual departments in a healthcare organization cannot improve healthcare services delivery and quality on their own, HRM or HIT cannot do so in isolation either. But, together, the power they can bring to service orientation and knowledge intensiveness can transform organizations. The inadequacy of HRM and HIT capabilities in healthcare organizations is not a challenge but an opportunity, a call for improvement rather than a drawback. Healthcare organizations should focus on the personnel growth and technological innovation that go beyond supporting day-to-day, shortterm operations to enable and elevate capabilities that sustain gains and reap long-term benefits. Such a transformation cannot be done in isolation in any functional area—be it clinical or HR or IT or operations. HRM and HIT personnel should acquire an in-depth understanding of the clinical and managerial aspects of the organization. For instance, the HR function should be involved in organizational change efforts to foster and grow clinical and interprofessional collaboration and teamwork. Using HRM capabilities in this way can enable health professionals to break down process barriers and organizational silos to see the broader picture of the process, the entire breadth of the service encounter between the patient and various clinicians. This creates organizational learning that relies on HIT capabilities to manage and disseminate knowledge throughout the healthcare organization.

Chapter Discussion Questions 1. What key features of service organizations differentiate them from manufacturing firms? What are the implications of these distinctions in managing healthcare organizations? 2. What are the underlying reasons for the failure of HIT initiatives? 3. In its prescription to reduce medical errors and improve quality of care, the Institute of Medicine report To Err Is Human (Kohn, Corrigan, and Donaldson 2000) put a lot of emphasis on HIT but not on HRM. Do you see any flaw in such an approach?

305

306

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

4. If you were the CEO of a hospital, what immediate steps would you take to improve the delivery of health services in your organization? 5. What does it mean to design a “learning-oriented HRM system”?

Case Study  University Hospital University Hospital is a 300-bed acute care facility that serves the residents of a large Midwestern state. It is part of a large university health system that employs more than 8,000 faculty and staff and has an annual operating budget of close to $1 billion. In addition to University Hospital, a medical school, nursing school, children’s hospital, regional hospital, and cancer center complete the health system. University Hospital operates the region’s only level 1 trauma center, the only burn and wound intensive care unit, and the only kidney transplant program. It receives more than 50,000 visits to its emergency department annually.

Human Resource Management Department Because University Hospital is part of a health system, recruitment and selection of physicians are done at the level of the medical school, and the allocation of research and clinical time is governed by the medical school in concert with the hospital’s administration. The hospital deals with most of the human resource management (HRM) duties for nurses and other employees, and its HR policies and programs are affected by the university’s HR policies. All of this makes HRM issues at the hospital complex, because three different entities—university, medical school, and hospital—need to work together to manage the hospital’s HR function. The hospital’s HRM department is headed by Luz Salvador, the vice president who was hired about four years ago to transform the HR function. Luz has more than 20 years of experience in healthcare management, although she does not have specific professional education or training in HRM. On average, the hospital has one HRM employee per 200 FTEs (full-time equivalents), as compared with one HRM employee per 100 FTEs in companies of similar size in other industries. This ratio shows that the hospital’s HRM department is grossly understaffed, which is reflected in the clinical staff’s complaints that vacant positions are filled too slowly. Recruiting employees takes several months, and often the hospital loses the best-qualified applicants to other healthcare organizations in the region. Employee turnover at the hospital during normal economic conditions is about 22 percent, which is above the industry average. Employee morale at the hospital is low, and many employees feel a sense of helplessness and frustration.

Cha p t er 1 4 :  T h e Ro l e o f Pe o p l e a n d In f o r ma t ion in D eliver ing Patient- C entered C are

Luz is expected to solve these problems, but most of the promises she made when she was hired have not been fulfilled. However, she thinks she has done an excellent job so far in reforming the department’s practices. While she thinks the department plays an important role in the hospital, most other department heads and managers are not aware of HRM’s role or any of its contributions and purported changes. These leaders think that the department is as slow in responding to their needs as it was before. Some of them prefer the department to assume a policing role rather than a strategic, supportive role. Clinicians in the hospital are the dominant power, not only in clinical decision making but also in management function. They expect a lot of support from the department, but they do not consider it an important strategic resource, do not provide it with sufficient resources, and do not give it enough authority to manage its own function and implement changes. As a result, the department remains weak, poorly led, and highly understaffed, and it fulfills only traditional personnel management tasks.

Health Information Technologies Ten years ago, with a view to taking the lead in implementing health information technology (HIT), University Hospital outsourced its IT function to a private company that specializes in health information management systems. For several years, the hospital maintained its own workforce of about a hundred IT workers and received supplemental services from the vendor. Maintaining a large IT workforce and paying an annual contractual amount to a vendor were expensive. Two years ago, the vendor agreed to absorb the hospital’s IT workers. The university still provides the IT function for clinical departments. The individuals who headed the hospital’s IT initiatives were physicians who did not use HIT much, did not understand it, and resisted efforts to learn about it or any emerging systems, not to mention its power to transform clinical processes. The former chief information officer (CIO) at University Hospital has a PhD in genetics and expertise in bioinformatics but knows little about organizational complexity and dynamics or about the transformation of clinical decision processes. Meanwhile, the point person at the vendor company excels at marketing and selling its services, making it easy for the vendor to mislead the hospital CIO and other leaders by promising IT enhancements that the company was not able to deliver. The hospital’s enthusiasm faded quickly as users found that the installed system was dysfunctional and woefully inadequate to handle the complexities of health services delivery. Most physicians complained about the serious problems, but the CIO and senior management insisted on using (continued)

307

308

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

the same system and the same vendor. Two years ago, the hospital appointed a new CIO—a senior nurse manager, who also had no IT background but a firm belief that clinical knowledge is the critical competency needed to understand clinical decision support and to oversee the planning, design, and implementation of HIT. Today, HIT at University Hospital remains in disarray and is holding the hospital back from delivering efficient, affordable, and high-quality patient care.

Case Study Discussion Questions 1. Is the scenario in this case unique to University Hospital, or do other academic medical centers approach HRM and HIT in a similar piecemeal, ad hoc fashion? 2. What suggestions would you make to the senior management at University Hospital to improve its HRM and HIT functions? 3. How might the HR function in an era of knowledge workers differ from the traditional personnel function? 4. How would you design a performance evaluation instrument for professional staff who are structured to work as multiprofessional teams? 5. What qualities would you look for in a chief HR officer?

Additional Resources Allscripts: www.allscripts.com. Cerner: www.cerner.com. Epic: www.epic.com/software. GE: www.ge.com/digital/industries/healthcare.

References Baluch, A. M., T. O. Salge, and E. P. Piening. 2013. “Untangling the Relationship Between HRM and Hospital Performance: The Mediating Role of Attitudinal and Behavioural HR Outcomes.” International Journal of Human Resource Management 24 (16): 3038–61. Bassellier, G., and I. Benbasat. 2004. “Business Competence of Information Technology Professionals: Conceptual Development and Influence on IT–Business Partnerships.” MIS Quarterly 28 (4): 673–94. Beltran-Martin, I., and Y. Roca-Puig. 2013. “Promoting Employee Flexibility Through HR Practices.” Human Resource Management 52 (5): 645–74.

Cha p t er 1 4 :  T h e Ro l e o f Pe o p l e a n d In f o r ma t ion in D eliver ing Patient- C entered C are

Black, A. D., J. Car, C. Pagliari, C. Anandan, and K. Cresswell. 2011. “The Impact of eHealth on the Quality and Safety of Health Care: A Systematic Overview.” PLoS Medicine 8 (1): 1–15. Blount, Y., T. Castleman, and P. M. C. Swatman. 2005. “E-commerce, Human Resource Strategies, and Competitive Advantage: Two Australian Case Studies.” International Journal of Electronic Commerce 9 (3): 73–89. Blumenthal, D. 2011. “Wiring the Health System—Origins and Provisions of a New Federal Program.” New England Journal of Medicine 365 (24): 2323–29. Blumenthal, D., and J. P. Glaser. 2007. “Information Technology Comes to Medicine.” New England Journal of Medicine 356 (24): 2527–34. Boselie, P., and J. Paauwe. 2005. “Human Resource Function Competencies in European Companies.” Personnel Review 34 (5): 550–66. Boudreau, J., and E. E. Lawler III. 2014. “Stubborn Traditionalism in HRM: Causes and Consequences.” Human Resource Management Review 24 (3): 232–44. Bowen, D. E., and C. Ostroff. 2004. “Understanding HRM–Firm Performance Linkages: The Role of the ‘Strength’ of the HRM System.” Academy of Management Review 29 (2): 203–21. Brandl, J., and D. Pohler. 2010. “The Human Resource Department’s Role and Conditions That Affect Its Development: Explanations from Austrian CEOs.” Human Resource Management 49 (6): 1025–46. Broadbent, M., and E. S. Kitzis. 2005. The New CIO Leader. Boston: Harvard Business School Press. Brynjolfsson, E., and L. M. Hitt. 2003. “Computing Productivity: Firm-Level Evidence.” Review of Economics and Statistics 85 (4): 793–808. Buchan, J. M. 2004. “What Difference Does (‘Good’) HRM Make?” Human Resources for Health 2 (1): 6. Burke, D., E. Randeree, N. Menachemi, and R. Brooks. 2008. “Hospital Financial Performance: Does IT Governance Make a Difference?” Health Care Manager 27 (1): 71–78. Chadwick, C., J. Super, and K. Kwon. 2013. “Resource Orchestration: CEO Emphasis on SHRM, Commitment-Based HRM, and Firm Performance.” Proceedings of the Academy of Management 1: 10338. Chadwick, C., S. A. Way, G. Kerr, and J. W. Thacker. 2013. “Boundary Conditions of the High-Investment Human Resource Systems–Small-Firm Labor Productivity Relationship.” Personnel Psychology 66 (2): 311–43. Chow, I. H. S. 2012. “The Roles of Implementation and Organizational Culture in the HR–Performance Link.” International Journal of Human Resource Management 23 (15): 3114–32. Crant, J. M. 2000. “Proactive Behavior in Organizations.” Journal of Management 26 (3): 435–62. Curry, A., and G. Knowles. 2005. “Strategic Information Management in Health Care—Myth or Reality?” Health Services Management Research 18 (1): 53–62.

309

310

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Davis, S. 1983. “Management Models for the Future.” New Management 1: 12–15. Dussault, G., and C. Dubois. 2003. “Human Resources for Health Policies: A Critical Component in Health Policies.” Human Resources for Health 1 (1): 1. Everhart, D., D. Neff, M. Al-Amin, J. Nogle, and R. Weech-Maldonado. 2013. “The Effects of Nurse Staffing on Hospital Financial Performance: Competitive Versus Less Competitive Markets.” Health Care Management Review 38 (2): 146–55. Feld, C. S., and D. B. Stoddard. 2004. “Getting IT Right.” Harvard Business Review 82 (2): 72–79, 122. Fernandopuule, R., and N. Patel. 2010. “How the Electronic Health Record Did Not Measure Up to the Demands of Our Medical Home Practice.” Health Affairs 29 (4): 622–28. Fink, L., and S. Neumann. 2007. “Gaining Agility Through IT Personnel Capabilities: The Mediating Role of IT Infrastructure Capabilities.” Journal of the Association for Information Systems 8 (8): 440–62. Frese, M., and D. Fay. 2001. “Personal Initiative: An Active Performance Concept for Work in the 21st Century.” In Research in Organizational Behavior, edited by B. M. Staw and R. M. Sutton, 23, 133–87. Amsterdam, Netherlands: Elsevier Science. Furukawa, M. F., T. S. Raghu, and B. B. M. Shao. 2010. “Electronic Medical Records, Nurse Staffing, and Nurse-Sensitive Patient Outcomes: Evidence from California Hospitals, 1998–2007.” Health Services Research 45 (4): 941–62. Gera, S., and W. Gu. 2004. “The Effect of Organizational Innovation and Information Technology on Firm Performance.” International Productivity Monitor 9: 37–51. Grossman, J. M., A. Gerland, M. C. Reed, and C. Fahlman. 2007. “Physicians’ Experience Using Commercial E-Prescribing Systems.” Health Affairs 26 (3): w393–w401. Guest, D. 2011. “Human Resource Management and Performance: Still Searching for Some Answers.” International Journal of Human Resource Management 21 (1): 3–13. Guest, D., and N. Conway. 2011. “The Impact of HR Practices, HR Effectiveness and a ‘Strong HR System’ on Organizational Outcomes: A Stakeholder Perspective.” International Journal of Human Resource Management 22 (8): 1686–702. Han, J., P. Chou, M. C. Chao, and P. M. Wright. 2006. “The HR Competencies–HR Effectiveness Link: A Study in Taiwanese High-Tech Companies.” Human Resource Management 45 (3): 391–406. Hansen, J. I., and C. A. Thompson. 2002. “Knowledge Management: When People, Process, and Technology Converge.” LIMRA’s MarketFacts Quarterly 2 (2): 14–21. Herzenberg, S. A., J. A. Alic, and H. Wial. 1998. New Rules for a New Economy: Employment and Opportunity in Postindustrial America. Ithaca, NY: Cornell University Press.

Cha p t er 1 4 :  T h e Ro l e o f Pe o p l e a n d In f o r ma t ion in D eliver ing Patient- C entered C are

Himmelstein, D. U., A. Wright, and S. Woolhandler. 2010. “Hospital Computing and the Costs and Quality of Care: A National Study.” American Journal of Medicine 123 (1): 40–46. Hyde, P., C. Harris, and R. Boaden. 2013. “Pro-Social Organizational Behavior of Health Care Workers.” International Journal of Human Resource Management 24 (16): 3115–30. Jaw, B. S., and W. Liu. 2003. “Promoting Organizational Learning and Self-Renewal in Taiwanese Companies: The Role of HRM.” Human Resource Management 42 (3): 223–41. Jones, S. S., P. S. Heaton, R. S. Rudin, and E. C. Schneider. 2012. “Unraveling the IT Productivity Paradox—Lessons for Health Care.” New England Journal of Medicine 366 (24): 2243–45. Kabene, S. M., C. Orchard, J. M. Howard, M. A. Soriano, and R. Leduc. 2006. “The Importance of Human Resources Management in Healthcare: A Global Context.” Human Resources for Health 4 (20): 1–17. Karreman, D., and M. Alvesson. 2004. “Cages in Tandem: Management Control, Social Identity, and Identification in a Knowledge-Intensive Firm.” Organization 11 (1): 149–75. Kaushal, R., and D. Blumenthal. 2014. “Introduction and Commentary for Special Issue on Health Information Technology.” Health Services Research 49 (1, Part II): 319–24. Kellermann, A. L., and S. S. Jones. 2013. “What It Will Take to Achieve the As-YetUnfulfilled Promises of Health Information Technology.” Health Affairs 32 (1): 63–68. Ketkar, S., and P. K. Sett. 2009. “HR Flexibility and Firm Performance: Analysis of Multi-level Causal Model.” International Journal of Human Resource Management 20 (5): 1009–38. Khatri, N. 2015. “Effective Implementation of Electronic Medical Records and Health Information Technologies.” Missouri Medicine 112 (1): 41–45. ———. 2006a. “Building HR Capability in Health Care Organizations.” Health Care Management Review 31 (1): 45–54. ———. 2006b. “Building IT Capability in Health Care Organizations.” Health Services Management Research 19 (2): 73–79. Khatri, N., A. Baveja, S. Boren, and A. Mammo. 2006. “Medical Errors and Quality of Care: From Control to Commitment.” California Management Review 48 (3): 115–41. Khatri, N., and V. Gupta. 2016. “Effective Implementation of Health Information Technologies in U.S. Hospitals.” Health Care Management Review 41 (1): 11–21. Khatri, N., V. Gupta, and A. Varma. 2016. “The Relationship Between HR Capabilities and Quality of Patient Care: The Mediating Role of Proactive Work Behaviors.” Human Resource Management 56 (4): 673–91.

311

312

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Khatri, N., J. Wells, J. McKune, and M. Brewer. 2006. “Strategic Human Resource Management Issues in Hospitals: A Study of a University and a Community Hospital.” Hospital Topics 84 (4): 9–20. Kivinen, T., and J. Lammintakanen. 2013. “The Success of a Management Information System in Health Care—a Case Study from Finland.” International Journal of Medical Informatics 82 (2): 90–97. Klauer, K. 2013. “Meaningful Use—Propping Up the EHR Market Without Improving Quality.” Emergency Physicians Monthly 20 (3): 22. Kohn, L. T., J. M. Corrigan, and M. S. Donaldson. 2000. To Err Is Human: Building a Safer Health System. Washington, DC: National Academies Press. Korczynski, M. 2002. Human Resource Management in Service Work. New York: Palgrave. Lawler III, E. E., and S. A. Mohrman. 2003. Creating a Strategic Human Resource Organization: An Assessment of Trends and New Directions. Stanford, CA: Stanford University Press. Leggat, S. G., T. Bartram, and P. Stanton. 2011. “High Performance Work Systems: The Gap Between Policy and Practice in Health Care Reform.” Journal of Health Organization and Management 25 (3): 281–97. Lu, Y., and K. Ramamurthy. 2011. “Understanding the Link Between Information Technology Capability and Organizational Agility: An Empirical Examination.” MIS Quarterly 35 (4): 931–54. MacDuffie, J. P. 1995. “Human Resource Bundles and Manufacturing Performance: Organizational Logic and Flexible Production Systems in the World Auto Industry.” Industrial and Labor Relations Review 48 (2): 197–221. Mandl, K. D., and I. S. Kohane. 2012. “Escaping the EHR Trap—the Future of Health IT.” New England Journal of Medicine 366 (24): 2240–42. McAfee, A., and E. Brynjolfsson. 2008. “Investing in the IT That Makes a Competitive Difference.” Harvard Business Review 86 (7–8): 99–107. McBride, A., and S. Mustchin. 2013. “Crowded Out? The Capacity of HR to Change Healthcare Work Practices.” International Journal of Human Resource Management 24 (16): 3131–45. McClean, E., and J. C. Collins. 2011. “High-Commitment HR Practices, Employee Effort, and Firm Performance: Investigating the Effects of HR Practices Across Employee Groups Within Professional Service Firms.” Human Resource Management 50 (3): 341–63. McCormick, D., D. H. Bor, S. Woolhandler, and D. U. Himmelstein. 2012. “Giving Office-Based Physicians Electronic Access to Patients’ Prior Imaging and Lab Results Did Not Deter Ordering of Tests.” Health Affairs 31 (3): 488–96. Mithas, S., N. Ramasubbu, and V. Sambamurthy. 2011. “How Information Management Capability Influences Firm Performance.” MIS Quarterly 35 (1): 237–56. Murphy, G. D., and G. Southey. 2003. “High Performance Work Practices: Perceived Determinants of Adoption and the Role of the HR Practitioner.” Personnel Review 32 (1): 73–92.

Cha p t er 1 4 :  T h e Ro l e o f Pe o p l e a n d In f o r ma t ion in D eliver ing Patient- C entered C are

Nanji, K. C., J. M. Rothschild, C. Salzberg, C. A. Keohane, K. Zigmont, J. Devita, T. K. Gandhi, A. K. Dalal, D. W. Bates, and E. G. Poon. 2011. “Errors Associated with Outpatient Computerized Prescribing Systems.” Journal of American Medical Informatics Association 18 (6): 767–73. O’Malley, A. S., J. M. Grossman, G. R. Cohen, N. M. Kemper, and H. H. Pham. 2009. “Are Electronic Medical Records Helpful for Care Coordination? Experiences of Physician Practices.” Journal of General Internal Medicine 25 (3): 177–85. Pearlson, K. E. 2001. Managing and Using Information Systems: A Strategic Approach. New York: Wiley. Peppard, J. 2010. “Unlocking the Performance of the Chief Information Officer (CIO).” California Management Review 52 (4): 73–99. Pines, J. M. 2013. “Making the Most of Your EHR.” Emergency Physicians Monthly 20 (3): 22. Quinn, R. W., and W. Brockbank. 2006. “The Development of Strategic Human Resource Professionals at BAE Systems.” Human Resource Management 45 (3): 477–94. Ross, J. W., C. M. Beath, and D. L. Goodhue. 1996. “Develop Long-Term Competitiveness Through IT Assets.” Sloan Management Review 38 (1): 31–45. Schneider, B., and S. S. White. 2004. Service Quality: Research Perspectives. Thousand Oaks, CA: Sage Publications. Sittig, D. F., and H. Singh. 2012. “Electronic Health Records and National PatientSafety Goals.” New England Journal of Medicine 367 (19): 1854–60. Stanton, P., S. Young, T. Bartram, and S. G. Leggat. 2010. “Singing the Same Song: Translating HRM Messages Across Management Hierarchies in Australian Hospitals.” International Journal of Human Resource Management 21 (4): 567–81. Towler, A. D., V. Lezotte, and M. J. Burke. 2011. “The Service Climate—Firm Performance Chain: The Role of Customer Retention.” Human Resource Management 50 (3): 391–406. Townsend, K., S. A. Lawrence, and A. Wilkinson. 2013. “The Role of Hospitals’ HRM in Shaping Clinical Performance: A Holistic Approach.” International Journal of Human Resource Management 24 (16): 3062–85. Townsend, K., and A. Wilkinson. 2010. “Managing Under Pressure: HRM in Hospitals.” Human Resource Management Journal 20 (4): 332–38. Tremblay, M., J. Cloutier, G. Simard, D. Chenevert, and C. Vandenberghe. 2010. “The Role of HRM Practices, Procedural Justice, Organizational Support and Trust in Organizational Commitment and In-Role and Extra-Role Performance.” International Journal of Human Resource Management 21 (3): 405–33. Tyagi, R. K., L. Cook, J. Olson, and J. Belohlav. 2013. “Healthcare Technologies, Quality Improvement Programs and Hospital Organizational Culture in Canadian Hospitals.” BMC Health Services Research 13: 413. Ulrich, D. 1996. Human Resource Champions. Boston: Harvard University Press.

313

314

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Woodrow, C., and D. E. Guest. 2014. “When Good HR Gets Bad Results: Exploring the Challenge of HR Implementation in the Case of Workplace Bullying.” Human Resource Management Journal 24 (1): 38–56. Wright, P. M., B. B. Dunford, and S. A. Snell. 2001. “Human Resources and the Resource-Based View of the Firm.” Journal of Management 27 (6): 701–21. Yeh, C. H., G. G. Lee, and J. C. Pai. 2012. “How Information System Capability Affects E-Business Strategy Implementation: An Empirical Study in Taiwan.” Business Process Management Journal 18 (2): 197–218.

CHAPTER

VALUATION AND FINANCING OF HEALTHCARE SERVICES AND INFORMATION TECHNOLOGY INFRASTRUCTURE

15

Kalyan S. Pasupathy and Gordon D. Brown

Learning Objectives After reading this chapter, you should be able to do the following: • Understand the structure for financing a health system. • Describe the challenges with existing financing mechanisms. • Conceptualize the valuation of healthcare services and information technology infrastructure. • Understand how healthcare financing has affected the clinical function. • Discuss and assess current financing and organizational models as they address issues of cost, quality, and continuity of clinical services.

Key Concepts • Beveridge, Bismarck, national health insurance, and out-of-pocket financing • Diagnosis-related group versus per diem rate payment • Patient-centered medical home structure • Accountable care organization structure • Cost-based, capitation-based, and value-based reimbursement • Valuation • Private and public investments in information technology

Introduction Healthcare is provided and financed through different systems structures across the world. There are four major models of care provision and financing: 315

316

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Beveridge, Bismarck, national health insurance, and out of pocket. Each model is employed in the United States for specific population groups. Active duty and veteran military members are under the Beveridge model, in which hospitals with salaried physicians are owned and financed by the federal government and the information system is centrally planned and managed. Most of the workforce is under a framework similar to the Bismarck model, in which employees and their employers pay into a pool to finance healthcare services and the information system is created and maintained by individual institutions. Medicare beneficiaries are under the national health insurance model, in which healthcare services are financed by state and federal governments but purchased from local private or public hospitals or clinics. People who are unemployed, not old enough to qualify for Medicare, not affiliated with the US armed forces, or not living below the poverty line to qualify for Medicaid are typically under the out-of-pocket model of healthcare financing. Healthcare services financed through insurance is often referred to as a three-party system, in which the patient (first party) receives the service from the provider (second party) who, in turn, is paid or reimbursed by the payer (third party). Often, the third party is a health insurance plan that provides and underwrites the coverage for a large group of people or members. The insurance premiums paid by all members are used to cover the few who claim reimbursement. The structures of these financing mechanisms have enabled and caused costs to spiral out of control. The cost of healthcare is higher in the United States than in other developed nations, with outcomes that are not proportional to the cost. Healthcare is technologically advanced in the United States, but fragmentation and lack of coordination pervade the system. This chapter walks the reader through healthcare financing and payment, focusing on key areas in which financing conflicts with the ability of integrated information systems to facilitate the coordination and quality of services. This chapter explains how coordination and communication (information flow and knowledge transfer) add value and how financing can enable such coordination to occur, realizing the potential of information technology (IT).

Financing Models Changes in the structure of the clinical function have occurred primarily as the result of changes in healthcare financing. The complex dynamics between how healthcare services are financed and how they are structured have been the traditional focus of much research and policy analysis and will continue to determine the nature of the health system and how quickly it is transformed around the IT logic. Investment in IT can enable improved financial

Ch a p t er 1 5 :  Va l u a ti o n a n d Fi n a n c i n g o f H e althc are Ser v ic es and IT Infrastr uc ture

performance, but many variables have to be sorted out to isolate this relationship (Wang and McLeod 2018). We explore how healthcare financing facilitates a transformed health system. The controversy and chaos caused by the passage of the Affordable Care Act (ACA) of 2010 (US Congress 2010) exemplify the complexity of and resistance to change in healthcare financing. The ongoing focus on changes in health insurance and access to healthcare services detracts from the policy debate on how the health system can be transformed by IT to improve quality and efficiency. The dynamic modeling of health systems (chapter 5) explores systems structures by analyzing the effect of individual functions and the ways they can be aligned to optimize quality, efficiency, and continuity and to tailor services to individuals. The logic needed to achieve this optimal state is IT, but IT must be considered from the perspective of health systems informatics, a transformative technology, and not from the perspective of automating traditional structures and processes. Models of financing that have shaped the evolution of the clinical function, healthcare organizations, and the health system are explored here.

Cost-Based Reimbursement Cost-based reimbursement was the initial form of insurance and continues to be dominant today. In the 1920s, the first private health insurance plans altered the clinical process by identifying specific services covered and the location for providing those services. Plans were based on a traditional insurance model of measuring and distributing risk among members of a defined population. Insurance paid for primarily high-cost surgeries and other complex inpatient interventions. This method of financing influenced where physicians treated patients and thus altered the doctor–patient relationship. Known as a third party, insurance plans were initially resisted by physicians but then were reluctantly accepted because patients and families needed help to pay for high hospital costs resulting from the growth in medical technology. The logic to this insurance model is that high-cost services pose the greatest financial risk to patients. Insurance plans de facto gave incentives to doctors to treat patients in hospitals because costs were covered by insurance. The incentives for inefficiency inherent in the insurance model continue to characterize today’s health system and were the main motivation behind the ACA. The cost of insurance continues to drive much of the policy debate, sometimes overriding the attention given to efficiency, quality, continuity, and patient-centered care. The role of federal and state governments as funders increases not only the availability of healthcare services but also unit costs and rates of utilization. The problem, however, is not the public funding of insurance but the funding of a dysfunctional system. The accounting and finance functions are ideally structured for electronic systems because finance has an extensive lexicon of standardized terms and measures that has been in existence globally for nearly a hundred years. Thus,

317

318

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

the finance department became the strategic driving force for early (1960s) computer applications in healthcare organizations. Later applications were in the payroll and personnel departments, which were both administrative functions with well-established standardized terms/language and large databases. This functional structure of information was incompatible with the development of the electronic health record (EHR), which is based on the logic of clinical decisions and processes (Wohlner 2017): The Gap Between Finance and Clinical Standardization. Standardized double-entry bookkeeping based on a system of debits and credits, ledgers, and journals was developed in 1492 and still used today. Cost accounting systems were spurred on by the industrial revolution in the 1700s. The impetus for such a system was to enable corporations to be accountable to individuals who had a vested interest in the corporation but were not part of management, i.e. the shareholders.

Capitation-Based Reimbursement Capitation-based financing has existed in the United States for decades and continues as a form of finance for some systems. It grew rapidly in the 1970s with the passage of the HMO Act of 1973 (Morrison and Luft 1990). Health maintenance organizations shifted financial risk from an illness-based cost reimbursement to a wellness-based population model. This model required both hospitals and clinicians to share risks with insurance companies for the health of a defined patient population and offered providers incentives to (1) keep members healthy and (2) develop an efficient and effective healthcare delivery system. Capitation-based financing provided incentives for prevention, efficiency, coordination, and quality but conflicted with the traditional roles of hospitals, health professionals, and patients by focusing on prevention and introducing controls on how healthcare services were consumed (Scott et al. 2000): Standardization and measurement of clinical process and outcomes became the nemesis of health professions and healthcare organizations, as did accountability for outcomes and coordination of clinical processes. Providers’ reluctance to take on these tasks resulted in insurance companies, as financial agencies, to step in and introduce control mechanisms (e.g., preadmission and preauthorization procedures, clinical gatekeepers) to “bludgeon” clinicians and hospitals to [help] manage risk. Healthcare organizations such as Kaiser Permanente [had, before the HMO Act,] embraced the assumptions of capitation and demonstrated the ability to transform clinical processes integrating specialties, institutions, and practice locations based on outcomes measures. Outcome measures linked clinical care with population-based health providing incentives to keep the enrolled population healthy.

Ch a p t er 1 5 :  Va l u a ti o n a n d Fi n a n c i n g o f H e althc are Ser v ic es and IT Infrastr uc ture

Capitation financing heightened the role of healthcare corporations and reduced the individual clinician’s decision-making autonomy, a radical change that was rejected by most physicians. Being accountable collectively for a population’s health, clinical quality, and operational efficiency required focusing on health outcomes, process–outcome relationships, and the coordination of care, which were feasible only through a profound restructuring of the health system. Capitation-based financing demonstrated that it was possible to align financing, organizational structure, and the clinical function to focus on health maintenance and the efficient use of medical services. It failed because of its inability to transform the system. Many of capitation’s qualities are enabled by an information-driven system, but the system must be restructured around these qualities. In principle, the basic design of accountable care organizations includes the logic of capitation but also retains the fee-for-service logic.

Value-Based Reimbursement In the early 1980s, hospital reimbursement strategies led by Medicare and Medicaid changed from a pure cost-based scheme to case-based diagnostic-related groups (DRGs). Healthcare organizations were given incentives to redesign cost measurement and information system structures to assign costs by diagnosis rather than by function. This was not a trivial task given that few hospitals had adequate accounting systems to measure cost (activity-based accounting) by clinical service. Reimbursement by Medicare and Medicaid became the financial impetus that moved the health system toward standardizing inpatient services based on DRGs. Standardization and costing of clinical services spread quickly as a basis for increasing efficiency without decreasing quality. Value-added payment based on classifying and managing by diagnostic category has emerged in recent years as bundled payments from the Centers for Medicare & Medicaid Services. However, currently structured practices that aim to better coordinate care have difficulty with the payment mechanisms (Lewis 2016). Viewing the clinical processes as integrated or bundled requires rewards to be based on outcomes and not processes. Value-based reimbursement will continue to be the direction followed by US policymakers. Bundled payments transcend the hospital to include the entire care process but by specific diagnosis and thus differ from capitation-based financing. In this regard, bundled payments remain based on diagnosis and not on health maintenance. The federal government has thus moved from the sustainable growth rate model to the merit-based incentive payment system (Bosko and Hawkins 2016). Under the merit-based incentive payment system, physicians are reimbursed for achieving target rates and levels of improvement, reducing use of resources, delivering quality, improving processes, and using EHR effectively.

319

320

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Value-based reimbursement does effectively use clinical outcomes drawn from the EHR. It is a significant recognition and use of IT and links the financing and information system functions. Case-based reimbursement also rewards the development of teams of health professionals. Process measures are still used as surrogates for outcomes and need to be refined. They also cause appropriate physician disagreement (Gupta, Karst, and Mendelson 2016).

Lack of Coordination and Information Sharing The current financing systems in healthcare tend to reinforce the fragmentation in the delivery of services through fragmented payment methods. Each type of provider is paid for separate services given to a patient, and payment is often based on volume of services. For example, the dominant form of paying physicians is a capped fee schedule, wherein a maximum fee is established for each billable service (although this fee varies by payer and may simply be a discounted usual, customary, or reasonable charge). Resource-based relative value scales often are the foundation for the fees set, but they are still a feefor-service system based on identifiable, distinct units of service and thus link provider incomes to volume and intensity of service. A fee-for-service system rewards a greater volume of services and services for which the payments are high relative to the resources used to produce the services. In addition, billable services such as patient education, care coordination, and comprehensive care planning tend to be underused—a circumstance known as voltage drop (Chung and Schuster 2004). In some cases, a number of related services are bundled into a single product and priced as a package of services, but the incentive is given for providing more of the bundled service. For ancillary services that are not a part of a bundled product, the incentive is reserved for higher volume and greater intensity. Hospitals are typically paid either a set amount for each admission (based on DRG) or a fee for each day a patient is in the hospital (per diem rate). Under either system, patients must be admitted for the hospital to receive revenue. Under the DRG system, there is an incentive to use resources efficiently while the patient is hospitalized and to discharge the patient as soon as possible, but no continuity of care is encouraged before or after discharge. Under the per diem system, there is an incentive to extend the length of stay of the patient within a necessary or “acceptable” limit. Exhibit 15.1 shows the configuration of the existing healthcare delivery system and its three major provider types— primary care practice, specialists, and hospitals and health systems. Each entity has three components—access, utilization, and payment. Insurance plans are at the center of the system because they insure and finance the three provider types. As shown in exhibit 15.1, the entities are fragmented.

Ch a p t er 1 5 :  Va l u a ti o n a n d Fi n a n c i n g o f H e althc are Ser v ic es and IT Infrastr uc ture

Utilization

Patients

Patients

Access

Access

Primary care practice

Insurance

Specialists Utilization

Payment

Patients

Access

Payment

Hospitals and health systems

Payment

Utilization

Redesigning Structure and Financing Models of healthcare services delivery have emerged with a goal of better aligning finance and organizational structure with the power of IT to improve efficiency, quality of services, and population health. The effect of financial models cannot be studied separately from organizational design and IT because they are highly interrelated and models are specific to designs. Here, we explore financing in the context of different designs. Does design optimize system performance or financial rewards? The critical point of analysis is to consider alternative designs and models that can facilitate or enable optimal performance. Different financial schemes and related structural designs are explored. Outcome measures generally include cost reduction, increased quality, and improved population health (Berwick, Nolan, and Whittington 2008). Stakeholders in the system must be financially rewarded at a level that is sufficient to attract capital and human resources (with a focus on future human resources models). System performance is measured against alternative finance, design, and IT models to produce an optimal system effect. The financing mechanisms analyzed may include fee-for-service, types of bundled payments, and capitation (Tanenbaum 2017). The design of the health system should not be based on optimizing financial return but rather on optimizing system performance. The challenge of financing models

321

EXHIBIT 15.1 Configuration of an InsuranceCentric, Fragmented Healthcare Delivery System

322

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

is ensuring that they reward system components in a manner that results in superb system performance.

Medical Home Financing It is difficult to generalize about the financial models for new organizational designs such as medical homes. Even though medical homes have adopted various financial reform models, such as bundled payments with pay-for-performance incentives and capitation based on per-member-per-month, the dominant payment model for patient-centered medical homes continues to be fee-forservice. Designed around the assumption of value-based reimbursement, medical homes thus include a combination of payment models within a single system. Traditional fee-for-service payments are the basis for many services, particularly those consultations provided outside the medical home “system.” In addition, extra compensation is paid to cover the cost of care coordination. If capitationbased financing is followed, extra compensation might also be provided above a per-member-per-month fee. Additional financial incentives may be provided if specific quality targets are achieved. Thus, medical homes are a combination of traditional fee-for-service models and value-based financing, rewarding the coordination of care and health maintenance in a team structure. Bundled payments are a form of reimbursement based on an illness episode, and financial incentives are given for managing costs by reducing inpatient care and physician consultations. Thus, they are less oriented to wellness and health maintenance of a population. The episodic nature of the payment mechanism is an inherent limitation of medical homes. Value is added by improving the coordination of care through the effective use of teams (which include dietitians, nurses, physical therapists, and rehabilitation services staff), with the expectation that costs can be reduced while maintaining or improving quality. If the organization is able to achieve target savings, it shares in the savings as an incentive. Two principles of the medical home (chapter 4) are most relevant to bundled-payment financing (American Academy of Family Practice et al. 2011). First, care coordination and integration across all elements of care delivery must be present in (not outside) the medical home. This coordination must be supported by, among other things, appropriate IT and a health information exchange (HIE) system. Second, a medical home needs to recognize value-added elements on the basis of the following payment framework (American Academy of Family Practice et al. 2011): • Reflect value added by physician and other care management staff even if the care falls outside the face-to-face visit. • Pay for services associated with coordination of care both within the specific practice and between various providers (e.g., primary care, specialist, hospital).

Ch a p t er 1 5 :  Va l u a ti o n a n d Fi n a n c i n g o f H e althc are Ser v ic es and IT Infrastr uc ture

• Support the adoption and use of health IT. • Provide enhanced communication access, such as secure e-mail and phone consultation. • Recognize the value of physician work associated with remote monitoring of clinical data using technology. • Allow for separate fee-for-service payments for face-to-face visits, an incentive that overrides coordination in evidence-based decision making. • Recognize case-mix differences in the patient group. • Allow physicians to share in the savings from reduced hospitalizations and other efforts. • Make additional payments for achieving measurable and continuous quality improvements. For this framework to be operational, sharing of information—including clinical, quality of care, outcomes, and other operational data—among all stakeholders is required. This framework also entails negotiating and distributing reimbursement on the basis of value-added services, such as face-to-face visits, clinical consultations, collaborative practice (through shared clinical guidelines and protocols), and case coordination and management. The medical home model is a good example of the power of IT when information, structure, and financing are aligned, and it can serve as a model for delivering care as well. However, this model is still considered a relatively closed system, with multiple payment mechanisms deployed to address specific service dimensions. When patients are traveling or go outside the medical home for referrals, sharing of information and coordination of services are diminished. Although they emphasize use of clinical information systems and the payfor-performance scheme, the National Committee for Quality Assurance and other organizations likely understate the importance of other factors that are essential to quality and efficiency. These factors include patient education, training to improve patient self-management of disease, collaborative physician networks, physician–patient partnerships, multiprofessional teamwork, and a shared commitment to quality and efficiency—all of which are principles of a medical home. The development of high-performing work teams, a culture dedicated to patientand family-centered care, and mission and values that support these efforts is an important part of creating the structure and function of a medical home. Medical homes align a number of elements of information, organizational design, clinical processes, and financing. Hospitals and physicians see improvements in the use of emergency services, specialist referrals, hospital beds, laboratory services, inpatient pharmacy, and other services (Salzberg et al. 2017). However, medical homes are still relatively closed systems that have multiple and conflicting payment mechanisms, some of which will prevent

323

324

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

further system improvement if the system is not transformed. Medical homes are designed to achieve a new equilibrium of financial incentives with efficiency and quality. They must continue to evolve their financial models to maximize quality, coordination, and efficiency with shifting financial reward structures. The patient-centered medical home model will likely become more common and play an important role in healthcare services delivery reform.

Accountable Care Organization Financing The accountable care organization (ACO) was designed with integrated information systems, payment reform, and population health management in mind. Under the ACO, a group of stakeholders—generally including physicians, other health professionals, hospitals, and health plans—voluntarily forms to provide coordinated, high-quality care for a population of patients. Information exchange is a required and key element in shared consultations and coordinated services between and among payers and providers within a closed regional system. Payment for services is primarily on a fee-for-service basis, with a financial incentive or gainshare for reducing costs and improving quality (Tanenbaum 2017). The incentive can be increased by setting higher efficiency and quality targets and thus taking on more risk. Because an ACO is responsible for the overall health of the population it manages, it is required to share not only EHRs but also quality-of-care data and patient-reported outcomes over an HIE system, which constitutes a limited form of information sharing (Casalino, Erb, and Shortell 2015). HIE is achieved, but shared consultations are not supported in an ACO because of the limitations of collaborative practice (i.e., sharing clinical knowledge and financial reward). Because an ACO frequently brings together more than one medical home or range of providers, it is often referred to as a medical neighborhood (Fisher 2008) but does not rise to the level of a community of practice (Hermanrud 2017). The concept of a medical neighborhood includes not only the traditional stakeholders (clinicians and hospitals) but also community-based and other organizations that offer social support. Such broad focus is vital to managing certain conditions, such as behavioral health, especially when the acute care setting lacks the capacity to provide timely and appropriate care. ACOs have been found to consistently improve performance over previous disaggregated systems of institutions and information systems by reducing costs and improving quality and continuity. However, improvement over a very disjointed and inefficient system is not a high bar, and acceptance of such a design is not ideal (Ouayogodé, Colla, and Lewis 2017). The effectiveness of ACOs in addressing and improving population health is a high bar that most systems do not reach, and current financing models serve as an impediment to this outcome, regardless of IT or ACO design. An ACO would likely have to adopt some form of capitation financing to align it with population health. This

Ch a p t er 1 5 :  Va l u a ti o n a n d Fi n a n c i n g o f H e althc are Ser v ic es and IT Infrastr uc ture

will be a big hurdle for the current system given the complexity of measuring and changing health risk in a defined population. For example, ACOs could be rewarded only for the changes that they achieved, a complex measure and a more complex intervention. Incentives for patients (or some population of patients) and for the system would have to be devised. A wide range of ACO designs and performance makes it difficult to reach firm conclusions regarding ACO financing models and performance, although emerging evidence supports various designs that can guide future policy direction. Financing for embedded care coordinators, the availability of local and regional health information systems, and timely feedback of performance data have correlated with improved quality and reduced costs among ACO models (D’Aunno et al. 2018). Interestingly, among the significant factors related to performance in ACOs, there are more design (structural) features than differences in financing models. ACOs can be considered a step in the process of developing an integrated health information network.

Strategic Implications of Structural Changes Redesigning financing requires an understanding of how it can be changed to support an improved or idealized system design. To what extent are such changes effective in improving quality-of-care outcomes, the efficiency of operations, and population health? Do improvements in quality justify the costs? Hospitals and health systems need to identify the required investments in capital, time, and resources before getting involved in a redesign. In an ACO, for example, the capital investment in IT by hospitals and clinics in rural and isolated areas is required because the coordination of services needs to focus on the system of care and not regional specialty institutions. Structural changes need to align incentives across payment methods and care settings. The payment system should encourage the production of higher-value care, not simply more care, and go beyond the individual episode of care to improve care over time and reduce fragmentation of care. Incentive payments should be linked to quality, efficiency, and patient experience, focusing on the coordination of services to ensure the appropriateness of care and care setting as well as encouraging prevention to avoid the need for costly services. Exhibit 15.2 illustrates the configuration of a medical home and the integration among the three provider types (in contrast with exhibit 15.1). The medical home encompasses not only the coordination of care but also the sharing of knowledge between specialists and primary care practice as well as access of patient information by both specialists and hospitals. As shown in exhibit 15.2, specialists and hospitals have access to and use patient information and shared clinical knowledge. Therefore, they are components of the integrated system and should undergo performance evaluation and be subject to the payment structure. Without their involvement, the system becomes

325

326

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

EXHIBIT 15.2 Configuration of an InformationCentric, Coordinated Medical Home

Patients

Access Utilization

Patients

Access

Primary care practice Payment

Specialists

Payment

Utilization

Payment

Utilization

Information on patients Kn0wledge Information on patients Access

Hospitals and health systems

Patients

Insurance

yet another fragmented effort at transforming healthcare and is poised to fail, leaving broken processes and adding more cost to the existing system. Beyond aligning service reimbursement methods to reward efficiency and quality is a more basic issue—investing in health IT systems that enable such exchange to take place. This is not a failure of the service reimbursement system but of a system investment strategy, discussed later in this chapter. Currently, there is no design or investment strategy for developing an integrated information architecture beyond a regional medical home or ACO structure (Dixon, Colvard, and Tierney 2015). This is a systems failure and cannot be understood or resolved with incremental strategies, no matter how well they are conceived. We are left with regional systems that do not function very well, and we lack a design and investment strategy for integrating services across regional systems.

Valuation of Healthcare Services Assigning value to healthcare services is complex in all societies because it assumes (to various degrees) that an individual’s health has intrinsic and not

Ch a p t er 1 5 :  Va l u a ti o n a n d Fi n a n c i n g o f H e althc are Ser v ic es and IT Infrastr uc ture

just economic value. This intrinsic value supports the arguments for a service economy (as opposed to a pure market economy, in which value is based on individual preferences and demand is based on price and ability to pay). In most countries, healthcare services are primarily the responsibility of the public sector, but in the United States, healthcare is provided by a combination of the public, private nonprofit, and for-profit sectors. Each of these sectors brings different assumptions about value, making the valuation process more complex—both in the United States and in other countries that are expanding their private sector investments in healthcare (chapter 11). The value premise in a service economy is achieved by public sector regulations and controls, such as restrictions on risky behavior (e.g., wearing a seatbelt); direct provision of clean water, sanitation, and other services; investment in infrastructure such as IT; and reimbursement for some level of clinical services. The United States has followed a unique course by providing social goods—particularly hospital services—through the private nonprofit sector and by generally allowing these private institutions to be exempt from taxes, depending on their social contribution. The basis for IT investment includes decisions on what services are to be made available and what value contribution IT can bring to these services. This has resulted in IT being conceptualized as a support service within the structure of a disaggregated private–public system. Historically, healthcare services in the United States have been provided and paid for primarily by the private sector, and the public sector has made considerable investment in facilities, research, manpower, and equipment. The market economy has defined and continues to influence the structure and culture of the health system. Private hospitals and clinics that function under the assumptions of a market economy undervalue some services but respond to financial incentives and other controls from state and federal governments. The public sector, meanwhile, subsidizes and stimulates the private sector and has shifted its assumptions from a market economy to a service economy through the increasing level of direct financing of healthcare services. The ongoing policy debate in the United States and other nations addresses how healthcare services should be valued and how the health system (including IT) should be designed and financed. The value premise is shifting in the United States, partly because of the high cost of care and poor population health outcomes but maybe primarily as a result of IT-enabled patient involvement in the caregiving process. In this regard, IT can be considered as a transformational technology. The responsibility for making insurance and thus healthcare services available has shifted to increased public sector financing through Medicare and Medicaid payments and the ACA. In 2009, the federal government enacted the Health Information Technology for Economic and Clinical Health (HITECH) Act to stimulate and augment private sector investment in IT. As a major purchaser of healthcare services, the public sector measures and assumes accountability for the quality, continuity, and cost of services

327

328

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

because it is expending public funds. The private sector, on the other hand, engages in measurement and accountability for the purpose of the market and competition, and it has notable advantages in timeliness of service availability, which is what consumers desire in a “market-based system.” The US health system has always deployed a layered strategy of direct public sector provision of services and purchase of services from the private sector, and it has operated under a robust market economy with private investment and consumption. Even with a strong market economy, however, many services are provided by nonprofit healthcare organizations and highly professionalized private clinicians dedicated to patients and communities, and these services are balanced with profits. The health system is thus a complex environment that negates any single set of assumptions about valuation and investment. Healthcare organizations and professionals function within this market complexity, and so do IT systems. The integrated manner in which IT must operate further complicates how the optimal investment in IT capacity should be made. The concept of dynamic capabilities of organizations is a framework for making such an investment decision from a market or modified market perspective.

Valuation of IT Infrastructure IT valuation and investment in the US health system are conceptually complex because several dimensions that add value to healthcare organizations need to be aligned and not diminished. The development of IT infrastructure has followed a three-stage trajectory: (1) IT as an operational strategy, (2) IT as an enterprise/market strategy, and (3) IT as a public/social good. Each stage serves different functions, but all must work together. An essential strategy is to design an infrastructure that supports individual clinicians and departments and at the same time anticipates a fully integrated environment in which patients can access medical information wherever they are located and health professionals can collaborate by drawing on shared clinical knowledge. Such an infrastructure can be tailored to meet everyone’s needs (exhibit 15.3). The level and source of IT investment depend on the value contribution of each strategy; the designs are different but must be aligned to meet all three purposes. Individual hospitals and clinics have historically concentrated their IT investment on transforming and aligning clinical and business functions (an operational strategy). Most organizations are now focusing on an enterprise/ market strategy and are finding that this strategy requires a different design. ACOs and other regional systems are investing in enterprise exchange, a type of HIE dedicated to the regional system. A regional IT system adds value by coordinating care and enabling the system to provide services more efficiently (Vest and Kash 2016). This type of HIE is complicated by the simultaneous

Ch a p t er 1 5 :  Va l u a ti o n a n d Fi n a n c i n g o f H e althc are Ser v ic es and IT Infrastr uc ture

People

Patients Departments and clinicians Hospitals ACOs Nation/world

pursuit of both an institutional/operational design and an ACO design. ACOs are currently focused on the exchange of patient information but, in the future, will add value by exchanging clinical knowledge through clinical consultations, evidence-based guidelines and protocols, and potentially large databases to be mined. Such a system will require ACOs to be designed more formally as a regional enterprise, making the regional strategy an enterprise strategy. Knowledge transfer benefits all patients in the community but adds value to the system only for those paid through some form of capitation, where all doctors and institutions share the risk and reward. There is an opportunity cost to clinicians for patients outside the capitation-based financing scheme, given that knowledge is transferred without all parties being rewarded for their value-added contribution. The conflicting motivations make the clinical process complex, limiting the development of a consistent IT design and investment strategy. Thus, IT is underdesigned, undervalued, and underinvested in existing systems because the health system itself is rewarded on underperformance. Another type of HIE is direct exchange, where limited patient information is exchanged directly between clinicians by some means of secure mail or between institutions with compatible IT systems. This type is consistent with the traditional model of medical consultation that is based on the value contributions of individual clinicians. It maintains many of the features of feefor-service reimbursement and is consistent with the traditional design of a highly inefficient and ineffective health system.

329

EXHIBIT 15.3 Whom Does IT Serve?

330

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

An HIE that serves a public good has been referred to as a community or public information exchange (Vest and Kash 2016). This strategy is currently not a characteristic of ACOs outside the patient population they directly serve. A community exchange is a dedicated HIE that links disparate ACOs, hospitals, and regional systems. It enables sharing of patient information and clinical knowledge across states and regions, and potentially across nations. Individual ACOs are not likely to design and invest in a community exchange under the current reimbursement models because doing so would siphon away financing and add costs that the return on investment cannot justify. Investment in a community exchange would likely require public funding because this type of exchange realizes a social benefit. However, such investment should enable a community/public information exchange and not subsidize an operational or enterprise strategy. Are current ACOs and other regional systems designed as “sand” or “cubes”? IT designers know that when a CEO instructs them to develop an IT infrastructure to look like a box, once the box has been created, the CEO will ask them to turn it into a globe. How do you make a box look like a globe? By filling the box with sand and not cubes. In other words, make the design flexible so that it can easily be changed as organizational functions change. IT as an operational strategy, an enterprise/market strategy, and a public/ social good consists of three types of infrastructure that perform distinct functions, embody different values, and follow different strategies. The meaningful use requirement of the HITECH Act does not specify or even consider how information systems should be designed to enable a public/social exchange and thus justify the investment of public funds. Not only is this a poor economics strategy, but it also distorts the understanding and potential of public sector investment in IT. The federal government’s current policy interests revolve around the government’s role as (1) a public insurer that pays for medical services and thus demands providers to exercise sufficient accountability for quality and efficiency; and (2) an investor in services that benefit the public at large, such as medical research, population health, and clean-air standards. The private health insurance sector’s lack of commitment to investing in greater efficiency and disease prevention (resulting in the excessive cost of services) may be explained, in part, by the ability of insurers to pass costs on to patients, corporations, and the government. As a public insurer, the government has a responsibility to integrate clinical information so that patients can access their medical records and other health information regardless of who is providing the healthcare services. This integration has the potential to improve effectiveness, efficiency, and satisfaction. In addition, patient access to health records and other health information is valued in its own right, beyond that justified and verified by improved efficiency and quality. Access to information constitutes

Ch a p t er 1 5 :  Va l u a ti o n a n d Fi n a n c i n g o f H e althc are Ser v ic es and IT Infrastr uc ture

a new dimension of quality, one valued in itself but already well established in other countries. This is the dominant logic that justifies the federal government’s IT investment, which is focused on developing a standardized technology architecture comprising a national HIE. It justifies the logic (if not the practice) of the HITECH Act by addressing private healthcare organizations’ underdesign of and underinvestment in IT, and has prompted the commitment of $27 million (disbursed over ten years) to stimulate and underwrite such investments. This underinvestment by the private sector stems from the assumptions of a cottage industry filled with small, independent, and marketdriven clinics and hospitals that use a dedicated technology architecture. The IT investment by the federal government should galvanize the transition of healthcare to a more integrated system focused on coordination of care, clinical quality, and patient centeredness.

Private Sector IT Investment A current major challenge for healthcare organizations is aligning enterprise strategy with their IT infrastructure (Morteza et al. 2017). Part of this challenge is first aligning the business and clinical functions in the healthcare corporation (chapter 4). As a result of this internal focus, IT in hospitals and clinics developed primarily as a technology that supported operations and not as a strategic asset. Healthcare organizations achieve greater alignment of the clinical and business functions by applying clinical decision support. Hospitals and clinics adopted and maintained their own IT system, as an operations technology, according to organizational needs. As a default strategy, IT systems were created to be incompatible with other systems. Currently, IT is being developed as an enterpise strategy—in other words, as a strategic asset. IT becomes a strategic asset if it enables the goals and strategies of the corporation. Once the appropriate levels and functions of the organization have been engaged in developing the enterprise strategy, value and risk can be articulated. The focus is prospective and visionary and based on markets, competition, consumers and their wants, clinical knowledge (translational research), and IT. Valuation is not based on traditional organizational or IT structures, professional domains, or reimbursement strategies; these considerations are obsolete and viewed as constraints. Instead, valuation is based on the dynamic variables that position the organization for the future and that contribute to corporate or clinical strategy. The strategic vision of the organization is more critical to this process than technical or clinical knowledge or experience per se. Leading healthcare organizations and systems have increasingly recognized information and knowledge as a valuable strategic asset. This recognition will bring about profound disruptions, including changing the competencies of those in the executive suite and aligning corporate strategy with IT design and investment (Shortliffe 2005). The strategic focus will be on patients, markets,

331

332

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Enterprise strategy Plan to develop and align the internal capabilities of an organization, including IT, with risks and opportunities in the external market

and competitors and the value assessment of services as the organization positions itself in the market. Internally, the focus will shift to the implications of the strategy (including the design and selection of appropriate IT architecture and infrastructure) on governance and leadership functions, organizational structure, work processes, and operations. Enterprise strategy has been extensively researched in the business community, yielding considerable knowledge and experience to guide the development of IT as a strategic resource (Bharadwaj 2000; Bradley et al. 2012; Chi, Ravichandran, and Andrevski 2010). Enterprise strategy is generally regarded as a plan for developing and aligning the internal capabilities of an organization, including IT, with the risks and opportunities in the external market, including potential competitors and collaborators; this strategy is sometimes called resource-based strategy. The effectiveness of such a strategy depends on the ability of an enterprise (organization or regional system) to design and assemble an appropriate combination of resources, such as funding, to respond to its environment, as well as its ability to change. The ability to change is referred to as “being nimble,” a key factor in complex, dynamic systems. As discussed, healthcare financing models have rewarded the investment in and design of enterprise strategies that continue to slowly shift from cost-based reimbursement to rewarding clinical processes and outcomes. This process is frequently characterized by periods of regressing to or combining with cost-based models and then reemerging with new outcome-based models, without abandoning the tried-and-true (if inefficient and ineffective) methods of the past. This strategic shift has resulted from advances in IT and applications of IT in innovative strategies. This shift has demonstrated improved outcomes and performance, but an improved market-driven enterprise strategy alone will not lead to optimal performance of the health system; the market strategy design is what perpetuates many of the system’s failures. Improving the performance of a dysfunctional system design is not an effective solution and should not continue to be pursued as an isolated strategy. We cannot continue to pursue the same strategy and expect a different outcome. Market-driven investment in IT by hospitals and clinics varies and has lagged because of the lack of incentives in most markets to achieve greater value, such as increased efficiency, improved quality, better integration, and more patient involvement. The desire to improve efficiency is offset by the ease with which providers and private insurers can pass on costs to the government, corporations, and patients in the form of higher premiums. Market demand for high-quality, integrated, and patient-involved services enabled by IT is blunted by providers’ general resistance to change and by insurance companies’ reluctance to provide incentives for change. This situation engenders a complacent market. The general principle of increased innovation and capital investment (including in IT) through private sector involvement is not well demonstrated in the US health system. Many models that have evolved

Ch a p t er 1 5 :  Va l u a ti o n a n d Fi n a n c i n g o f H e althc are Ser v ic es and IT Infrastr uc ture

can provide direction and be used to test the effectiveness of IT valuation and strategy. Additional research by healthcare institutions is needed, however, to reveal how these models value IT and guide investment strategy.

Public Sector IT Investment Another current major challenge for healthcare organizations is aligning enterprise strategy with public good. This is a particular concern in the US health system because most factors of healthcare production—hospitals, physicians, insurance plans, pharmaceuticals, and the like—are in the private sector and thus pursue an independent enterprise strategy. As discussed throughout this book, however, the health system does not have to be publicly owned to be responsive to social needs. However, such an argument might be made if the private sector and policymakers are not able to design and implement a system that responds to legitimate social demands. Societal expectations of the health system include the availability of highquality services that are efficiently provided; accessibility to EHRs based on the location of the patient and not the record; and evidence-based, shared clinical decision making throughout the system. Clinical consultations and referrals are expected to be based on knowledge systems and not the location of health professionals. These IT-driven solutions require investments in and development of a community or public HIE that complements the enterprise HIE. The private sector asks (legitimately) why it should underwrite the portion of the IT system that serves the public good. However, the same applies to the public sector’s (government’s) investment in operational or enterprise capacity. The public sector should not subsidize the private sector’s investment in IT that serves enterprise interests, which is the current pattern under the HITECH Act. The problem is that the public sector continues to subsidize private IT systems that are designed around the logic of an enterprise strategy, which lacks a vision or concern for a community information exchange. The public sector itself has many policy interests in IT—for example, developing and maintaining national databases as instruments of surveillance for purposes of both public health and national security. Direct social good can be derived from national databases and is a justification for the government’s investment in IT, although little work has been done to estimate and plan for this public benefit (Boles and Cook 2005; Ozdemir, Barron, and Bandyopadh­ yay 2011). Funding health registries and public health surveillance projects reflects a strong commitment to install these databases. These projects would curtail or eliminate the obsolete and costly practice of extracting information from hospital databases to develop national registries. In addition, society benefits from investments in IT that enable a national information exchange. Such investments support IT capacity and design that are based not on the logic of operational and enterprise strategies (the needs of health institutions and health professionals) but rather on the needs of patients and communities.

333

334

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Value measurement methodology (VMM) Framework and toolkit for estimating the effects of a given IT investment on society, values, costs, risks, and tangible and intangible returns

The essential requirements of such public investments are measured not only in terms of their value potential but also in terms of the design requirements of the IT infrastructure (Dixon, Colvard, and Tierney 2015). Access by patients and health professionals to integrated health information is a public good, and its value contribution is the rationale for public sector investment in IT. Sufficient capital will not be allocated by the private sector, nor will it come from reimbursement rates for medical services; this lack of funding will result in an underinvestment in IT. This does not mean that the government should design and control the health information infrastructure but only that it would invest in it and set standards to ensure public benefit was served. Medicare and Medicaid programs and ACA provisions are expressions of the public sector’s investment in healthcare services for the public—specifically, the poor, elderly, young, uninsured, and high-risk populations. Just as the public sector provides access to medical care and ensures continuity, efficiency, and quality, so must it also invest in IT infrastructure that supports that system of care. Public insurance differs from private insurance in that the expenditure of public funds requires provider accountability for quality, continuity, and efficiency. The government cannot allow the continued underinvestment in IT because it is a critical resource of healthcare production. This investment is not a subsidy to private healthcare organizations and health professionals but rather a factor cost of production to achieve the social (or public good) purpose. The method many governments use to assess the value of, and thus justify, public IT investments is different from the method deployed in the private sector (Mataracioglu 2015). The US government, for example, uses the value measurement methodology (VMM), which was developed by the Federal Chief Information Officer Council in 2002. The VMM is a framework and a toolkit for estimating the effects of a given IT investment on society, values, costs, risks, and tangible and intangible returns. It uses a scoring system to determine values in both quantitative and qualitative measures (Federal Chief Information Officer Council 2002). It calculates return on investment (ROI), although ROI is not based solely on financial return but also on factors such as social benefits, access and use of data by other agencies, and strategic or political goals. Public investment then relies on the valuation of IT investment related to the ROI of alternative public investments. The VMM in itself does not differ substantially from the logic of private sector IT investment, but the private and public values differ, resulting in different decision choices. Both public and private investment processes start by defining and estimating value, risk, and cost parameters. If the logic of public sector investment is limited to enterprise strategy, the government will inappropriately subsidize private interests and IT underinvestment will result, because not all social benefits are included. The public sector should not expect the full investment to come from the private sector and its IT investment logic. Public investment must be made in community exchanges. For example, the government might fund IT investments in small

Ch a p t er 1 5 :  Va l u a ti o n a n d Fi n a n c i n g o f H e althc are Ser v ic es and IT Infrastr uc ture

and rural hospitals and clinics to ensure that local services are accessible and integrated within the larger grid of the region’s health system. These institutions lack the capital and technical expertise to develop an information system, but their viability is important to patient care and system efficiency. An IT system with the ability to interface with other IT systems throughout the nation also fulfills a societal need. An investment in such a system cannot be justified by local private organizations because it serves a greater social need. Coye and Bernstein (2003) offer a model for estimating value and allocating public funds for private sector IT investment in public infrastructure development. This approach is interesting in that it provides private organizations access to public capital through a revolving fund, which in turn stimulates private investment. The provision to repay the fund recognizes the importance of private sector investment. Another model is awarding “needy” organizations, such as safety-net hospitals, grants to stimulate IT investment. Such an approach has merit because it requires organizations to develop their own IT strategy and submit proposals for review by health IT corporations that have the authority to award grants or loans. In addition, the approach adds rigor to the concept of meaningful use as a basis for allocating IT funds.

Conclusion Healthcare faces various challenges with lack of coordination and information sharing. Traditional financing mechanisms do not foster information flow. Primary care, specialists, hospitals, and other providers are forced to provide care and store information in department silos. Such fragmentation has led to uncontrollable increases in healthcare costs. Fundamental changes to payment mechanisms and to the organizational structure (e.g., the extension and connection of medical homes and ACOs) are needed to rein in the cost increases. Currently, Amazon, Berkshire Hathaway, and J. P. Morgan—titans in logistics and technology, insurance, and finance, respectively—are joining forces to address healthcare’s cost issues (Young 2018). The formation of such alliances is exactly the jolt needed to bring about changes in US healthcare financing. Valuation of and investment in healthcare services use a complex mix of public and private funds allocation. Valuation of and investment in IT architecture are similarly complex in that different designs afford different capacities, each of which has a value basis. The logic of most IT architecture in healthcare organizations today reflects the tradition of independent clinicians and institutions, whereas the value of IT is derived primarily from integration of processes and services. Thus, deriving the value potential from an IT investment entails making a set of assumptions about how IT will be applied—that is, how IT will enable an enterprise strategy, which includes redesigning the structure of clinical processes. For this reason, health IT and health informatics are considered to

335

336

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

be a transformational science. Building IT capacity without using it to its full potential is certainly a failed strategy. The increased role of the public sector in the US health system will likely increase the pressure for structural change. The challenge is to stimulate private clinicians and organizations to change without adding to the significant costs required by regulations and public bureaucracy. The US health system cannot continue to pursue the same strategies and expect a different outcome. Policy issues are complex but solvable. Healthcare needs leaders who can conceptualize a patient and systems perspective and who are not focused on benefits accruing to their delivery system.

Chapter Discussion Questions 1. What are the major challenges to healthcare financing in the United States? 2. What are the interrelationships among the stakeholders of medical homes and ACOs related to shared information and knowledge as well as coordination of services? 3. What are the necessary elements of a strong IT infrastructure in a medical neighborhood? 4. Why is valuation of healthcare services and IT infrastructure important? 5. How can investment in private and public IT infrastructure support the achievement of healthcare goals? 6. What payment model would enable the development of an effective HIE system oriented on maintaining population health?

Case Study  Med City’s Diabetes Management Care Group The prevalence of diabetes in populations throughout the world is increasing, and its management is a challenge. The incidence of diabetes in a population is directly correlated with obesity, although many other risk factors are at play, including genetics. Health departments and health ministries engage in ongoing debates about the effectiveness and practicality of reducing risk by aggressively managing diet and exercise, managing risk more effectively by training more health professionals, and establishing specialized diabetes clinics, among other interventions. As a chronic condition, diabetes requires patients to have regular visits to primary care physicians and consultations with specialists. Furthermore, diabetes puts patients with other diseases at higher risk for hospitalization for any acute condition.

Ch a p t er 1 5 :  Va l u a ti o n a n d Fi n a n c i n g o f H e althc are Ser v ic es and IT Infrastr uc ture

Med City Hospital is a well-recognized institution that has been delivering care in Med City for many years. The hospital uses an electronic health record (EHR) system developed by HRecord, one of the largest vendors in the market. Some patients with diabetes visit the hospital’s emergency department (ED) when their diabetes is not properly managed and causes complications. The hospital is reimbursed by insurance plans for patients who are admitted through the ED or by Medicare or Medicaid for the specific ED visit. ED visits have been on the rise, and so have repeat visits, which sometimes tax the already busy department. The ED staff have been calling for an expansion. The hospital’s leadership is focused on finding a solution. Most primary care physicians in the Med City region practice in groups, although some practice independently. They have built their practices on the strength of access, reputation for quality, and patient loyalty. They are reimbursed primarily on a per-visit basis through a range of insurance plans, including Medicare. An ACO has been initiated, which gives physicians incentives to base treatment on value—but only for a defined population. These physicians have built long-lasting relationships with their patients and provide some of the diabetes care; they also refer patients to endocrinologists and other specialists. Some physicians fear their patients will retain the specialists for treatment using their treatment protocols. Some physicians have access to the hospital EHR, whereas others are developing information systems that link through a health information exchange (HIE) system. Med City forms a study group to give guidance on providing personalized, integrated, quality care for patients who have diabetes and other chronic care needs. A number of physicians are invited to join the study, but, after an initial meeting, mostly young generalists attend. A few of them point out the benefits of involving other health professionals in developing care plans, such as dietitians and nurses. These discussions generate considerable debate about the merits of hiring registered dietitians, nutritionists, nurse educators, diabetes nurse practitioners, and physician assistants. These discussions are spirited and consistent with the culture of the innovation institute at the hospital, whose motto is “Everyone comes to the ‘commons’ and is heard.” Everyone agrees that allied health professionals bring knowledge that is important for treating this patient population. The chief objection is that these professionals are expensive, and reimbursement rates do not factor them into the payment scheme. The primary care clinics in the area do employ nurses, most of whom left the hospital because of its long and irregular hours. Because they will be part of a developing ACO in the region, some physicians point out that they will have access to EHRs that enable them to obtain (continued)

337

338

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

patient information. Others argue that they cannot effectively manage chronic cases through the current HIE system. There is considerable discussion on what reimbursement rates will be and how services can be optimized and financially rewarded. Despite the challenges ahead, the hospital is interested in forming a care network; it is working on identifying key outcomes and negotiating a bundled-payment scheme based on the cost of providing evidence-based, effective, and coordinated care. Through such a care network, the hospital hopes to coordinate diabetes care management and be rewarded for it.

Case Study Discussion Questions 1. What are the implications of employing a registered dietitian versus a nutritionist? Describe and discuss the recruitment strategy for registered nurses in clinics. 2. Who are the key stakeholders in diabetes care management, and how do they coordinate their work? Does a primary care physician manage the care team, or is it a team effort? 3. How should medical information and knowledge of care management be shared among the stakeholders? 4. What payment models are barriers to collaborating and sharing information, and how can these barriers be overcome? What form of payment would align finance with the best practice of medicine? 5. Who controls payment models, and why are they difficult to change? Why is it so difficult to align payment models with care delivery models—particularly for complex, chronic diseases?

Additional Resources Information Technology Infrastructure Library: www.itil-officialsite.com/home/ home.aspx. ISACA: www.isaca.org/about-isaca/Pages/default.aspx. US CIO Council: www.cio.gov/about/.

References American Academy of Family Practice, American Academy of Physicians, American Academy of Pediatricians, and American Osteopathic Association. 2011. “Guidelines for Patient-Centered Medical Home Recognition and Accreditation Programs.” Published February. www.aafp.org/dam/AAFP/documents/practice_management/ pcmh/initiatives/PCMHJoint2011.pdf.

Ch a p t er 1 5 :  Va l u a ti o n a n d Fi n a n c i n g o f H e althc are Ser v ic es and IT Infrastr uc ture

Bharadwaj, A. S. 2000. “A Resource-Based Perspective on Information Technology Capability and Firm Performance: An Empirical Investigation.” MIS Quarterly 24 (1): 169–96. Berwick, D. M., T. W. Nolan, and J. Whittington. 2008. “The Triple Aim: Care, Health and Cost.” Health Affairs 27 (3): 759–69. Boles, K. E., and M. J. Cook. 2005. “Investing in Information Technology.” In Strategic Management of Information Systems in Healthcare, edited by G. D. Brown, T. T. Stone, and T. B. Patrick, 195–221. Chicago: Health Administration Press. Bosko, T., and C. Hawkins. 2016. “Evolving Physician Reimbursement Structures: Moving the Medical Group to Value-Based Success.” Journal of Healthcare Management 61 (3): 176–80. Bradley, R. V., R. M. E. Pratt, T. A. Byrd, C. N. Outlay, and D. E. Wynn Jr. 2012. “Enterprise Architecture, IT Effectiveness, and the Mediating Role of IT Alignment in US Hospitals.” Information Systems Journal 22 (2): 97–127. Casalino, L. P., N. Erb, and S. Shortell. 2015. “Accountable Care Organizations and Population Health Organizations.” Journal of Health Politics, Policy and Law 40 (4): 821–37. Chi, L., T. Ravichandran, and G. Andrevski. 2010. “Information Technology, Network, Structure and Competitive Action.” Information Systems Research 21 (3): 543–70. Chung, P. J., and M. A. Schuster. 2004. “Access and Quality in Child Health Services: Voltage Drops.” Health Affairs 23 (5): 77–87. Coye, M. J., and W. S. Bernstein. 2003. “Improving America’s Health Care System by Investing in Information Technology.” Health Affairs 22 (4): 56–58. D’Aunno, T., L. Broffman, M. Sparer, and S. Kumar. 2018. “Factors That Distinguish High-Performing Accountable Care Organizations in the Medicare Shared Savings Program.” Health Services Research 53 (1): 120–37. Dixon, B. E., C. Colvard, and W. M. Tierney. 2015. “Identifying Health Facilities Outside the Enterprise: Challenges and Strategies for Supporting Health Reform and Meaningful Use.” Informatics for Health and Social Care 40 (4): 319–33. Federal Chief Information Officer Council. 2002. “Value Measuring Methodology: How-To Guide.” Published October. https://s3.amazonaws.com/sitesusa/ wp-content/uploads/sites/1151/2016/10/ValueMeasuring_Methodology_ HowToGuide_Oct_2002.pdf. Fisher, E. S. 2008. “Building a Medical Neighborhood for the Medical Home.” New England Journal of Medicine 359 (12): 1202–5. Gupta, D., I. Karst, and E. B. Mendelson. 2016. “Value-Based Reimbursement: Impact of Curtailing Physician Autonomy in Medical Decision Making.” American Journal of Roentgenology 206 (2): 276–79. Hermanrud, I. 2017. “Best Fit KM: Linking Communities of Practice to an Innovation Strategy.” Electronic Journal of Knowledge Management 15 (1): 28–36. Lewis, N. 2016. “Independent Practices Seek Care Coordination Strategies.” Medical Economics 93 (9): 24–29.

339

340

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Mataracioglu, T. 2015. “On the Technical Description of Value Measuring Methodology.” International Journal of Managing Value and Supply Chains 6 (2): 1–12. Morrison, E. M., and H. S. Luft. 1990. “Health Maintenance Organization Environments in the 1980s and Beyond.” Health Care Financing Review 12 (1): 81–90. Morteza, A., H. Asgari, A. Gharibi, and M Rashidi. 2017. “Leveraging Business–IT Alignment Through Enterprise Architecture—an Empirical Study to Estimate the Extents.” Information Technology Management 18 (1): 55–82. Ouayogodé, M. H., C. H. Colla, and V. A. Lewis. 2017. “Determinants of Success in Shared Savings Programs: An Analysis of ACO and Market Characteristics.” Healthcare 5 (1–2): 53–61. Ozdemir, Z., J. Barron, and S. Bandyopadhyay. 2011. “An Analysis of Adoption of Digital Health Records Under Switching Costs.” Information Systems Research 22 (3): 491–503. Salzberg, C., A. Bitton, S. R. Lipsitz, C. Franz, S. Shaykevich, L. P. Newmark, J. Kwatra, and D. W. Bates. 2017. “The Impact of Alternative Payment in Chronically Ill and Older Patients in the Patient-Centered Medical Home.” Medical Care 55 (5): 483–92. Scott, W. R., M. Ruef, P. J. Mendel, and C. A. Caronna. 2000. Institutional Change and Healthcare Organizations: From Professional Dominance to Managed Care. Chicago: University of Chicago Press. Shortliffe, E. H. 2005. “Strategic Action in Health Information Technology: Why the Obvious Has Taken So Long.” Health Affairs 24 (5): 1222–33. Tanenbaum, S. J. 2017. “Can Payment Reform Be Social Reform? The Lure and Liabilities of the ‘Triple Aim’.” Journal of Health Politics, Policy and Law 42 (1): 53–71. US Congress. 2010. Public Law 111-148, Patient Protection and Affordable Care Act (PPACA), 111th Congress, March 23. www.gpo.gov/fdsys/pkg/PLAW111publ148/pdf/PLAW-111publ148.pdf. Vest, J. R., and B. A. Kash. 2016. “Differing Strategies to Meet Information-Sharing Needs: Publicly Supported Community Health Information Exchanges Versus Health Systems’ Enterprise Health Information Exchanges.” Milbank Quarterly 94 (1): 77–108. Wang, T., and A. McLeod. 2018. “Do Health Information Technology Investments Impact Hospital Financial Performance and Productivity?” International Journal of Accounting Information Systems 28: 1–13. Wohlner, R. 2017. “Accounting Basics: History of Accounting.” Accessed November 21. www.investopedia.com/university/accounting/accounting1.asp. Young, J. 2018. “Jeff Bezos, Warren Buffett and Jamie Dimon Vaguely Promise to Fix Health Care.” Published January 31. www.huffingtonpost.com/entry/jeff-bezoswarren-buffett-jamie-dimon-health-care_us_5a70f74ae4b0a6aa487441be.

CHAPTER

DATA AND INFORMATION SECURITY IN THE HEALTHCARE ENTERPRISE

16

Dixie B. Baker and Timothy B. Patrick

Learning Objectives After reading this chapter, you should be able to do the following: • Be aware of the prevalence of constant threats to privacy and security and be familiar with resources that track such threats. • Conceptualize the logical relationship among privacy, security, and safety. • Understand the interactions between safety and security from the perspective of both individual healthcare and public health. • Identify and develop arguments to support the principles of fair information practices. • Identify the technical safeguards required by the HIPAA Security Rule and the risks each safeguard is designed to address. • Identify the security and privacy risks of applying data-mining analytics. • Understand the threats to privacy and security on the Internet of Things, specifically in the healthcare domain.

Key Concepts • • • • • • •

Privacy, security, risk, and trust Data breach HIPAA Privacy and Security Rules Fair Information Practices Authentication Encryption Internet of Things

341

342

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Introduction Imagine a desktop or laptop computer that contains important and potentially sensitive data for tracking a specific communicable disease in society. Suppose the data contain the names and addresses of individuals in a particular community who have HIV infection. It is easy to imagine a trade-off between keeping the data secure and realizing the potential value of those data. Suppose that the computer is kept in a sealed room that has no windows and that no one can access, and that the computer is not connected to any network. Clearly, the data will be secure because no unauthorized access to or dissemination of the data is possible. But the data will also be practically useless—no one can use the data for surveillance of disease in that community (or for any other purpose) because no one has access to the data. If we loosen the security a bit—say, allow a select individual access to the computer and its data—then we will increase the value of the data but, at the same time, introduce some risk of inappropriate dissemination of the data. The person with authorized access might have an eidetic memory or might save the data on a flash drive and then remove the drive from the secured room to disseminate the data in an inappropriate manner later. We might allow even greater access to and sharing of the data among responsible researchers and epidemiologists by attaching the computer to a network, thus increasing the practical value of the data still further. The concomitant cost of loosening security, however, is the increased risk of unauthorized access and dissemination of the data. The point is that a trade-off always exists between keeping data and information secure and getting the most value from them. The ongoing question is, what is the most appropriate set of restrictions and protections for any given operational context? The focus of this chapter is data and information security as they relate to the operational context of patient care and the electronic health record (EHR). Data and information security in the context of patient care and the EHR are important at both the individual and the organizational level and from both a healthcare and a financial perspective.

Health and Financial Risks Related to Data and Information Security Individual Risks Data and information security are essential to the services provided to support the continuing health of the individual. Decision-making processes in healthcare require data and information that must be available in uncorrupted form to clinical decision makers and care providers when needed. A lack of access to

Ch ap te r 16 :   D a ta a n d In f o r ma ti on Sec ur ity in the H ealthc are Enter p r ise

appropriate data and information as a result of a security breach can directly affect those decision-making processes, diminishing the quality of healthcare. Identity theft is an ongoing problem with extremely serious ramifications for the quality of life of individuals. The 2017 Equifax breach (O’Brien 2017) and ongoing smaller-scale breaches have put the financial security, happiness, and—indirectly at least—health of millions of individuals at risk. Medical identity theft represents a serious risk to the ability of individuals to gain access to appropriate healthcare.

Organizational Risks Healthcare organizations that are targets of data and information security breaches may be prevented from providing appropriate care to both individual patients and populations of patients. Security breaches in healthcare organizations may also have dire financial consequences, including (CDW 2017) • • • •

direct financial costs, damage to reputation, loss of patient trust, and penalties and fines for violating federal security standards.

Resources for Tracking Breaches Several online resources are available for tracking security and privacy breaches of data and information, both in general and in healthcare specifically. One resource is the Privacy Rights Clearinghouse (2018) Chronology of Data Breaches. As of January 9, 2018, according to this source, 3,900 security breaches in healthcare were publicly reported since 2005. These breaches involved medical providers and insurance services and affected 227,638,510 records. The various types of breaches tracked by the Chronology of Data Breaches include credit or debit card fraud, hacking, and unauthorized insider access. Including industries outside of healthcare, 7,870 breaches were publicly disclosed since 2005, involving 10,059,986,194 records. The impact of these breaches is underscored by the fact that the Chronology of Data Breaches reports publicly disclosed breaches. In many cases, the date of the breach is not listed—only the date of public disclosure is available. Some breaches are not disclosed, and a delay between the occurrence of a breach and its disclosure can be dangerous. For example, at Penn Medicine in King of Prussia, Pennsylvania, a laptop was stolen on November 30, 2017, but the theft was not publicly disclosed until January 8, 2018. Another resource for tracking the occurrence of healthcare breaches is the Breach Portal of the US Department of Health and Human Services (HHS)

343

344

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Office for Civil Rights (2018). According to the Breach Portal, there were 397 reported breaches in 2016 and 2017 that affected 17,251,769 individuals, although this number may include double counting of some individuals. One difficulty in using these resources relates to their metadata schemata (chapter 12). The typology or categories of breaches used by one resource may not match that used by another. For example, the categories of breaches used by the Breach Portal are Hacking/IT Incident, Improper Disposal, Loss, Theft, Unauthorized Access/Disclosure, Other, and Unknown. These categories are significantly different from those used by the Chronology of Data Breaches. The practical result of these differences is that data concerning breaches may not be easily aggregated across tracking resources.

Technical Issues in the Protection of Data and Information To understand the technical aspects of breaches requires an understanding of the basic concepts of privacy, security, risk, and trust. The terms privacy and security are often used synonymously, but the two are distinct. Security is certainly necessary to protect individual privacy, but it is also required to safeguard the integrity of data and the availability of critical information and services, to ensure the authenticity of identities and data, and to maintain accountability of system and user actions. Most important, security is necessary to protect data and information and the associated computing systems that are critical to the health of individual patients and populations of patients.

Privacy More than a century ago, US Supreme Court Justice Louis Brandeis referred to privacy as “the right to be let alone” and characterized it as “the most comprehensive of rights and the right most valued by civilized men” (Olmstead v. United States, 277 U.S. 438 (1928)). Essentially, privacy is the state of being free from intrusion or disturbance in one’s private life or affairs. In the United States, citizens have the constitutional right to such a state. The delivery of safe, high-quality healthcare necessarily involves the collecOpportunity for Interprofessional Education tion, use, retention, and sharing Different types of clinicians, administrators, information technology of individual consumers’ most staff, and patients may have differing views and understanding of private information. The safety issues of privacy, security, risk, and trust. Organize a session with and quality of care depend on fellow students from these different professional groups and a guest the provider team’s ability to to represent the patient’s point of view. Discuss the issues, and docuaccess detailed, accurate, and ment the points of view represented in the group. complete information about an individual’s lifestyle and

Ch ap te r 16 :   D a ta a n d In f o r ma ti on Sec ur ity in the H ealthc are Enter p r ise

345

medical and psychological state. Still, the individual has a right to expect that only the information required is collected, used, retained, and shared; that the information shared is used only for the intended purposes; and that privacy rights are respected and honored. In the healthcare context, we can define privacy as the assurance that one’s health information is collected, accessed, used, retained, and shared only when necessary and only to the extent necessary and that the information is protected throughout its life cycle by means of fair privacy practices that are consistent with applicable laws and regulations as well as the preferences of the individual.

Security Security is the state of being free from danger or harm, or the set of defensive measures that collectively ensure that state. Information security is a specialized area aimed at protecting the confidentiality of information, the integrity of data, and the availability of information and system services. These measures generally include mechanisms for validating that the people, software applications, and information systems seeking access are who and what they claim to be and for keeping a record of actions taken by users and the system. The administrative, physical, and technical standards and implementation specifications in the Health Insurance Portability and Accountability Act of 1996 (HIPAA) Security Rule require such mechanisms (Centers for Medicare & Medicaid Services 2003). As health information is exchanged between organizations, consolidated in EHRs, and shared with personal health records (PHRs), security measures that ensure the integrity and provenance of data and metadata become increasingly important. In the healthcare context, security is defined as the protection of the confidentiality of private, sensitive, and safety-critical information; the integrity of health data and metadata; and the availability of information and services through measures that authenticate data provenance and user and system identity and that maintain an accounting of actions taken by users, software programs, and systems. The relationship between privacy and security is evident. Security measures that protect the confidentiality of personal health information contribute to privacy protection. However, as discussed later, protecting personal privacy involves more than securing confidential information. The importance of security to the safety and quality of care should also be apparent, as the availability of accurate, authenticated information and critical services at the point and time of care is essential to both patient safety and care quality. Deciding what privacy and security protections are necessary and appropriate is a matter of assessing and managing risk.

Risk Risk is a probability function that involves three variables: (1) threat, (2) vulnerability, and (3) valued asset. Risk, then, is the probability that a threat will

Security Protection of the confidentiality of private, sensitive, and safety-critical information; the integrity of health data and metadata; and the availability of information and services through measures that authenticate user and system identity and data provenance and that maintain an accounting of actions taken by users, software programs, and systems

346

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

exploit a vulnerability to damage, destroy, or harm a valued asset. To identify and manage risk, risk assessment is necessary. Risk assessment is a disciplined process of identifying valued assets (e.g., buildings, computers, information, people), threats (e.g., disgruntled employees, spyware), and vulnerabilities (e.g., storage of patient information on a laptop or flash drive) and determining the probability that the threat will exploit the vulnerability to cause harm (e.g., unauthorized access to patient information). Risk assessment provides the basis for deciding how to manage each risk—that is, whether to eliminate, moderate, or accept it (perhaps with insurance to reduce liability). Recognizing the importance of risk assessment in implementing appropriate security protections, the HIPAA Security Rule requires both risk analysis (to identify and assess potential risks and vulnerabilities) and risk management (to develop and execute security measures to counter the risks identified).

Trust Trust is the level of comfort, belief, or assurance one senses on the basis of the evidence available. Trust lies at the heart of modern medicine. Providers must trust that the information and software services they need will be available when and where needed and that the clinical decision support integrated with the EHR is accurate, reliable, and safe. In addition, clinicians must trust that the laboratory test result they receive is actually from the laboratory to which the test was sent for processing and that no modification or corruption occurred during the transmission. Conversely, individuals must trust that their providers keep their most private health information confidential, disclosing and using the information only to the extent necessary and in ways that are legal, ethical, authorized, and in keeping with their personal expectations and preferences. Both providers and consumers must trust that the technology used to provide care will “do no harm.” In this context, trust can be defined as the evidence-based confidence that the people, organizations, data and information, and information systems involved in healthcare delivery are what they claim to be and behave as expected. As shown in exhibit 16.1, trust involves a delicate balance among transparency, consent, technology, and laws and regulations. Transparency means that consumers and providers are told how information is being used, protected, and shared. In some cases, the individual’s consent must be (or should be) obtained before her information is collected, used, or disclosed. Technology guards against the unauthorized disclosure, modification, or use of information and records how the information is used. Some laws and regulations, such as the HIPAA Security Rule and the Common Rule that protects human subjects (HHS 1991), mandate the implementation of certain policies and protective measures. Other laws, such as the Genetic Information Nondiscrimination Act of 2008 and the Affordable Care Act of 2010, protect consumers from unfair

Ch ap te r 16 :   D a ta a n d In f o r ma ti on Sec ur ity in the H ealthc are Enter p r ise

Consent

EXHIBIT 16.1 Balance Among the Components of Trust

Laws and regulations

Pol ic

y G ove rna nce

Transparency

San ctio

ns

Technology

347

Trust Fair information practices

discrimination should their genetic or other health information be disclosed. Fair information practices serve as the foundation for trust.

Fair Information Practices Fair information practices (FIPs) are the foundation of information security and privacy laws and regulations in the United States and throughout the world. FIPs constitute fair and responsible information stewardship, which is essential to establishing and maintaining public trust when collecting, using, disclosing, and sharing personal information. The heritage of FIPs is the Code of Fair Information Practices, published in 1973 by the US Department of Health, Education, and Welfare (HEW; now the HHS), which sets forth the principles of openness, disclosure, secondary use, record correction, and security (HEW 1973). These original FIPs provided the framework for the Privacy Act of 1974, which protects certain types of personal information held by federal agencies. Further refinements and customizations to the FIPs include the following: • 1980: “Guidelines on the Protection of Privacy and Transborder Flows of Personal Data,” a consensus document published by the Organisation for Economic Co-operation and Development (1980) involving 24 countries, including the United States

Fair information practices (FIPs) Foundation of information security and privacy law and regulations in the United States and throughout the world; constitute fair and responsible information stewardship, which is essential to establishing and maintaining public trust when collecting, using, disclosing, and sharing personal information

348

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

• 1998: “Privacy Online: A Report to Congress,” published by the US Federal Trade Commission (FTC 1998) • 2001: “Standards for Privacy of Individually Identifiable Health Information,” developed pursuant to HIPAA (HHS 2001) • 2007: “Fair Information Practice Principles,” published by the FTC (2007) In 2008, the Office of the National Coordinator for Health Information Technology (ONC) released the “Nationwide Privacy and Security Framework for Electronic Exchange of Individually Identifiable Health Information,” which established FIPs specifically for electronic health information (ONC 2008). The FIPs set forth in this document are shown in exhibit 16.2, which also identifies how each principle is translated into healthcare laws and regulations.

Security Implementation in EHR Technology HIPAA’s Privacy and Security Rules, further strengthened by the American Recovery and Reinvestment Act (ARRA) of 2009, are the primary sources of standards and implementation specifications for health information security and privacy. Although technically the HIPAA Rules apply only to “protected health information” that is generated, used, and exchanged by specific entities covered under HIPAA, these standards are widely referenced and applied to protect all individually identifiable health information. The Privacy Rule essentially says that an individual’s health information may be used or disclosed only as explicitly permitted by law or as authorized by that individual (HHS 2000). The Security Rule defines administrative, physical, and technical safeguards that a healthcare organization covered under HIPAA must implement to protect the confidentiality, integrity, and availability of health information. Enforcement of compliance with the HIPAA Privacy and Security Rules is the responsibility of the Office of Civil Rights of the HHS. To be certified by an accredited certification body, EHR technology must be implemented according to the standards and certification criteria specified in ARRA (ONC 2010). A number of these criteria ensure that certified EHR technology offers the technical security safeguards that the organization using that technology will need to comply with the HIPAA Security Rule. Exhibit 16.3 shows how each of the HIPAA technical security safeguards is implemented in certified EHR technology.

Authentication Essential to all security safeguards is authentication. Authentication is the verification that the true identity of a person or entity (e.g., web server, health

Ch ap te r 16 :   D a ta a n d In f o r ma ti on Sec ur ity in the H ealthc are Enter p r ise

Principle

Definition

Legal Codification

Individual access Individuals are given simple • HIPAA Privacy Rule provides individuals and timely means to access and the right to request and obtain a copy of obtain their individually identheir health information. tifiable health information in a • ARRA extends this rule to include the right readable form and format. to request and obtain an electronic copy. Correction

Individuals should be provided • HIPAA Privacy Rule gives individuals the with a timely means to dispute right to request their health information to the accuracy or integrity of their be amended. individually identifiable health information and to have erroneous information corrected or to have a dispute documented if their requests are denied.

Openness and transparency

There should be openness and transparency about policies, procedures, and technologies that directly affect individuals and/or their individually identifiable health information.

• HIPAA Privacy Rule requires that each covered entity provide individuals with written notice of the organization’s privacy practices, including uses and disclosures of health information; the individual’s rights; and the organization’s legal duties to protect health information.

Individual choice Individuals should be provided a reasonable opportunity and capability to make informed decisions about the collection, use, and disclosure of their individually identifiable health information.

• HIPAA Privacy Rule requires that covered entities obtain the individual’s consent before using or disclosing that individual’s identifiable health information for purposes other than treatment, payment, healthcare operations, or other uses permitted by law.

Collection, use, and disclosure limitation

Individually identifiable health • HIPAA Privacy Rule limits the sharing of information should be collected, health information to the minimum necesused, and/or disclosed only to sary for the intended purpose. the extent necessary to accom- • GINA prohibits the improper use of genetic plish a specified purpose(s) information in health insurance and and never to discriminate employment. inappropriately. • ACA prohibits insurance companies from refusing to insure individuals with preexisting health conditions.

Data quality and integrity

Persons and entities should take • HIPAA Privacy Rule gives individuals the reasonable steps to ensure that right to review their health information individually identifiable health and to request an amendment. information is complete, accu• HIPAA Security Rule requires integrate, and up-to-date to the extent rity protection for health information necessary for the person’s or transmissions. entity’s intended purposes and has not been altered or destroyed in an unauthorized manner.

Safeguards

Individually identifiable health • HIPAA Privacy and Security Rule requires information should be protected administrative, technical, and physical with reasonable administrative, safeguards to protect health information. technical, and physical safeguards to ensure its confidentiality, integrity, and availability and to prevent unauthorized or inappropriate access, use, or disclosure. (continued)

349

EXHIBIT 16.2 Fair Information Practices

350

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

EXHIBIT 16.2 Fair Information Practices (continued)

Accountability

These principles should be • HIPAA Privacy Rule requires an accounting implemented, and adherence of all disclosures between organizations to them should be ensured, other than those required for treatment, through appropriate monitorpayment, and healthcare operations; ARRA ing, and other means should be eliminates this exception. in place to report and mitigate • HIPAA Security Rule requires the collection nonadherence and breaches. of audit records and periodic audit review. • ARRA requires notification of individuals whose health information may have been exposed through a breach incident.

Note: ACA = Affordable Care Act of 2010; ARRA = American Recovery and Reinvestment Act of 2009; GINA = Genetic Information Nondiscrimination Act of 2008; HIPAA = Health Insurance Portability and Accountability Act of 1996. Source: Adapted from ONC (2008).

EXHIBIT 16.3 HIPAA Technical Security Safeguards

HIPAA Technical Security Safeguards

EHR Standards

EHR Certification Criteria

Implement procedures to verify that a person or entity seeking access to electronic PHI is the one claimed

Verify that a person or entity seeking access to electronic health information is the one claimed and is authorized to access such information

Implement technical policies and procedures that maintain electronic PHI to allow access only to those persons or software programs that have been granted access rights. Policies and procedures include the following:

Establish controls that permit only authorized users to access electronic health information

• Unique user identification

Assign a unique name and/or number for identifying and tracking user identity

• Emergency access procedure

Permit authorized users (who are authorized for emergency situations) to access electronic health information during an emergency

• Automatic logoff

Terminate an electronic session after a predetermined time of inactivity

• Encryption and decryption

Any encryption algorithm identified by the NIST as an approved security function in Annex A of the FIPS Publication 140–42

Encrypt and decrypt electronic health information in accordance with the standard, unless the secretary determines that the use of such algorithm would pose a significant security risk for certified EHR technology

Ch ap te r 16 :   D a ta a n d In f o r ma ti on Sec ur ity in the H ealthc are Enter p r ise

Implement hardware, software, and/or procedural mechanisms that record and examine activity in information systems that contain or use electronic PHI

Record actions related to elec1. Record actions related to electronic health information; the tronic health information date, time, patient identification, 2. Enable a user to generate an and user identification must be audit log for a specific period recorded when electronic health and to sort entries in the information is created, modified, audit log according to any of accessed, or deleted, along with the elements specified in the an indication of which action(s) standard occurred and by whom

An individual has a right to receive an accounting of disclosures of PHI made by a covered entity in the six years before the date on which the accounting is requested, except for disclosures specifically listed in the regulation as exceptions, which include disclosures to carry out treatment, payment, and healthcare operations

Record disclosures made for treatment, payment, and healthcare operations purposes; include date, time, patient identification, user identification, and description of the disclosure

Optional. Record disclosures made for treatment, payment, and healthcare operations in accordance with the standard

Added by ARRA: If a covered entity uses or maintains an electronic health record, the HIPAA exceptions for treatment, payment, and healthcare operations do not apply; an accounting of disclosures for the previous three years must be made available to the individual Implement policies and procedures to protect electronic PHI from improper alteration or destruction, including mechanism to corroborate that electronic PHI has not been altered or destroyed in an unauthorized manner (addressable)

A hashing algorithm with a secu- 1. Create a message digest in rity strength equal to or greater accordance with the standard than SHA–1, as specified by the 2. Upon receipt of electronically NIST in FIPS Publication 180–83, exchanged health information, must be used to verify that elecverify that the information has tronic health information has not not been altered been altered in transit 3. Detect the alteration of audit logs

Implement technical security Any encrypted and integritymeasures to guard against protected link unauthorized access to electronic PHI that is being transmitted over an electronic communications network, including integrity controls and encryption

Encrypt and decrypt electronic health information when it is exchanged

Note: ARRA = American Recovery and Reinvestment Act; EHR = electronic health record; FIPS = Federal Information Processing Standards; HIPAA = Health Insurance Portability and Accountability Act; NIST = National Institute of Standards and Technology; PHI = protected health information; SHA = secure hash algorithm. Source: Adapted from ONC (2008).

351

EXHIBIT 16.3 HIPAA Technical Security Safeguards (continued)

352

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

clinic) attempting to perform a security-sensitive action (such as logging into a system) is the same as the asserted identity. Authentication addresses the risk that a person or entity masquerading as someone else will perform actions or access information for which he or it is not authorized. After the person or entity has asserted an identity that is recognized by the system, the authentication process starts. For example, a person may type in a user name that is easily guessable, such as an actual name. Authentication then asks the user to prove her identity by providing something she knows, such as a password; something she has, such as a random number generated by a hardware token; or something she is, such as a fingerprint. If the action the user is attempting to perform carries a very high risk, two-factor authentication may be needed. In this case, to prove his identity, a user must produce two pieces of evidence, such as a password and a fingerprint. For example, the Drug Enforcement Administration requires two-factor authentication for electronically prescribing controlled substances (US Department of Justice 2010). Authentication is foundational in that many other security safeguards depend on it. If authentication fails, any access-control decisions made on the basis of that identity, any documents that are digitally signed by that user, and any audit records of that user’s actions will be based on a bogus identity.

Access Control Access-control mechanisms mediate requests for access to protected resources and capabilities in a system to ensure that users are able to access only and all of the information they are authorized to access and to perform only and all of those actions they are authorized to perform. Access-control mechanisms generally make access decisions on the basis of the rules related to the identity and role of the requester or on the basis of a comparison between the trust attributes (e.g., security clearance) of the requester and the sensitivity label of the resource being requested (e.g., SECRET). Some access-control mechanisms also take into consideration context variables, such as the time the access is attempted and the location from which the request is received. Access-control mechanisms address the risk that an unauthorized person or entity is able to retrieve, read, write, or modify sensitive or safety-critical health information or to perform actions that the person or entity is not authorized to perform. The HIPAA Security Rule includes data encryption as an access control. Although closely related to access control, encryption is really a different protection mechanism. Whereas access control mediates attempts to access containers of data, such as files and database records, encryption garbles the data (bits) within those containers so that, even if the bits within the container are accessed, only authorized individuals are able to derive meaning (information) from those bits. Encryption is discussed again later in this chapter.

Ch ap te r 16 :   D a ta a n d In f o r ma ti on Sec ur ity in the H ealthc are Enter p r ise

Security Auditing and Accounting for Disclosures

353

Like Moths to a Flame

It seems some people find news and gossip about celebrities so enticSecurity auditing is the process ing that they are willing to violate laws to obtain it. All too often, of recording security-relevant employees in hospitals inappropriately access the records of celebriactivities that occur in an inforties who are undergoing treatment. When the breach is discovered, mation system, such as logthe employees typically are suspended or fired, and in some cases ging into the system, launchthe hospital is required to pay fines. Here is a list of recent celebritying a software application, and related breaches at healthcare facilities. opening a PHR. Recording an audit trail, along with an audit-record review process Date Celebrity Location Result and technology, helps ensure Palisades Medical 27 employees susOctober 2007 George Clooney that security safeguards are Center, New Jersey pended for a month without pay operating correctly. The audit UCLA Medical Cen- 13 employees fired March 2008 Britney Spears trail enables an organization and 6 suspended ter, California to monitor system actions and November 2008 Richard Collier Jacksonville Medical 20 employees fired Center, Florida to detect potential misuse and June 2010 Michael Jackson UCLA Medical Cen- Hospital fined intrusions so that appropriate ter, California $95,000; 4 employees fired actions can be taken. The audit trail also enables an organizaJuly 2013 Kim Kardashian Cedars-Sinai 6 employees fired Medical Center, tion to investigate breaches in California security policy enforcement. September 2014 Dr. Rick Sacra Nebraska Medical 2 employees fired (Ebola victim) Center, Nebraska Security auditing July 2015 Jason Pierre-Paul Jackson Memorial 2 employees fired addresses the risk that a sysSues Hospital, Florida tem administrator is unable to April 2016 Prince Moline, Illinois Unknown (emergency medical ascertain whether security safetechnicians) guards are operating correctly, whether authorized users are Sources: Information from Ornstein (2015) and ReminderCall.com (2017). acting in accordance with the organization’s security policy, and whether any unauthorized users have bypassed security safeguards to gain access to protected resources. Security auditing also serves as a deterrent in that the mere knowledge that their actions are being recorded may discourage some potential intruders and misusers from taking unauthorized or inappropriate actions. Closely aligned with auditing is ARRA’s requirement that organizations account for disclosures of health information. An accounting of disclosures is a record of every instance in which electronic health information is released, transferred, provided, opened up for access, or divulged in any other manner to an entity outside the holder of the information. An audit trail may help construct an accounting of disclosures, but it is not intended to be such an accounting. Audit trails are system records that contain detailed information a

a

a

a

a

a

b

b

a

b

354

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

of activity in a system, including attempts to take actions that are not allowed. For example, an audit trail would include activities such as attempting to log on, attempting to open a file, running a software application, reading a database record, and creating a folder. In contrast, an accounting of disclosures is a record of actual releases of health information to another entity. It addresses the risk that a trusted healthcare entity releases a patient’s record to a third party without the patient’s knowledge or consent.

Integrity Assurance No security mechanism can prevent data from being damaged or corrupted. Access controls can help ensure that only authorized individuals and entities can write to, modify, or delete resources that contain protected data, but access controls cannot prevent data from being accidentally corrupted or destroyed. Integrity assurance, or corroboration, guarantees that data that have been stored or transmitted are the same when they are used as when they were stored or transmitted. Integrity mechanisms address the risk that stored or transmitted data may be corrupted or changed in an unauthorized manner and then later used without the user realizing that the data are not the same as when they were originally stored or transmitted. The most commonly used integrity assurance mechanism is called a hash function, a mathematical formula executed against a block of data or a data stream to generate a number that represents that block or stream. Then, before the data are used, the same hash function is executed again. If the number the function produces is different from the original number, the user knows that the data have been changed—but not what elements have changed.

Encryption Another mathematical function used to protect data is encryption, which is simply the use of a mathematical algorithm (cipher) containing a secret variable (key) to scramble a block or stream of data such that the bits are no longer intelligible. The only way to make encrypted data intelligible again is to use the same cipher and a matching key to decrypt them. Encryption algorithms are either symmetric (for which the same key is used to both encrypt and decrypt the data) or asymmetric (for which one key is used to encrypt and another to decrypt the data). Asymmetric encryption is also known as public key encryption because one of the keys is openly made public and the other is kept private. Symmetric encryption is used only to encrypt and decrypt data, and it does that efficiently no matter what amount of data are involved and regardless of whether the data are at rest or in motion. Asymmetric encryption generally is used to secure exchanges between two parties, referred to as “Alice” and “Bob” in the following scenarios. Which key is used to encrypt and which to decrypt depends on the objective to be achieved.

Ch ap te r 16 :   D a ta a n d In f o r ma ti on Sec ur ity in the H ealthc are Enter p r ise

1. Authenticate the identity of a person or entity. In this scenario, to authenticate that Bob is, in fact, Bob, Alice encrypts some block of data using Bob’s public key. Bob decrypts the data using his private key. 2. Digitally sign a message or document. Here, sender Alice signs her message to Bob using her private key. Bob verifies that the message is actually from Alice by decrypting the signature using Alice’s public key. 3. Share a secret (symmetric) key to be used to exchange private or sensitive information. In this case, sender Alice encrypts the secret key using Bob’s public key. Receiver Bob decrypts the key using his own private key. Secure e-mail brings these scenarios together along with an integrity hash. To send an encrypted and digitally signed message to Bob, Alice first constructs her message and digitally signs it by generating an integrity hash value on the message and encrypting that value using her private key. Then Alice encrypts the message content using a secret key that she chooses. To keep the key secret, she encrypts it using Bob’s public key so that only Bob can decrypt it using a private key that only he knows. Alice sends the encrypted message to Bob, along with her digital signature and the encrypted secret key. When Bob receives the message, he uses Alice’s public key to decrypt the hash value, thus validating that the digital signature is hers. Bob uses his private key to decrypt the secret key and the hash value to verify that the message has not been changed. Encryption is used to address a number of risks. Both symmetric and asymmetric encryption are used to help ensure that even if an unauthorized entity gains access to a container of data, such as a file or network transmission, the “information” represented by the “data” will not be intelligible to anyone not possessing the secret (or private) key. Asymmetric encryption also addresses the risk that someone claims an identity other than her own or that someone repudiates having taken some action (such as sending a message).

New and Developing Opportunities and Challenges The standards specified in the HIPAA Privacy and Security Rules reflect principles that have been used for decades to protect the privacy of individuals, the confidentiality of information, the integrity of electronic data, and the availability of essential system resources and services. A healthcare organization that implements the FIPs set forth in the Privacy Rule and the administrative, physical, and technical safeguards required by the Security Rule will have a solid foundation of policies, processes, and safeguards for effectively managing its privacy and security risks.

355

356

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

However, risk is a continually evolving construct. We observe the changing nature of threats over time, as former friends become foes and business partners become competitors. New technologies emerge almost daily, each with its own set of vulnerabilities to be exploited. Even our values change over time, sometimes driven by technology changes. For example, consider the pay phone, valued as essential in years past and now almost a relic. In this section, we explore some of the emerging technologies that pose new opportunities to improve healthcare as well as new challenges to personal privacy and information security.

Data Mining As defined in chapter 5, data mining is the use of sophisticated search capabilities and analytical techniques on large databases to discover patterns, correlations, and trends that can be leveraged to produce new knowledge. The potential for applying data-mining techniques to make new biomedical discoveries, support decision making, and improve health outcomes is tremendous. At the same time, data mining represents a new threat to individual privacy in that by aggregating and extracting inferences from large volumes of clinical information, even information that has been deidentified in accordance with the HIPAA Privacy Rule can become individually identifiable. One need only consider today’s powerful search engines and personalized web services to realize that data mining makes remaining anonymous very difficult for any individual. New approaches to deidentification, as well as explicit privacy protections, are needed to enable healthcare to reap the benefits offered by data mining while protecting individual privacy.

Federated Identity To increase efficiencies and lower the operational overhead attendant to forcing users to separately log into each system, application, and database, single signon has become common practice in healthcare organizations. Single sign-on enables a user to log in once and then access all of the software applications and data needed across an enterprise. Single sign-on can be implemented in several ways, but the most common, standards-based approach is to pass a token called a security assertion to each system to which the user requests access. This approach also can be used to pass assertions between entities, allowing a user who has logged into a system in one organization to access resources in “federated” systems managed by other organizations. As discussed earlier, user authentication is the weakest link on which all security protections associated with individual users (e.g., access control, digital signature, audit) depend. This dependency poses a huge vulnerability for federated systems.

Virtualization Virtualization is the creation and use of a virtual (rather than an actual) computing resource, such as a software application, operating system, or storage device. Just about anything can be virtualized, producing software-as-a-service,

Ch ap te r 16 :   D a ta a n d In f o r ma ti on Sec ur ity in the H ealthc are Enter p r ise

platform-as-a-service, or infrastructure-as-a-service—also known by the generic term cloud computing. Virtualization offers significant security and privacy benefits because by virtualizing a resource, it is put under central control where security and privacy policies can be uniformly enforced. For example, installing an EHR software application on a workstation in a small health clinic may mean that the health information of all patients who come to that clinic for services is physically stored on that workstation, which may be protected by little more than the office door key and a smoke detector. If the clinic chooses instead to use the same EHR software application, but subscribes to it as a service instead of installing and maintaining it locally, the clinic’s data likely will reside in a large data center that is continuously protected at a high level, as explicitly defined in a service-level agreement. At the same time, cloud computing brings new risks because critical health data and essential applications are outside the physical and operational control of the subscriber. Organizations need to exercise due diligence, including carefully reviewing service-level agreements, in deciding which service providers are sufficiently reliable and trustworthy to provide virtualized services.

Social Technology Social technology refers to technologies that encourage and facilitate webbased social interactions by providing individuals with tools and services that enable them to create and manage a personal web identity. Twitter, Instagram, Facebook, and YouTube are well-known examples of social technology. Social technology offers significant opportunities to engage consumers in keeping themselves and their families healthy and in supporting patients and families who are dealing with medical conditions. Social technology also offers benefits to medical researchers both in recruiting research participants and in supporting the research itself. Personal privacy has been an ongoing issue in social media, and social networking sites have had to adjust their privacy practices in response to public concern. This concern frequently relates to a lack of transparency—that is, a social network’s practice of collecting or using private information without the user’s knowledge or consent. In the wake of the Cambridge Analytica crisis at Facebook, for example, users have begun to realize that their “use of social media platforms is not free, that they pay with attention to personalized ads and through loss of privacy about their personal information and social interactions” (Granados 2018). Healthcare providers and medical researchers can capitalize on the potential of social technology but must be vigilant in adhering to the FIPs discussed earlier.

Smartphones Smartphone applications (or apps) that help users manage their health or condition and connect with their physicians or specialists are being developed

357

358

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

at a rapid pace. In 2017 alone, 325,000 health/fitness and medical apps were available on all major app stores (Research 2 Guidance 2017). Smartphones are equipped with still and video cameras as well as geolocation technology. Such capability offers tremendous value for improving health and health outcomes. At the same time, the ability to locate an individual who does not want to be located or to surreptitiously make a video recording of an individual (and post it to the web) presents privacy risks that existing policy and fair practices have not anticipated.

Internet of Things The Internet of Things (IoT) is “the concept of basically connecting any device with an on and off switch to the Internet (and/or to each other). This includes everything from cellphones, coffee makers, washing machines, headphones, lamps, wearable devices and almost anything else you can think of” (Morgan 2014). In healthcare, the IoT is a network of connected devices, sensors, monitors, and persons capable of receiving and exchanging data, including “bedside monitors, smartwatches and fitness trackers, implanted medical devices, and any other object that transmits or receives a signal containing data that must be accessed or stored somewhere else” (Bresnick 2014). Connecting medical devices to a network (and ultimately to the Internet) may increase the value of the data and information collected by those devices. For example, it may allow a more timely use of those data for decision making by both providers and patients. However, as discussed earlier, increasing the accessibility of data on computing systems is associated with increased risk of security breaches. In this case, it may leave medical devices vulnerable to dangerous tampering and corruption, which could result in harm to patients. A recent incident involving the WannaCry ransomware software demonstrated these vulnerabilities. On May 12, 2017, the National Health Service (NHS) in the United Kingdom was attacked by WannaCry ransomware. Ransomware is a type of malware that holds a person’s or a company’s data and information hostage until a ransom is paid. The ransom demanded of the NHS was £415,000 (roughly $566,000) (Davis 2017). Around the same time, WannaCry also attacked 74 other countries. The attack exploited a vulnerability in the Microsoft Windows operating system. A patch to fix the vulnerability was released, but it is possible the patch has not been installed by many users. Fortunately, the United States was largely prepared for the attack, and only a few US-based websites were affected. Significantly, however, a Bayer Medrad device that is used to improve magnetic resonance imaging scans was infected by the malware (Fox-Brewster 2017). The WannaCry and other previous incidents have raised the specter of malware that could corrupt and cause the malfunction of medical devices—from pacemakers to insulin pumps to monitors—leading to harm or even deaths.

Ch ap te r 16 :   D a ta a n d In f o r ma ti on Sec ur ity in the H ealthc are Enter p r ise

The US Food and Drug Administration (FDA) has taken an active role in protecting the security of medical devices. Among its recent actions are the following (FDA 2018): On December 27, 2016, the FDA released the guidance: Postmarket Management of Cybersecurity in Medical Devices. . . . On May 18–19, 2017, the FDA partnered with the National Science Foundation (NSF) and Department of Homeland Security, Science and Technology (DHS, S&T) to hold a public workshop, Cybersecurity of Medical Devices: A Regulatory Science Gap Analysis. On January 12, 2017, the FDA held a webinar on the guidance: Postmarket Management of Cybersecurity in Medical Devices. . . . On August 29, 2017, the FDA issued a Safety Communication informing patients and health care providers about the release of a firmware update to address cybersecurity vulnerabilities identified in Abbott’s (formerly St. Jude Medical) implantable cardiac pacemakers.

Authentication Redux Systems of patient care are part of a larger environment that may affect the quality of patient care. In fact, the character and content of the data and information produced and used in patient care may be affected by that larger environment. Failures of authentication in systems other than patient care may affect the quality of that care and its associated data and information. Here, we consider two systems related to patient care—health sciences education and health sciences research.

Authentication Issues Related to Health Sciences Education Obviously, the quality of healthcare delivery is based on the antecedent education and training of clinicians, technicians, and managers. Persons hired by a healthcare organization who do not have the knowledge and competencies they claim to possess may compromise the quality of care provided by that organization and, by extension, the quality of the data and information collected and maintained by that organization. For a variety of reasons, online education is increasingly a standard feature of higher education. Online education is cost-effective for the educator and provides flexibility to the student. Unfortunately, the growth of online education has ushered in a burgeoning industry of online “resources” that enable students to cheat in their online classes. Many (but not all) academic institutions require only a user ID and password to log in to online courses. Online services such as Boostmygrade.com and Takeyourclass.com take advantage of that lax

359

360

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

requirement. A student provides his user ID and password and, for a fee, the company has an impostor log in to take the course on the student’s behalf. This situation stems, in part, from the less-than-rigorous standards imposed by the US Department of Education (DOE) on agencies that accredit institutions of higher education. Accrediting agencies are recognized formally by the DOE (2009) to ensure that [they] are, for the purposes of the Higher Education Act of 1965, as amended (HEA), or for other Federal purposes, reliable authorities regarding the quality of education or training offered by the institutions or programs they accredit.

Recognition of an accrediting agency matters because only highereducation institutions accredited by these agencies are eligible to dispense Title IV funds. DOE-recognized agencies include The Higher Learning Commission and the Middle States Commission on Higher Education. To be DOE recognized, an agency must satisfy the requirements of 34 CFR Part 602—The Secretary’s Recognition of Accrediting Agencies, which includes Section 602.17(g). It states that, as part of accreditation, a university or college that offers (DOE 2009) distance education . . . [is required] to have processes in place through which the institution establishes that the student who registers in a distance education . . . course or program is the same student who participates in and completes the course or program and receives the academic credit. The [accreditation] agency meets this requirement if it requires institutions to verify the identity of a student who participates in class or coursework by using, at the option of the institution, methods such as—     (i) A secure login and pass code;   (ii) Proctored examinations; and (iii) New or other technologies and practices that are effective in verifying student identity.

The easiest, and presumably cheapest, way for a university or college to meet this requirement is to install a system of secure logins and passwords. Unfortunately, this way is also the least secure, and any online course using such a method is prone to failures of authentication and fraud by students enrolled in the course. Hiring graduates of programs affected by such fraud may conceivably have a negative effect on the quality of care provided by the healthcare institution, as well as on the quality of data and information collected in the course of delivering that care.

Authentication Issues Related to Health Sciences Research A key characteristic of high-quality patient care is that it is evidence based: Diagnoses made and treatments carried out are supposed to be based on results of

Ch ap te r 16 :   D a ta a n d In f o r ma ti on Sec ur ity in the H ealthc are Enter p r ise

rigorous scientific study, including basic science research, clinical research, and healthcare services research. A fundamental characteristic of science is the process of peer review. Publication and dissemination of the results of research are subject to meticulous review by experts in the field. Typically, this involves the editor of a scientific journal or the chair of a scientific conference assigning a submitted paper to three or four experts for review, and the paper is then accepted or not accepted for publication or presentation depending on the results of that review. Scientific papers that have been published are sometimes retracted or removed from publication. They may be retracted for a variety of reasons. For example, the reported results may have been based on data that were manufactured. Or the paper may have been supported by faked peer review. The Center for Scientific Integrity, a group that champions transparency in scientific publications, maintains Retraction Watch, a database of scientific papers that have been retracted. In its 2017 Retraction Watch Year in Review, the center addressed the instances of faked peer reviews in one journal (Center for Scientific Integrity 2017): Springer is retracting 107 papers from one journal after discovering they had been accepted with fake peer reviews. . . . To submit a fake review, someone (often the author of a paper) either makes up an outside expert to review the paper, or suggests a real researcher—and in both cases, provides a fake email address that comes back to someone who will invariably give the paper a glowing review. In this case, Springer, the publisher of Tumor Biology through 2016, told us that an investigation produced “clear evidence” the reviews were submitted under the names of real researchers with faked emails. Some of the authors may have used a third-party editing service, which may have supplied the reviews.

Faked peer review is the result of failure of authentication of peer reviewers by the journal or publication. Such faulty science, if translated into clinical practice, may have negative consequences on patient care and on the quality of data and information collected in the course of that care. Retraction Watch is currently in beta testing and contains more than 16,000 reports of retracted scientific papers.

Conclusion This chapter discusses the general risks associated with data and information security breaches in healthcare, as well as the prevalence of such breaches. Concepts associated with healthcare-related privacy and security are introduced, including risk, trust, FIPs, and assurances necessary to attain and maintain information security and personal privacy. FIPs are shared throughout the world and

361

362

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

serve as the framework for protecting individual privacy. We review the technical security safeguards required by the HIPAA Security Rule and the risks that each measure is designed to address, and we identify how these measures are supported by implemented standards and features in certified EHR technology. Some leading-edge technologies are discussed that offer both opportunities for and challenges to protecting electronic health information and individual privacy. Authentication threats to evidence-based healthcare are also examined.

Chapter Discussion Questions 1. Draw a Venn diagram that represents the logical relationship among privacy, security, and safety. What risks might be representative of the overlap between privacy and safety? 2. Consider a physician practice that is transitioning from paper records to an electronic system. The receptionist area contains both hanging folders and the desktop computer on which the practice management system runs. During the transition, a patient’s health information (a valued asset) exists in two states: on paper in a hanging folder and in an EHR on a computer. Characterize the risks (low, medium, or high) for each state in terms of threats, vulnerabilities, and probability of a breach. 3. Privacy is a human value that depends on context. For example, people who share private personal information with Facebook friends may not want to share that same information with their physician. Or an HIV test result may be considered more sensitive than a cholesterol test result. What factors do you think contribute to how individuals assign privacy value to health information? 4. Suppose a miscreant manages to capture the password of a physician in a hospital that is part of a large integrated delivery network (IDN), which shares identity assertions across all hospitals and clinics. Knowing that the IDN typically uses lastname_first-initial as user identifiers, the miscreant then remotely logs into the EHR system using the physician’s user ID and captured password. Discuss the implications of this authentication failure on the access-control mechanisms, audit trail, and secure e-mail application in place. 5. A common practice in the peer review of scientific papers is for the editor or conference chair to request that the submitting author provide the names and contact information of potential reviewers. How might this practice put the entire edifice of evidence-based patient care at risk? 6. Explain why the health Internet of Things may be particularly vulnerable to security breaches.

Ch ap te r 16 :   D a ta a n d In f o r ma ti on Sec ur ity in the H ealthc are Enter p r ise

Case Study  Heinz Children’s Health Dixie B. Baker Heinz Children’s Health is a small pediatric practice serving the healthcare needs of children in a small, rural community. Twenty-five years ago, Dr. Helen Heinz founded the practice, which now includes two physician assistants, two registered nurses, a home health nurse, an office manager, and a receptionist. The practice has always used paper records, but when Dr. Heinz learns that, under the American Recovery and Reinvestment Act (ARRA) of 2009, the Centers for Medicare & Medicaid Services is offering significant incentives to eligible professionals who adopt electronic health record (EHR) technology and demonstrate its meaningful use, she sees an opportunity to move her practice into the electronic age. Once she decides to adopt an EHR system, Dr. Heinz is faced with another decision: which system to adopt. She discovers that there are not only several EHRs available from which to choose but also several ways to adopt them. The first option is to license the software and run it on a server that would be installed in her office. However, this would mean hiring an information technology (IT) person to set up the system, configure it, install upgrades, and keep it running. The second option is to license the software and run it on a server that would be installed in a data center. This option is not much different from the first option, except that the machine would be located somewhere else. Dr. Heinz still would need to hire IT staff. The third option is to subscribe to an EHR software-as-a-service (SaaS), where she and her staff would just log in to the EHR over the Internet to get access to their patients’ records. This option would not require Dr. Heinz to hire an IT person to install, configure, upgrade, or maintain any software and hardware. The SaaS provider would configure the software for the practice, create accounts for everyone in the office, provide 24/7 access, install updates, and keep it running smoothly. Plus, several SaaS vendors have pointed out that by subscribing to an EHR service, Dr. Heinz would not have to worry about HIPAA (Health Insurance Portability and Accountability Act of 1996) compliance; the provider would take care of that for her. This sounds too good to be true (as it later turns out to be). The SaaS vendor that Dr. Heinz selects is Fleet Software, a small, local company that for the past ten years has been developing and implementing custom software for businesses in the area. The owner, Jake Fleet, is friends with Dr. Heinz’s son; he is charming and widely considered to be a “guru” in computers and software. When he reads about the ARRA’s (continued)

363

364

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

incentives, his entrepreneurial mind sees an opportunity to get into the EHR software business. He has developed a lot of business software but has never developed EHR software. To avoid having to develop software for business processes he’s not familiar with and to enable him to get to the market sooner, he decides to license an EHR product and make it available through a SaaS subscription model. Jake checks out some of the leading EHR and practice-management products, both commercial and open source, designed for small practices. He hires his cousin, a retired physician, to help with product selection and training. After calling several references who have done business with Fleet Software and are delighted with its work, Dr. Heinz signs up for Fleet’s SaaS offering. Because Fleet Software will be providing a service involving protected health information, Heinz Children’s Health signs a business associate agreement with Jake, as required by HIPAA. Fleet Software creates accounts for all of the practice’s staff and, using the existing paper records, sets up EHR records for patients with chronic conditions and for those who have appointments scheduled over the next two months. Jake and his cousin train the staff on how to use the software, and Jake gives Dr. Heinz a number to call if her staff run into any problems with the software. Two months later, Heinz Children’s Health goes “live” with its new EHR. It takes some time for Dr. Heinz and her staff to get accustomed to using the system instead of the paper record, but they quickly see some real advantages. Information in the EHR is organized and always easy to find. It provides reminders when lab tests and vaccinations are due, and it is equipped with pop-up calendars to help in scheduling a patient’s next appointment. Graphs show each patient’s growth in comparison with mean growth patterns. Jake and his cousin have done a good job selecting and customizing the EHR application for the practice. Things go smoothly until one morning when Dr. Heinz and her staff discover that all of their patient records are gone from the EHR. They can log in, but no records are showing up. In a panic, Dr. Heinz calls Jake. After doing some checking, he discovers that the system upgrade they rolled out the night before has inadvertently overwritten the storage partition containing the records. Fortunately, he has backed up the records, and he promises to have the backup reloaded within a couple of hours. Meanwhile, Dr. Heinz and her staff revert back to paper until the EHR is restored. Several months later, Dr. Heinz receives an irate phone call from the mother of a child who has been diagnosed with sickle cell disease. The mother’s anger was triggered when a neighbor expressed sympathy even though the

Ch ap te r 16 :   D a ta a n d In f o r ma ti on Sec ur ity in the H ealthc are Enter p r ise

mother has not mentioned the diagnosis to anyone. Dr. Heinz questions her staff and learns that the receptionist, after seeing the child with his distraught mother, checked his medical record in the EHR and saw the diagnosis. The receptionist became upset and discussed the information with the neighbor, who in turn approached the mother. Dr. Heinz is surprised that the receptionist could even view the patient information, particularly given that the EHR is supposed to be HIPAA compliant, as Fleet Software had promised. She is also astounded when she walks out into the reception area and sees the receptionist’s screen displaying another patient record in full sight of those waiting for their appointments. After closing the exposed record and reprimanding the receptionist privately, she calls Jake. He explains that the company just creates the accounts, but Heinz Children’s Health is responsible for telling them about user privileges and any access restrictions that should be set up for those accounts. Upon further discussion, Dr. Heinz learns that Jake considers the clinic responsible for a number of other HIPAA requirements. Later, Fleet Software’s server suffers a malicious attack. The attacker takes advantage of a known vulnerability in the server software, allowing the attacker to bypass user authentication and thus gain unauthorized access to all data stored on the server. Jake reports the incident to Dr. Heinz and provides a list of patients whose information may have been exposed. When Dr. Heinz asks him what the company is doing about the problem, he explains that the company investigated the incident and found that the attacker had exploited a known vulnerability in the server–software vendor’s critical security update. Fleet Software then generated a list of individuals whose information could have been exposed and reported this information to its clients, including Heinz Children’s Health. He declares that these steps complete the company’s regulatory and contractual obligations as a business associate, leaving the clinic responsible for notifying the individuals whose information may have been exposed. Dr. Heinz is now rethinking her decision to adopt an EHR to qualify for the incentive payment. The clinic may be better off using paper records until she retires.

Case Study Discussion Questions 1. Heinz Children’s Health has experienced risks to information confidentiality, data integrity, service availability, and the business itself. Identify the consequences, the vulnerabilities exploited, the threats that exploited them, and the ways these risks could have been mitigated. (continued)

365

366

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

2. Think about a clinical practice that subscribes to an EHR service rather than implementing its own system. How would you assign responsibility for the following HIPAA requirements? a. Deciding what access rights and privileges each user or user group should be assigned b. Configuring system accounts to restrict user access and privileges c. Screening, supervising, and training personnel on appropriate use of patient information d. Installing, maintaining, testing, and managing configuration of server security hardware and software e. Installing and updating malicious code–detection software f. Developing plans for emergency-mode operations and disaster recovery g. Reporting and responding to a security incident h. Providing security and privacy awareness training and reminders i. Selecting and naming an individual to be responsible for HIPAA privacy and security compliance j. Providing online data storage and secure backup k. Securing paper copies of patient records l. Ensuring that user access devices are positioned so that the screens cannot easily be viewed by patients in the waiting room m. Securing the transport link between the system and users’ browsers so that the integrity and confidentiality of all health information are protected n. Auditing system activities o. Providing, at the patient’s (or patient representative’s) request, an accounting of all disclosures of the patient’s health information p. Notifying patients whose health information may have been exposed through a breach of the system security protections 3. What are some of the risks that are not addressed by HIPAA but that a SaaS subscriber may need to consider?

Additional Resources Genetic Information Nondiscrimination Act of 2008: www.eeoc.gov/laws/statutes/gina.cfm. Privacy Act of 1974: www.justice.gov/opcl/privacy-act-1974. The Center for Scientific Integrity Retraction Watch: http://retractiondatabase.org/ RetractionSearch.aspx.

Ch ap te r 16 :   D a ta a n d In f o r ma ti on Sec ur ity in the H ealthc are Enter p r ise

References Bresnick, J. 2014. “What Does the ‘Internet of Things’ Mean for Healthcare?” HealthITAnalytics. Published November 5. https://healthitanalytics.com/ news/internet-things-mean-healthcare. CDW. 2017. “The Cost of a Data Breach for Healthcare Organizations.” Published October 26. www.cdw.com/content/cdw/en/articles/security/2017/10/26/ the-cost-of-a-data-breach-for-healthcare-organizations.html. Center for Scientific Integrity. 2017. “2017 Retraction Watch Year in Review.” Published December 29. http://retractionwatch.com/2017/12/29/2017-retractionwatch-year-review-hint-database-nearly-done/. Centers for Medicare & Medicaid Services (CMS). 2003. “Health Insurance Reform: Security Standards; Final Rule.” 45 CFR Parts 160, 162, and 164 (SecRule 2003). Federal Register 68 (34): 8334–81. Davis, J. 2017. “Updated: Hospitals in UK National Health Service Knocked Offline by Massive Ransomware Attack.” Healthcare IT News. Published May 12. www. healthcareitnews.com/news/updated-hospitals-uk-national-health-serviceknocked-offline-massive-ransomware-attack. Fox-Brewster, T. 2017. “Medical Devices Hit by Ransomware for the First Time in US Hospitals.” Forbes. Published May 17. www.forbes.com/sites/thomas brewster/2017/05/17/wannacry-ransomware-hit-real-medical-devices/. Granados, N. 2018. “Transparency: How Facebook Can Recover from the Cambridge Analytica Crisis.” Forbes. Published March 26. www.forbes.com/sites/nelson granados/2018/03/26/transparency-how-facebook-can-recover-from-thecambridge-analytica-crisis/. Morgan, J. 2014. “A Simple Explanation of ‘The Internet of Things.’” Forbes. Pub­ lished May 13. www.forbes.com/sites/jacobmorgan/2014/05/13/ simple-explanation-internet-things-that-anyone-can-understand/. O’Brien, S. A. 2017. “Giant Equifax Data Breach: 143 Million People Could Be Affected.” CNN. Published September 8. http://money.cnn.com/2017/09/07/technology/business/equifax-data-breach/index.html. Office of the National Coordinator for Health Information Technology (ONC). 2010. “Health Information Technology: Initial Set of Standards, Implementation Specifications, and Certification Criteria for Electronic Health Record Technology; Final Rule.” Federal Register 75 (144): 44589–654. ———. 2008. “Nationwide Privacy and Security Framework for Electronic Exchange of Individually Identifiable Health Information.” Published December 15. www.healthit.gov/sites/default/files/nationwide-ps-framework-5.pdf. Olmstead v. United States, 277 U.S. 438 (1928). Organisation for Economic Co-operation and Development. 1980. “Guidelines on the Protection of Privacy and Transborder Flows of Personal Data.” Accessed March 1, 2012. www.oecd.org/sti/ieconomy/oecdguidelinesontheprotection ofprivacyandtransborderflowsofpersonaldata.htm.

367

368

Hea lt h I n fo r ma ti c s: A Sy ste ms Pe r s p e c tive

Ornstein, C. 2015. “Celebrities’ Medical Records Tempt Hospital Workers to Snoop.” Published December 10. www.propublica.org/article/clooney-to-kardashiancelebrities-medical-records-hospital-workers-snoop. Privacy Rights Clearinghouse. 2018. “Data Breaches.” Accessed January 24. www. privacyrights.org/data-breaches. ReminderCall.com. 2017. “Top 15 Celebrity HIPAA Fails and Their Consequences.” Published January 13. www.remindercall.com/celebrity-hipaa-fails/. Research 2 Guidance. 2017. “mHealth App Economics 2017—Current Status and Future Trends.” Accessed January 24, 2018. https://research2guidance.com/product/ mhealth-economics-2017-current-status-and-future-trends-in-mobile-health. US Department of Education (DOE). 2009. “Accreditation in the United States.” Accessed January 24, 2018. www2.ed.gov/admins/finaid/accred/accreditation_pg11.html. US Department of Health and Human Services (HHS). 2001. “Standards for Privacy of Individually Identifiable Health Information.” Published July 6. http:// aspe.hhs.gov/admnsimp/final/pvcguide1.htm. ———. 2000. “Standards for Privacy of Individually Identifiable Health Information. Final Privacy Rule Preamble.” Published December 28. https://aspe.hhs.gov/ report/standards-privacy-individually-identifiable-health-information-finalprivacy-rule-preamble. ———. 1991. “Federal Policy for the Protection of Human Subjects; Notices and Rules.” Published June 18. www.fda.gov/ScienceResearch/SpecialTopics/ RunningClinicalTrials/ucm118862.htm. US Department of Health and Human Services (HHS) Office for Civil Rights. 2018. “Breach Portal: Notice to the Secretary of HHS Breach of Unsecured Protected Health Information.” Accessed January 24. https://ocrportal.hhs.gov/ocr/ breach/breach_report.jsf. US Department of Health, Education, and Welfare (HEW). 1973. “Records, Computers, and the Rights of Citizens.” Published July 1. http://aspe.hhs.gov/ DATACNCL/1973privacy/tocprefacemembers.htm. US Department of Justice. 2010. “Electronic Prescriptions for Controlled Substances; Final Rule.” Federal Register 75 (61): 16236–319. www.deadiversion.usdoj. gov/fed_regs/rules/2010/fr0331.pdf. US Federal Trade Commission (FTC). 2007. “Fair Information Practice Principles.” Accessed January 24, 2018. https://web.archive.org/web/20090331134113/ http://www.ftc.gov/reports/privacy3/fairinfo.shtm. ———. 1998. “Privacy Online: A Report to Congress.” Accessed January 24, 2018. www.ftc.gov/sites/default/files/documents/reports/privacy-online-reportcongress/priv-23a.pdf. US Food and Drug Administration (FDA). 2018. “Cybersecurity.” Updated April 17. www.fda.gov/MedicalDevices/DigitalHealth/ucm373213.

APPENDIX: PROFESSIONAL SOCIETIES, ACCREDITING AGENCIES, AND ADDITIONAL INSIGHTS IN HEALTH INFORMATICS

Timothy B. Patrick The information in this appendix is taken directly from the organizations’ websites. The insights presented are those of the authors and do not represent the views of the organizations featured here.

Professional Societies American Health Information Management Association (AHIMA) www.ahima.org Mission AHIMA works to transform “healthcare by leading HIM, Informatics, and Information Governance.” History AHIMA traces its history back to 1928, when the American College of Surgeons established the Association of Record Librarians of North America to “elevate the standards of clinical records in hospitals and other medical institutions.” In 1938, the association changed its name to the American Association of Medical Record Librarians to more accurately reflect its membership. The association changed its name in 1991 to the American Health Information Management Association. Its current name captures the expanded scope of clinical data beyond the single hospital medical record to “health information comprising the entire continuum of care.”

369

370

A p p en dix

AMIA www.amia.org Mission AMIA supports education and research in these five domains: 1. Translational bioinformatics 2. Clinical research informatics 3. Clinical informatics 4. Consumer health informatics 5. Public health informatics Core Purposes • Advance the science of informatics • Promote the education of informatics • Assure that health information technology is used most effectively to promote health and healthcare • Advance the profession of informatics • Provide services for our members such as networking and opportunities for professional development History AMIA was created as a merger of several different professional groups in the 1980s—the American Association for Medical Systems and Informatics, the American College of Medical Informatics, and the Symposium on Computer Applications in Medical Care. Originally, “AMIA” was merely an acronym of the American Medical Informatics Association’s name. However, during the 1990s, AMIA membership objected to the name, arguing that “medical informatics” was physician-centric and thus inappropriate for a society that included nurses, librarians, and allied health professionals as members. As a result, the organization’s name was officially changed to “AMIA.”

Association for Information Science and Technology (ASIS&T) www.asist.org Mission The purpose of ASIS&T is “to advance the information sciences and related applications of information technology by providing focus, opportunity, and support to information professionals and organizations.” Medicine and nursing are among the disciplines served by ASIS&T.

A p p endix

History ASIS&T was founded on March 13, 1937, as the American Documentation Institute (ADI), a service organization made up of individuals nominated by and representing affiliated scientific and professional societies, foundations and government agencies. Its initial interest was in microfilm as an aid to information dissemination. In 1968, responding to the increasing production and demand for information, ADI changed its name to ASIS. In 2000, the name was changed to ASIS&T.

Healthcare Information and Management Systems Society (HIMSS) www.himss.org Mission HIMSS exists to “globally, lead endeavors optimizing health engagements and care outcomes through information and technology.” History HIMSS was founded in 1961 at the Georgia Institute of Technology.

Medical Library Association (MLA) www.mlanet.org Mission MLA “fosters excellence in the professional practice and leadership of health sciences library and information professionals in order to enhance the quality of health care, education, and research throughout the world.” History MLA was founded as the Association of Medical Librarians in 1898. This name was changed to Medical Library Association in 1907. MLA is known for its cooperation with AMIA, as well as for its influence on health informatics—not surprising given that medical librarianship and health informatics are related to the National Library of Medicine. In 2001, for example, the MLA Donald A. B. Lindberg Research Fellowship was established “to fund research that extends the knowledgebase of health sciences librarianship and informatics and improves the practice of the profession.”

371

372

A p p en dix

Accreditation and Other Organizations Association of Specialized and Professional Accreditors (ASPA) www.aspa-usa.org ASPA collaborates with higher education and government officials to enhance education and accreditation. Both the Commission on Accreditation for Health Informatics and Information Management and the Commission on Accreditation for Health Management Education (see below) are ASPA members.

Commission on Accreditation for Health Informatics and Information Management (CAHIIM) www.cahiim.org CAHIIM is the accrediting agency for master’s programs in health information management and health informatics. The two customers of CAHIIM are AHIMA and AMIA. Both AHIMA and AMIA specify the competencies or subject matters that define their respective professions, and they work with CAHIIM to develop the accreditation standards for associated master’s programs. Master’s programs in health information management are accredited according to AHIMA specifications, whereas the accreditation of master’s programs in health informatics is based on competencies specified by AMIA.

Commission on Accreditation for Health Management Education (CAHME) www.cahme.org CAHME is the accrediting agency for master’s programs in healthcare management and health administration. The role of CAHME is to improve health management education by doing the following: • Setting measurable criteria for excellent healthcare management education • Supporting, assisting, and advising programs that seek to meet or exceed the criteria and continuously improve • Accrediting graduate programs that meet or exceed the criteria • Making this information easily available to interested constituencies

Council for Higher Education Accreditation (CHEA) www.chea.org

A p p endix

CHEA comprises 3,000 institutions of higher education. It recognizes institutional and programmatic accrediting agencies, such as CAHIIM and CAHME.

Health Professions Accreditors Collaborative (HPAC) http://healthprofessionsaccreditors.org Formed in 2014, HPAC aims “to enhance accreditors’ ability to ensure graduates of health profession education programs are prepared for interprofessional collaborative practice.” CAHIIM is a member of HPAC.

Additional Insights Natural Affinity Between Health Informatics and Health Administration These two disciplines have much in common, as shown by the following report on winners of the 2018 CAHME Cerner Award for Excellence in Healthcare Management Systems Education (Business Wire 2018): The University of Alabama at Birmingham Master of Science in Health Administration for integrating experiential learning in informatics with the UAB health system, and incorporating UAB alumni in knowledge-sharing and professional development; and the University of Missouri Department of Health Management and Informatics for its informatics program that focuses on lifelong leadership development, an integrated dual degree or certificate in health informatics, application of QI/PI methods, integrated Six Sigma Green Belt certification, and peer based learning.

Whether this trend that unifies the two professions will continue is unknown.

AMIA and AHIMA Competencies At this time, CAHIIM does not accredit master’s programs in health informatics that follow AHIMA-specified competencies, but these programs are accredited if they follow AMIA-specified competencies. AHIMA itself provides certification in health informatics to individuals. This certification can be obtained after passing the Certified Professional in Health Informatics (CPHI) examination. One way to qualify to take this exam is to graduate with a “Master’s degree or higher in health informatics from a regionally accredited academic institution” (AHIMA 2018). This master’s degree in health informatics is not required to come from a CAHIIM-accredited health informatics program (which is based on the AMIA competencies). Some attempts have been made to provide a crosswalk or mapping between the AHIMA CPHI exam and the AMIA competencies. CAHIIM has yet to issue an official publication that demonstrates this consistency. This situation may cause confusion for those

373

374

A p p en dix

interested in pursuing health informatics education and training as well as for professionals in the field.

References American Health Information Management Association (AHIMA). 2018. “Certified Professional in Health Informatics.” Accessed February 10. www.ahima.org/ certification/CPHI. Business Wire. 2018. “CAHME Announces 2018 Award Winners.” Published January 29. www.businesswire.com/news/home/20180129005809/en/CAHMEAnnounces-2018-Award-Winners.

GLOSSARY Accountable care organization (ACO): “Group of doctors, hospitals, and other health care providers, who come together voluntarily to give coordinated high quality care to their Medicare patients” (Centers for Medicare & Medicaid Services 2018; see chapter 4) Agent-based (AB) modeling: Type of modeling used to study the behavior of systems on the basis of the interactions between agents or entities Big Data: In healthcare, very large collections of data, typically represented in a variety of ways, on individual patients, populations of patients, animals, biological objects, and other health-related matters Bioinformatics: Discipline that combines the biological sciences (microbiology, biochemistry, physiology, genetics) with computational fields (e.g., statistics, computer science) Clinical decision support system (CDSS): Software that presents users with a knowledge base, patient-specific data, and related information at the point of care to enhance healthcare provision and management Clinical guidelines: Evidence-based clinical information that guides clinicians during clinical encounters; also used to represent alerts and reminders embedded in the EMR Clinical protocols: Evidence-based clinical information that informs clinicians in a clinical process, as opposed to a clinical encounter; similar to critical pathways, a concept developed in industrial engineering Common data element (CDE): Pair consisting of a variable and a set or domain of permissible values for the variable; used by different studies or healthcare settings Community of practice: Clinical team that is both self-organizing and structured and rewarded on the basis of the needs of the patient, thus transcending individual clinicians, organizations, or systems Complex adaptive system: Organization with a large number of interdependent parts or agents that have their own pattern relationships, present interaction complexity, and are self-organizing but can adapt to their environments and help create those environments Consumer health informatics: Area of health informatics that focuses on the implementation and evaluation of system design to ensure that it interacts directly with the consumer, with or without the involvement of healthcare providers

375

376

Glo s s ar y

Controlled terminology: Set of domain-specific multiword terms selected from natural language and organized by hierarchical and associative relationships Data mining: Use of sophisticated search capabilities and analytical techniques on large databases to discover patterns, correlations, and trends that can be leveraged to produce knowledge Data warehouse: Data from a broad range of sources linked together and stored for easy retrieval, reporting, analysis, and decision making Decision science: Systematic analysis of the complexity and dynamic nature of decision making Discrete event (DE) modeling: Type of modeling used primarily to study processes, streamline them, and reduce bottlenecks through better resource allocation, capacity utilization or standardization, and mechanization of routine processes E-health: Use of telecommunication platforms, mobile and ubiquitous hardware and software, and advanced information systems to support and facilitate healthcare delivery and education Electronic health record (EHR): Documentation of the clinical workflow; provides alerts, reminders, therapy plans, and medication orders Electronic medical record (EMR): Person’s electronic clinical information created, integrated, managed, and accessed by authorized clinicians, nurses, coders, and other health professionals Enterprise strategy: Plan to develop and align the internal capabilities of an organization, including IT, with risks and opportunities in the external market Fair information practices (FIPs): Foundation of information security and privacy law and regulations in the United States and throughout the world; constitute fair and responsible information stewardship, which is essential to establishing and maintaining public trust when collecting, using, disclosing, and sharing personal information GenBank: “A comprehensive database that contains publicly available nucleotide sequences for almost 260,000 formally described species” (Benson et al. 2013; see chapter 9) Genomics: “A branch of biotechnology concerned with applying the techniques of genetics and molecular biology to the genetic mapping and DNA sequencing of sets of genes or the complete genomes of selected organisms, with organizing the results in databases, and with applications of the data (as in medicine or biology)” (Merriam-Webster 2018; see chapter 9) Global institutional network: Legal arrangement between two or more healthcare institutions in two or more countries for the purpose of collaborating on clinical care, education, or development of shared evidence-based clinical decision support systems Health information exchange (HIE): Framework that enables the movement of patient health data and information across organizations that are geographically dispersed by using nationally recognized standards

Glossar y

Health systems informatics: Application of multidisciplinary sciences to transform (not just automate) the structure and behavior of systems, organizations, and individuals who interact to provide personalized care Information management strategy: Plan to acquire, manage, use, and deliver information through products or services to internal or external customers Information technology (IT) architecture: Framework of information system infrastructure that aligns IT with enterprise strategy; could be hierarchical in a functional organization and focused on efficiency, or it could be strategic and form the basis for health system design Internet of Things (IoT): “Interconnection via the Internet of computing devices [including smart home devices] embedded in everyday objects, enabling them to send and receive data” (Höller and Höller 2014; see chapter 8) Knowledge-based clinical decision support system: Collection of scientific evidence and expert and experiential knowledge (including that generated through artificial intelligence) used to inform clinical decisions Knowledge management: Process of harnessing all the relevant knowledge assets of an organization and system and then deploying them to achieve optimal system performance Knowledge organization: Organization in which leaders engage in innovative and creative pursuits involving health professionals and other knowledge workers Meaningful use: Measure of the level of application of IT in clinical decision support systems Medical home: Model of care in which the services provided by a team of health professionals are coordinated by a primary care physician and involve the patient Medical informatics: Discipline that deals with the structure and properties of clinical information generated from clinical trials and medical records; generally includes imaging informatics and clinical informatics Metadata: Data about data and information Metadata schema: Collection of fields or data elements, names for the fields, definitions for the fields, and specifications of what constitute permissible values for the fields; also known as data dictionary Modeling: Method of studying, understanding, and then replicating the complexities of the real world to design, change, and improve systems Natural language processing (NLP): Computational approach to processing human language Nursing informatics: “A specialty that integrates nursing science with multiple information management and analytical sciences to identify, define, manage and communicate data, information, knowledge and wisdom in nursing practice” (American Nurses Association 2015, 1; see chapter 7)

377

378

Glo s s ar y

Online analytical processing (OLAP) architecture: IT setup that enables the user to “slice and dice” the data in multiple dimensions to provide insights Open system: System with an environment that can affect its state and with which the system interacts Open systems theory: Theory that views organizational and clinical functions in terms of their relationship with and contribution to overall system performance Patient portal: “Secure online website that gives patients convenient, 24-hour access to personal health information from anywhere with an Internet connection” (Office of the National Coordinator for Health Information Technology 2017; see chapter 7) Personal health record (PHR): Person’s digital medical record that conforms to nationally recognized interoperability standards and is managed, shared, and controlled by the individual Postcoordinated concept expressions: Expressions that combine precoordinated concepts and semantic relationships Precision medicine: “Innovative approach to disease prevention and treatment that takes into account individual differences in people’s genes, environments, and lifestyles” (White House 2015; see chapter 9) Precoordinated concept expressions: Expressions that form the basis of controlled terminology; the building blocks of more complicated expressions of meaning Privacy: In the healthcare information context, assurance that one’s health information is collected, accessed, used, retained, and shared only when necessary and only to the extent necessary and that the information is protected throughout its life cycle using fair privacy practices consistent with applicable laws and regulations and the preferences of the individual Representational science: Branch of information science that studies how data and information are represented to support information storage and retrieval Security: Protection of the confidentiality of private, sensitive, and safety-critical information; the integrity of health data and metadata; and the availability of information and services through measures that authenticate user and system identity and data provenance and that maintain an accounting of actions taken by users, software programs, and systems Smart home: Personal living quarters with an automated network of devices and systems that operate together by sharing data System dynamics (SD) modeling: Type of modeling used to model complex nonlinear relationships between components and to study the dynamics of the system over time Systems thinking: Viewing parts or agents of a social or biological system as interdependent

Glossar y

Terminology mapping: Process of matching a term from one controlled terminology (or natural language) to a term in another controlled terminology Value measurement methodology (VMM): Framework and toolkit for estimating the effects of a given IT investment on society, values, costs, risks, and tangible and intangible returns Web 2.0 technology: Technology that replaced the traditional, static World Wide Web by enabling more community-based input, interaction, content sharing, and collaboration

379

INDEX Note: Italicized page locators refer to figures or tables in exhibits. ABI/INFORM, 263 AB modeling. See Agent-based modeling ACA. See Affordable Care Act Access-control mechanisms, 352 Accessibility: e-health applications and, 176–77 Access logs, 275 Accountability: safety and, 80–81 Accountable care organizations, 86–88, 98, 215, 330; best practices in, 87; definition of, 86; disruptive innovation and, 87; financing of, 324–25; as form of community practice, 107; future outlook for, 88; knowledge management in: case study, 41–44; knowledge socialization in, 38; purpose of, 86; as regional enterprises, 329; structuring and managing change in, 87 Accreditation: of medical homes, limitations with, 85 Accreditation agencies/organizations, 77, 360, 372–73 Ackoff, Russell, 216, 217, 218 Ackoff’s decision model: modification of, 217–18 ACOs. See Accountable care organizations ACS. See American College of Surgeons Active monitoring, 170 Actor-network theory: guidelines developed under, 62

ADAPTABLE distributed clinical trial: Greater Plains Collaborative Shared Health Research Information Network and, 32 ADI. See American Documentation Institute Administrative function: traditional design, 76, 76 Admissions: medication reconciliation and, 155; patient-centered care and, 153 Advanced practice nurses, 161 “Advancing Clinical Decision Support: Key Lessons in Clinical Decision Support Implementation” (Byrne), 129 Adverse events: detection of, 156 Advertising: prescription drug, 64 Affective decisions, 59, 63 Affordable Care Act, 219, 317, 327, 346; accountable care organizations and, 86; adjusted hospitalbased reimbursements and, 134; National Quality Strategy and, 123; population health provisions in, 209–11, 221; public sector information technology investment and, 334 Agency for Healthcare Research and Quality, 29, 209 Agent-based modeling: approach in, 113, 113; definition of, 111; strength of, 111–12; theories underlying, 112

381

382

I n d ex

AgileMD, 134–35 AGREE. See Appraisal of Guidelines for Research and Evaluation AHIMA. See American Health Information Management Association AHRQ. See Agency for Healthcare Research and Quality AI. See Artificial intelligence AI/RHEUM, 134 Alert fatigue: Clinical decision support systems and, 133 Algorithms: association-rule mining, 105, 106; classification, 105–6; clustering, 105, 106; encryption, 354; genetic, 126, 127; machinelearning, 30; mathematical, 101 Alignment: of health systems functions, 4; of multiple systems, 9; systems thinking, optimizing networked electronic health records, and, 32 “All of Us” research program: Precision Medicine Initiative and, 193 Allscripts, 134, 304 Alphabet, 89 Alphanumeric identifying codes, 254–55 Amazon.com, 104, 335 American Academy of Family Physicians, 84 American Academy of Pediatrics, 84 American Association for Medical Systems and Informatics, 370 American Association of Colleges of Nursing, 160 American Association of Health Plans, 29 American College of Medical Informatics, 370 American College of Physicians, 84 American College of Surgeons: National Surgical Quality Improvement Program, 275 American Documentation Institute, 371 American Health Information Management Association:

competencies, 373–74; mission and history, 369–70 American Medical Association, 29, 67 American Medical Informatics Association, 129 American Nurses Association: definition of nursing informatics, 148, 150 American Osteopathic Association, 84 American Recovery and Reinvestment Act, 349, 350, 363; accounting for disclosures of health information and, 353; Health Insurance Portability and Accountability Act’s Privacy and Security Rules and, 348; Health Information Technology for Economic and Clinical Health Act of, 299; passage of, 123 Amgen, 198 AMIA: competencies, 373–74; mission, core purposes, history, 370 ANA. See American Nurses Association Analytical modeling, 98 Analytics, 98; business intelligence and, 99; data mining and, 98–101, 100; for disease management and wellness: case study, 115–17; modeling and, 107, 108; solutions, informed decision making and, 286. See also Predictive analytics in knowledge management Ancker, J., 174 Anderson, R. M., 168 ANNs. See Artificial neural networks Apomediation, 174 Appraisal of Guidelines for Research and Evaluation, 239 Aramco, 242 Arena software, 111 ARLNA. See Association of Record Librarians of North America ARRA. See American Recovery and Reinvestment Act Artificial intelligence, 29–30, 56, 57; clinical decision support systems

Index

and, 133; computational clinical decisions and, 56; explainable, 30. See also Machine learning Artificial neural networks, 126–27; applications with, 127; architecture, 127; definition of, 126; layers in, 126–27 Ashley, Robert, 196 ASIS& T. See Association for Information Science and Technology ASPA. See Association of Specialized and Professional Accreditors Assembly lines, 82 Assessment: of clinical decision support system readiness, 130 Assets: knowledge, 22 Association for Information Science and Technology: mission and history, 370–71 Association of Record Librarians of North America, 369 Association of Specialized and Professional Accreditors, 372 Association-rule mining algorithms, 105, 106 Asymmetric encryption, 354, 355 ATM. See Automated teller machine Audit logs, 275 Audit trails, 353–54 Australia: centralized information technology architecture in, 231; electronic database system in, 232; electronic medical records in, 228; integrated health information system in, 233–34, 235 Authentication, 348, 352; health sciences education and, 359–60; health sciences research and, 360–61 Automated teller machine: transformational change in banking and, 10–11 Automating processes: transforming processes vs., 10 Autonomy: of medical faculty, 52; professional vs. individual, 50

Banking: transformational change in, 10–11 Bank One, 92 Barclay, Iowa: missed technological opportunities for, 13 Barclay, James, 13 Barcode medication administration, 154 Barcoding, 98 Bartram, T., 295 Bates, D. W., 128 Batini, C., 284 Bayer, 198 BCMA. See Barcode medication administration Begley, Glenn, 198 Behavioral determinants of health, 214, 215, 221 Behavioral Risk Factor Surveillance System surveys, 213 Benchmark data, 275 Berkshire Hathaway, 335 Bernstein, W. S., 335 Bertalanffy, Ludwig von, 9 Berwick, D. M., 210, 212 Beveridge model of healthcare financing, 316 Big Data, 32, 98–99, 198, 199; crisis of reproducibility and, 199; data representation and, 281; definition of, 195; “mother of all databases” pitfall and, 285; precision medicine and, 194–97 Bioinformatics, 4–5, 10, 10; definition of, 5; evolution of, 5; focus of, 5; medical informatics comparison, 5; methods, techniques, and theories, 6, 6 Biomedical data and information: management of, 252 Biomedical devices: data movement and, 280 Biomedical informatics: extended model, 9–10; knowledge base, 6; methods, techniques, and theories, 6; use of term, 5

383

384

I n d ex

Biomedical Information Science and Technology Initiative, 9; core competency, 6; formation of, 5 Biomedical literature: constructing surrogate representations of, 257 BIPMed. See Brazilian Initiative on Precision Medicine Bismarck, Otto von, 219 Bismarck model of healthcare financing, 316 BISTI. See Biomedical Information Science and Technology Initiative Blended information technology architecture, 231–32 Blondlot, René, 197–98 BMJ Quality and Safety, 238 Bohm Dialogue, 38 Bokhari, Q., 124 Boostmygrade.com, 359 Boundaries: system, 216 Boyd, D. M., 172 Brandeis, Louis, 344 Brazilian Initiative on Precision Medicine, 193, 194 Breach Portal, 343–44 BRFSS surveys. See Behavioral Risk Factor Surveillance System surveys Bundled payments, 215, 319; medical home financing and, 322–23 Burr, Richard, 196 Business function: basis of, 78; characteristics of, 77 Business intelligence: and analytics, 99 caDSR. See Cancer Data Standards Registry and Repository CAHIIM. See Commission on Accreditation for Health Informatics and Information Management CAHME. See Commission on Accreditation for Health Management Education Cambridge Analytica: crisis at Facebook, 357 Cambridge Health Alliance, 171

Canada: centralized information technology architecture in, 231 Cancer Data Standards Registry and Repository, 260 Cancer patients: pilot health social networks for, 173 Cancer Treatment Centers of America, 194 Capitation: medical home financing and, 322 Capitation-based reimbursement, 318–19 Capped fee schedule, 320 Care transitions: admissions, 152, 153; discharges, 153–54; documentation and, 152; transfers, 152, 153 Carle Health System: in partnership with University of Illinois at Urbana-Champaign, 67 Carle-Illinois College of Medicine: disruptive innovation caused by, 52; transforming physician education at (case study) 67–69 Case-based reimbursement, 320 Case studies: analytics for disease management and wellness, 115–17; diabetes management care group, 336–38; display codes problem, 265–66; effective clinical decision support system implementation, 137–40; electronic health record system adoption, 363–66; electronic health records: where does the system end?, 15–17; envisioning a global community, 246–47; genetic and genomic patient information, 200–203; guiding a merger, 287– 88; human resources management and health information technology functions, 306–8; knowledge management in accountable care organizations, 41–44; not all innovation created equal in transition to value-based care, 91–93; Pemiscot County, Missouri population health,

Index

222–23; personal health records integrated into telehealth infrastructure, 182–85; question of evidence, 162–63; redesigning futures: firstever engineering-driven college of medicine, 67–69 CBHD. See Citizens Basic Health Data CDC. See Centers for Disease Control and Prevention CDEs. See Common data elements CDSSs. See Clinical decision support systems Celebrity healthcare breaches, 353 Center for Scientific Integrity, 361 Center of excellence: definition of, 283 Centers for Disease Control and Prevention, 157, 211 Centers for Medicare & Medicaid Services, 153, 270, 363; bundled payments from, 319; meaningful use program, 123, 137 Central control, 3 Centralized repository design, 231 CEOs. See Chief executive officers Cerner Corporation, 102, 134, 304 Champions: clinical decision support system implementation and, 130 Change from within: shortfalls with, 91–92 Charles Schwab, 92 CHEA. See Council for Higher Education Accreditation Chernichovsky, D., 218 Chief clinical (or medical) information officer, 90 Chief data officer, 283 Chief executive officers: business education and orientation of, 77; enlightened view and support of human resources, 296, 297, 298 Chief human resources officers: vision and competence of, 296, 297, 298 Chief information officers: health information technology capabilities

and vision and competence of, 301, 302, 303 Chief operating officers: business education and orientation of, 77 Children’s Medical Center University of Wisconsin, 32 Christensen, C. M., 14 Chronic disease management: e-health applications, 169–70; healthcare expenditures and, 135 Churchill, Winston, 14 CIOs. See Chief information officers CIS. See Clinical information system Citizens Basic Health Data: in Greece, 231 Cityblock, 89 Civan, A., 173 Clancy, C. M., 239 Classen, D. C., 156 Classification algorithms, 105–6 Cleveland Clinic, 240 Climate change: health and, 244 Clinical buy-in: clinical decision support system implementation and, 130 Clinical care: informatics, nursing, and transformation of, 149–50 Clinical data: electronic medical records and types of, 26–27 Clinical decision making: bioinformatics and, 5; as complex, dynamic process, 28; evolution of, 78; interactive components in, 55. See also Clinical decision support systems Clinical decision making, science of, 54–65; context of clinical decisions, 59–61; types of clinical decisions, 56–59; types of decision context, 61–65 Clinical decisions: complexity of processes in, 53; context of, 59–61; decision science applied to, 55 Clinical decisions, types of, 56–59; affective decisions, 59; computational decisions, 56–57; evidencebased decisions, 56; intuitive

385

386

I n d ex

decisions, 59; judgment-based decisions, 57–58 Clinical decision support systems, 4, 6, 24, 81, 114, 121–40; acceptance of, 52; advances and expanded uses for, 136–37; alert fatigue and, 133; building, 29; challenges and barriers related to, 131–33; clinical domain examples, 134–36; definition of, 23, 121, 122; design and implementation of, 129–31; design and use of, 55; domain-specific, 61; effective, critical analysis of, 54; effective implementation of (case study), 137–40, 138; electronic health record–based, 98, 100; evaluation of, 131; evolution of, 122, 124, 124; Five Rights and, 129; free-text technology and, 133; global limits to application of, 227–28; global sources of evidence for, 237–39; guides to more effective, 128–29; history and national policies, 122–24; human factors and, 132; interoperability and, 132–33; introducing, challenges in, 27–28; judgment-based decisions and, 57, 58; knowledgebased, 126, 237; knowledge brokering and, 38, 39; meaningful use stage 3 requirements, 123; nonknowledge-based, 126–27; quality and safety of care and, 155–56; standardized terminologies, 123; tailoring to type of decision context, 61; Ten Commandments and, 128–29; types of, 124–27, 125; urgent care and role of, 65 Clinical decision support tools: electronic medication administration record integrated with, 154 Clinical domains, 61 Clinical evidence: generation of, 28 Clinical function, transformation of, 50–54; clinical decision process and decision support, 52–54;

effective clinical decision support systems, 54; health professionals and, 51–52 Clinical guidelines, 81, 83, 84; application of, 29; definition of, 29; global dissemination of, 238–39; modifications to, 58; outcomes improvement and, 52; reduction of clinical errors and, 24 Clinical informatics, 5, 6, 10, 10 Clinical information: standardizing, early initiatives, 8 Clinical information system: state immunization registries integrated into, 155 Clinical information systems: traditional corporate structures as logic for, 75–78 Clinical knowledge management, 25–33; electronic health records, 27–30; electronic medical records, 25–27; health information exchange, 30–33 Clinical nurse specialists, 161 Clinical outcomes: human resources management capabilities and, 295; standardized, 80–82 Clinical protocols: definition of, 29 Clinical trials: accessing and integrating knowledge from, 28–29; assessing objectivity of, 239; global, scope of, 238–39; randomized, 28–29, 30 ClinicalTrials.gov, 28 Clinical workflows: integrating clinical decision support systems into, 130–31; nursing work, informatics, and, 151–52 Clinical work processes: evolution of, 78; structuring, 82–83; unstandardized, 79–80 Clinics: design/interorganizational relationships among hospitals, nursing homes, and, 75–78 Closed-system concept: primary care medical homes and, 85

Index

Closed systems: evolution of, 31 Cloud computing, 31, 357 Clustering algorithms, 105, 106 CMS. See Centers for Medicare & Medicaid Services Code of Fair Information Practices, 347 Coding variation: quality measurement and, 263–64 Cohorts: population health management and, 212, 213 Collaborative curriculum, 62 Collaborative team environment, 62 Collins, F. S., 199 Commercial systems: health systems vs., 74 Commission on Accreditation for Health Informatics and Information Management, 372 Commission on Accreditation for Health Management Education, 372 Committee on Accounting Procedures, 12 Common data elements: definition of, 260; repositories and, 260; sample, 261 Common Rule, 346 Commonwealth Fund, 236 Communities of practice, 88–89, 107; definition of, 88; emerging models, 89; knowledge management in, 40–41; training for health professionals in, 90; value-based reimbursement and, 88–89 “Communities of Practice: The Organizational Frontier” (Wenger & Snyder), 88 Community or public information exchange, 330 Competition: human resources management and, 295 Complex adaptive systems: characteristics of, 2; definition of, 2; expansion of biomedical informatics and, 10; in healthcare, 2–4

Complexity leadership, 108 Complexity science, 53 Complexity theory: system dynamics modeling and, 112 Components: in population health system, 218 Computational decisions, 56–57 Computational linguistics, 5 Computerized physician-order entry, 81, 121–22; Health Information Technology for Economic and Clinical Health Act and, 299; reduction in medication errors and, 154; at University of Illinois Hospital and Health Sciences System, 137 Concept-based controlled biomedical vocabulary. See Controlled terminology(ies) Concepts: unique, in controlled terminology, 254 Confidentiality: e-health applications, 175, 181; healthcare security and, 345. See also Privacy; Security Connections: in population health system, 218 Consumer engagement, 157–60; patient education, patient portals, and self-management, 158–59; telenursing and, 159; virtual monitoring and, 159–60 Consumer health informatics, 180; applications, 169; definition of, 168–69; social media and, 171–74 Context: in population health system, 218 Controlled substances: state laws and monitoring of, 157 Controlled terminologies: addressing need for, 264; alphanumeric identifying codes, 253, 254–55; application rules, 256; case study, 265–66; components of, 253–54; conflicting interests, 252; definition of, 252; metadata and, 258–59; postcoordinated concept expressions,

387

388

I n d ex

255–56; precoordinated concept expressions, 254; preferred concept expressions, 254; semantic relationships, 254; terminology problem and, 261; uses for, 257–58 Coolidge, Calvin, 52 COOs. See Chief operating officers Corcoran, S., 148 Corcoran-Perry, S., 151 Core data, 275 Corporate presence: growth in, US health system and, 73–74 Corporate structures: traditional, as the logic for clinical information systems, 75–78 Cost-based reimbursement, 317–18 Council for Higher Education Accreditation, 372–73 Coye, M. J., 335 CPOE. See Computerized physicianorder entry CPT. See Current Procedural Terminology Credit/debit card fraud, 343 Crimean War, 148, 149 Crisis of reproducibility, 198–99 Cronin, K., 239 Cross-domain integrative knowledge, 263 Crossing the Quality Chasm (Institute of Medicine), 81 CTCA. See Cancer Treatment Centers of America Current Procedural Terminology, 162, 163 Current state: understanding, 273–74, 274 Cybersecurity threats: medication systems and, 155 Darling v. Charleston Community Memorial Hospital, 76 Darwin, Charles, 127 Data: accessibility of, 283; benchmark, 275; categorized and inventoried, 277–78; as competitive advantage,

286; core, 275; current, 277, 280; design, 278; encrypted, 354–55; external, 276, 281; fluidity, 277, 280; high-value, 281–82; interoperable, 278; late binding of, 281, 282–83; master, 274–75, 276, 279, 284; metadata vs., 260; mobility, 277, 280; movement of, 280–81; operational, 275, 276; quantifying value of, 271; reference, 275, 276; registries, 8; representation of, 281–83; security, 278; standardization, 252, 276, 279; stored, 101; third-party, 275, 277; truth and beauty in, 285; valuable, determining, 259. See also Data and information protection, technical issues in; Data mining Data analysis: data-mining methods vs., 104 Data and information protection, technical issues in, 344–47; privacy, 344–45; risk, 345–46, 361; security, 345; trust, 346–47, 361 Data and information security: individual risks related to, 342–43; organizational risks related to, 343 Data architecture: definition of, 278 Data as assets, 270–72, 277; determining the data facets, 272; quantifying value of data, 271; understanding informatics and information, 272 Databases: NoSQL, 101, 103; objectoriented, 101, 103; prerelational, 101–2; relational, 101, 102; resource descriptive format, 101, 103; types of, and their impact on data mining, 101–4 Database size: defining, 99 Data collection: by nurses, 150 Data-driven algorithms, 105 Data-driven trials (e-trials), 30, 33 Data facets: determining, 272 Data governance, 286; formalizing in healthcare organization, 283–84,

Index

287; level of ownership and, 277; roles and responsibilities, 279 Data lake, 281 Data map: sample, 276 Data mining, 98, 285; analytics and, 98–101, 100; database types and impact on, 101–4; definition of, 99; of electronic health records, 98, 99–101; Internet and impact on, 104–5; potential value and technical challenges of, 101; privacy threats and, 356; value-added potential of, exploring, 100–101 Data-mining methods, 104–6; association-rule mining, 106; classification algorithms, 105–6; clustering, 106 Data principles: common, 277–78; establishing, 276–77, 287 Data quality program, 284 Data repositories: centralized vs. distributed, 231 Data value protection, 284–85 Data vaults, 31 Data visualization: Nightingale as pioneer in, 148 Data warehouse: definition of, 102 DDBJ. See DNA Data Bank of Japan Deaths, resulting from medication and medical errors, 80, 154 Decision context, types of, 61–65; patient involvement context, 62–64; team decision context, 61–62; urgent decision context, 64–65 Decision making: knowledge management in, 22–24; patient-centered transformational change and, 35; patient involvement in, 60. See also Clinical decision making Decision science: applications, 55; definition of, 55 Decision types: logic and basis for, 55 Decomposed systems, 110 Decryption, 354 Defense Intelligence Agency, 196 “Deification” of the customer, 34

DE modeling. See Discrete event modeling Denmark: centralized information technology architecture in, 231; electronic medical records in, 228; health system design in, 235 Department administrators: human resources management tasks and, 296 Department of Veterans Affairs: primary care practitioners and alert fatigue survey, 133 Description logic rules, 106 Destination healthcare, 239–40 Diabetes: incidence of, 336; in United States, 208 Diabetes management care group (case study), 336–38 Diagnosis-related groups: diagnosisrelated group system vs. per-diem rate payment system, 320; Medicare/Medicaid reimbursement and, 319 Digital signatures, 355 Direct exchange, 329 Direct-to-consumer genetic testing kits, 271 Discharges, 153–54; medication reconciliation and, 155 Discrete event modeling: applications with, 110–11; approach in, 113, 113; definition of, 110; skills required for, 111 Disease management and wellness analytics: case study, 115–17 Disease models: dynamic modeling and, 244 Disintermediation, 174 Display codes problem: case study, 265–66 Disruptive innovation, 14; accountable care organizations and, 87; in physician education, 52 Distributed repository design, 231 DL rules. See Description logic rules DNA Data Bank of Japan, 195

389

390

I n d ex

DNA-sequencing experiments, 198 Documentation: clinical informatics and, 152 DOE. See US Department of Education Doherty, I., 172 DRGs. See Diagnosis-related groups Drone deliveries, 215 Drug safety alerts, 133 Drug spending: estimates for, 135 Dublin Core Metadata Initiative, 260 Dubuque and Pacific Railroad, 13 Dunn, J. R., 213 Dutch Nationwide Electronic Health Record, 231 Dynamic systems: transformational nature of, 2 Dynamic systems modeling, 106–13, 114, 244, 245, 317 ED. See Emergency department Edit checks: data quality and, 284 e-education, 242 e-health, 180; accessibility and, 176– 77; blended architecture and, 231; case study, 182–85; challenges with, 175–77; cost of, 179; definition of, 167; home-based applications, 169–70; networks, 241–42; outcomes, 177–78; patient acceptance, 179–80; personal health records, 170–71; platforms, 168; privacy/confidentiality and, 175– 76; processes, 178–79; provider acceptance, 180; purpose of, 167– 68; success factors with, 177–80 EHRs. See Electronic health records Eisenhower, Dwight, 247 e-learning: global corporate systems and, 242–43 Electronic health records, 26, 27–30, 34, 82, 87, 98, 99, 114, 270, 318, 342; accountable care organizations and, 324; adverse events detection and, 156; AgileMD and integration of, 134–35; basis for,

7; case study, 162–63; clinical decision support systems and, 121; clinical guidelines, 7; data mining of, 98, 99–101; definition of, 27; discharges and, 154; evolutionary development of, 33; Health Insurance Portability and Accountability Act technical security safeguards, 350–51; identifying patient patterns and, 99, 100; Infobutton and, 136; integrated, accountable care organizations and, 86; limits on data mining and, 101; linking and integrating across institutions, 100; meaningful use of, 299; medication reconciliation and, 155; networked, 31, 31–33; nursing care plans and, 152; nursing informatics and access to, 161; patient access to, 63; patient education materials and, 158; personal health records vs., 234; prerelational format in, 101; relational databases and use in, 102; security and, 345; simulation and, 160; transfers of patients and, 153; transformation of electronic medical record and, 27; transition-in-care workflows supported by, 153–54; value-based reimbursement and, 319, 320; vendors, 230. See also Electronic medical records; Personal health records Electronic health record system adoption: case study, 363–66 Electronic health records case study, 15–17; community-connected care, 16–17; noninstitutionalized care, 17; personalized care, 16; Vision of Care approach, 15–16, 17 Electronic health record technology, security implementation in, 348, 352–55; access control, 352; authentication, 348, 352; encryption, 354–55; integrity assurance,

Index

354; security auditing and accounting for disclosures, 353–54 Electronic medical records, 25–27, 26, 34, 81, 99; basis of, 7; definition of, 25; digitization and use of, in clinical practice, 26; display codes problem (case study), 265–66; early, 25; evolutionary development of, 33; expanded use of, 29; as foundation of health information technology systems, 228–29; guiding a merger (case study), 287–88; health information exchange and, 30; Health Information Technology for Economic and Clinical Health Act and, 299; implementing, 76–77; as knowledge sources, 27; maturing, 286; standardized clinical vocabulary and, 80; universal access to, 230–31. See also Electronic health records; Personal health records Electronic medical record system: selecting and investing in, 27 Electronic medication administration record, 154 Ellison, N. B., 172 eMAR. See Electronic medication administration record EMBL-EBI. See European Molecular Biology Laboratory’s European Bioinformatics Institute Emergency department: predicting volumes in, 273–74 EMRs. See Electronic medical records Encryption, 354–55; asymmetric, 354, 355; Health Insurance Portability and Accountability Act Security Rule and, 352; symmetric, 354, 355 Enrollment: population health management and, 212 Enterprise: in knowledge management, 107 Enterprise architecture/business architecture, 278 Enterprise data warehouses, 281, 282

Enterprise exchange, 328–29 Enterprise-market strategy: information technology as, 328, 330 Enterprise strategy: definition of, 332 Epic Systems Corporation, 101, 134, 171, 304 Equifax breach (2017), 343 Equifinality principle, 3, 9, 229, 233 Errors: handoffs and, 80; medical, 80, 123; medication, 145 e-trials, 30, 33 European Commission: State of Health in the European Union report, 235 European Molecular Biology Laboratory’s European Bioinformatics Institute, 195 Evidence-based decisions, 56 Evidence-based healthcare: authentication threats to, 360–61, 362 Evidence-based patient care, 360–61 Excel, 282 Exomes: sequencing of, in clinical practice, 192–93 Experiential knowledge: clinical decision making and, 22; evidencebased decisions and, 56 Explainable artificial intelligence, 30 Explicit knowledge: in clinical decision making, 23; definition of, 23 External data, 276, 281 Eysenbach, G., 174 Facebook, 174, 175, 176; Cambridge Analytica crisis at, 357 Facebook Connect, 173 Fair information practices, 361–62; definition of, 347; principle, definition, legal codification, 349–50; refinements and customizations to, 347–48; trust and, 347 Faked peer review, 361 Farr, William, 148 Fast Healthcare Interoperability Resources: Health Level Seven International specification, 134, 136, 282

391

392

I n d ex

FDA. See US Food and Drug Administration Feature maps: self-organizing, 106 Federal Chief Information Officer Council, 334 Federal government: information technology investment, 331; policy interests of, 330 Federal Health IT Strategic Plan, 270 Federated identity, 356 Feedback loops: systems thinking, population health, and, 218, 221 Fee-for-service payment system, 320; accountable care organizations and, 324; medical home financing and, 322 Ferris, T. G., 88, 91 Feste, C., 168 FHIR. See Fast Healthcare Interoperability Resources Financing: of accountable care organizations, 86; of medical homes, 86 Financing, redesigning structure and, 321–26; accountable care organization financing, 324–25; medical home financing, 322–24; strategic implications of structural changes, 325 Financing models, 316–20; Beveridge model, 316; Bismarck model, 316; capitation-based reimbursement, 318–19; cost-based reimbursement, 317–18; national health insurance model, 316; out-ofpocket model, 316; value-based reimbursement, 319–20 Fines: for Health Insurance Portability and Accountability Act privacy violations, 271 FIPs. See Fair information practices First-order logic rules, 106 Five Rights: for designing successful clinical decision support systems, 129 Fleischmann, Martin, 198 Flexner Report, 27 Foege, W. H., 213

Folksonomy, 174 FOL rules. See First-order logic rules Ford Motor Company, 284 Foresite Healthcare, 34 For the Public’s Health (Institute of Medicine), 220 Foundations of Informatics (Mikhailov), 4 Fowler, J. W., 110 Fragmented payment methods, 320 France: electronic database system in, 232 Free-text technology: clinical decision support systems and, 133 Freidman, T. L., 243 FR systems. See Fuzzy rule systems Fuzzy rule systems, 106 GA4GH. See Global Alliance for Genomics and Health Ganguli, I., 88, 91 Garfield Innovation Center at Kaiser Permanente, 36 GAs. See Genetic algorithms GE, 304 GenBank: definition of, 195 Gene Expression Omnibus, 198–99 Genetic algorithms, 127 Genetic and genomic patient information (case study), 200–203 Genetic Information Nondiscrimination Act, 346 Genome, 192 Genomic-based precision medicine, 194 Genomic data sets: growth in, 195, 198 Genomics: definition of, 195 Genomic science: precision medicine and, 192–93 GEO. See Gene Expression Omnibus George, Lloyd, 219 Germany: blended information technology architecture in, 232 Global Alliance for Genomics and Health, 193

Index

Global community: envisioning (case study), 246–47 Global corporate systems, 240–43; collaborative networks, 240–41; e-health networks, 241–42; e-learning, 242–43; joint ventures, 242 Global Digital Health Index, 244 Global health systems development, collaborative systems, 237–43; destination healthcare, 239–40; global corporate systems, 240–43; global sources of evidence for clinical decision support, 237–39 Global health systems informatics, 227–47; case study, 246–47; development of global health systems, 237–43; global health policy and population health, 243–45 Global institutional networks: definition of, 240 Goad, Walter, 195 Governing boards: principal responsibility of, 77 GPC. See Greater Plains Collaborative Graves, J. R., 148 Greater Plains Collaborative: medical centers participating in, 32 Greece: semidistributed information technology architecture in, 231 Grimmelmann, J., 175 Groupe d’Etude sur l’Information Scientifique, 261 Gupta, V., 296, 299, 303, 304 Hacking, 343 Haenlein, M., 172 Handoffs: errors and, 80 Hardcastle, L. E., 215, 218 Hash function, 354 Hayes, M. V., 213 Health: behavioral determinants of, 214, 215, 221; climate change and, 244; intrinsic value of, 326– 27; lifestyle determinants of, 214, 215, 221; social determinants of, 213–14, 215, 221

Health administration: affinity between health informatics and, 373 Healthcare: care provision and financing models for, 315–16; complex adaptive systems in, 2–4; as dataintensive industry, 270, 286; decision science in, 55; integrating public health and, through systems design, 220–21; integration of public health and, push for, 214– 15; lack of information and information sharing in, 320, 321, 335; optimal system performance for, achieving, 90; reported breaches in, 343 Healthcare Cost and Utilization Project, 104 Healthcare delivery services: at consumer level, 74; at policy level, 74–75; transformational strategy for, 36–37; valuation of, 326–28 Healthcare delivery system: insurancecentric and fragmented, 320, 321 Healthcare expenditures: total, and as percentage of gross domestic product, 135 Healthcare Information and Management Systems Society: board of directors position statement on nurses, 149; definition of interoperability, 132; electronic medical record adoption model, 15; mission and history, 371; Nicholas E. Davies Award of Excellence, 137 Healthcare lawyers, 168 Healthcare organizations: management information science in, 12; meaningful use and, 75; Mintzberg’s model and, 79–80; nimble structure of, need for, 74; security breaches in, 343 Health data vaults: networked, 114 Health in All Policies approach, 215 Health informatics, 98; affinity between health administration and,

393

394

I n d ex

373; representational science and, 252–53. See also Health systems informatics Health information: accessing, 174 Health information exchanges, 30–33, 34, 99; accountable care organizations and, 86–87; data mining through, 101; definition of, 30; evolutionary development of, 33; Institute of Medicine’s view of, 81; networked, 114; networked electronic health records and, 31, 31–33 Health information exchanges, types of: direct exchange, 329; enterprise exchange, 328–29; public information exchange, 330 Health information technology: adoption by health systems, 123; benefits with, 121; fundamental role of, in healthcare, 291; investment choices, 270; knowledge intensiveness and, 293; proactive work behaviors and, 294; University Hospital case study, 307–8 Health information technology capabilities, 299–304; inadequacy of, as opportunity, 305; overview, 299– 300; relationship between quality of patient care and, 303, 303–4; three dimensions of, 301–3, 302; transformational change and, 300–301 Health Information Technology for Economic and Clinical Health Act, 330, 331, 333; meaningful use provision in, 75, 123, 270, 299, 300; private sector investment in information technology and, 327 Health information technology infrastructure: health information technology capabilities and, 301, 302 Health information technology staff: professionalism of, 301–2, 302 Health information technology systems: electronic medical records

foundational to, 228–29; global development of, 227; outsourced vs. homegrown, 304 Health insurance: financing mechanisms, 315–16; objectives of, 219; public vs. private, 334 Health Insurance Portability and Accountability Act: challenges, 175; fines for privacy violations, 271; Privacy Rule, 278, 348, 349, 350, 355, 356; Security Rule, 345, 346, 348, 349, 350, 352, 355, 362; technical security safeguards, 350–51 Health InterNetwork Access to Research Initiative: World Health Organization and launch of, 238 Health Level Seven International, 123, 134, 136, 282 Health literacy: definition of, 158 Health maintenance organizations, 318 Health outcome indicators: United States as outlier in, 208–9, 209 Health professionals: changing role of, 49–50; clinical judgment and, 58; context of clinical decisions and, 60; culture of, 51; integrated teams of, 84; information technology provisions and, 49; traditional role of, 51, 58, 81 Health professions: protected role of, in society, 4 Health Professions Accreditors Collaborative, 373 Health sciences education: authentication issues related to, 359–60 Health sciences research: authentication issues related to, 360–61 Health services delivery process: proactive work behaviors in, 293–94 Health services delivery process, major features of, 292–93; knowledge intensiveness, 293; service orientation, 292–93 Health social networks/networking: case study, 182–85; definition of,

Index

173; goals of, 173; privacy issues related to, 176; value of, 174 Health system design: determinants of, 3; health systems informatics and, 235–36 Health system ownership: increased diversification of, 234, 235 Health systems: commercial systems vs., 74; growth in corporate presence and, 73–74 Health systems informatics, 1–2, 66, 114; assumptions underlying, 4; case study, 15–17; definition of, 1; extended model, 9–10; focus of, 9, 11–12; health system design and, 235–36; knowledge brokering and, 40; management information science vs., 12; methods, techniques, and theories, 10, 10; restructuring health systems according to logic of, 232–36; systems theory perspective and, 2; transformational change and, 8–12; transformation of clinical function and, 78; transformative power of, 9; transformed clinical decision process, system design, and, 53; values and tradition related to, 13 Health systems informatics, comparative analysis of, 228–32; design of information technology architecture, 229–30; electronic medical records, 228–29; types of information technology architecture, 230–32 Health 2.0: definition of, 172 Heuristic reasoning, 57 HHS. See US Department of Health and Human Services HiAP approach. See Health in All Policies approach Hierarchical clustering, 106 Hierarchical functional structures: ineffectiveness of, 77–78 Hierarchical organizations: lack of fit with information age, 83

HIEs. See Health information exchanges Higher Learning Commission, 360 HIMSS. See Healthcare Information and Management Systems Society HIPAA. See Health Insurance Portability and Accountability Act Hippocratic oath, 51, 201, 202 HIT. See Health information technology HITECH Act. See Health Information Technology for Economic and Clinical Health Act HL7. See Health Level Seven International HMO Act of 1973, 318 HMOs. See Health maintenance organizations Home health care: blended information technology architecture and, 231, 232; extended healthcare models for, 273; telenursing and, 159 Hospital merger: guiding (case study), 287–88 Hospitals: deaths from medical errors in, 80; design/interorganizational relationships among nursing homes, clinics, and, 75–78; vertical structure of, 77 HPAC. See Health Professions Accreditors Collaborative HRM. See Human resource management Hubble Space Telescope, 195 Human factors: success of clinical decision support systems and, 132 Human Genome Project, 195, 198 Human resource management: fundamental role of, in healthcare, 291; proactive work behaviors and, 294; University Hospital case study, 306–7 Human resource management capabilities, 295–99; building, research developments and, 296;

395

396

I n d ex

importance of, in healthcare, 295; inadequacy of, as opportunity, 305; relationship between patientcentered care and, 299, 299; three dimensions of, 296, 297–98, 298–99 Human resource management staff: professionalism of, 296, 297–98, 298–99 Human Variome Project (HVP), 193, 194 Hwang, J., 14 i2b2. See Informatics for Integrating Biology and the Bedside IBM, 92 ICD-10-CM. See International Classification of Diseases, 10th Revision, Clinical Modification Idealized design, 218 IDEATEL study, 169 Identity theft, 343 Illness and treatment model: moving to wellness model, 84 Imaging informatics, 5, 6, 6, 10, 10 Immunization registries: statesponsored, 156 Immunization status: public health surveillance of, 155 Immunization surveillance: prescription drug monitoring programs and, 156–57 Implicit knowledge: in clinical decision making, 23; definition of, 24 Incentives: for home-based care, 159; pay-for-performance, 322 Indiana University Medical Center, 32 Individual autonomy: professional autonomy vs., 50 Industrial Revolution, 82 Infectious diseases: global impact of, 243, 244 Inflows: system, 217–18 Infobutton, 136 Informatics: applications, 5; definition of, 5; origin of term, 4–5;

understanding meaning of, 272. See also Consumer health informatics; Global health systems informatics; Health informatics; Health systems informatics; Medical informatics; Nursing informatics; Population health informatics; Public health informatics Informatics for Integrating Biology and the Bedside, 32, 282 Information: access to, quality and, 330–31; application of, 23; digitizing, 7; nurses as creators of, 150; understanding meaning of, 272. See also Big Data; Data and information protection, technical issues in Information facets: determining, 274, 274–76 Information management strategy: as component of information technology strategy, 270; data accessibility and, 283; data as assets in, 270–72; data as competitive advantage in, 286; data governance program and, 283–84; data movement and, 280–81; data quality program and, 284; data representation and, 281–83; data value protection and, 284–85; definition of, 269; perils and pitfalls, 285–86 Information management strategy development, steps in, 272–80; create reference data architecture, 274, 278–80, 287; determine data principles, 274, 276–78, 287; determine information facets, 274, 274–76, 287; understand current state, 273–74, 274 Information model, 260 Information system: fragmented, 74 Information technology: advanced, transformation of clinical work processes and, 84; commercial and business applications, 3; as core function in knowledge

Index

management, 21–22; destination healthcare and, 240; evolution of nursing informatics and, 148; global challenges with, 247; initial applications of, in health systems, 13; integrating across organizations and professions, 8; internal capabilities, 301; Institute of Medicine’s recommendations, 81, 82; knowledge organizations and power of, 25; medical home model and power of, 323; nurses as users of, 150; paradox, 303; precision medicine and, 196; private sector investment in, 331–33; public sector investment in, 333–35; rapid development of, to support clinical decision making, 97; as transformational technology, 327; transformation linked to, 8; transformative power of, in healthcare, 75; understanding transformative power of, 83 Information technology architecture: definition of, 230; design of, 229– 30; obstacles in design of, 230; security of accessible data and, 230; types of, 230–32 Information technology infrastructure: designing, 328, 329; three-stage trajectory in development of, 328; valuation of, 328–35; private sector information technology investment, 331–33; public sector information technology investment, 333–35 Information technology strategy: information management strategy as component of, 270 Infrastructure: in population health system, 218 Innovation: information technologyenabled, 236–37; lack of equality in transition to value-based care (case study), 91–93 Innovators Network, 236

Innovator’s Solution (Christensen & Raynor), 91, 92 Instagram, 357 Instances, systems of record, 275 Institute for Healthcare Improvement, 236; Triple Aim, 210 Institute for Technology Assessment, 54 Institute of Medicine, 60, 82; Crossing the Quality Chasm, 81; To Err Is Human, 80, 123, 154; For the Public’s Health, 220; U.S. Health in International Perspective: Shorter Lives, Poorer Health, 208 Integrated health systems: advent of, 74 Integrated information system, 75 Integrated systems, perspectives on, 83–90; accountable care organizations, 86–88; communities of practice and future designs, 88–89; medical homes, 84–86; overview, 83–84; summary of integrated systems, 89–90 Integrity assurance (or corroboration), 354 Interactive communication software, 38 Intermediate lexicon, 261, 262 Intermediation, 174 International Classification of Diseases, 162, 163; 9th and 10th revisions of, 123; development of, 8 International Classification of Diseases, 10th Revision, Clinical Modification, 256, 258, 262, 274 International Health Terminology Standards Development Organisation, 255 International Nucleotide Sequence Database Collaboration, 195 International Statistical Congress, 8 Internet: impact on data mining, 104–5; patient involvement context and, 62 Internet of Things, 33; definition of, 170, 358; precision medicine and, 197; security breaches and, 358

397

398

I n d ex

Interoperability: clinical decision support systems and, 132–33; data, 278; Healthcare Information and Management Systems Society’s definition of, 132; need for, 9; terminology problem and, 261–62 Interoperable global systems: e-health networks and, 241–42 Interprofessional education: opportunities for, 9, 60 Intrinsic value: of health, 326–27 Intuitive decisions, 59 IOM. See Institute of Medicine IoT. See Internet of Things IPE. See Interprofessional education IT. See Information technology iThink software, 112 Jacob, Howard, 192 Johns Hopkins Aramco Healthcare Company (Dhahran, Saudi Arabia), 242 Joint Commission, 153 Joint ventures: global corporate systems and, 242 J.P. Morgan, 335 Judgment-based decisions, 57–58 Kaiser Permanente, 171, 318 Kaplan, A. M., 172 Khatri, N., 295, 296, 298, 299, 303, 304 Kindig, D., 211 KLAS Research, 134 k-means, 106 Knee magnetic resonance imaging interpretation: machine learning and, 30 Knowledge: application of, 23; assets, 22; decision-making use of, 56; experiential, 22, 56; explicit, 23; implicit, 23, 24; latent, 23, 24; nurses as creators of, 150–51; tacit, 23, 24; types of, 23–24 Knowledge-based clinical decision support systems, 126, 237

Knowledge-based decision making, 21–44; clinical knowledge management, 25–33; definition and use of knowledge in decision making, 22–24; knowledge brokering, 38–40; knowledge organizations, 24–25; knowledge socialization, 37–38; patient-centered care, 33–36; transformational strategy, 36–37 Knowledge-based health systems design, 236–37; global diversity of, 236; information technologyenabled innovations, 236–37 Knowledge-based information technology systems: types of, 99 Knowledge brokering, 22, 38–40; basis of, 38–39; for clinical care, 39; definition of, 38; integrated decision support systems and, 40 Knowledge management, 293; applications, 22; clinical, 25–33; in community of practice, 40–41; comprehending basics of, 41; in decision making, 22–24; definition of, 21, 106–7; in dynamic enterprise system, 107; dynamics and complexity of, 22; performing, clinical decision support systems and, 131; scope of, 25; systems perspective, 22–23. See also Predictive analytics in knowledge management Knowledge management in accountable care organizations (case study), 41–44; background, 41–42; evidence-based strategy, 42–43 Knowledge organizations, 53; characteristics of, 24–25; definition of, 24; introduction of, 24 Knowledge socialization, 22, 37–38; definition of, 37; reinforcing, 38 Knowledge systems: clinical decision support in, 26; patient-oriented, 31 Krames Patient Education, 158

Index

Late binding of data, 281, 282–83 Latent knowledge: in clinical decision making, 23; definition of, 24 Leaders: clinical decision support system implementation and, 130 Leadership: complexity, 108 Learning: modeling and, 109 Learning organizations, 53 Learning-oriented human resources management systems: designing, 293 Ledley, R. S., 122 Lee, J., 22 Leggat, S., 295 Leibowitz, A. A., 218 Life expectancy: at birth, and health spending per capita, 2011, 209; in the United States, 208, 213 Lifestyle determinants of health, 214, 215, 221 Logical Observation Identifiers Names and Codes (LOINC), 278 Lorenzi, N. M., 173 Los Alamos National Laboratory, 195 Los Alamos Sequence Database, 195 Los Angeles County Medical Association, 4 Ludwig, W., 231, 232 Lui, K., 138 Lusted, L. B., 122 Machine learning, 126; applications with, 29–30; computational clinical decisions and, 56. See also Artificial intelligence Machine-learning algorithms: black box nature of, 30 Malware, 358 Managed care, 215 Management information science, 12–13; business vs. clinical functions, 12–13; definition of, 12; health systems informatics vs., 12 Management information systems, 78 Manufacturers: service organizations vs., 292, 292

MAR. See Medication administration record Market economy: service economy vs., 327 Marshfield Clinic Research Foundation, 32 Massachusetts General Hospital, 102 Mass production, 82 Master data, 274–75, 276, 279, 284 Mastrian, K., 151 Mathematical algorithms, 101 Mayo Clinic: Center for Innovation, 36; destination healthcare at, 240 McCormick, D., 304 McGinnis, J. M., 213 McGonigle, D., 151 MD Anderson Cancer Center: Project ECHO hosted by, 240 MD Anderson Cancer Network, 241 Meadows, D. H., 218 Meaningful use, 75, 123, 137, 270, 299, 300 Mechanistic models, 37 Medicaid, 316, 327; public sector information technology investment and, 334; reimbursement by, 319 Medical College of Wisconsin, 32 Medical devices: security of, 358–59 Medical drones, 57 Medical errors: deaths related to, 80; health information technology and prevention of, 123 Medical faculty: autonomy of, 52 Medical function: traditional design, 76, 76 Medical homes, 84–86; accreditation of, limitations with, 85; assessments of, 85; definition of, 84; financing of, 86, 322–24; information-centric, coordinated, 325, 326; promotion and sponsoring of, 84–85 Medical identity theft, 343 Medical informatics, 5–7; bioinformatics comparison, 5; definition of, 5; focus of, 5; origin of, 6

399

400

I n d ex

Medical Library Association: mission and history, 371 Medical neighborhood, 324 Medical schools: new, 67 Medical Subject Headings, 28, 256, 257, 263 Medical tourism, 239–40 Medicare, 316, 327; public sector information technology investment and, 334; reimbursement by, 319; Shared Savings Program, 42 Medicare Access and CHIP Reauthorization Act of 2015: passage of, 134 Medication administration record, 154 Medication delivery: nursing care, safety, and, 154–55 Medication errors, 145 Medication reconciliation: safety and, 155 MedlinePlus, 158 MedSocket, 135 MedSocket Rx, 135 Mergers and acquisitions, 273 Merit-based incentive payment system, 319 Merrill Lynch, 92 MeSH. See Medical Subject Headings Meta-analysis, 198, 256 Metadata: controlled terminology and, 258–59; crosswalks, 263; data vs., 260; definition of, 258; organization of, 259 Metadata schema: definition of, 260 Metathesaurus, 262 Metrics: population health management and, 220, 221, 222 Metzger, J. B., 125 Middle States Commission on Higher Education, 360 Middleton, B., 122, 131 Mihalas, Barbara, 195, 196 Mikhailov, A. I., 4 Mintzberg’s model: standardization of clinical work processes and, 79–80 MIS. See Management information science

Misinformation, 174 MLA. See Medical Library Association Mobile apps, 168 Modeling, 98, 106–13; analytics and, 107, 108; approach comparison, 113, 113; data-mining techniques vs., 109; definition of, 107; dynamic, 107; learning and, 109; limitations with, 110; process, 109, 109; types of, 110–13. See also Dynamic systems modeling Mokdad, A. H., 214 Molecular Biology Database Collection, 196 MongoDB, 103 Moore’s Law, 198 Mortality: premature, 213–14, 219, 221 Multiprofessional teams: knowledge socialization and, 38 MUMPS, 101 Munro, Henry S., 219 Mutual adjustment: inherent limitations of, 79 MYCIN, 30, 134 MyHealtheVet, 171 My Health Record (myHR): in Australia, 234 Narayan, S., 151 National Alliance for Health Information Technology, 170 National Center for Biomedical Information, 198 National Center for Quality Assurance, 85 National Center for Supercomputing Applications, 195 National Committee for Quality Assurance, 323 National Committee on Vital and Health Statistics, 134 National Guideline Clearinghouse: database, 238; mission of, 29 National health insurance model of healthcare financing, 316

Index

National health policies: evidencebased, 244; global forces and, 243–44 National Health Service (United Kingdom): telemonitoring study for chronic obstructive pulmonary disease, 169 National health system: of Greece, districts in, 231 National health systems, 228 National Institutes of Health: Biomedical Information Science and Technology Initiative formed by, 5; funds development of Informatics for Integrating Biology and the Bedside, 32 National Interstate and Defense Highways Act, 246–47 National Library of Medicine, 199, 256, 260, 371; funding for 1-Search, 135; MedlinePlus, 158; PubMed portal, 28 National Prevention Council, 210 National Prevention Strategy, 210–11, 213, 219, 221 National Provider Identifier, 274 National Quality Strategy: aims of, 209–10; establishment of, 123 National Security Agency, 196 “Nationwide Privacy and Security Framework for Electronic Exchange of Individually Identifiable Health Information,” 348 Natural language processing: clinical decision support systems and use of, 133, 137; definition of, 103 Natural selection theory, 127 Netherlands: electronic medical records in, 228; patient portals and integrated patient information in, 235 Networked electronic health records, 31–33 Neural networks, 105 New Zealand: primary care practitioners and use of health information technology in, 229

Next-generation sequencing platforms, 198 NGC. See National Guideline Clearinghouse NGS platforms. See Next-generation sequencing platforms Nielsen, P., 229, 230 Nightingale, Florence, 147, 149 Nightingale Pledge, 51 NLM. See National Library of Medicine NLP. See Natural language processing Nolan, T. W., 210, 212 Nonaka, I., 37 Non-knowledge-based clinical decision support systems, 126–27; artificial neural networks, 126–27, 127; genetic algorithms, 126, 127 Nonprofit (or plural) sector, 3 North, M., 202 Norway: electronic medical record in, 228 NoSQL database, 101, 103 NPC. See National Prevention Council Nucleic Acids Research journal, 196 Nurse anesthetists, 161 Nurse informaticists: crucial leadership roles of, 149–50; patient-centered care and central role of, 157–58 Nurse practitioners, 161 Nurses: human resources management tasks and, 296; as largest workforce in healthcare, 161 Nurses, role in informatics, 150–51; as creators of information, 150; as creators of knowledge, 150–51; as innovators, 151; as users of information technology, 150 Nurse scientists, 160 Nursing care planning, 152–54 Nursing diagnoses, 152 Nursing education and research, 160 Nursing homes: design/interorganizational relationships among hospitals, clinics, and, 75–78 Nursing informatics, 5, 147–63; case study, 162–63; consumer

401

402

I n d ex

engagement, 157–60; definition of, 148–49; evolution of, 148; historical roots of, 147–48; modern, 148; nursing education and research, 160; quality and safety of care, 154– 57; recognizing overall impact of, 161; research and practice, 160–61; roles of nurses in, 150–51; transformation of clinical care and, 149–50 Nursing profession: Nightingale Pledge and, 51 Nursing work and information system applications, 151–54; care planning, 152–53; care transitions, 153–54; clinical workflows, 151– 52; documentation, 152 Obesity: diabetes and, 336; in the United States, 208 Object-oriented databases, 101, 103 Office of Civil Rights (US Department of Health and Human Services), 348 Office of the National Coordinator for Health Information Technology, 129, 348 Okie, S., 192 OLAP architecture. See Online analytical processing architecture Olmstead v. United States, 344 1-Click Decision Support (1-CDS), 135 1-Search: National Library of Medicine funding for, 135 Online analytical processing architecture, 103 Online education: authentication issues related to, 359–60 Online health communities, 174 Online support groups, 168 Open Group Architecture Framework, 278 Open system: definition of, 216 Open systems theory: community of practice concept and, 88; development and definition of, 9; health system assumption underlying, 89

Operating budget, salaries/wages as percentage of, 295 Operating model, 273 Operational data, 275, 276 Operational strategy: information technology, 328, 330 Ophthalmology: machine-learning technologies in, 29 Opioid epidemic: combating, 157 Osheroff, J. A., 129 Outcomes: clinical, 80–82, 295; e-health impact, 177–78; improvement, clinical guidelines and, 52; standardized, 80–82 Outflows: system, 217–18 Out-of-pocket model of healthcare financing, 316 Pain medications, 157 Pan-African e-Network, 241 Papier, Art (dermatologist), 135–36 Passive monitoring: definition of, 170; sensors, 168 Pathology: machine-learning technologies in, 29 Patient-centered care, 33–36; admissions and, 153; development of, 33; discharges and, 153–54; nursing informaticists’ central role in, 157–58; patient-centered transformational change, 34–36; personal health record and, 33–34; relationship between human resources management capabilities and, 299, 299; transfers and, 153 Patient-centered information systems, 168 Patient-centered medical homes, 84–86; financing of, 322–24; promotion and sponsorship of, 84–85 Patient-Centered Outcomes Research Institute, 32 Patient-Centered Outcomes Research Network Common Data Model, 282 Patient-centered systems, review of, 169–71; home-based e-health

Index

applications, 169–70; personal health records, 170–71 Patient decision support system, 137 Patient education materials: accessing, 158 Patient empowerment model: e-health and, 168 Patient involvement context, 62–64 Patient-oriented knowledge system, 31 Patient portal: definition of, 158 Patients: role transition for, 168. See also Consumer engagement; Patient-centered care Patrick, T. B., 198 Pattern recognition, 105. See also Data mining Pay-for-performance incentives: medical home financing and, 322 PCEHR. See Personally controlled electronic health record PCORNet Common Data Model. See Patient-Centered Outcomes Research Network Common Data Model PCPs. See Primary care practitioners PDMPs. See Prescription drug monitoring programs PDSS. See Patient decision support system Peer review: faked, 361 Pemiscot County, Missouri: population health case study, 222–23 Per-diem rate payment system: diagnosis-related group system vs., 320 Personal health records, 33–34, 82, 98, 99, 170–71; case study, 182–85; definition of, 170–71; electronic health records vs., 234; electronic medical record integration, 171; initiatives, 171; security and, 345. See also Electronic health records; Electronic medical records Personalized medicine: new era of, 168 Personally controlled electronic health record: in Australia, 234

Pharmaceutical companies, 74 Pharmacogenomic biomarkers, 194 PHI. See Protected health information Philips’s IntelliSpace Portal 9.0, 30 PHRs. See Personal health records Physicians: communities of practice and, 89; drug safety alerts and, 133; employed, 74; enhanced role of, 4; fee-for-service vs. valuebased care models and, 91; global, electronic medical record use by, 228–29; individual autonomy of, 50; role of, 52 Plural (or nonprofit) sector, 3 Pons, Stanley, 198 Popular culture: precision medicine and, 194 Population health, 84; accountable care organizations and, 324–25; case study, 222–23; definition of, 211–12; exploring, from open systems perspective, 207; global health policy and, 243–45; improving with systems thinking, 218–19; integrative systems as integrative technology for, 207–23; public health vs., 211–12; status of, in the United States, 208–9; as a system, 215–19; transformation to health systems informatics and, 8 Population health informatics: applying to community population, 220–21; applying to patient population, 220 Population health management in the United States, 212–15; integration of healthcare and public health, 214–15; premature mortality, lifestyle and behavioral factors, and social determinants of health, 213–14 Population health provisions in Affordable Care Act, 209–11; National Prevention Strategy, 210–11; National Quality Strategy, 209–10

403

404

I n d ex

Postcoordinated concept expressions, 255–56 Pragmatics, 256 Precision medicine, 181, 191–203; Big Data and, 194–97; challenges with, 199–200; definition of, 168, 191; genomic science and, 192–93; initiatives, 193–94; popular culture and, 194; scientific reproducibility and, 197–99; from systems perspective, 197 Precision Medicine Initiative: launch of, 193 Precoordinated concept expressions, 254 Predictive analyses: of global threats, 244 Predictive analytics in knowledge management, 97–117; case study, 115–17; data mining and analytics, 98–101, 100; data-mining methods, 104–6; data types and their impact on data mining, 101–4; dynamic systems modeling, 106–13 Preferred concept expressions, 254 Prerelational databases, 101–2 Prescription drug monitoring programs: immunization surveillance and, 156–57 Prescription drugs: television advertising of, 64 Primary care medical homes: populations served by, 85 Primary care practitioners: alert fatigue among, 133 Privacy: breaches, resources for tracking, 343; data value protection and, 284; e-health applications, 175, 181; Internet of Things and, 197; security relationship, 344, 345; social media and, 175–76. See also Confidentiality; Security Privacy Act of 1974, 347 Privacy Rights Clearinghouse: Chronology of Data Breaches, 343

Private insurance: public insurance vs., 334 Private key encryption, 354 Private sector: healthcare services and, 327, 328 Private sector investment: in global health systems, 234–35; in information technology, 331–33 Proactive behaviors: health information technology capabilities, quality of patient care, and, 303 Proactive work behaviors, 293–94, 299 Probabilistic decisions, 56, 57 Profession: definition of, 51 Professional autonomy: individual autonomy vs., 50; standardization vs., 53, 80 Professionals: definition of, 51; role of, 51. See also Health professionals Professional societies, 369–71 Project ECHO, 240 ProModel software, 111 Prostate cancer detection: machine learning and, 30 Protected health information: fines for improper disclosure of, 271 Providers: e-health acceptance by, 180 Proxy decision makers, 59 Public health: definition of, 211; integrating healthcare and, through systems design, 220–21; population health vs., 211–12 Public health informatics, 5, 6, 7–8, 10, 10 Public insurance: private insurance vs., 334 Public key encryption, 354 Public–private partnerships: integrated health information systems and, 234–35 Public sector: information technology investment by, 333–35; as major purchaser of healthcare services, 327–28 Public/social good: information technology as, 328, 330

Index

PubMed, 257, 263; sample record, 257, 259, 260 Quality: health information technology systems and issues related to, 300 Quality and safety measures: clinical informatics, nursing care, and, 152 Quality and safety of nursing care, 154–57; adverse events and, 156; clinical decision support systems and, 155–56; immunization surveillance and prescription monitoring programs, 156–57; medication delivery and safety, 154–55; medication reconciliation and safety, 155 Quality improvement techniques, 40–41 Quality measurement: coding variation and, 263–64 Quality of patient care: outsourced vs. internal health information technology systems and, 304; relationship between health information technology capabilities and, 303, 303– 4. See also Patient-centered care Quantified self, 271 Radiology: machine-learning technologies in, 29, 30 Randomized clinical trials, 28–29, 30, 169, 177–78 Ransomware, 358 RCTs. See Randomized clinical trials RDF databases. See Resource descriptive format databases Recurrent networks, 127 Reference data, 275, 276 Reference data architecture, 286, 287; creating, 278–80; sample, 280; technology aspects of, 279–80 Regional medical centers: knowledge brokering and, 39 Registries: federally mandated, 8; population health management and, 212

Reimbursement: capitation-based, 318–19; case-based, 320; costbased, 317–18; value-based, 319– 20, 322 Relational databases, 101, 102 Remote patient monitoring, 33–34 Repositories: common data element, 260 Representational science: definition of, 253; health informatics and, 252–53 Reproducibility in science: precision medicine and, 197–99 Resource-based relative value scales, 320 Resource-based strategy, 332 Resource descriptive format databases, 101, 103 Retraction Watch, 361 Retrospective review, 58 Risk assessment, 346 Risks, 361; changing nature of threats, 356; variables tied to, 345 RNA-sequencing experiments, 198 Roadmap Development Steering Committee: establishment of, 129 Robert Wood Johnson Foundation: County Health Rankings & Roadmaps website, 222 Robots, 57 Rogers, Michael, 196, 197, 198 Rosenkötter, N., 235 Ross-Loos Group (Los Angeles), 4, 76 Rule-based systems, 30 Rural communities: telehealth and, 159 Russell, Jonathan, 197 Sæbø, J. I., 229, 230 Safety: accountability and, 80–81. See also Quality and safety of nursing care Salaries: as percentage of operating budget, 295 Saudi Arabia: Johns Hopkins Aramco Healthcare Company in, 242;

405

406

I n d ex

national and provincial health systems in, 235 Scale: in population health system, 218 Scannapieca, M., 284 Schömig, A. K., 110 Science: crisis of reproducibility in, 198–99 Science, 195 Scientific reproducibility: precision medicine and, 197–99 SD modeling. See System dynamics modeling Secure-messaging interface: patient portals and, 159 Security: auditing, 353–54; data, 278; data value protection and, 284; definition of, 345; Internet of Things and, 197; information technology architecture and, 230; logs, 275; medication systems and, 155; patient portals and, 158; privacy relationship, 344, 345. See also Confidentiality; Data and information security; Electronic health record technology, security implementation in; Privacy; Security breaches Security assertion, 356 Security breaches: in healthcare organizations, 343; resources for tracking, 343–44 Self-management theories, 158 Semantic relationships: example of, 255; hierarchical or nonhierarchical, 254 Semantics: data representation and, 281 Semidistributed information technology architecture, 231 Senge, P. M., 24 Sensmeier, J., 149 Sensor technologies: blended information technology architecture and, 231, 232 Sepsis, 156

Service economy: market economy vs., 327 Service organizations: manufacturers vs., 292, 292 Service orientation: proactive work behaviors and, 294 Shared Health Research Information Network, 32, 99 Shortliffe, E. H., 30 SHRINE. See Shared Health Research Information Network Silo architecture, 229–30 Simio software, 111 Simulation: in nursing education, 160 Single sign-ons, 356 Sittig, D. F., 122, 124 Six Sigma, 40 Skeels, M. M., 173 Skills: standardization of, 79 Smart home: definition of, 170; technologies, 170 Smartphone applications (apps): privacy risks with, 357–58 Smart-pump technology, 154 Smart systems, 57 SNOMED CT. See Systematized Nomenclature of Medicine Clinical Terms Snyder, W. M., 88 Snyder-Halpern, R., 151 Social determinants of health, 213–14, 215, 221 Socialization skills: teaching and transferring, 37 Social media, 168; consumer health informatics and, 171–74; explanation of, 171–72; privacy issues and, 175–76, 357 Social networking, 171; definitions and characteristics of, 172–73; health implications of, understanding, 173 Social technology: privacy issues and, 357 Social Transformation of American Medicine (Starr), 73

Index

Software: clinical decision support system, regulation of, 123–24; for discrete event modeling, 111; for system dynamics modeling, 112 Software contracts: data rights, sharing, and, 284–85 Source vocabularies, 262 Spreadsheets, 282 Stakeholders: accountable care organization financing and, 324; clinical decision support system implementation and, 129–30; financing mechanisms and, 321 Standardization: professional autonomy vs., 53, 80; of skills, 79; structured work processes and, 53 Standardization of clinical work processes, 78–83; Mintzberg’s model, 79–80; overview, 78–79; standardized clinical outcomes, 80–82; structuring clinical work processes, 82–83 Standardized terminologies: importance of, in clinical care, 252; standardized clinical outcomes and, 80 Standards: strictly enforced, compromised clinical quality and, 83 Stanton, P., 295, 298 Starr, Paul, 73 State of Health in the European Union report, 235 STELLA software, 112 Stoddart, G., 211 Stored data, 101 Stoto, M. A., 213 Structural changes: strategic implications of, 325–26 Supervision: direct, inherent limitations of, 79 Suppliers: market influence of, 74 Supply robots, 57 Support vector machines, 105 Suprasystems (or super systems), 217 Surrogate representations: constructing, controlled terminology and, 257; controlled terminology components and, 253–56

Surveillance: immunization, 155, 156–57; public health, 8 Survival-of-the-fittest theory, 127 Swan, M., 173, 176 Sweden: electronic medical records in, 228 Switching language: intermediate, terminology problem and use of, 262 Symmetric encryption, 354, 355 Symposium on Computer Applications in Medical Care, 370 Systematic reviews, 29; clinical decision process and, 238; of clinical trials, 239 Systematized Nomenclature of Medicine Clinical Terms, 123, 163, 255, 258, 262, 278 System dynamics modeling, 112–13; applications of, 112–13; approach in, 113, 113; complexity theory and, 112; definition of, 112 Systems: inflows and outflows, 217– 18; properties of, 216; super, 217 Systems of record (or gold source of data), 275, 276, 279; data movement and, 280; reference data architecture and, 280 Systems perspective: focus of, 2–3; transformation to, 36 Systems science: inflows and outflows in, 217 Systems theory, 236; comprehending basics of, 41; healthcare dynamics recognized by, 2–3 Systems thinking: basis of, 84; definition of, 215; using to improve population health, 218–19 Tabak, L. A., 199 Tacit knowledge: in clinical decision making, 23; definition of, 24 Takeuchi, H., 37 Takeyourclass.com, 359 Tax exemption status: private nonprofit sector and, 327 Team competency practicum, 62

407

408

I n d ex

Team decision context, 61–62 Teams and teamwork: for clinical decision support system implementation, 130; in medical homes, 85; multiprofessional, 38; transforming clinical and organizational designs and, 83–84 Team socialization, 37 Tele-ERA study (Minnesota), 169 Telehealth: case study, 182–85; expanded services in, 159 Telemedicine: evaluation methods, 179; optimal use of, 180 Telemedicine Perception Questionnaire: domains of, 180 Telenursing, 158, 159 Teleologic design, 3 Ten Commandments for Effective Clinical Decision Support, 128–29 Terminology mapping, 262–63 Terminology problem, 264; interoperability and, 261–62; summary of, 261 Third-party data, 275, 277 Three-party system, 316 TMPQ. See Telemedicine Perception Questionnaire To Err Is Human (Institute of Medicine), 80, 123, 154 TOGAF. See Open Group Architecture Framework Training: clinical decision support system implementation and, 131 Transactional change: transformational change vs., 11 Transfers: medication reconciliation and, 155; workflows involved in, 153 Transformational change: in banking, 10–11, 300; health systems informatics and, 8–12; health information technology capabilities and, 300–301; leading in clinical decision making, 3–4; patient-centered, 3436; transactional change vs., 11 Transformational strategy, 36–37

Transformation of clinical care: consumer engagement and, 157–58; informatics, nursing, and, 149–50 Transforming processes: automating processes vs., 10 Transitions in care. See Care transitions Transparency: in scientific publications, 361; trust and, 346, 347 Triple Aim, 210, 212 Trust, 361; balance among components of, 346, 347; definition of, 346; fair information practices and, 347 Tuskegee syphilis study, 252 21st Century Cures Act: enactment of, 123 Twitter, 357 Type 2 diabetes, 84 UGC. See User-generated content UI Health. See University of Illinois Hospital and Health Sciences System Ulrich, D., 298 Umbilical cord blood transplant, 192 UMLS. See Universal Medical Language System UNESCO, 194 United States: cost of healthcare in, 316; healthcare sectors in, 327; health spending per capita in, 208, 209, 221; information system design in, 229; life expectancy in, 208, 209, 213; population health system in, 217, 217; road construction/maintenance in, 246–47; status of population health in, 208–9 United States Health Information Knowledgebase, 260 Universal Medical Language System, 28 University of Illinois at UrbanaChampaign: in partnership with Carle Health System, 67 University of Illinois Hospital and Health Sciences System: clinical decision support system committee

Index

governance structure, 137, 138; effective clinical decision support system implementation at, 137–40 University of Iowa, 32 University of Kansas Medical Center, 32 University of Minnesota Academic Health Center, 32 University of Missouri: eldercare monitoring project at, 34; Hospitals and Clinics, 240 University of Missouri–Columbia, 32 University of Missouri Health Care: smartphone app for patients with depression, 62 University of Nebraska Medical Center, 32 University of Texas Health Science Center at San Antonio, 32 University of Texas Southwestern, 32 Urgent care: use of clinical decision support systems in, 65 Urgent decision context, 64–65 US Department of Education, 360 US Department of Health and Human Services, 347; adoption of Health Level Seven International standards, 134; Office of Civil Rights, 348 US Department of Veterans Affairs: electronic health record of, 101 User-generated content: definition of, 172 US Food and Drug Administration, 194; cybersecurity of medical devices guidance, 359; regulation of clinical decision support system software and devices, 123 U.S. Health in International Perspective: Shorter Lives, Poorer Health (Institute of Medicine), 208 USHIK. See United States Health Information Knowledgebase US Public Health Service, 84 Valuation: of healthcare services, 326–28; of information technology infrastructure, 328–35

Value-based care: not all innovation created equal in transition to (case study), 91–93 Value-based decision making, 59 Value-based reimbursement, 319–20; communities of practice and, 88–89; medical home financing and, 322 Value measurement methodology, 334 Varma, A., 296, 299 Vendors: data use agreements with, 285; health information technology, 299 Vensim software, 112 Vertical structure, 77 Veterans Health Administration: MyHealtheVet system, 171 Videoconferencing, 180 Virtualization, 356–57 Virtual monitoring, 159–60 Virtual visits, 178–79 Visual Dx, 135–36 VMM. See Value measurement methodology Vocabularies: clinical, development and linkage of, 80, 81; controlled, data-mining frameworks and, 104; establishing, information technology development and, 97; source, 262; standardized, 80. See also Controlled terminologies; Standardized terminologies Volker, Nicholas, 192, 194, 195, 199 Voltage drop, 320 Wages: as percentage of operating budget, 295 Waldrop, M. M., 196 WannaCry ransomware, 358 Warfarin protocols: precision medicine and, 192 Watson Assistant (IBM), 38 Wearable devices, 168 WebMD, 174 Web portals, 168 Web 2.0: definition of, 172; Health Insurance Portability and

409

410

I n d ex

Accountability Act compliance issues and, 175; online health environments and, 174 Weiss, J. B., 173 Wellness, analytics for: case study, 115–17 Wellness model: integrated systems perspective and, 84 Wenger, E. C., 88 Whittington, J., 210, 212 WHO. See World Health Organization Winslow, C. E. A., 219 Woolco, 92 Woolworth, F. W., 92 Wootton, R., 169 Work behaviors: proactive, in health services delivery process, 293–94

Workflow(s): clinical, 151–52; definition of, 151; transition-in-care, 153–54 World Economic Forum, 244 World Health Organization, 8; Global Digital Health Index, 244; Health Academy, 242–43; Health InterNetwork Access to Research Initiative, 238; One Health approach championed by, 197; public health defined by, 211; social determinants of health defined by, 214 Worthey, Elizabeth, 194, 195 Wright, A., 122, 124 YouTube, 357

ABOUT THE AUTHORS/EDITORS Gordon D. Brown, PhD, is professor emeritus in the Department of Health Management and Informatics at the University of Missouri School of Medicine, where he served as department chair for 28 years. He has worked as a consultant on health system development and global health systems; as a scientist for the World Health Organization’s Division of Epidemiology and Communications Science, he developed integrative health system models. Dr. Brown was also chair of the Accrediting Commission on Education for Health Services Administration (now the Commission on Accreditation for Health Management Education) and a founding director and faculty member of the National Center for Managed Health Care Administration. He is a recipient of the Gary L. Filerman Prize for Educational Leadership from the Association of University Programs in Health Administration and was awarded the Public Service Award by the American College of Health Care Administrators. Dr. Brown is the author of numerous articles and books, including Strategic Management of Information Systems in Healthcare and Health Informatics: A Systems Approach. He has held faculty appointments at the Pennsylvania State University and the Universidad del Valle in Cali, Colombia. He was a faculty member in the Executive Program in Health Administration of the University of Alabama at Birmingham, which was taught in the Kingdom of Saudi Arabia. He earned his MHA and PhD degrees from the University of Iowa. Kalyan S. Pasupathy, PhD, is an associate professor in the Department of Health Sciences Research at the Mayo College of Medicine in Rochester, Minnesota. He is the founding scientific director of the Clinical Engineering Learning Laboratories and leads the Information and Decision Engineering program in the Mayo Clinic Kern Center for the Science of Health Care Delivery. His research advances the science and technology at the intersection of people, processes, and information to transform care delivery systems. His practice translates innovative engineering and informatics principles into awardwinning solutions that improve care delivery. He is frequently consulted by organizations and federal agencies and invited to lecture internationally. He received his MS and PhD degrees in industrial and systems engineering from Virginia Polytechnic Institute and State University.

411

412

A bo u t t h e A u th o r s / E d i to r s

Timothy B. Patrick, PhD, is an associate professor in (and was founding chair of) the Department of Health Informatics and Administration in the College of Health Sciences at the University of Wisconsin–Milwaukee. He received an MS degree in computer science and a PhD degree in philosophy/ logic at the University of Missouri–Columbia School of Medicine, where he also completed postdoctoral studies in medical informatics through a National Library of Medicine postdoctoral fellowship. His current research focuses on metadata, data representation, and informatics to improve the reproducibility of biomedical science. He is currently a commissioner for the Health Informatics Accreditation Council of the Commission on Accreditation for Health Informatics and Information Management Education.

ABOUT THE CONTRIBUTORS Dixie B. Baker, PhD, is a senior partner at the healthcare consulting firm Martin, Blanck and Associates, where she provides consulting services in the areas of health information technology, electronic health records technology, privacy and security technology, and the sharing and protection of genomic and clinical data. She serves on several national and international health- and genomics-related advisory boards. In 2017, Health Data Management recognized Dr. Baker for her thought leadership by naming her one of the “most powerful women in health information technology.” She holds academic degrees from The Ohio State University, Florida State University, and the University of Southern California. James D. Buntrock, MS, is a vice chair of information technology at Mayo Clinic. He leads an organization for enterprise technology services that supports enterprise platforms of data/analytics, application programming interface, identity, and knowledge and shared services for software development and innovation. He is currently an instructor in biomedical informatics at the Mayo Clinic College of Medicine and Science. George Demiris, PhD, is a Penn Integrates Knowledge University Professor at the University of Pennsylvania with joint appointments in the School of Nursing and the Perelman School of Medicine. His research focuses on the design, implementation, and evaluation of patient-centered technologies and tools to empower older adults and their families and to redesign care services in home and hospice care. He is a fellow of the American College of Medical Informatics and of the Gerontological Society of America. He was a professor of nursing at the University of Washington and holds a PhD in informatics from the University of Minnesota. Pavithra I. Dissanayake, DO, is a postdoctoral fellow at the Informatics Institute of the University of Alabama School of Medicine. She obtained her medical degree from Edward Via College of Osteopathic Medicine, completed pathology training at Baptist Health Systems, and completed surgical pathology and clinical informatics training at the University of Illinois at Chicago. Her research interests include clinical decision support and application of clinical ontology in the electronic health record. 413

414

A bo u t t h e Con tr i b u to r s

Carmelo Gaudioso, MD, MBA, PhD, is an assistant member of clinical research at the Roswell Park Comprehensive Cancer Center. He has a PhD in medicine and surgery and a PhD in biomedical informatics. His research work is in the areas of clinical data management; standard terminologies; biomedical informatics and research networks; and the design, evaluation, and adoption of multidisciplinary knowledge management and decision support systems to enable evidence-based decision making by multidisciplinary cancer care teams. Brian K. Hensel, PhD, MSPH, PhD, is an assistant professor in the Department of Health Management and Informatics at the University of Missouri School of Medicine. He completed a National Library of Medicine postdoctoral fellowship in health informatics and an Aging Policy/American Political Science Association Congressional Fellowship, working on legislation related to palliative care and advance planning as well as on health information technology. He teaches strategic planning and marketing and decision making for healthcare organizations, and his research and consulting interests include the application of health informatics and media to improve health, chronic disease management, and long-term and palliative care. He holds an MSPH from the Department of Health Management and Informatics and a PhD in journalism from the University of Missouri. Mark A. Hoffman, PhD, is chief research information officer at Children’s Mercy Hospital of the University of Wisconsin–Madison. His specialties include medical informatics, bioinformatics, personalized/precision medicine, translational medicine, and outcomes research. He was vice president of genomics and research at the Cerner Corporation. He was also on the faculty in the Department of Biomedical and Health Informatics and the Department of Pediatrics at the University of Missouri–Kansas City, where he launched the Center for Health Insights. He earned his PhD in bacteriology from the University of Wisconsin–Madison. Julie M. Kapp, MPH, PhD, is an associate professor in the Department of Health Management and Informatics at the University of Missouri School of Medicine. She is a Fellow of the American College of Epidemiology and a 2014 Baldrige Executive Fellow. Her research interests include epidemiology, public health, health program and policy evaluation, health promotion, and disease prevention. She has served as director of the Health and Behavioral Risk Research Center; as a Margaret Proctor Mulligan Faculty Scholar; and as director of the Partnership for Evaluation, Assessment, and Research. She is a National Institutes of Health grant reviewer, an associate editor for BMC Cancer, and a reviewer for the International Journal of Cancer and The Lancet. She holds an MPH and a PhD in public health with a concentration in epidemiology from Saint Louis University.

A b out the C ontr ib utor s

Naresh Khatri, PhD, is an associate professor in the Department of Health Management and Informatics at the University of Missouri School of Medicine. He has been recognized as one of the top 50 healthcare professors in the United States by healthcareadministrator.org, received the Emerald Literati Network’s Award for Excellence for a research article, and received grant funding from the Agency for Healthcare Research and Quality. Dr. Khatri’s research interests include strategic human resource management, transformational leadership, and cross-cultural organizational behavior. He earned a PhD in organizational behavior and human resource management from the State University of New York and an MBA from the Indian Institute of Management. Carol G. Klingbeil, DNP, is a clinical assistant professor of nursing at the University of Wisconsin–Milwaukee College of Nursing. She is the lead for the DNP Informatics Track and teaches clinical informatics for graduate students at the University of Wisconsin–Milwaukee. Dr. Klingbeil is an advanced practice nurse at the Children’s Hospital of Wisconsin and an urgent care pediatric nurse practitioner. She has worked for more than 35 years in practice and education settings, teaching nursing and other disciplines. She started her career in nursing at the University of Michigan in 1980 and received her DNP in 2015 from the University of Wisconsin–Milwaukee. Karl M. Kochendorfer, MD, FAAFP, is an assistant vice chancellor for health affairs, a chief health information officer, and an associate chief medical officer at the University of Illinois Hospital & Health Sciences System, as well as an associate professor of clinical family medicine at the University of Illinois at Chicago. He is board certified in clinical informatics and family medicine and has a degree in computer science from the University of Illinois at UrbanaChampaign. At the University of Illinois Hospital & Health Sciences System, Dr. Kochendorfer is the executive sponsor for the meaningful use/advancing care information program, is chair of the patient portal and data governance committees, is immediate past chair of the electronic medical records and clinical guideline and protocol oversight committees, and oversees the clinical decision support and clinical documentation improvement teams. He holds several informatics-related patents and founded the company MedSocket. Norma M. Lang, PhD, has authored numerous publications. Her pioneering work in identifying standards and measures to evaluate the quality of nursing care has served as a basis for nursing policy throughout the world. Her research expertise includes quality assurance, nursing standards and outcome measures, peer review, the Nursing Minimum Data Set, and the International Classification for Nursing Practice. She is a fellow of the Institute of Medicine, the American Academy of Nursing, and the College of Physicians of

415

416

A bo u t t h e Con tr i b u to r s

Philadelphia; an honorary fellow of the Royal College of Nursing in London; and an honorary member of the American Association of Colleges of Nursing. She was named to the governor’s board of directors for the Wisconsin Relay of Electronic Data for Health (now called the Wisconsin Statewide Health Information Network) and currently serves on its policy committee. She is a member of the Office of the National Coordinator for Health Information Technology’s Quality Measurement Workgroup as well as the Professional Leadership Group of the Robert Wood Johnson Foundation’s Aligning Forces for Quality Initiative in Wisconsin. Dr. Lang has served in several advisory capacities for the American Nurses Association and the American Medical Informatics Association workgroups. She received the American Nurses Association’s First President’s Award, the Distinguished Membership Award, and the Jessie M. Scott Award. In addition, she was recognized by the North American Nursing Diagnosis Association with an Outstanding Leadership Award and by The Joint Commission with the prestigious Ernest A. Codman Individual Award. In 2010, her lifetime achievements were honored when she was inducted as a Living Legend in the American Academy of Nursing. Currently, she leads the knowledge-based nursing initiative at the University of Wisconsin–Milwaukee. Dr. Lang received a BSN from Alverno College and an MSN and a PhD from Marquette University. She also holds honorary doctorates from the State University of New York and Marquette University and an honorary master’s from the University of Pennsylvania. Ryan Marling, MPH, currently works at the Clayton Christensen Institute, where he investigates potentially disruptive healthcare delivery models and the technologies that will enable their success. He is particularly interested in health information technology and is currently researching disruptive innovation in the space of electronic health records. He earned a BS in biochemistry from the University of Iowa and an MPH in health policy and management from the Boston University School of Public Health. He has worked in health policy research at the Massachusetts Medical Society, which publishes the New England Journal of Medicine. Aurash A. Mohaimani, BS, is currently a doctoral candidate in the biomedical and health informatics PhD program at the University of Wisconsin–Milwaukee. He is currently the sole bioinformatician at the Great Lakes Genomics Center. He received a BS in molecular biochemistry and biophysics from the Illinois Institute of Technology. Mihail Popescu, PhD, is an associate professor in the Department of Health Management and Informatics at the University of Missouri School of Medicine. His current research interests include eldercare, pattern recognition

A b out the C ontr ib utor s

technologies, image processing, computational intelligence, agent-based modeling, and mass behavior modeling. He obtained a BS in electrical engineering from the Polytechnic Institute of Bucharest, as well as an MS in medical physics, an MS in electrical engineering, and a PhD in computer science from the University of Missouri–Columbia. Blaine Reeder, PhD, is a faculty member at the University of Colorado College of Nursing. He conducts informatics research to connect the contexts of personal health and public health with a focus on three areas: aging in place, organizational information systems, and research tools. His aims are to improve information interactions at individual levels, communication and coordination at group levels, and health outcomes at population levels. Dr. Reeder’s research is inherently interprofessional. He has collaborated with nurses, physicians, epidemiologists, computer scientists, gerontologists, public health administrators, and other informaticians in pursuit of his research goals. Peter J. Tonellato, PhD, is director of the Center for Biomedical Informatics Research and professor in the Department of Health Management and Informatics at the University of Missouri School of Medicine. His research focuses on biomedical informatics, mathematical modeling, and simulations. Pei-Yun Tsai, PhD, RN, is an associate scientist at the University of Wisconsin–Milwaukee (UWM) College of Nursing. She implements and manages electronic health records for two community nursing centers as part of the UWM Automated Community Health Information System data repository. She also provides support to an informatics research program in collaboration with the UWM community nursing center director, faculty, and students. Phyllis M. Wise, PhD, is CEO of the Colorado Longitudinal Study, an organization committed to creating a definitive biobank and accompanying data on the social determinants of health and using community health data to facilitate a paradigm shift in disease prevention, early detection, optimized treatment, and health equity. She served as the chancellor of the University of Illinois at Urbana–Champaign and as interim president and provost of the University of Washington. Dr. Wise is a member of the National Academy of Medicine and the American Academy of Arts and Sciences, and she is a fellow of the American Association for the Advancement of Science. She serves on the board of directors of the Robert Wood Johnson Foundation and the board of directors of RAND Health. She received her bachelor’s degree from Swarthmore College and her doctoral degree from the University of Michigan.

417