Secondary Findings in Genomic Research [1 ed.] 0128165499, 9780128165492

Table of contents :
Cover
Secondary
Findings in
Genomic Research
Copyright
Contributors
In Memoriam: Pia Erdmann
Introduction
Unexpected findings in medicine
Our general approach in this book
Clarifying the issues by defining the terms—Different types of additional findings
The chapter structure of the book—Topics and approaches
References
1
Concept, history, and state of debate
Introduction: “Additional” findings in medical research
History of debate: A four-layer model
Lessons learned from population-based imaging
Quality of information
To tell or not to tell?
What to tell? Actionability and clinical utility
How to tell?
How to prepare participants?
How to design an appropriate research protocol
In search of an appropriate ethical framework
Conclusion
References
2
Oversight, governance, and policy for making decisions about return of individual genomic findings
Introduction
International legal and ethical principles
Planning ahead: Return of results protocols
Consent
What to return (or not)
How to return
Children and adults who lack decisional capacity: Special considerations
Return of secondary findings versus the right of access
International data sharing and return
Conclusion
Acknowledgments
References
3
Selecting secondary findings to report: Creating a list that suits your study
Types of genomic findings
Multifactorial disease risks
Pharmacogenomics
Mendelian disease
Carrier status
Criteria for a reportable secondary finding
Confirmation and reporting
Genes
Actionability
Participant and study factors
Variants
Balancing risks and benefits
Lists
Conclusion
References
4
How secondary findings are made
Secondary findings in clinical sequencing
Chapter overview
Current methods for large-scale DNA sequencing in clinical laboratories
Sequence read alignment, variant calling, and annotation
Use of variant filtration to detect or avoid secondary findings
Potential categories of secondary findings
Variant validation via alternative sequencing assay
Conclusion
References
5
Informed consent and decision-making
Introduction
Ethical issues
Respect for autonomy and informed consent
Providing information
Offering options and eliciting preferences
Deciding what to disclose
Counseling challenges
Addressing barriers to understanding
Accounting for changing information
Managing therapeutic and diagnostic misconceptions
Approaches for addressing ethical issues and counseling challenges
Genetic counseling
Decision-making without formal counseling
Consent models
Staged consent
Modular consent
Family consent
Broad consent
Conclusion
References
6
Reporting of secondary findings in genomic research: Stakeholders’ attitudes and preferences
Introduction and background
Why measure preferences? Normative ethical questions
Methodological answer: It depends on the validity of the empirical research
Epistemological answer: It depends on the intended use of the empirical data
Gaining insights into current moral stances or into moral behavior
Identifying ethical problems or aspects
Solidifying or contextualizing (specifying) accepted norms or principles
Justifying norms: Information regarding acceptability/practicality
Justifying norms/single actions: As a component in coherentism
Justifying norms/single actions: As (more or less) direct justification
Relevance of background premises for guideline development
Eliciting preferences for developing policies
How to measure preferences? Strategies and tools for eliciting preferences on secondary findings
Whose preferences need to be elicited?
Factors influencing preferences: Personal and disease related
Tools or instruments to elicit stakeholders’ preferences
When to elicit stakeholders’ preferences?
What has been found? Stakeholders’ attitudes and perspectives
Professionals’ attitudes toward disclosing or not disclosing: Clinicians, researchers, and IRB members
Preferences for secondary findings
Impacts and implications of secondary findings and the question of understanding and literacy
Information process before consent to genetic research and process after disclosure of secondary findings
Rights, responsibilities, policies, and practices
Participants’ and lay attitudes and preferences
Preferences for secondary findings
Impacts and implications of secondary findings
Information process before consent to genetic research and process after disclosure of secondary findings
Rights, responsibilities, policies, and practices
Summary
References
Further reading
7
Disclosing genomic sequencing results
Context of the research: Managing participant expectations
Characteristics of the study participants
Notification to participants of a sequence result
Disclosure modality and content
Laboratory reports of genomic findings
Who will return results
Summary
References
Further reading
8
Implications of secondary findings for clinical contexts
Introduction
International approaches to genomics in clinical care and translational medicine
United States
United Kingdom
Australia
Germany
France
Canada (Quebec)
Singapore
Estonia
Japan
Summary
Emerging and future scenarios
Pediatric, neonatal, and prenatal genomics
Wellness genomics
Storage and return of raw sequence data in the clinical and research settings
Economic dimensions of returning secondary findings from genomics in clinical routine care
Discussion and conclusions
References
9
Secondary findings: Building a bridge to the future of ELSI
Secondary findings as sentinel debate
Secondary findings as cultural crossroads
Secondary findings and the future of ELSI scholarship
Reanalysis for clinical purposes
Polygenic risk scores
References
Index
A
B
C
D
E
F
G
H
I
J
L
M
N
P
Q
R
S
T
U
V
W
Back Cover

Citation preview

Secondary Findings in Genomic Research

Translational and Applied Genomics Series

­Secondary Findings in Genomic Research Volume Editors

Martin Langanke

Department of Social Work, Protestant University of Applied Sciences, Bochum, Germany

Pia Erdmann†

Faculty of Theology, University of Greifswald, Greifswald, Germany

Kyle B. Brothers

Endowed Chair of Pediatric Clinical and Translational Research, University of Louisville, Louisville, KY, United States

Series Editor

George P. Patrinos

Department of Pharmacy, University of Patras, School of Health Sciences, Patras, Greece; United Arab Emirates University, College of Medicine and Health Sciences, Department of Pathology, Al-Ain, United Arab Emirates; Erasmus University Medical Center, School of Medicine and Health Sciences, Department of Pathology—Bioinformatics Unit, Rotterdam, The Netherlands

Academic Press is an imprint of Elsevier 125 London Wall, London EC2Y 5AS, United Kingdom 525 B Street, Suite 1650, San Diego, CA 92101, United States 50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, United Kingdom © 2020 Elsevier Inc. All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Details on how to seek permission, further information about the Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions. This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein). Notices Knowledge and best practice in this field are constantly changing. As new research and experience broaden our understanding, changes in research methods, professional practices, or medical treatment may become necessary. Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein. In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility. To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. Library of Congress Cataloging-in-Publication Data A catalog record for this book is available from the Library of Congress British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library ISBN: 978-0-12-816549-2 For information on all Academic Press publications visit our website at https://www.elsevier.com/books-and-journals

Publisher: Andre Gerhard Wolff Acquisitions Editor: Peter Linsley Editorial Project Manager: Kristi Anderson Production Project Manager: Punithavathy Govindaradjane Cover Designer: Christian Bilbow Typeset by SPi Global, India

Contributors Jonathan S. Berg Department of Genetics, School of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, NC, United States Gabrielle Bertier Center for Genomic Health, Charles R. Bronfman Institute for Personalized Medicine, Icahn School of Medicine at Mount Sinai, Manhattan, NY, United States Kevin M. Bowling HudsonAlpha Institute for Biotechnology, Huntsville, AL, United States Kyle B. Brothers Endowed Chair of Pediatric Clinical and Translational Research, University of Louisville, Louisville, KY, United States James Buchanan Health Economics Research Centre (HERC), Nuffield Department of Population Health, University of Oxford, Oxford, United Kingdom Alena Buyx Institute for History and Ethics of Medicine, Technical University of Munich (TUM), Munich, Germany Yasmin Bylstra SingHealth Duke-NUS Institute of Precision Medicine (PRISM), Singapore, Singapore Eva De Clercq Institute for Biomedical Ethics, University of Basel, Basel, Switzerland Gregory M. Cooper HudsonAlpha Institute for Biotechnology, Huntsville, AL, United States Pia Erdmann† Faculty of Theology, University of Greifswald, Greifswald, Germany Ann Katherine Major Foreman Department of Genetics, School of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, NC, United States Clara Gaff Melbourne Genomics Health Alliance, Parkville; Faculty of Medicine, Dentistry and Health Sciences, The University of Melbourne, Melbourne, VIC, Australia Leigh Jackson University of Exeter Medical School, Royal Devon & Exeter Hospital, Exeter, United Kingdom

xi

xii

Contributors

Kazuto Kato Department of Biomedical Ethics and Public Policy, Graduate School of Medicine, Osaka University, Suita, Japan Elke Kaufmann National Center for Tumour Diseases (NCT), German Cancer Research Center (DKFZ), Heidelberg, Germany Susan Kelly Egenis Centre for the Study of the Life Sciences, University of Exeter, Exeter, United Kingdom Bartha Maria Knoppers Centre of Genomics and Policy, McGill University, Montreal, QC, Canada Martin Langanke Department of Social Work, Protestant University of Applied Sciences, Bochum, Germany Gabriel Lázaro-Muñoz Center for Medical Ethics and Health Policy, Baylor College of Medicine, Houston, TX, United States Liis Leitsalu Estonian Genome Center, University of Tartu, Tartu, Estonia Wenke Liedtke Faculty of Theology, University of Greifswald, Greifswald, Germany Marcel Mertz Working Group Research/Public Health Ethics & Methodology, Institute for History, Ethics and Philosophy of Medicine, Hannover Medical School, Hannover, Germany Lili Milani Estonian Genome Center, University of Tartu, Tartu, Estonia Michael Morrison Centre for Health, Law and Emerging Technologies (HeLEX), Faculty of Law, University of Oxford, Oxford, United Kingdom Gesine Richter Institute of Experimental Medicine, Division of Biomedical Ethics; Institute of Epidemiology, Kiel University, University Hospital Schleswig-Holstein, Kiel, Germany G. Owen Schaefer Centre for Biomedical Ethics, Yong Loo Lin School of Medicine, National University of Singapore, Singapore, Singapore Christoph Schickhardt National Center for Tumour Diseases (NCT), German Cancer Research Center (DKFZ), Heidelberg, Germany

Contributors

Sebastian Schleidgen Institute of Philosophy, Faculty of Humanities and Social Sciences, FernUniversität in Hagen, Hagen, Germany Mahsa Shabani Centre for Biomedical Ethics and Law, Katholieke Universiteit Leuven, Leuven, Belgium Harriet Teare Centre for Health, Law and Emerging Technologies (HeLEX), Faculty of Law, University of Oxford, Oxford, United Kingdom; HeLEX Melbourne, Melbourne Law School, The University of Melbourne, Carlton, VIC, Australia Michelle L. Thompson HudsonAlpha Institute for Biotechnology, Huntsville, AL, United States Adrian Thorogood Centre of Genomics and Policy, McGill University, Montreal, QC, Canada Erin Tutty Victorian Clinical Genetics Services, Murdoch Children’s Research Institute, Royal Children's Hospital, Parkville, VIC, Australia Janet L. Williams Senior Research Genetic Counselor, Genomic Medicine Institute, Geisinger, Danville, PA, United States Eva C. Winkler National Center for Tumour Diseases (NCT), German Cancer Research Center (DKFZ), Heidelberg, Germany Sarah Wordsworth Health Economics Research Centre (HERC), Nuffield Department of Population Health, University of Oxford, Oxford, United Kingdom Ma’n Zawati Centre of Genomics and Policy, McGill University, Montreal, QC, Canada

xiii

In Memoriam: Pia Erdmann When we embarked on the journey to stitch this book together, we were a team of three. Martin and Pia were happily married and living in Stralsund, Germany. Kyle lived thousands of miles away in Kentucky. Despite the distance, we had been colleagues and friends for nearly a decade: our sons about the same age, our vision for the good life remarkably similar. We first met through our scholarship. Martin, an ethicist with a background in philosophy, was visiting Nashville to learn about BioVU, a biorepository at Vanderbilt University. Kyle, who had just completed his training as a pediatrician, had started his career working on ethical issues raised by BioVU. Martin and Kyle struck up a friendship, and soon, Kyle was in Germany where he met Pia, a social scientist and an ethicist in her own right. This trip was followed by numerous other trips, including one where our sons were able to meet and play video games together, despite their language barrier. It was as old friends, then, that we decided several years ago to work on this book together. However, in the summer of 2018—long before this book was complete— we lost Pia: wife to Martin and old friend to Kyle, a mother, scholar, saxophonist, and saint. Many of us struggle through our lives, looking for something indefinable that seems to be missing. We feel that we have a hole somewhere inside, and we could be comfortable with ourselves if only we could find that something to fill in the hole. Pia was one of those rare humans who never struggled in this way. She always knew who she was. She always knew the right direction to go; she always did exactly what she knew she must do. She was an ethicist whose life was a master class on ethics. Pia lived each day as a gift that should not be wasted. Perhaps, it was because of her training as a nurse, or perhaps, it was a result of her earlier experience as a breast cancer survivor. She loved jazz music, so she methodically made time in her life to play. She wanted to find peace in her soul, so she took time to meditate each day. She wanted her son to know her love, so she never missed an opportunity to show him. She wanted the world to be filled with kindness, so she approached everyone with her own unique style of kindness. When she visited Kyle’s family in Kentucky, Kyle’s wife Becky observed, “It was like she brought light into our home.” That was no fluke: she brought light into the world every day, everywhere she went. And she died as she lived. She accepted her destiny and decided to face it with peace and love. She continued playing jazz. She continued meditating. She c­ ontinued pouring out kindness into the world. As we later learned, she continued writing sections of this book so the burden would not be so great on her husband and friend

xv

xvi

In Memoriam: Pia Erdmann

when she was gone. She comforted others about her inevitable passing, and she prepared those she loved to endure her death. We miss her sorely, but the pain is much less thanks to the lessons she taught us before she left. Even though we lost Pia along the way, this book reflects a great deal of work on her part. The arrangement and organization of the book reflects her ideas, and the chapters bearing her name reflect both her ideas and her words. She played a major role in finding the best authors available and worked with some of them on the content of their chapters. When we finally submitted the last chapter to the publisher, she was as much a part of this book as when we started. We dedicate this book to the memory of Pia Erdmann: with gratitude for what we learned from her life and in thanks for the too-brief time that her light has shown into this world. Kyle B. Brothers Endowed Chair of Pediatric Clinical and Translational Research, University of Louisville, Louisville, KY, United States Martin Langanke Department of Social Work, Protestant University of Applied Sciences, Bochum, Germany

Introduction Kyle B. Brothersa, Martin Langankeb, Pia Erdmannc,† a

Endowed Chair of Pediatric Clinical and Translational Research, University of Louisville, Louisville, KY, United States b Department of Social Work, Protestant University of Applied Sciences, Bochum, Germany c Faculty of Theology, University of Greifswald, Greifswald, Germany

The question of how to handle secondary genomic findings is one of the most enduring ethical challenges of the past decade. Even from the beginning of the Human Genome Project, scientists, clinicians, ethicists, and research participants have been asking the same question: What do we do if we discover something we weren’t expecting?

1 ­Unexpected findings in medicine Physicians have been struggling with this question for decades, even centuries. When ordering a laboratory or imaging test in the care of a patient, it is surprisingly common to discover something unexpected. Serum chemistry tests often uncover findings that were not expected. These findings are often not requested by the healthcare provider, but show up nonetheless because of the technique used by the laboratory machine. Even the physical exam, a focal point for medicine extending back to Greek antiquity, frequently turns up findings that were not expected and are unrelated to the reason the patient was seeking care. Some of the most sacred axioms of medicine are intended to help manage the uncertainty created by unexpected findings: “Never order a test unless you know what you are going to do with the result.” Or “Always use the most focused test available that will answer your clinical question.” The challenge of uncovering unexpected results that might be relevant to the health of participants has also been a part of medical research for decades. This problem was first examined thoroughly in the field of imaging research. All medical imaging technologies can expose changes in the bodies of unsuspecting research participants. With the introduction of magnetic resonance imaging (MRI), however, the ability to detect such changes increased significantly. The first sustained consideration of how to manage incidental findings in research, therefore, emerged from the use of MRI for research on the brain (cf. Refs. 1–6). In this context the challenge routinely faced by physicians—“How do I handle this result I wasn’t expecting?”—takes on an †

Deceased.

xvii

xviii

Introduction

a­ dditional layer. Researchers are often not clinicians, and their duties to their research participants are different from the duties that physicians owe their patients. These challenges from clinical care and imaging research had set the stage when researchers started utilizing multiplex genetic tests for research purposes. A multiplex genetic test is one designed to provide information about multiple locations in the genome. This type of test therefore raises the same types of issues as an MRI of the brain or a physical exam: because the technology is not narrowly focused, it is possible to generate results that were not expected (for the early stage of this discussion cf. Refs. 7–11). The number and types of potential results that can be generated by a multiplex genetic technology vary depending on the specific technology. Gene panels or ­single-nucleotide polymorphism (SNP) panels are designed to interrogate a subset of the genome. Typically, this includes as few as about five genes or as many as several thousand genes. Exome sequencing (also referred to as whole exome sequencing) involves nearly all of the human genome’s 20,000 genes, while genome sequencing (also referred to as whole genome sequencing) covers the entire human genome, including all genes and the regions between. The discourse over how to handle unexpected genetic findings started with the introduction of gene panels and SNP panels and has expanded as exome and genome sequencing have become more widely available. In fact the discourse over the management of unexpected findings has become so broad and complex, and it would be extremely difficult for a newcomer to catch up. Over time, language has changed, controversies have evolved into consensus, and new controversies have sprung up in their place. Various groups around the world have issued recommendations, and these recommendations have been superseded by newer ones (e.g., Refs. 12–19). In short, those new to this field will find it difficult to know where to start. That is why we put together this book. We hope that this book will prove useful to anyone who is new to thinking about the ethical, legal, and social issues (ELSI) raised by genomic research. This audience could certainly include researchers who are starting to use genomic technologies in their research or graduate students who hope to pursue a scientific career in genomics. Graduate students pursuing a career in ethics through disciplinary training in fields like philosophy, social sciences, or bioethics will also find this book helpful. Finally, we hope well-established colleagues in ethics and bioethics will find this book as a useful entrée to the world of ELSI research. There is still a great number of problems to solve in this area, and new ideas and perspectives are needed to solve them.

2 ­Our general approach in this book In light of this goal, we have adopted a “handbook” format. Like any other handbook, this book is not necessarily meant to be read from beginning to end. Each chapter provides an overview of a specific topic area within the larger discourse on secondary findings in genomics research, and in this way, each chapter is free

Introduction

standing. It can be read on its own as an introduction to a specific topic. Although the chapter authors are world-renowned experts in their field, they have not sought to defend specific theses or introduce new insights. Rather, each chapter provides the background information that would be needed by a newcomer to understand how the discourse has evolved over time, as well as a discussion of the current state of the discourse. As a result of this strategy, certain background facts and ideas are repeated in multiple chapters. We believe this redundancy will be useful to readers who just want to read specific chapters or in classroom settings where only a few chapters may be assigned. The apparent redundancy across chapters may provide another useful opportunity, however. Readers will find that, even though chapter authors will sometimes cover territory that has already been addressed in other chapters, this territory will be presented from a slightly different perspective. The chapter authors come from numerous disciplinary backgrounds, including philosophers, social scientists, laboratory scientists, clinicians, and even multidisciplinary bioethicists. This book therefore provides the reader an opportunity to see some of the issues that are most critical to this debate from a variety of perspectives. We consider this to be a “feature” rather than a “bug.” As with any controversial issue, the best way to grasp points of disagreement is to understand the various points of view.

3 ­Clarifying the issues by defining the terms—Different types of additional findings Although we have allowed certain inconsistencies and uncertainties to persist across chapters so that readers will have an opportunity to understand different perspectives, we have endeavored to harmonize the use of certain key terms throughout the book. So far in this introduction, we have been speaking primarily of “unexpected” findings. This term is useful as a starting point because it is a familiar word that is being used in a familiar way. However, it lacks the precision that we will need to effectively discuss the problems raised by genomics results in research settings. First, the extent to which a finding can be thought of as “unexpected” can vary widely. It is difficult to argue that a result was unexpected when only five genes were being tested. When 20,000 genes are being tested, however, that claim holds more water. Second, the distinction between expected and unexpected is not particularly clear. Those who conduct imaging research know that they will sometimes uncover results that are not related to the purpose of their research. Researchers studying the effects of stress on the functioning of the brain, for example, will perform imaging studies to fulfill the purpose of their study. They know, however, that they will sometimes uncover a finding, like a brain mass, that is not related to that purpose. In that respect, such a finding would not be entirely unexpected by the researcher. However, if the research participant has no symptoms of a brain mass, both the researcher and the participant will experience this finding as a surprise. In some ways, then, these results are often both expected and unexpected at the same time.

xix

xx

Introduction

Rather than focusing on whether a result is expected, it is more useful to speak of results in relation to the purpose for doing the study, whether it be a research purpose or a clinical purpose. In clinical practice a physician treating a child for fever and respiratory distress might order a chest X-ray. If the X-ray were to reveal that the child has pneumonia, this would be considered a relevant explanation for the child’s symptoms and thus a primary finding. Findings that would not explain fever or respiratory distress, such as the discovery that someone has broken the child’s ribs, would not be considered a primary finding. In general, we will refer to such findings as additional findings, since a recent study showed that this term is preferred because it is neutral and easily understood.20 However, not all additional findings are the same. One important distinction to consider is whether a result was generated incidentally. An incidental result is one that could not have been avoided; it must be generated for the clinician or researcher to reach their goal. In imaging, most additional findings are best classified as incidental findings. Because imaging technologies produce images that cover a specific part of the body, unexpected changes in that part of the body will unavoidably be “seen.” Incidental findings are certainly produced by multiplex genetic tests, as well. For example, to confirm that samples have not been switched or data have not been incorrectly transposed, genetics laboratories frequently compare the reported sex of the individuals being tested with the genetic sex revealed by the genetic testing technology. This quality control measure will incidentally also reveal sex chromosome aneuploidies—differences in the number of X chromosomes or Y chromosomes that cause medical conditions like Turner syndrome or Klinefelter syndrome. The unexpected discovery of one of these medical conditions could be said to be incidental to the quality control method utilized in the laboratory. To fulfill the intended purpose of the clinical or research test, it is unavoidable that such findings will sometimes be generated. Many of the findings generated by multiplex genetic tests, however, are not truly unavoidable. Once the raw data from a genetic test are generated, it becomes necessary to analyze that data using computerized tools. These tools can be designed to uncover only the minimum number of results needed to fulfill the test’s primary purpose, or it can be designed to reveal other findings as well. In genetics and genomics, it has become common to refer to these kinds of results—results that could have been avoided but were intentionally uncovered—as secondary findings. This book, then, is intended to illuminate the ethical, legal, and policy issues that are created by additional findings generated through the use of multiplex genetic technologies in research settings. Some of these findings are incidental to the research process, so the most important questions relate to issues like how to prepare participants for these results and how to manage them once they are generated. For secondary results, however, we need to consider the additional question of whether they should be generated in the first place. Is there an obligation to modify the research process to intentionally uncover secondary findings?

Introduction

While the authors of this book’s chapters will use these terms consistently, it is important to note that the use of these terms has evolved over time. A bioethicist writing a decade ago might have used the phrase incidental findings when today we would say that they were referencing secondary findings. In considering how these debates have evolved over time, therefore, it is necessary to keep in mind that the earlier works that are referenced and discussed in this book might not have used terms in the same way we have chosen to define them.

4 ­The chapter structure of the book—Topics and approaches In keeping with the general strategy for this book, we have decided to cover the complex issues related to the management of secondary findings in genomics and genetics research by dividing the relevant issues into eight thematic clusters. Working with experts in each area, this book comprises a chapter that covers each collection of issues, as well as a closing chapter that provides a summary of the book and some thoughts on future directions for work on secondary findings. Chapter  1 analyzes how the discourse on secondary findings in genomics research emerged from an earlier discourse on “incidental findings” in imaging research, where many of the most important issues were first addressed in the context of magnetic resonance imaging of the brain. By revisiting the roots of this debate, the chapter authors demonstrate that the field of genomics can still learn important lessons by investigating the debate on incidental findings in the context imaging. With regards to the management of additional findings in genomics research, the chapter highlights six of these lessons: First, researchers should adopt practices and procedures that emphasize quality. Second, researchers need to decide whether to disclose findings to participants. Given the fact that a rigid nondisclosure policy does not work well in most circumstances, researchers must, third, decide which types of results to disclose. In most cases, researchers should determine prospectively which results will be returned based on the anticipated clinical utility or “actionability” of findings. Fourth, it is critical to determine how additional findings will be disclosed to participants. This includes who will actually disclose results to participants, as well as how other resources like multimedia information will be utilized. Fifth, the authors emphasize the importance of giving participants the opportunity to consider relevant information so that their decisions to receive additional findings are adequately informed. Sixth, in the spheres of both imaging and genomics research, a protocol should be produced reflecting all of the practical and ethical issues mentioned. This protocol should be summited to a research ethics committee (REC) or institutional review board (IRB) for review and approval. Overall, Chapter 1 argues that the management of additional findings in imaging research as well as in genetics and genomics research needs to be approached carefully, based on a plan that is ethical, responsible, and feasible.

xxi

xxii

Introduction

Chapter 2 focuses on legal and regulatory issues relevant to the management of secondary findings in genomics. The authors of this chapter demonstrate how laws and policies are increasing the disclosure of individual findings in research contexts. IRBs and RECs are also giving increased attention and scrutiny to how findings are handled. Nevertheless, it appears that expanding regulation has not created greater certainty for researchers. This is partly because fundamental ethics and policy issues remain unresolved. An inherent difficulty faced by those seeking to draft policy is the diversity of contexts where genomics is being conducted. In addition, policies tend to focus on scientific, medical, and ethical considerations (such as quality and utility), and generally provide little guidance on how to balance these considerations with competing concerns (such as limited expertise, time, and resources). It is increasingly clear that researchers, IRBs, and RECs cannot be expected to address individual findings on their own. They need dedicated funding and institutional support to develop and review return plans, to assess the quality and usefulness of results, to effectively communicate results, and to ensure the participant can act on the information they receive. A trend affecting both research and research oversight generally is the rise of participant engagement. The return of individual findings is sometimes considered a form of engaging with and showing respect for participants. Participants can also be engaged in designing research governance, including weighing in on project-specific plans for governing individual findings. This amounts to a shift away from policies that set out rules for researchers and toward an approach that involves the cocreation of governance. On one hand, this offers opportunities to ensure individual findings are handled in an effective and respectful manner. On the other hand, the replacement of substantive rules with engagement processes can decrease certainty and uniformity of practices over time. Emphasis on the return of secondary findings also raises concern about a shift toward “routine genomic screening,” especially given the large size of modern research cohorts. Fortunately, there are research projects like Geisinger’s MyCode exploring the prevalence of secondary findings and clinical utility, as well as the costs, benefits, and risks of such routine screening. Ultimately, participants will only realize the value of secondary findings if there is a health system ready to accept a “hand off” from researchers. But health systems are awaiting evidence of clinical utility before adopting genomics. This remains one of the fundamental “chicken and egg” problems for genomics. Chapter 3 explores potential criteria for selecting secondary findings that might be disclosed in genetic and genomic research studies. The authors deal with questions related to the process for confirming findings, presenting results in a report, assessing the clinical relevance and actionability of genetic variants, and factors related to individual studies and participants’ interests. The authors conclude that researchers who will generate large-scale sequencing data on their participants must consider what types of findings should be reported as secondary findings. The answer may depend to a great extent on the nature of study, the proximity between the participants and the research team, the perspective of institutional bodies providing oversight, the ease of analysis and reporting, and the resources available for results disclosure and genetic counseling. Such decisions should be made as part of the study design

Introduction

(and indeed will be required for an IRB or REC protocol approval), and should be reflected in study budgets. Chapter 4 provides insight into the generation of secondary findings in genomic and genetic research. More specifically, the authors discuss secondary findings in the context of exome and genome sequencing that is being conducted to identify genetic variation in individuals suspected of having a genetic disease. The authors summarize how secondary findings are discovered in clinical large-scale DNA sequencing as well as the sequencing technologies that are currently being employed by clinical and research laboratories. Moreover, they detail the downstream analytical pipelines that are utilized to detect potentially medically relevant genetic variation, and discuss places within this pipeline where secondary genetic findings can be proactively sought out, or even avoided. Chapter  5 examines the important role a robust informed consent process can play in ensuring that secondary findings are managed in ways that meet research participants’ needs and preferences. The authors explore why pretest genetic counseling is considered the gold standard for helping patients and research participants make good decisions, and why a great deal of contemporary genomics research is focused on replacing genetic counselors with alternative approaches. They point out how the complexity of genomics has created a fascinating dynamic over the past decade in which the conventional principles and strategies of research ethics have needed to be revisited in light of new genomic study designs, including challenges related to secondary findings, that had not previously been envisioned. This dynamic relationship between innovation in genomic research and innovation in research ethics has proven to be extremely generative to both. The authors conclude that in coming years, it is likely that genomics researchers and research ethicists will continue to learn from one other, creating new opportunities in both fields. Chapter 6 summarizes the main points from the debate on whether and why empirical evidence should be integrated into decision making and policy development related to secondary findings in genomic research. It provides an overview of tools and instruments appropriate for eliciting stakeholder preferences in this field. The chapter presents an overview of recent empirical studies into stakeholder preferences on secondary findings in genomic research. The authors point out that different groups of stakeholders differ in their attitudes toward secondary findings, including in their understanding and literacy, in their preferences concerning the information process before consenting to genetic research and the process after disclosure of secondary findings and, finally, in their attitudes concerning participant rights and responsibilities. Although the authors emphasize that the literature on stakeholders’ attitudes and preferences should to be interpreted with caution, they conclude that empirical knowledge on the attitudes and preferences of stakeholders can help to implement feasible recommendations and strategies for dealing with secondary findings from genomics research. These policies should create a balance between individual preferences and practicability. Chapter 7 analyzes how the decision to offer participants choices about receiving secondary findings influences the way studies should design and carry out the

xxiii

xxiv

Introduction

d­ isclosure process. According to the chapter author, the approach to disclosure should account for several important factors, including the study context, characteristics of the study population, type of results planned for return, participant notification, disclosure modality, laboratory reporting, and decisions about who will disclose findings. The author highlights the gap created by the fact that no professional practice guidelines have been developed to address these different aspects of the return of results process. In light of this situation, she argues that laboratory reports can support the disclosure process through clear and concise language to relay the specific genomic finding and implications for clinical practice. Overall the author concludes that proper planning will facilitate a robust program for disclosure of genomic findings in research settings. Chapter  8 analyzes different strategies for dealing with “additional” genomic findings in the care of adult patients in clinical contexts. Examples of clinical and translational genomics from the USA, UK, Australia, Germany, France, Japan, Singapore, Estonia, and the Canadian province of Quebec illustrate a range of different approaches to managing secondary or additional findings. Furthermore, the authors highlight other clinical situations, including testing in neonatal, pediatric, and prenatal contexts, to illustrate the wide range of clinical contexts in which secondary or additional findings must be considered. In each of these cases, practical, organizational, economic, legal, and ethical aspects must be taken into account when deciding how best to proceed with the management of secondary findings. In Chapter 9, we seek to pull together the threads addressed throughout the book. We highlight how scholars new to the study of the ethical, legal, and social implications (ELSI) of genomics can come to understand a great deal about this field through the lens of secondary findings. We highlight how the discourse on secondary findings has come to shape ELSI research, and how the debate on secondary findings reflects a number of cultural trends that are central to other ELSI issues. We close by suggesting some future directions that the discourse on secondary findings might take in coming years.

­References 1. Illes J, Desmond JE, Huang LF, et al. Ethical and practical considerations in managing incidental findings in functional magnetic resonance imaging. Brain Cogn. 2002;50:358–365. 2. Illes J, Rosen AC, Huang LF, et al. Ethical consideration of incidental findings on adult brain MRI in research. Neurology. 2004;62:888–890. 3. Illes  J, Kirschen  MP, Edwards  E, et  al. Incidental findings in brain imaging research. Science. 2006;311:783–784. 4. Check E. Brain scan ethics come under the spotlight. Nature. 2005;433:185. 5. Kirschen  MP, Jaworska  A, Illes  J. Subjects’ expectations in neuroimaging research. J Magn Reson Imaging. 2006;23:205–209. 6. Hentschel F, Klix W-E. Management inzidenter Befunde in der bildgebenden Diagnostik und Forschung. Fortschr Neurol Psychiatr. 2006;74:651–655.

Introduction

7. Knoppers BM, Joly Y, Simard J, Durocher F. The emergence of an ethical duty to disclose genetic research results: international perspectives. Eur J Hum Genet. 2006;14:1170–1178. 8. Ravitsky V, Wilfond BS. Disclosing individual genetic results to research participants. Am J Bioeth. 2006;6:8–17. 9. Cho MK. Understanding incidental findings in the context of genetcs and genomics. J Law Med Ethics. 2008;36:280–285. 10. Berg JS, Khoury MJ, Evans JP. Deploying whole genome sequencing in clinical practice and public health. Meeting the challenge one bin at a time. Genet Med. 2011;13:499–504. 11. Bredenoord  AL, Onland-Moret  NC, van Delden  JJM. Feedback of individual genetic results to research participants: in favor of a qualified disclosure policy. Hum Mutat. 2011;32:861–867. 12. Presidential Commission for the Study of Bioethical Issues. ANTICIPATE and COMMUNICATE. Ethical Management of Incidental and Secondary Findings in the Clinical, Research, and Direct-to-Consumer-Contexts. http://bioethics.gov/node/3183; 2013. Accessed 12 January 2016. 13. Deutscher Ethikrat. Die Zukunft der genetischen Diagnostik. https://www.ethikrat.org/fileadmin/Publikationen/Stellungnahmen/deutsch/stellungnahme-zukunft-der-genetischendiagnostik.pdf; 2013. (Accessed 22 March 2019). 14. Green RC, Berg JS, Grody WW, et al. ACMG recommendations for reporting incidental findings in clinical exome and genome sequencing. Genet Med. 2013;15(7):565–574. 15. German Society of Human Genetics (GfH). Stellungnahme der Deutschen Gesellschaft für Humangenetik zu genetischen Zusatzbefunden in Diagnostik und Forschung. https://www. gfhev.de/de/leitlinien/LL_und_Stellungnahmen/2013_05_28_Stellungnahme_zu_genetischen_Zufallsbefunden.pdf. Accessed 22 March 2019. 16. National Health and Medical Research Council. Principles for the Translation of ‘Omics’Based Tests From Discovery to Health Care. Canberra: National Health and Medical Research Council; 2015. 17. Project EURAT Ethical and Legal Aspects of Whole Human Genome Sequencing. Cornerstones for an Ethically and Legally Informed Practice of Whole Genome Sequencing: Code of Conduct and Patient Consent Models. Heidelberg: Marsilius-Kolleg der Universität Heidelberg; 2016. 18. National Academies of Sciences, Engineering, and Medicine. Returning Individual Research Results to Participants: Guidance for a New Research Paradigm. Washington, DC: The National Academies Press; 2018. 19. American College of Medical Genetics and Genomics. ACMG Updates Recommendation on “Opt Out” for Genome Sequencing Return of Results. https://www.acmg.net/docs/ Release_ACMGUpdatesRecommendations_final.pdf; 2018. (Accessed 26 July 2018). 20. Tan N, Amendola LM, O’Daniel JM, et al. Is “incidental finding” the best term?: a study of patients’ preferences. Genet Med. 2017;19(2):176–181.

xxv

CHAPTER

Concept, history, and state of debate

1

Martin Langankea, Pia Erdmannb,†, Wenke Liedtkeb, Kyle B. Brothersc a

Department of Social Work, Protestant University of Applied Sciences, Bochum, Germany b Faculty of Theology, University of Greifswald, Greifswald, Germany c Endowed Chair of Pediatric Clinical and Translational Research, University of Louisville, Louisville, KY, United States

1 ­Introduction: “Additional” findings in medical research As the title makes clear, this book deals with “secondary findings” in genomics research. Since terms like “secondary findings,” “additional findings,” and “incidental findings” are often used interchangeably in the literature, it is important to begin by clarifying the terminology used in this book. In accordance with the framework proposed by Janet Williams in Chapter 7, page 133, we will use the term “secondary finding” in the following sense: Secondary findings in genomics are genomic sequencing results not directly related to a participant’s indication for research sequencing, but identified based on intentional interrogation of the research sequence, often based on specified criteria.

This definition highlights that the genomic sequencing results of primary interest in this book are results that are not created “incidentally” in the research process. In other words the investigators performing sequencing in the research laboratory chose to generate secondary results for reasons distinct from the research question. Our decision in this book to specifically focus on the production and handling of these additional, nonincidental findings in genomics research is paradoxically the result of our long-term research on the ethical issues raised by findings that truly are incidental from whole-body imaging in population-based research.1–4 But this paradox can be easily resolved: incidental findings from imaging in population-based research share one key feature with findings from genomics research that are “secondary” in the sense proposed earlier. They are “additional” insofar as they are not the primary target of the research performed. In this sense, “additional findings” is a collective term that includes findings that are generated “incidentally” in research †

Deceased.

Secondary Findings in Genomic Research. https://doi.org/10.1016/B978-0-12-816549-2.00001-1 © 2020 Elsevier Inc. All rights reserved.

1

2

CHAPTER 1  Concept, history, and state of debate

and findings that are intentionally produced “secondary” to the research aims of a study. In the context of population-based imaging research, incidental findings are not ethically important simply because they are “additional” to the primary target of the research performed. These types of incidental findings are of ethical interest because they have two important features: (a) They have the potential to be medically relevant. (b) They are not expected by the research participant. With respect to (a), incidental findings demand ethical attention specifically because they potentially carry medical relevance. In other words, it is this feature that makes them ethically relevant. Findings of this sort can be very helpful for the person receiving them either because they shed light on symptoms they have been experiencing or because they reveal a condition that is not yet symptomatic. Even in the unfortunate case that such a condition turns out to be untreatable and ultimately lethal, participants may value learning about it while they are still asymptomatic because it may inform decisions related to lifestyle or finances, especially with respect to the effects on family members and children. Despite these potential benefits, disclosing incidental findings of this sort to research participants may create various types of suffering: confusion, anxiety, grief, etc. Much of this suffering will be linked to the way more severe findings are disclosed to participants. This highlights that the harm potentially created by disclosing this information can be reduced by doing an effective job of disclosing this information (or increased by doing an ineffective job). In this way the question of potential benefits and potential harms created by the disclosure of results is not only a practical question but also an ethical question. After all, one of the central principles of research ethics is that research should be conducted such that the harms created for participants are minimized. This principle informs not only the issue of how results should be disclosed to participants but also whether they should be disclosed. Withholding incidental findings from participants is risky from an ethical perspective, since failure to disclose information about an incidental finding can lead to significant harms if it turns out that the disease detected by the finding would have been treatable. From an ethical perspective, then, we can see that incidental findings can lead to a number of situations that raise ethically important challenges, including when results of poor quality are returned, when results are disclosed in a way that causes harm, or when results are withheld that might have proven medically relevant for the participant. In many cases, participants in imaging studies undergo a consent process that includes an explicit discussion that incidental findings may be generated. They may even receive a list of the types of incidental findings that will be disclosed to them if they are detected.1, 4 As noted in (b), however, even when participants are informed in this way that they may receive incidental findings, they typically are not aware that they carry such an abnormality and thus tend to experience the specific finding as a surprise.

2 ­ History of debate: A four-layer model

2 ­History of debate: A four-layer model The debate on practical and normative challenges related to handling “additional” findings in biomedical research began in the early 2000s and developed in progressive layers or phases as time progressed (see Fig. 1). It was during the first phase that, initially in the US context, practical and normative questions were raised in the bioethics literature about the processing and disclosure of incidental findings from neuroimaging studies. Much of this early discussion can be attributed to a trilogy of papers by Illes et al. that appeared in high-impact journals between 2002 and 2006.5–7 Publications in this first phase of the debate focused exclusively on magnetic resonance imaging (MRI) of the brain (for other important contributions, see Refs. 8–15). In the second phase of the debate, a new layer of discourse started to explore the more general topic of normative problems associated with the use of imaging techniques in medical research (see Fig.  1). Papers dealing with neuroimaging— including pediatric neuroimaging research—continued to appear,16–20 but increasingly the literature took a broader perspective on incidental findings from research imaging.21–28 In particular, two papers led by Susan Wolf appeared in 200821, 22 that set the stage for this second period of the debate, including the first general recommendations for responsibly managing incidental findings generated through imaging research.22 Also appearing during this period was a series of papers exploring the normative and practical aspects associated with the use of whole-body MRI in population-based research (e.g., Refs. 26–28) and philosophically grounded work examining the meaning of the term “incidental findings.”27 The third phase of this discourse was marked by the first attempts to transfer best practices and knowledge standards from research imaging contexts to nonimaging

FIG. 1 History of the debate.

3

4

CHAPTER 1  Concept, history, and state of debate

contexts, especially to genetics and genomics. Papers by Knoppers et al. (2006)29 and Ravitsky and Wilfond (2006)30 appeared as early forerunners of the developing discourse on the management of “additional findings” in genetics and genomics (at that time almost exclusively referred to as “incidental findings” by analogy with imaging). However, this third layer of discourse characterized by knowledge transfer from research imaging to research in genetics and genomics became more robust from 2008 to 2015, with contributions from a range of disciplines and national contexts (e.g., Refs. 31–38). This third phase culminated in the appearance of recommendations from around the world that were intended to guide the responsible disclosure of additional findings in the context of genetics and genomics research (e.g., Refs. 39–48). Toward the end of this third phase, the discourse about “incidental” research findings in genetics and genomics began to diverge from the discourse on imaging in terms of both terminology and methodology. While the literature in imaging contexts was increasingly focused on the special case of incidental findings from populationbased whole-body MRI,1–4 in the area of genetics and genomics, the term “incidental finding” was gradually substituted by terms like “unsolicited findings,” “secondary findings,” or “additional findings.” The rationale for this divergence in terminology involved several factors: (a) Incidental findings in the literal sense of the word, that is, findings that are detected unintentionally, are relatively rare in genomics and genetics research.49 While incidental findings from imaging occur in the research context with a frequency of up to 30%,2 genomic research findings that are literally “incidental” are relatively rare, limited primarily to the discovery of conditions like Klinefelter’s syndrome and Turner syndrome when the sex chromosomes in genetic samples are being compared with the reported gender of research participants to ensure that sample mix-ups have not occurred. (b) The vast majority of additional research findings in genetics and genomics research are produced intentionally (see Chapter 4), and therefore the usage of the term “incidental” is misleading or even incorrect. With this shift, we find ourselves in the current and fourth phase of discourse. It is in this fourth phase that this book now appears, with an overall goal of summarizing this extensive and complex debate into a single book that is accessible to those who are entering this conversation when it is already late in Act 4, as it were. To that end, it will be useful to recap the most important issues that overlap both secondary findings in genomics and incidental findings in imaging, before continuing the conversation in subsequent chapters with an exclusive focus on the practical and normative issues that are specific to secondary findings.

3 ­Lessons learned from population-based imaging As we observed earlier, incidental findings from whole-body imaging in population-­ based studies potentially carry medical relevance and are, for the most part,

3 ­ Lessons learned from population-based imaging

e­ xperienced as unsolicited and unexpected. Given that these features are shared with secondary findings in genetic and genomic contexts, it is somewhat intuitive that the legal and ethical discourse on incidental findings from imaging would be highly relevant to comparable debates about the management of genomic findings that are not directly related to the primary focus of the research.36–38 In the sections that follow, we will explore issues raised by incidental findings in the context of imaging research that are also relevant to secondary findings in the context of genomics. These issues will be discussed separately, but in truth they tend to overlap with one another and are fundamentally connected. The purpose of this discussion is not only to explore how layer 3 of this discourse (see Fig. 1) informs the modern debate but also to set the stage for the rest of the book by laying out the practical and ethical challenges that need to be addressed in all types of biomedical research that generate additional findings of potential medical relevance and, therefore, of potential interest to participants.

3.1 ­ Quality of information Although the process of generating radiological findings is quite different from the process used to generate genetic or genomic findings, the issue of quality is fundamental to both. That is, to effectively deal with additional findings in these contexts, it is critical to consider both the quality of the underlying procedures of data production and the quality of the interpretation that must be performed to make medical sense of this data. In this way the term “quality” encompasses several dimensions: First, scientific data can only be considered “high quality” if the procedures that are used to generate the data are executed according to the best available standards and practices. This means that the test itself was performed correctly using appropriate equipment and that the preparatory steps that lead up to the test were done correctly. In the setting of genetics and genomics, this means that biosamples were stored and handled correctly. In the context of imaging, this might mean, for example, that the appropriate sequences and contrast agents were selected in preparing for the ­imaging.2, 3, 9 Second, quality can be influenced by the use of innovative, nonstandard, and/ or unvalidated tests or methods that are used for research purposes. One example from imaging is the use of nonstandard MRI sequences that are not used for regular diagnostics.2, 3 Data generated by these types of investigational procedures cannot be directly transferred to the treatment context, as their quality for clinical applications is often not known. These findings must therefore be replicated using standard tests and methods that have been validated for diagnostic and treatment purposes. Failure to conduct such confirmatory testing risks the return of information to participants and their healthcare providers that is misleading and thus of dubious quality. As we have already discussed, the practice of returning incidental findings or secondary findings to research participants or their healthcare providers must be grounded in a commitment to minimizing potential harms. Since both false positives and false negatives have the potential to create harms, the results from innovative

5

6

CHAPTER 1  Concept, history, and state of debate

tests or methods (that is to say, diagnostics that are not validated for clinical uses) should only be returned to patients and their providers in conjunction with diagnostics that are appropriately validated and licensed for clinical uses. This is because the sensitivity, specificity, and reliability of innovative tests and methods are not fully validated, even if there is some initial evidence that they produce better quality results than existing methods.2, 3 In some cases, clinical confirmation of a research result may be very difficult. This can occur when innovative tests generate results that are not typically used in clinical care or when there are limited options for clinical laboratories that have the expertise and the equipment to confirm such a finding. Another challenge can arise if the research results are generated in settings where the scientists do not have the clinical resources or experience necessary to effectively and safely disclose results to participants. In both of these cases, it may not be possible to ensure that a participant will experience benefits that outweigh the potential harms of receiving a result. These factors stand to emphasize that the quality of raw data generated through innovative, nonstandard tests and methods is a crucial issue in the handling of research findings. Third the quality of data depends on the skill of the person interpreting the data, such as a laboratory scientist, radiologist, or physician scientist.2, 3, 28 While interpretation may seem straightforward, it is in fact an extraordinarily complex task. A “good” interpretation is not only one that is correct and reliable but also one that addresses how that information might be used in some context. Imagine, for example, that a whole-body imaging study performed for research purposes reveals mild anatomical abnormalities in the spine or structural differences in the fallopian tubes of a female participant.2, 3, 28 In the context of a particular research question, either of these might be relevant and useful for the research. But from the perspective of interpreting these observations as possible incidental findings, it would be necessary for the person interpreting the imaging data to consider whether they meet other standards for “quality” including actionability, nonredundancy, or pathological relevance. For example, abnormalities in the spine might be observed, but that does not mean that they correlate with back pain. In fact, they might not have an association with any medical symptoms or conditions. “Normal variation” of this type, then, might not meet quality standards related to actionability (i.e., there is nothing to do about it) or pathological relevance (i.e., there is no medical “pathology” associated with the finding). Structural differences in the fallopian tubes might face similar questions about quality. If the person interpreting the findings expects that the difference could cause infertility, this expectation would not necessarily be relevant information for an elderly participant who is no longer concerned about fertility. And it might be redundant information for a participant who has already undergone clinical evaluation for fertility issues, where this difference had already been recognized. Actionability may also be impacted by societal factors. An “actionable” result is defined as a result that can inform the use of clinical diagnostics, screening, prevention measures, or treatment interventions (see Refs. 32, 50). Typically, actionability is treated as a feature of a gene and an associated medical condition. A genetic result indicating that a patient will develop Huntington’s disease is often regarded as not

3 ­ Lessons learned from population-based imaging

a­ ctionable because the genetic result cannot be used to inform efforts to prevent or treat the condition (since there is not a way to prevent Huntington’s disease in a person who has inherited the predisposition to this condition). However, actionability could also be a factor determined by the social conditions faced by the research participant. Patients living in developing countries, or in more rural or underserved regions of developed countries, may not have access to certain medical technologies that would be necessary for realizing the actionability of an incidental or secondary finding. The discovery that a research participant carries a genetic variant that causes colon cancer risk, for example, would only be actionable if the research participant was able to access colonoscopy, preventive surgery, etc. In this way, factors associated with quality (actionability, nonredundancy, and pathological relevance) certainly do have “objective” components, in the sense that the person performing the interpretation can accurately describe what is being observed on the imaging. But they remain highly context specific insofar as they can differ in their relevance from participant to participant and depending on personal or even societal circumstances. These factors thus play an important part in determining the “quality” of an incidental or secondary finding. This examination of quality makes clear that the findings from a clinical or research test, including all types of additional findings, should not be thought of as simply “data.” They are the result of a complex and context-sensitive process of interpreting data in a particular context. A significant difference between the research and clinical settings is that in the clinical context, the patient history typically provides physicians with relevant background information (including information not only on age and sex but also on comorbidities, patient preferences, societal circumstances, etc.). This contextual information enables healthcare providers to enrich the “pure” results of a test with context-specific information. This type of important information is not necessarily available in the research context.3 This lack of contextual information can reduce significantly the quality of information for the research participant, and this effect is even more pronounced in cases where research involves innovative tests or methods that are not directly comparable with the tests or methods typically used in clinical settings. The quality of a result, and the influence of context on this quality, is thus a critical issue in examining the ethical implications of incidental and secondary results.

3.2 ­To tell or not to tell? Studies have repeatedly demonstrated that most research participants and members of the public believe that researchers are obliged to disclose any research findings that might be relevant to their health.3, 51, 52 Many participants even want to receive results of no clinical significance, a fact that likely reflects the desire to maintain autonomous control over one’s own personal health information. The idea that patients and research participants are passive recipients of information is being rapidly replaced with the realization that many patients and research participants want to play an active role in gathering their health information and interpreting its importance in

7

8

CHAPTER 1  Concept, history, and state of debate

their personal context.53 In fact, this “information-seeking” trait seems to be the main reason that many individuals choose to participate in research studies.3 This phenomenon of participating in research to receive health information creates a number of challenges for both participants and researchers. One important effect is that it can create a “diagnostic misconception,” the misunderstanding that research is being conducted to provide diagnostic benefit to the participants rather than its primary goal of benefiting society, including future patients.3, 54 The diagnostic misconception is concerning from an ethical perspective because it threatens the quality of informed consent for research participation. Ideally, potential research participants will have an opportunity to carefully consider the risks and benefits of research participation and voluntarily choose to participate (or not participate) in the research. If they misunderstand the goals of the research as an opportunity to receive diagnostic information, they are also likely to overestimate the chances that they will receive this kind of information. In many genetic studies, for example, less than 2% of participants will receive a secondary finding.55 Perhaps even more importantly, they may also fail to give due consideration to a range of negative effects that can potentially be created by the return incidental or secondary findings: (a) Psychological distress of various types, from mild to serious. (b) Discrimination in the context of personal finances, insurance, and employment. For example, merely the suspicion that one might suffer from a certain disease can affect opportunities for employment and life insurance policies. This can cause a variety of challenges.40, 56 (c) Additional financial costs created by the need for confirmatory testing or follow-up screening. In some countries (such as Germany), the costs associated with clinically verifying a research finding are usually covered by publicly funded health insurance. In some countries, however, this process may not be covered by insurance, and even when the tests themselves are covered by insurance, the research participant may end up paying other types of costs.52 Even when potential research participants are aware of these risks, they may nonetheless choose to receive additional findings. This may be because they assume that these outcomes are unlikely to occur or, in weighing the risks and benefits, they may conclude that the potential advantages outweigh the potential disadvantages. And in fact, there are a number of important possible benefits from receiving incidental or secondary findings. If a research finding creates a concern that the participant has or is at risk for developing a medical condition and clinical testing confirms this result, the participant may receive health benefits, such as the opportunity to start treatment or prevention measures earlier than they might otherwise have been available. Another reason that researchers may want to return additional findings to participants, separate from their potential health benefits, is that disclosure of this information could be viewed as a way to demonstrate respect for the person. Withholding information in opposition to the preference of the individual to receive information could be considered paternalistic. The right of a research participant to make decisions for themselves, and the obligation of researchers to respect the participant’s

3 ­ Lessons learned from population-based imaging

autonomy, is a potentially powerful argument for providing research results to those who want them, even when the results relate to risk for disease that is not preventable or treatable. In this sense it makes no difference whether a treatment exists or the participant is in a social or economic situation that would prevent them from benefiting from a treatment. Being aware that one suffers from a life-threatening and untreatable disease might enable a person to put their personal affairs in order and in ideal cases might even allow them to adopt a lifestyle and frame of mind that will allow them to live out their remaining life span in the way they choose.57, 58 Many scholars have also suggested that the disclosure of incidental or secondary findings may increase societal trust in research.52 In deciding whether incidental or secondary findings should be disclosed to participants, it is essential to consider the view of various stakeholders. In the setting of incidental findings from imaging, the views of participants have been examined and widely discussed in the literature (e.g., Refs. 3, 35). However, the perspectives of other stakeholders in this context have been examined less exhaustively. In at least one study, Cole et al. provided a consideration of the perspectives of investigators, IRB members, and physicians, among others. These results show that stakeholders feel a moral obligation to report incidental findings from imaging research to participants, but on the whole they are more circumspect than participants about the potential benefits of disclosure, and they tend to keep the disadvantages of disclosure in perspective. Specifically, they focus on the waste of time and other resources, the worthlessness of certain information, the potential psychological burdens (including unnecessary worry) that can be created, and also the detrimental effects that an unrestricted disclosure of incidental findings would pose, both to the healthcare system in general and to specific research efforts.52 Another issue relevant to the ethics of disclosing results is the effect of this practice on the research itself. If, in fact, disclosing incidental or secondary findings to research participants allows them to seek effective prevention or treatment earlier than individuals who have not participated in the research, then this may itself introduce bias into research. This effect could be most prominent in longterm cohort research, where the long-term outcomes of disclosure of secondary results could cause research participants to do better than the general population, thus leading to scientific conclusions that are less generalizable to the general population.59 Finally, strategies for managing uncertainty may also influence decisions about whether to disclose incidental or secondary findings to participants. In conditions where the relevance of a research finding might be uncertain, researchers might choose to disclose it on the grounds that it is “better to be safe than sorry.” This is an especially attractive approach in contexts where researchers are concerned that they may have legal liability if they choose not to disclose results that, in hindsight, might have prevented an adverse outcome. As the Presidential Commission for the Study of Bioethical Issues correctly pointed out in their report on the management of incidental and secondary findings in research, this strategy for managing uncertainty might be counterproductive.40 The report correctly notes that the diagnostic tests or

9

10

CHAPTER 1  Concept, history, and state of debate

procedures that are needed to confirm and flesh out a finding might lead to new— sometimes even life-threatening—risks or adverse psychological effects, including the risk of generating even more incidental findings. Just as a suspicious finding might help improve a patient’s health or even save his or her life, it might instead lead to unnecessary and harmful tests and unexpected costs.40

3.3 ­What to tell? Actionability and clinical utility In the preceding section we considered a variety of risks and benefits that could result from disclosing additional findings to participants or their healthcare providers. From that brief discussion we can begin to enumerate the most important and relevant risks associated with disclosing or withholding additional findings: (a) the risk that a patient might not receive relevant health information if research findings are withheld, (b) the risk that research subjects might be burdened by unnecessary, costly, or even risky procedures to clarify the relevance of a finding, (c) the risk that research participants might experience psychological distress and/ or financial disadvantages as a result of either disclosure strategies that are either too restrictive or too lenient. Given these potential risks, we can see that, whenever possible, additional findings should only be disclosed in situations where they are likely to provide benefits to the participant. Although this conclusion sounds trivial, the operationalization of this basic advice is a highly complex issue. This is demonstrated clearly in the ongoing debate about the management of incidental findings from imaging. Due to the complexity of this issue, recommendations and guidelines on this issue are often quite vague. A recommendation that is used widely in the United Kingdom, for example, points out that, at the moment, it is not possible to formulate one single optimal strategy. It recommends that study centers should continue to review their practices as circumstances and available resources change and that a range of strategies are acceptable as long as they exceed a specified minimum standard.56 Similarly the report of the Presidential Commission for the Study of Bioethical Issues in the United States allows room for a wide range of practices to be adopted and gives the very general advice that patients and participants should be informed “about the plan for disclosing and managing incidental and secondary findings, including what findings will and will not be returned.”40 The report does not provide concrete decision aids or criteria that could be helpful deciding if additional findings should be disclosed at all. Instead the report points to a need for further investigations: Professional representative groups should develop guidelines that categorize the findings likely to arise from each diagnostic modality; develop best practices for managing incidental and secondary findings; and share these guidelines among practitioners in the clinical, research, and direct-to-consumer contexts. Ref. 40

3 ­ Lessons learned from population-based imaging

Although the recommendation makes this type of work sound relatively straightforward, it is in fact extraordinarily difficult to definitively identify and categorize the findings that are likely to arise using different diagnostic modalities, much less obtain consensus from a variety of stakeholders on highly specific best practices. Instead, we are left with a range of both conflicting and overlapping guidance that provides frameworks for thinking about the disclosure of incidental or secondary findings. Although a comprehensive discussion of this literature is probably not necessary, it will be helpful to briefly consider the different algorithms that might be used for determining which results can and should be returned. In this discussion we will focus on the context of imaging; Chapter 3 provides an in-depth discussion of these frameworks in the context of genomics. The first “algorithm” for managing research findings was to simply not disclose any research findings. This strategy does carry certain benefits, including being the most effective way to avoid the diagnostic misconception. However, this strategy is not widely used in the context of imaging. One common reason for this is that legal restrictions in some jurisdictions preclude this practice. In Germany, for example, the criminal code specifies that there is an obligation to render necessary and reasonable assistance in case of emergencies §323c StGB (Strafgesetzbuch).60 As a result, if an incidental finding is generated that obviously requires immediate treatment, the researcher would be obliged to initiate appropriate steps regardless of whether they are a healthcare provider. The only way to adopt a complete nondisclosure policy under this type of legal framework is to quarantine imaging results long enough that any findings could no longer be considered emergent.28 As far as we know, however, this is only a theoretical solution that has never actually been implemented in research protocols. Another framework for specifying which findings might be disclosed is the so-called positive list; this is the approach adopted in the German NaKo National Cohort study.1 In this approach, investigators generate a list of findings that could foreseeably be generated through the research and would also carry significance for participants’ health. This strategy is attractive because it is highly transparent to participants; it lets them know exactly the kinds of results they may receive. This is particularly important, because it helps avoid false expectations by the participant, including the impression that a disclosure that no incidental findings were found equates to a “clean bill of health.” This approach is also useful because it helps constrain the total number of incidental findings that might be generated, thus conserving the study’s work force and budget. A critical challenge of this approach is that findings may be uncovered that were not anticipated, were not listed on the “positive list,” but nonetheless carry significance and relevance for participants. When such findings are identified, the investigators are still faced with a difficult decision about whether to circumvent their own positive list. This challenge can be addressed by the creation of an advisory board that is tasked with continually discussing new cases and updating, that is, expanding the list based on its decisions. While this is probably an indispensable part of the positive list approach in most circumstances, it does decrease the transparency to participants, since the list they were provided at the time of their enrollment might change over time.

11

12

CHAPTER 1  Concept, history, and state of debate

The extreme version of the positive list is to offer to return any result of potential relevance to participants. As we have discussed earlier, however, defining which results are relevant in which circumstances and which results are of sufficient quality given the context is a remarkably complex problem. It can therefore be helpful to substitute specific lists of conditions with a clear framework that specifies how decisions will be made about disclosing results. One such framework, developed by Susan Wolf and colleagues and included in the UK report discussed earlier, is intended to guide professionals in their work to appraise the relevance and significance of results and decide in which circumstances they should be returned.22 This framework is suitable for both imaging and genetics contexts (Table 1). Although this is a very helpful framework, it does leave several questions unanswered. What happens, for example, if a participant chooses not to receive incidental findings, but the research imaging reveals urgent or potentially life-threatening information? This situation creates important questions about whether researchers have a “duty to communicate” that could supersede the stated wishes of the participant, on the grounds that the information could benefit the patient or others, such as their family members.61 One way to avoid conflicts of this kind is to exclude individuals who decline incidental findings from participating in research likely to generate these findings. Although an in-depth discussion of this issue is beyond the scope of this chapter, it is worth noting that this policy would prevent individuals who want to contribute to research for altruistic reasons from making this contribution. Not only this seems punitive to potential participants, but also it could affect the research itself by complicating recruitment and introducing bias.

Table 1  Recommended classification of IFs according to Wolf et al.22 Class of benefit Strong net benefit

Possible net benefit

Unlikely net benefit

Type of additional finding

Recommended action

• Additional finding indicating a condition likely to be life-threatening • Additional finding indicating a condition likely to be grave that can be avoided or ameliorated • Additional finding indicating a nonfatal condition that is likely to be grave or serious but that cannot be avoided or ameliorated • Additional finding indicating a condition that is not likely to be of serious health or reproductive importance • Additional finding whose health or reproductive importance cannot be ascertained

Disclose to research participant, unless she/he elected not to know

May disclose to research participants, unless she/he elected not to know Do not disclose to research participants

3 ­ Lessons learned from population-based imaging

3.4 ­How to tell? A great deal of work has been done in recent years to identify the best ways to return genomic secondary findings to research participants. However, this remains a largely unexplored area in the context of incidental findings generated through imaging research. In truth, this is a significant oversight, since there is evidence that inappropriate disclosure methods represent one of the most important factors in creating stress for participants who receive incidental findings.3, 62 In considering how best to disclose incidental findings and secondary findings to research participants, it will be helpful to consider three specific questions: (a) Who should be responsible for the disclosure of additional findings in the research context? (b) How should additional findings be disclosed? (c) Who should receive additional findings as a part of the disclosure process? With respect to the first question, there are three groups of people who might play a part in disclosing additional findings of the research context: scientists, study staff, or healthcare providers. In clinical research the healthcare providers involved in a study might also be involved in providing clinical care to a patient/participant. In some cases, however, there are no healthcare providers directly involved in a research study, or they may have no direct contact with a participant. In this case a participant’s primary care provider, or another clinician with whom they have a relationship, might be pursued as an appropriate person to disclose results to a participant. In this case the question of who should receive the findings could become a two-step process, with disclosure first to a clinician not associated with the study and then to the participant or others. For obvious reasons, there are many who believe that only healthcare providers have the proper training to return incidental or secondary findings to a research participant. However, it is worth considering whether nonclinician scientists or research staff might provide some advantages. First, for results that are unfamiliar to clinicians, the researcher conducting a study is likely to be the “local expert” on a particular health issues or a particular testing technology. Second, keeping the responsibility for disclosing results within the nonclinician study staff might help prevent the diagnostic and therapeutic misconceptions. Third, this approach might help constrain the costs of disclosing results. In the case of genomic secondary findings, genetic counselors are frequently responsible for disclosing results to participants. Genetic counselors may be associated with the research team, or may practice exclusively as clinicians. Either way, genetic counselors’ knowledge of genetics and training in communication makes them well suited for the role of disclosing genomic secondary findings. Study nurses may also be utilized for the disclosure of incidental findings in imaging research or secondary findings in genetics research. This can be more economical than the use of physicians or genetic counselors, but in some jurisdictions this approach may not be possible due to laws or policies requiring that physicians perform this kind of role.

13

14

CHAPTER 1  Concept, history, and state of debate

With respect to the second question listed earlier, additional findings may be disclosed verbally through a telephone call or face-to-face conversation or in written form by letter or email. In many cases results are disclosed through a combination of verbal and written communication, such as an initial telephone call followed by a letter to reinforce the information that was discussed. While a telephone call can be an easy and cost-effective method for disclosing additional findings, it does have several disadvantages. The most important of these is that the participant receiving a finding may become overwhelmed or distressed as a result of the disclosure, but the inability to see nonverbal cues may prevent the person disclosing the finding to recognize this problem. The challenge is even more problematic in the context of disclosure via letters or email. Nevertheless, some studies make use of this option because it is cost-effective and because it allows researchers to choose language carefully and present information in a visual format, both of which may help reduce stress. While face-to-face disclosure is the best option with respect to recognizing and addressing distress, it is more expensive and requires more effort on the part of the research team in comparison with the other options. Nevertheless, from an ethical standpoint, face-to-face disclosure represents the gold standard for the disclosure of additional findings. This is particularly effective when the person disclosing findings utilizes evidence-based strategies for breaking bad news, such as those used in oncology.63 The third question, addressing the issue of who should receive the additional finding, is not as straightforward as it may initially seem. Some studies do not disclose additional findings directly to participants, but instead to the participant’s healthcare provider who is then tasked with communicating with the participant. This indirect method for communicating results to a participant can help address questions of quality, like whether a finding is actionable for a particular patient, but it requires that the researchers identify the correct provider with confidence. Healthcare providers may also object to being asked to bear the burden of disclosing findings which they did not order or may not feel confident in interpreting.64 For a more detailed analysis of the normative and practical advantages and disadvantages of the various options for disclosing findings, see Table 2. One could argue that some of the more elaborate strategies for disclosing incidental findings are unnecessary and that unverified findings can be disclosed in a letter as preliminary, inconclusive findings that merely indicate a need for conventional testing. For most research participants, however, this distinction between an unverified research result and a standard of care result will not be evident, especially when they have not had an opportunity to ask questions. Even attempts to avoid anxiety-­ inducing words like “tumor” or “aneurysm” are insufficient here. Participants who receive results in a written form and therefore cannot benefit from asking questions of a suitable provider will often resort to consulting the internet for answers. Although anxiety-inducing words might have been avoided in the written report, they will certainly be encountered in online sources. The most dependable way to mitigate this fear and anxiety is to communicate results in a face-to-face conversation, which

3 ­ Lessons learned from population-based imaging

Table 2  How to disclose additional findings? Advantages and disadvantages of different approaches to disclosure of incidental findings to participants. “How to tell?” Person responsible for disclosing additional findings Scientists

Healthcare providers

Study staff

• Clinicians associated with the study • Clinicians not associated with the study, including the participants’ own primary care provider • Genetic counselors Study nurses

Technical staff

Administrative staff

Advantages • Best available scientific expertise • Expertise on how data were generated • Prevention of diagnostic/therapeutic misconception Disadvantages • Lack of medical expertise • Lack of professional communication skills Advantages • Medical expertise • Professional communication skills • Skills with techniques for breaking bad news Disadvantages • Relatively expensive • Risk of exacerbating diagnostic/ therapeutic misconception

Advantages • Medical expertise • Professional communication skills • Relatively inexpensive Disadvantages • Lack of physician-specific medical expertise • Risk of exacerbating diagnostic/ therapeutic misconception Advantages • Technical know-how: expertise on how data were generated • Prevention of diagnostic/therapeutic misconception • Relatively inexpensive resource Disadvantages • Lack of medical expertise • Lack of professional communication skills Advantages • Prevention of diagnostic/therapeutic misconception • Relatively inexpensive resource Disadvantages • Lack of medical expertise • Lack of professional communication skills Continued

15

16

CHAPTER 1  Concept, history, and state of debate

Table 2  How to disclose additional findings? Advantages and disadvantages of different approaches to disclosure of incidental findings to participants—cont'd “How to tell?” How results are disclosed Verbal

Phone call

Face-to-face conversation

Written form

Letter or email

Advantages • Fast communication • Allows for dialogue about finding and its implications • Inexpensive • Allows some opportunity to utilize techniques for breaking bad news Disadvantages • Reduces ability to identify and address distress • Reduces ability to identify and address misunderstandings Advantages • Allows for dialogue about finding and its implications • Excellent opportunity to identify and address distress • Excellent opportunity to identify and address misunderstandings • Excellent opportunity to utilize techniques for breaking bad news Disadvantages • Difficult and time-consuming to arrange • High administrative burden • Relatively expensive Advantages • Inexpensive • Enables precise and accurate description of findings • Enables careful choice of words to avoid stress • Enables use of visual aids Disadvantages • Precludes dialogue about finding and its implications • Reduces ability to identify and address distress • Reduces ability to identify and address misunderstandings

3 ­ Lessons learned from population-based imaging

Table 2  How to disclose additional findings? Advantages and disadvantages of different approaches to disclosure of incidental findings to participants—cont'd “How to tell?” Person receiving additional finding Participant’s physician

Participant

Advantages • Medical expertise to (more) accurately interpret findings • Background knowledge regarding the participant (comorbidities, personal situation, living conditions, etc.) • Professional communication skills • Access to further resources to clarify research findings Disadvantages • Reliable contact data often unknown by the study • Lack of expertise regarding experimental methods/technologies • Data protection issues raised by transfer of participants’ data to a third party • Patients may object to physician serving as gatekeeper to information Advantages • Person for whom the information is most relevant • Contact data more likely to be known to the study • Reduced data protection issues Disadvantages • Lack of medical expertise • Increased risk of diagnostic/therapeutic misconception • Increased risk for misunderstandings • Highly vulnerable regarding distress

provides the opportunity to explain, for example, the possibility that a result might representing a false-positive finding or that the finding requires confirmation with a clinically validated test. While researchers understandably need to account for challenges like the need to return a large volume of results or difficulty with reaching some participants to discuss results directly, the first priority should always be to minimize foreseeable stress. One approach to balancing logistical challenges with ethical obligations is to adjust the strategy for disclosure based on how much distress a result is anticipated to create. This adaptive approach is utilized widely in clinical care settings. In some research and direct-to-consumer testing contexts, however, findings ­indicating a

17

18

CHAPTER 1  Concept, history, and state of debate

life-threatening condition have been disclosed exclusively in writing. A common justification for this strategy is that the respective findings are just provisional and eventually turn out to be false positive. For the person receiving such a finding, however, the difference between a diagnosis and a suspicion might be not perceivable in that moment.3 Face-to-face communication at least offers the chance to carefully explain this distinction and undertake other efforts to mitigate fear and anxiety. Appropriately trained providers might also be able to offer options and next steps within a face-to-face communication. This could reduce the feelings of hopelessness and helplessness that participants experience. Some recommendations for imaging centers have gone one step further, recommending that unverified findings should not be disclosed until an action plan has been developed and appropriate specialists have been identified to help with providing advice and suggesting a path forward. Without these types of strategies, participants are left to seek clarification of the finding on their own, an odyssey that can last weeks or months without appropriate guidance. Patients and participants often experience this time of ambiguity as even more stressful than the diagnosis of a severe disease itself, since coping processes are somewhat suspended.3

3.5 ­How to prepare participants? It is critical that potential research participants have an opportunity to weigh the pros and cons of receiving additional findings as a part of their decision to participate (or not participate) in research. This means that the informed consent process needs to include detailed information about a number of factors: • • • •

examples of the types of findings that might be revealed, the likelihood that incidental findings will be detected, the risk that false-positive and false-negative results will be disclosed, negative consequences that might result from the disclosure (e.g., possible effects on employment or insurance coverage), • the possibility that the process of clarifying a finding could itself create risks and psychological distress, • the options available for declining certain findings, including by declining research participation. When research is being conducted in clinical settings, it is also important to help potential research participants understand which procedures are being carried as a part of clinical care and which are being conducted for research purposes only. This can help avoid misunderstandings that can arise from the therapeutic or diagnostic misconceptions. These efforts to inform potential research participants are supported by the ethical principles of respect for persons and beneficence and the principles of justice and fairness. The report produced by the Presidential Commission for the Study of Bioethical Issues40 emphasizes the responsibility of the person providing information on additional findings to provide guidance and support so that the patient,

4 ­ In search of an appropriate ethical framework

r­esearch participant, or direct-to-consumer client can make informed choices (see, e.g., recommendations 5, 7, and 15 in this report). In most cases, effective communication about incidental or secondary findings will require more than just verbal communication. So information should typically include decision aids and graphical or multimedia representations. In certain cases, effective information might also mean explicitly informing opting out of certain examinations tests, or electing not to receive certain results, might be the more beneficial decision for an individual.

3.6 ­How to design an appropriate research protocol The Presidential Commission, among other authorities, have recommended that study protocols should define which additional findings are foreseeable and specify a detailed plan for handling them. In cases where subsequent tests would be necessary to clarify findings, the plan should provide information about how these tests will be paid for (i.e., by the study or through health insurance). The Presidential Commission also recommends that the plan for managing additional findings should be submitted to an IRB for review and approval.40 Another important issue related to planning a study is to consider carefully which testing modalities will be utilized in a study. In general the choice of testing modalities will be driven by the hypothesis or purpose of the research. For example, a whole-body MRI might be chosen for a population-based cohort study, while an MRI limited to the brain might be utilized for a research focused more narrowly on the nervous system. Although the purpose of the study is the primary consideration, investigators also need to consider the potential effects these decisions will have on the generation of additional findings. It is anticipated that the rate of false-positive and false-negative findings will depend on the testing modalities selected for a study. There is thus controversy about how researchers should weigh this risk of falsepositive and false-negative results in deciding which testing modality will be used for a study.56

4 ­In search of an appropriate ethical framework As we have demonstrated, the debate over incidental findings has been underway for a number of years, and during this time a number of guidelines have been published. In many cases these guidelines state that consensus could not be reached among those developing the guidance. When internal consensus is reached and definitive recommendations are made, these sometimes conflict with guidance provided in other guidelines. One reason consensus has not been achieved, perhaps, is that these discussions lack an organizing ethical framework. When stakeholders are engaging in debate, it can be helpful to identify a framework that everyone can agree upon. This allows for stakeholders who disagree on specific policies to make convincing arguments to one another, since they have already agreed on guiding principles.

19

20

CHAPTER 1  Concept, history, and state of debate

Important work has been conducted, however, to identify ethical frameworks that might serve this purpose. One example is the SHIP study, a population-based cohort study conducted in the state of Mecklenburg-Vorpommern.65 This study involved the collection of whole-body MRI from a sample representing the population. In this context the preliminary results from whole-body MRI resulted in the generation of incidental findings in about 30% of participants.3, 4 Empirical work with participants who received whole-body MRI through this study informed conceptual work conducted by Erdmann and Langanke to devise a framework of ethical principles well suited to this context.3, 66, 67 This framework proposed a moderate contractualism as an appropriate basis for the management of incidental findings. Contractualism refers to the idea that agreements between parties need to be specified prospectively, and once those involved have agreed to the terms, the contract needs to be followed. Moderate contractualism in this context simply means that all contracts need to have some degree of flexibility to allow those involved to do the right thing. However, in general, it holds that the terms of a research study agreed to through the informed consent process are binding for the investigators. This framework also proposed that decisions about the management of incidental findings should be based on the principle of fairness in the sense that decisions should be made in a transparent way and in the sense that investigators obligated to minimize foreseeable risks and harms to participants.3, 66, 67 The Presidential Commission report40 also set out ethical principles that were selected to underpin their recommendations. These principles were intended to provide the recommendations with a degree of universality that would allow them to carry weight in the settings of medical care, research, and direct-to-consumer testing contexts. The Commission admitted, however, that the principles might need to be applied in different ways in these three contexts.40 The principle proposed by the Presidential Commission included respect for persons, beneficence, justice and fairness, and intellectual freedom and responsibility.40 These principles were derived from two classic resources: Principles of Biomedical Ethics by Tom Beauchamp and James Childress68 and the so-called Belmont Report.69 In the Belmont Report the principle of respect for persons is divided into two subissues. The first specifies that the autonomy of potential research participants should be respected by ensuring that they have the opportunity to receive information about research participants and make a voluntary decision to participate or not participate. The second specifies that researchers are obligated to take measures to protect vulnerable individuals whose autonomy is limited, such as those with cognitive impairment. The Presidential Commission, building on this framework, concludes that respect for persons also implies the principles of prudence and deliberateness, since these principles are needed to ensure that participation decisions are appropriately informed and are undertaken voluntarily.40 The Presidential Commission also emphasizes beneficence, which holds that professionals should take measures to ensure the well-being of others. Nonmaleficence, often considered a corollary of beneficence, is interpreted as an obligation to avoid imposing harm on others. Nonmaleficence is relatively uncontroversial, and there

4 ­ In search of an appropriate ethical framework

appears to be broad consent that professionals working in research, medical care, and direct-to-consumer settings are obligated to take measures to avoid causing harm to research participants, patients, and consumers. The proposed obligation to maximize benefits is more controversial, however, except perhaps in the context of clinical care.40 Although most would agree that professionals should avoiding causing harm to research participants, patients, and consumers, this principle can prove surprisingly difficult to implement in practice, especially in research settings. To avoid causing harm, it becomes necessary to dedicate resources such as time and money to this effort. Even though it is foreseeable that in-person disclosure of results will minimize harms to participants, this approach is the most expensive (although it is likely that many overestimate the associated costs). The need to dedicate resources to avoid causing harm therefore becomes a question of both what is appropriate and what is reasonable given the costs. Since stakeholders might evaluate both of these questions differently, identifying measures to avoid harm creates some need to engage in deliberation with stakeholders about what should be considered appropriate and reasonable. Another complication related to the principle of beneficence is the extent to which benefits to the public should be considered. In discussing this issue the Presidential Commission refers to their first report New Directions: The Ethics of Synthetic Biology and Emerging Technologies in which benefit to society, including weighing costs with benefits, was considered an inherent dimension of the principle of beneficence.40 It is perhaps not surprising that the ethical obligations of researchers should include an obligation to conduct the best possible research practice so that society may benefit. It is also clear that professionals working in the healthcare system need to weigh disproportionate costs and burdens when making policy decisions. However, it is not clear whether direct-to-consumer testing companies carry an obligation to benefit society. Nonetheless, although benefit to society is important in at least two of these domains, individual beneficence should never be disregarded. When personal risks cannot be avoided because, for example, the broader costs to society would be too great, participants have to be informed of these risks to ensure that their consent to proceed is valid. The third principle advanced by the Presidential Commission is the principle of justice and fairness, which demands fair and equitable treatment of all.40 In the Presidential Commission’s discussion of justice and fairness, it is clear that justice and fairness are linked with beneficence; not only should benefits be maximized and harms minimized, but also the distribution of these benefits and harms should be spread evenly among affected parties. The principle also suggests that resources should be allocated equitably and ethically similar cases should be treated similarly.40 The Presidential Commission points out that justice and fairness imply that research and clinical care should be undertaken in such a way that it maximizes the number of persons whose basic healthcare needs are met. This undoubtedly reflects the origins of this report in the United States, where access to affordable healthcare services is not guaranteed for all. Taken as a whole the application of justice and fairness to

21

22

CHAPTER 1  Concept, history, and state of debate

the context of incidental and secondary findings implies that overtesting should be avoided since its costs restrict the resources available to others (and is thus unfair).40 The fourth and final principle proposed by the Presidential Commission recommends the principle of intellectual freedom and responsibility. This principle, which has no equivalent in either the principles of Beauchamp and Childress 68 or the Belmont report,69 represents the interests of researchers and scientific progress. This principle is based on the observation that professionals are in a unique position to shape practice and are thus obligated to use this authority appropriately. The Commission emphasizes that intellectual freedom and responsibility support the rejection of the “technological imperative.”40 The technological imperative refers to the tendency for technological tools to be used simply because they have become available. But the principle of intellectual freedom and responsibility constrains this phenomenon by emphasizing that not everything that is technically possible is also useful. The technological imperative thus creates a risk that limited resources will be wasted on the generation and follow-up of results that benefit neither the individuals receiving them nor society as a whole. As we noted earlier, one benefit of agreeing on ethical principles is that when everyone agrees on the underlying principles, it is much easier to identify appropriate policies and practices and convince others that they should be followed. Implicit in this observation is that recommendations based on ethical principles will not motivate action unless stakeholders agree on the principles. Recognizing this the Presidential Commission likely evoked the well-established principles of Beauchamp and Childress68 and the Belmont Report69 because they are already seen as generally accepted. The fourth principle of intellectual freedom and responsibility, however, was likely added to more well-established principles to address the specific challenges professionals encounter in managing incidental findings. Erdmann and Langanke took a similar approach in developing their ethical framework.3, 66, 67 Empirical research conducted by Erdmann3 revealed that some ethical issues believed to be important turned out to create fewer challenges than anticipated. On the other hand a number of unanticipated issues arose that proved particularly challenging, especially the mode of disclosing incidental findings to participants. The moderate contractualism of Erdman and Langanke, therefore, was intentionally developed to address both theoretical considerations and empirical evidence from real-world situations. The point here is simply that different procedures for developing an ethical framework may result in the development of different frameworks. That is not to say, however, that the framework proposed by the Presidential Commission is incompatible with the one developed by Erdmann and Langanke. Moderate contractualism can be understood as an application of the principles of both autonomy (in that research, participants should have the opportunity to choose whether they wish to enter into a contract) and justice (in that, breaking a contract is unjust). The principle of fairness captures another dimension of justice. The imperative for transparency can be seen as an expression of the principle of respect for persons. Finally the obligation to minimize stress emphasizes a particular dimension of the principle of nonmaleficence.3, 66, 67

5 ­ Conclusion

Although these systems largely overlap, the value of the Erdmann-Langanke framework is that it emphasizes dimensions that, based on empirical research, create the greatest challenges.

5 ­Conclusion Despite the initial tendency for genetic and genomics researchers to borrow the term incidental findings, time has proven that purely incidental findings are relatively uncommon in this field. However, secondary findings—those additional findings that are created intentionally—create a number of normative and practical challenges that are similar to those created by incidental findings in imaging (and in particular neuroimaging) research. In this chapter, then, we have examined the history of incidental findings in research imaging as a way to lay the groundwork for the remainder of this book, which will focus exclusively on the context of genetics and genomics. In this chapter, not only we presented a model that allows for a deeper understanding of the history of both areas of discourse, but also we highlighted the lessons learned through the debate on incidental findings in research imaging that can be applied to the management of secondary findings in genetics and genomics. In particular, we have shown that there are six areas where research imaging has laid the groundwork for developing policies and practices in biomedical studies, including genetics and genomics, where there is a chance that additional research findings potentially relevant to research participants will be generated. First, it is critical that when additional findings might be disclosed to participants or their providers, researchers must adopt practices and procedures that emphasize quality. This includes not only the quality of the technical processes used to produce findings but also the quality of the communication processes. If available, information about the participant, such as medical history or preferences for receiving additional findings, can significantly improve the quality of information. Second, researchers must decide whether or not to disclose findings to participants. In most circumstances a rigid nondisclosure policy will not work well, given that most participants wish to receive results and many researchers would feel uncomfortable not returning findings to participants if they might have an impact on health. However, there are a variety of other strategies for determining which results will be returned. Third, researchers must decide which types of results to disclose. In most cases, researchers should determine prospectively which results will be returned based on the anticipated clinical utility/actionability of findings. This strategy can and probably should be used in both imaging and genetics/genomics research. Fourth, it is critical to determine how additional findings will be disclosed to participants. This includes who will actually disclose results to participants and how other resources like multimedia information will be utilized. The disclosure of results by a physician (or a genetic counselor, in the case of genetics and genomics research) via a face-to-face conversation is the current gold standard for disclosing additional findings from research.

23

24

CHAPTER 1  Concept, history, and state of debate

Fifth, measures must be taken to ensure that participants’ decisions to receive additional findings are informed by ensuring they have the opportunity to consider relevant information. This information should include (a) examples of the types of findings that might be revealed, (b) the likelihood that incidental findings will be detected, (c) the risk that additional findings will be disclosed, (d) the basic algorithms according to which additional findings will be handled, (e) the risk that false-positive and false-negative results will be generated, (f) negative consequences that might result from the disclosure (e.g., possible effects on employment or insurance coverage), (g) the possibility that the process of clarifying a finding could itself create risks and psychological distress, (h) the options available for declining certain findings, including by declining research participation. All of those are important for ensuring that the diagnostic and therapeutic misconceptions are minimized, but it may also be useful to discuss the distinction between research and clinical care explicitly. All of these considerations should be included in the information consent process to the extent possible. Sixth a protocol should be produced reflecting all of the practical and ethical issues mentioned earlier. This protocol should be summited to a research ethics committee or IRB for review and approval. Overall the management of incidental findings in imaging research and secondary findings in genetics and genomics research needs to be approached carefully, with a plan that is ethical, responsible, and feasible. This outcome is most likely to be accomplished if the approach is informed by widely accepted ethical principles including (a) respect for persons, (b) beneficence, (c) nonmaleficence, (d) justice and fairness, and (e) intellectual freedom and responsibility. Two additional principles more specific to research ethics should also be considered: (1) transparency and (2) minimizing the distress experienced by participants.

­References 1. Hegedüs P, von Stackelberg O, Neumann C, et al. How to report incidental findings from population whole-body MRI: view of participants of the German National Cohort. Eur Radiol. 2019; https://doi.org/10.1007/s00330-019-06077-z. [Epub ahead of print]. 2. Erdmann P. Incidental findings – ethical aspects. In: Weckbach S, ed. Incidental radiological findings. London: Springer; 2017:9–24. 3. Erdmann P. Zufallsbefunde aus bildgebenden Verfahren in populationsbasierter Forschung – Eine empirisch-ethsiche Untersuchung. Münster: Mentis; 2015. 4. Schmidt CO, Hegenscheid K, Erdmann P, et al. Psychosocial consequences and severity of disclosed incidental findings from whole-body MRI scanning in a general population study. Eur Radiol. 2013;23:1343–1351.

­References

5. Illes  J, Desmond  JE, Huang  LF, et  al. Ethical and practical considerations in managing incidental findings in functional magnetic resonance imaging. Brain Cogn. 2002;50:358–365. 6. Illes J, Rosen AC, Huang LF, et al. Ethical consideration of incidental findings on adult brain MRI in research. Neurology. 2004;62:888–890. 7. Illes  J, Kirschen  MP, Edwards  E, et  al. Incidental findings in brain imaging research. Science. 2006;311:783–784. 8. Check E. Brain scan ethics come under the spotlight. Nature. 2005;433:185. 9. Kirschen  MP, Jaworska  A, Illes  J. Subjects’ expectations in neuroimaging research. J Magn Reson Imaging. 2006;23:205–209. 10. Hentschel F, Klix W-E. Management inzidenter Befunde in der bildgebenden Diagnostik und Forschung. Fortschr Neurol Psychiat. 2006;74:651–655. 11. Kumra S, Ashtari M, Anderson B, Cervellione KL, Kan L. Ethical and practical considerations in the management of incidental findings in pediatric MRI studies. J Am Acad Child Adolesc Psychiatry. 2006;45:1000–1006. 12. Vernooij MW, Ikram MA, Tanghe HL, et al. Incidental findings on brain MRI in the general population. N Engl J Med. 2007;357:1821–1828. 13. Heinemann T, Hoppe C, Listl S, Spickhoff A, Elger CE. Zufallsbefunde bei bildgebenden Verfahren in der Hirnforschung. Ethische Überlegungen und Lösungsvorschläge. Deutsches Ärzteblatt. 2007;104(27, Internetsupplement 1):A-1982–A-1987. 14. Nelson CA. Incidental findings in magnetic resonance imaging (MRI) brain research. J Law Med Ethics. 2008;36:315–319. 15. Heinemann T, Hoppe C, Weber B, et al. Ethisch angemessener Umgang mit Zufallsbefunden bei bildgebenden Verfahren in der Hirnforschung. Leitlinienvorschlag. http://www.ewipsy.fu-berlin.de/einrichtungen/kommissionen/ethik-kommission/Leitlinienvorschlag_ Zufallsfunde.pdf?1338976880; 2009. (Accessed July 2, 2013). 16. Hoggard N, Darwent G, Capener D, Wilkinson ID, Griffiths PD. The high incidence and bioethics of finding on magnetic resonance brain imaging of normal volunteers for neuroscience research. J Med Ethics. 2010;35:194–199. 17. Chow A, Drummond KJ. Ethical considerations for normal control subjects in MRI research. J Clin Neurosci. 2010;17:1111–1113. 18. Borra RJ, Sorensen AG. Incidental findings in brain MRI research: what do we owe our subjects? J Am Coll Radiol. 2011;8:848–852. 19. Shoemaker JM, Holdsworth MT, Aine C, et al. A practical approach to incidental findings in neuroimaging research. Neurology. 2011;77:2123–2127. 20. Underwood E. When a brain scan bears bad news. Science. 2012;338:455. 21. Wolf SM. The challenge of incidental findings. J Law Med Ethics. 2008;36:216–218. 22. Wolf SM, Lawrenz FP, Nelson CA, et al. Managing incidental findings in human subject research: analysis and recommendations. J Law Med Ethics. 2008;36:219–248. 23. Royal JM, Peterson BS. The risks and benefits of searching for incidental findings in MRT research scans. J Law Med Ethics. 2008;36:305–314. 24. Siddiki H, Fletcher JG, McFarland B, et al. Incidental findings in CT colonography: literature review and survey of current research practice. J Law Med Ethics. 2008;36:320–331. 25. Woodward  CI, Toms  AP. Incidental findings in “normal” volunteers. Clin Radiol. 2009;64:951–953. 26. Morin SHX, Cobbold JFL, Lim AKP, et al. Incidental findings in healthy control research subjects using whole-body MRI. Eur J Radiol. 2009;72:529–533.

25

26

CHAPTER 1  Concept, history, and state of debate

27. Hoffmann  M, Schmücker  R. Die ethische Problematik der Zufallsbefunde. In: Puls  R, Hosten  N, eds. Ganzkörper-MRT-Screening. Befunde und Zufallsbefunde. Berlin: ABW Wissenschaftsverlag; 2010:1–16. 28. Puls R, Hamm B, Hosten N. MRT ohne Radiologen – ethische Aspekte bei bevölkerungsbasierten Studien mit MR-Untersuchung. Fortschr Röntgenstr. 2010;182:469–471. 29. Knoppers BM, Joly Y, Simard J, Durocher F. The emergence of an ethical duty to disclose genetic research results: international perspectives. Eur J Hum Genet. 2006;14:1170–1178. 30. Ravitsky V, Wilfond BS. Disclosing individual genetic results to research participants. Am J Bioeth. 2006;6:8–17. 31. Cho MK. Understanding incidental findings in the context of genetics and genomics. J Law Med Ethics. 2008;36:280–285. 32. Berg JS, Khoury MJ, Evans JP. Deploying whole genome sequencing in clinical practice and public health. Meeting the challenge one bin at a time. Genet Med. 2011;13:499–504. 33. Bredenoord  AL, Onland-Moret  NC, van Delden  JJM. Feedback of individual genetic results to research participants: in favor of a qualified disclosure policy. Hum Mutat. 2011;32:861–867. 34. Tabor HK, Berkman BE, Hull SC, et al. Genomics really gets personal: how exome and whole genome sequencing challenge the ethical framework of human genetics research. Am J Med Genet A. 2011;155:2916–2924. 35. Christenhusz GB, Devriendt K, Dierichs K. To tell or not to tell? A systematic review of ethical reflections on incidental findings arising in genetic contexts. Eur J Hum Genet. 2013;21:248–255. 36. Brothers KB, Langanke M, Erdmann P. The implications of the incidentalome for clinical pharmacogenomics. In: Pharmacogenomics. 14. 2013:1353–1362. 37. Rudnik-Schöneborn  S, Langanke  M, Erdmann  P, Robienski  J. Ethische und rechtliche Aspekte im Umgang mit genetischen Zufallsbefunden, Herausforderungen und Lösungsansätze. Ethik Med. 2014;26:105–119. 38. Langanke M, Erdmann P, Robienski J, Rudnik-Schöneborn S, eds. Zufallsbefunde bei molekulargenetischen Untersuchungen—Ethische, rechtliche und medizinische Perspektiven. Berlin/Heidelberg: Springer; 2015. 39. Burke W, Antommaria AH, Bennet R, et al. Recommendations for returning genomic incidental findings? We need to talk!. Genet Med. 2013;15:855–859. 40. Presidential Commission for the Study of Bioethical Issues. Anticipate and Communicate. Ethical Management of Incidental and Secondary Findings in the Clinical, Research, and Direct-to-Consumer-Contexts. http://bioethics.gov/node/3183; 2013. (Accessed January 12, 2016). 41. Deutscher Ethikrat. Die Zukunft der genetischen Diagnostik. https://www.ethikrat.org/fileadmin/Publikationen/Stellungnahmen/deutsch/stellungnahme-zukunft-der-genetischendiagnostik.pdf; 2013. (Accessed March 22, 2019). 42. Green RC, Berg JS, Grody WW, et al. ACMG recommendations for reporting incidental findings in clinical exome and genome sequencing. Genet Med. 2013;15(7):565–574. 43. Hall A, Hallowell N, Zimmern R. Managing incidental and pertinent findings from WGS in the 100,000 Genome Project. PHG Foundation, 2013. ISBN: 978-1-907198-12-0. 44. German Society of Human Genetics (GfH). Stellungnahme der Deutschen Gesellschaft für Humangenetik zu genetischen Zusatzbefunden in Diagnostik und Forschung. https://www. gfhev.de/de/leitlinien/LL_und_Stellungnahmen/2013_05_28_Stellungnahme_zu_genetischen_Zufallsbefunden.pdf; 2013. Accessed 22nd March 2019.

­References

45. National Health and Medical Research Council. Principles for the translation of ‘omics’based tests from discovery to health care. Canberra: National Health and Medical Research Council; 2015. 46. Project EURAT Ethical and Legal Aspects of Whole Human Genome Sequencing. Cornerstones for an ethically and legally informed practice of whole genome sequencing: code of conduct and patient consent models. Heidelberg: Marsilius-Kolleg der Universität Heidelberg; 2016. 47. National Academies of Sciences, Engineering, and Medicine. Returning individual research results to participants: guidance for a new research paradigm. Washington, DC: The National Academies Press; 2018.https://doi.org/10.17226/25094. 48. American College of Medical Genetics and Genomics. ACMG updates recommendation on “opt out” for genome sequencing return of results. https://www.acmg.net/docs/ Release_ACMGUpdatesRecommendations_final.pdf; 2014. (Accessed July 26, 2018). 49. Schuol S, Schickhardt C, Wiemann S, et al. So rare we need to hunt for them: reframing the ethical debate on incidental findings. Genome Med. 2015;7(1):83. 50. Berg JS, Foreman AK, O'Daniel JM, et al. A semi-quantitative metric for evaluating clinical actionability of incidental or secondary findings from genome-scale sequencing. Genet Med. 2016;18:467–475. 51. Bjugn  R. Research findings with clinical implications. Tidsskr Nor Legeforen. 2015; https://doi.org/10.4045/tidsskr.14.0861. 52. Cole  C, Petree  LE, Phillips  JP, Shoemaker  JM, Holdsworth  M, Helitzer  DL. ‘Ethical responsibility’ or ‘a whole can of worms’: differences in opinion on incidental finding review and disclosure in neuroimaging research from focus group discussion with participants, parents, IRB members, investigators, physicians and community members. J Med Ethics. 2015; https://doi.org/10.1136/medethics-2014-102552. 53. Brothers  KB. Wie konnte das passieren? Die US-amerikanische Perspektive auf Zufallsbefunde in den ACMG-Empfehlungen. In: Langanke M, Erdmann P, Robienski J, Rudnik-Schöneborn  S, eds. Zufallsbefunde bei molekulargenetischen Untersuchungen. Medizinische, juristische und ethische Perspektiven. Heidelberg: Springer; 2015:178–190. 54. Appelbaum  PS, Lidz  CW, Grisso  T. Therapeutic misconception in clinical research: frequency and risk factors. IRB Ethics Hum Res. 2004;26(2):1–8. https://doi. org/10.2307/3564231. 55. Hart MR, Biesecker BB, Blout CL, et al. Secondary findings from clinical genomic sequencing: prevalence, patient perspectives, family history assessment, and health-care costs from a multisite study. Genet Med. 2019;21(5):1100–1110. 56. The Royal College of Radiologists. Management of incidental findings detected during research imaging. A report by Representatives of Research Imaging Centres, Professional Societies, Regulatory Bodies, Funding Organisations, Royal Colleges involved in research imaging and Patient Organisations, in the UK https://www.rcr.ac.uk/sites/default/files/ publication/BFCR%2811%298_ethics.pdf; 2011. (Accessed January 13, 2016). 57. Charmaz K. Good days, bad days: the self and chronic illness in time. New Brunswick: Rutgers University Press; 1993. 58. Erdmann P, Langanke M. The ambivalence of early diagnosis – returning results in current Alzheimer research. Curr Alzheimer Res. 2018;15:28–37. 59. Hoffmann M. Zufallsbefunde in der epidemiologischen Forschung. In: Lenk C, Duttge G, Fangerau H, eds. Handbuch Ethik und Recht der Forschung am Menschen. Heidelberg: Springer; 2014:305–311. https://doi.org/10.1007/978-3-642-35099-3_52.

27

28

CHAPTER 1  Concept, history, and state of debate

60. Strafgesetzbuch in der Fassung der Bekanntmachung vom 13. November 1998 (BGBl. I S. 3322), das durch Artikel 15 des Gesetzes vom 4. Juli 2013 (BGBl. I S. 1981) geändert worden ist. http://www.gesetze-im-internet.de/stgb/BJNR001270871.html#BJNR001270 871BJNG000102307; (Accessed July 3, 2016). 61. Schmücker  R. Zufallsbefunde – was gebietet die Menschenwürde? http://www.unimuenster.de/imperia/md/content/kfg-normenbegruendung/intern/publikationen/schmuecker/36_-_schm__cker_-_zufallsbefunde_und_menschenw__rde.pdf; 2012. (Accessed July 3, 2016). 62. Levine PA. In an unspoken voice. How the body releases trauma and restores goodness. Berkley: North Atlantic Books/ERGOS Institute Press; 2010. 63. Baile  WF, Buckmann  R, Lenzi  R, Glober  G, Beale  EA, Kudelka  AP. SPIKES—a sixstep protocol for delivering bad news: application to the patient with cancer. Oncologist. 2000;5:302–311. 64. Pet  DB, Holm  IA, Williams  JL, et  al. Physicians' perspectives on receiving unsolicited genomic results. Genet Med. 2019;21(2):311–318. 65. Völzke H, Alte D, Schmidt CO, et al. Cohort profile: the study of health in Pomerania. Int J Epidemiol. 2011;40:294–307. 66. Langanke M, Erdmann P. Die MRT als wissenschaftliche Studienuntersuchung und das Problem der Mitteilung von Zufallsbefunden. Probandenethische Herausforderungen. In: Theißen H, Langanke M, eds. Tragfähige Rede von Gott. Festgabe für Heinrich Assel zum 50. Geburtstag am 9. 2011. Hamburg: Verlag Dr. Kovač; February 2011:197–240. 67. Langanke  M, Fasold  J, Erdmann  P, Lorbeer  R, Liedtke  W. Informed consent in Gani_ Med—a sectional design for clinical epidemiological studies within individualized medicine. In: Fischer  T, Langanke  M, Marschall  P, Michl  S, eds. Individualized medicine. Ethical, economical and historical perspectives. Heidelberg: Springer; 2015:183–208. 68. Beauchamp  TL, Childress  JF. Principles of biomedical ethics. New York: Oxford University Press; 2009. 69. The National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research. The Belmont Report: ethical principles and guidelines for the protection of human subjects of research. Washington, DC: Department of Health, Education, and Welfare; 1978. http://www.hhs.gov/ohrp/humansubjects/guidance/belmont.html; (Accessed July 3, 2016).

CHAPTER

Oversight, governance, and policy for making decisions about return of individual genomic findings

2

Adrian Thorogood, Ma’n Zawati, Bartha Maria Knoppers Centre of Genomics and Policy, McGill University, Montreal, QC, Canada

1 ­Introduction This chapter considers the regulation of the return of individual genomic findings in health research, potentially encompassing primary, incidental, and secondary findings. Regulation of research encompasses laws, professional as well as ethics policies guiding researcher conduct the latter often referred to as “soft law,” and finally, oversight by Institutional Review Boards (IRB). For complex and rapidly changing issues such as the handling of individual findings, soft law is an important source of guidance for health professionals.1 Regulations may be international, regional (e.g., European Union), national, or more local in application. They are distinguished from soft law by their binding force: law implies liability, legislation implies legal sanctions (e.g., fines) while funder or institutional policies imply administrative sanctions. Legal regulations relevant to return of individual findings may apply to human subject research broadly, including research with identifiable samples and data, or to more specific research activities such as clinical trials, biobanks, or genetic research. Identifying what regulations apply can be challenging considering that genomic sequencing is now used in many different research contexts. In population-based studies, genomes are compared across many people to better understand the relation between genetic variation and human disease. In precision clinical trials, genomes are used to explore if a drug is more (or less) efficacious or safe for subsets of participants. In translational genomics, researchers study if sequencing patients in clinical contexts improves patient outcomes. Because of the tendency of genomics to blur distinctions between research and care, clinical policies guiding the reporting of genomic test results may also—­ directly or indirectly—influence research practice. IRBs, also commonly referred to internationally as Research Ethics Committees, are groups “who undertake the ethical review of research protocols involving humans, applying agreed ethical principles”.2 Oversight by IRBs is an important element of research regulation. They perform a gatekeeping role: projects cannot begin Secondary Findings in Genomic Research. https://doi.org/10.1016/B978-0-12-816549-2.00002-3 © 2020 Elsevier Inc. All rights reserved.

29

30

CHAPTER 2  Oversight, governance, and policy for making decisions

until they have undergone ethics review and received approval. When IRBs review a genomic research protocol, they often consider how individual findings will be handled. In doing so, IRBs interpret and apply relevant laws, policies, and ethics principles. Unfortunately, this does not always result in consistent and predictable outcomes. This is partly because of the diversity of research contexts employing genomics. In any case, researchers who fail to address individual genomic findings in their protocols are likely to face delays in approval. After approval, IRBs may continue to play an advisory role, especially when researchers are faced with unexpected ethical issues that arise during research (e.g., handling findings that could affect family members). In these cases, their role will be to help researchers find the most ethical way of managing any potential return. This chapter focuses primarily on policies specific to health research and the role of IRBs. We do not consider in detail the wide range of legal duties, professional codes, privacy laws, and clinical or laboratory regulations that bear on the handling of individual findings. We adopt an international perspective, discussing common approaches illustrated with jurisdiction-specific examples. We begin by introducing ethical and legal principles influencing the handling of individual findings. Next, we review common policy elements and requirements: development of an ethicsapproved plan, participant consent, rules about what should be returned (or not), procedures for return, and adaptations for children and adults with limited capacity. Finally, we consider how legal regulations deal with individual findings in the context of international sample and data sharing in genomics. Collaboration is essential in genomics but creates complexity over who is responsible for identifying and returning individual findings.

2 ­International legal and ethical principles The return of individual findings is guided by medicolegal principles. Generally, physicians have a fiduciary duty to act in the interests of patients. Under a legal approach, this normally includes a duty to disclose the existence of a medical condition to a patient.3 Sequencing is adopted in different research contexts, so researcher-­ participant relationships in genomics are diverse. The basic medicolegal concern is that a failure to return results could trigger liability under the legal theory of negligence.3 Negligence is a conduct that exposes others to an unreasonable risk of harm. A claimant generally has to prove that the defendant engaged in negligent conduct, resulting in harm to the claimant. For health professionals, negligent conduct is that that falls below a “standard of care,” reflecting prevailing professional practice and established in court through expert testimony. The standard of care may be influenced by applicable laws and professional guidance. We were unable to find case law in the United States, the United Kingdom, or Europe that addresses a failure to return incidental or secondary findings from research or clinical settings.4, 5 If asked whether or not a researcher owes a legal duty to return individual findings, a court would likely consider the type of research (population vs. familial or individual

2 ­ International legal and ethical principles

s­ tudies), data collection and storage practices (coded vs. irreversibly delinked), and the researcher-participant relationship (e.g., is the researcher also the patient’s physician?).6 Returning individual findings might also result in liability if, for example, individual findings are interpreted or communicated in a negligent manner.6 Another theoretical basis for a legal duty to return individual findings is a duty to assist others in imminent danger. Common law jurisdictions usually limit this duty to situations where there is a preexisting “special relationship”.6 In contrast, the duty to rescue in civil law jurisdictions may apply more broadly. For example, Canadian province of Quebec (civil law jurisdiction), for example, a duty to rescue is stated in the Charter of human rights and freedoms. It is triggered when an individual’s life is in peril and requires every person to provide necessary and immediate physical assistance.7 The familial nature of genetic information also raises the prospect of a “duty to warn” at-risk family members. Professionals generally have a duty to warn identifiable third parties at risk of serious, foreseeable, and imminent harm. This duty is a commonly recognized exception to the professional duty of confidentiality. It may be relevant when a patient refuses or is unable to share genetic risk information with family members. The primary legal uncertainty, debated in lawsuits in the United States, the United Kingdom, and Canada, is whether genetic risk information meets the “imminence” criteria.8, 9 The UK case ABC v. St. George10 suggests physicians in that country may have a duty to warn family members despite patient refusal. In that case a physician failed to warn a patient’s pregnant daughter of a 50% risk of inheriting Huntington’s disease. The appeal court found that harm from failure to inform was reasonably foreseeable, which could mean that physicians in the United Kingdom have a duty to override patient confidentiality to warn family members of genetic risks. Data protection laws govern the collection, use, and disclosure of personal (health) data, which are data about an identifiable individual. These laws may also complicate the return of results. In the United States, the Health Insurance Portability and Accountability Act (HIPAA) governs the privacy of personal health information held in covered healthcare entities, including genetic information.11 In Europe the General Data Protection Regulation (GDPR) applies to personal data held by data controllers in all sectors.12 Both laws impose legal duties of confidentiality over personal data. This does not necessarily preclude the return of individual findings to family members, however, as there are exceptions where information affects the vital interests of a third party. We discuss the individual right to access genetic results in a later section. The professional duty to return individual findings also has roots in the human “right to know” information of health relevance, recognized particularly in European human rights documents.13 A problem with the right to know is that researchers cannot presume participants want to receive all forms of clinically relevant information revealed through sequencing. Participants (or their family members) may have privacy, informational autonomy, economic, and even health interests in not knowing their genetic risks. A “right not to know” is recognized by some human rights ethics instruments.13,14 The right not to know is explicitly enshrined in Germany’s Genetic

31

32

CHAPTER 2  Oversight, governance, and policy for making decisions

Diagnostics Act (Gendiagnostikgesetz [GenDG]), although this act only applies to clinical and not research contexts.15 As a result, most policy approaches condition the return of individual findings on the participant’s consent (see in the succeeding text). Beyond legal duties, there are also numerous ethical and policy bases for returning individual findings. The principle of beneficence supports an ethical duty of researchers to return findings, though there are countervailing nonmaleficence concerns of exposing individuals to discrimination. Many participants are interested in or expect researchers to return individual findings.16

3 ­Planning ahead: Return of results protocols A common requirement is that researchers establish a proactive plan for handling individual findings as part of their research protocol, subject to review, and approval by an IRB or funding body (e.g., Refs. 17, 18). This gives researchers an opportunity to carefully consider the likelihood that research will reveal genetic findings of clinical relevance and to prepare in advance to assess and communicate such findings to participants in an ethical and feasible manner.18 This plan can in turn be reflected in the informed consent documents (see next section). That being said the creation of a plan for handling individual findings does not necessarily mean that these findings will actually be returned. All it means is that a decision was made on whether or not such findings could be returned, and if they are, how this will be managed. The Canadian Tri-Council Policy Statement, for example, creates an exception to the obligation of returning “material” findings (i.e., those with severe welfare implications) if the researchers can show that doing so is either impracticable or impossible. The burden of proof lies on the shoulders of the researcher (2018, Application of art. 3.4). Under the US Common Rule, researchers must disclose their plans of whether or not clinically relevant findings will be returned to participants.19

4 ­Consent Voluntary and informed consent is a basic ethical and legal requirement for health research involving human subjects.20 Individual findings present numerous ethical challenges relating to consent, discussed in Chapter 5. Generally, consent is a precondition for returning findings to participants. Regulations therefore encourage or require researchers to inform prospective participants if individual findings will be returned and under what conditions (e.g., Refs. 21, 22). An exception is Spain, where researchers are legally required to return findings to avoid serious harm to participants or their family members.23 Danish guidelines express concern about the ability of participants to understand the choice whether or not to receive results, so they recommend return of serious results unless there is an “unequivocal” wish not to know.24 In addition to respecting participant autonomy, consent also addresses risks associated with genetic testing, such as economic harms resulting from insurer or

5 ­ What to return (or not)

employer discrimination or psychological harms (e.g., anxiety). Under Germany’s Genetic Diagnostics Act, it is mandatory to inform patients about the possibility of incidental findings.15 Regulations may go beyond the principle of informed participation and also offer individuals choices to receive results or not. On a consent form, choices may be framed as an opt-in or opt-out from receiving results. Some policies recommend offering participants tiered choices (e.g., Ref. 21) or even opportunities to update their choices over time (e.g., Ref. 25). An emerging concern is that the mention of potential return of findings in the consent could lead to a “therapeutic misconception.” Research participants could believe they are receiving clinical genomic testing and will assume, if they do not receive results, that nothing is wrong. For this reason the US Common Rule requires researchers to disclose any possibility that clinically relevant results might NOT be returned (HHS, 2017). Another available option is to ask participants if they have a treating physician that can be contacted should any findings be returned to them to ensure that the physician (and not the research team) takes on the clinical follow-up.

5 ­What to return (or not) A central question for researchers and policy makers is determining what types of findings should be returned to participants (discussed in Chapter  3). The primary considerations when returning individual findings are accuracy and usefulness, often described according to the following criteria: Analytical validity refers to how well the test predicts the presence or absence of a particular gene or genetic change. Clinical validity refers to how well the genetic variant being analyzed is related to the presence, absence, or risk of a specific disease. Clinical utility refers to whether the test can provide information about diagnosis, treatment, management, or prevention of a disease…,26 also commonly referred to as “medical actionability.”27 The quality of genomic data is an important concern both for rigorous science and individual clinical decision-making.18 But sequencing approaches in research contexts may involve innovative technologies with uncertain validity, and research laboratories do not have the same quality controls required by clinical laboratory regulations. The response in the United States is to prohibit or discourage the return of individual findings where sequencing is not carried out in a clinically accredited laboratory.18 Recognizing that not all research data can be generated in regulated clinical laboratories, NASEM recommends that standards or regulation (e.g., certification) be developed for research laboratories. Alternatively, if permitted, researchers could clinically validate findings before return or could recommend the participant or their physician seek to validate the finding.28 In the case of innovative assays, validation may not be possible, and researchers and IRBs may need to consider validity and

33

34

CHAPTER 2  Oversight, governance, and policy for making decisions

usefulness on a case-by-case basis. Policies in other countries reflect concern about limited clinical utility. Germany’s professional guidelines for clinical and research genomics prohibit the return of findings that only slightly increase disease risk or cannot be treated or prevented.29 The criterion of “clinical utility” (i.e., “medical actionability”) is contested by ethicists. We summarize four debates here. First, should utility be judged against standards of evidence-based medicine? On one hand the medical standard of care is a relatively clear and objective standard and reflects a traditional vision of physician-­ researchers’ duties toward “patients” in clinical research. A clinician would be reluctant to administer a “treatment” (here the return of information) without clear evidence that doing so would benefit the patient. On the other hand the standard of care typically lags well behind science. It may be clear to genomic researchers that even a novel finding is likely to be useful to the health of the participant. Moreover, participants generally expect to be provided with such information.18 Second, should carrier status be returned? Information about hereditary risks may not directly implicate the health of the participant but could (depending on the person) have important implications for reproductive choices. Some laws and policy explicitly mention such information in the list of types of results to return.29 Third, should determinations of utility incorporate personal or subjective considerations? (see Chapter  3). Why should this assessment be the exclusive domain of professional researchers, clinicians, or expert panels? Genetic information can “inform decisions, actions, or selfunderstanding, which are personal in nature”.16 A number of policies now adopt a broader definition of utility (e.g., Refs. 16, 18). There is little guidance, however, to help researchers determine a priori what is subjectively useful to the participant. Fourth, should the obligation to return results extend beyond participants to their biological family members, given the familial nature of genetic risks30? Interpreting the usefulness of findings for family members before testing them, however, is challenging. There is a potential conflict with the participant’s confidentiality, if he or she refuses to or is unable to consent to sharing the information. Returning findings to family members also raises questions about how to determine their informational preferences and respect their right not to know. Beyond criteria of accuracy and usefulness used to determine what to return, another important question is who should assess quality and usefulness. Policies often leave the decision to return to the researcher to make on a case-by-case basis, according to the criteria described earlier. Recognizing that researchers often lack clinical and ethical expertise, not to mention time and resources, policies may also require or recommend that a physician, IRB, or even an expert (interdisciplinary) committee participate in an assessment (e.g., Ref. 24). Policies are often distinguished between “individual research results,” “incidental findings” (serendipitous findings unrelated to the research question), and “secondary findings” (findings actively analyzed but unrelated to the research question). There are two normative distinctions here. First the significance of individual research results relates to the general outcomes of a study and is therefore inherently less certain. The significance of incidental or secondary findings is independent from

6 ­ How to return

the research results. Second, intentionality distinguishes secondary from incidental findings. In other words the question is not only what to return but also what to look for. The potential for such findings in genomics is largely “a function of technical and analytic choices”.31 How these choices affect the prevalence of secondary findings is discussed in more detail in Chapter 4. There is ongoing debate over analytic choices in clinical testing contexts. Some clinical policies encourage the use of “gene panels” in clinical and research sequencing contexts. The European Society of Human Genetics and the Canadian College of Medical Genetics both recommend that genomic data be filtered, so that clinicians or researchers only inspect a panel of genes relevant to a clinical or research question. This filtering approach essentially aims to reduce the potential for incidental findings (e.g., Refs. 32, 33). By contrast the ACMG policy recommends a uniform gene panel of secondary findings, established and updated by an expert panel, be actively analyzed and returned whenever patients or participants undergo WGS.34 National sequencing project policies in Europe have recommended a list approach. The 100,000 Genomes Project has established a list, which will be updated over time, of secondary findings including ­cancer- and hypercholesterolemia-­related genes.35 The French Society of Predictive and Personalized Medicine recommended return of secondary findings for 36 cancer-­ associated genes in the clinic, noting their guidelines could also guide research.36 Similar lists span from 10 to 100 genes, but it is widely expected that such lists will expand as our knowledge of clinically actionable genetic variation improves.37 While secondary finding approaches in clinical contexts are likely to influence research, important differences must be kept in mind.38

6 ­How to return Laws and policies often address aspects of how results should be returned to participants. The “how” of return is very much concerned with the effectiveness and safety of returning results, i.e., in ensuring the benefit is realized, while avoiding harms such as unnecessary anxiety. Achieving these goals boils down to participant understanding: results “should be returned in ways and at times that maintain the integrity of the research, insofar as the safety and welfare of the research participants are not at risk”.16 Approaches to disclosing results more generally are discussed in detail in Chapter  7. Briefly, findings are usually returned to participants or their physicians. Indeed, clinical standards are an important reference point. Germany’s Genetic Diagnostics Act, for example, requires incidental findings to be reported directly to patients by a physician.15 Effective strategies for communicating with participants include providing a clear synopsis, providing opportunities for follow-up, and tailoring information (e.g., literacy) to different communities or individuals.18 It is often recommended or required that researchers ensure participants have access to a clinician, geneticist, or genetics counselor.21 Returning results directly to the participant’s physician can help to ensure appropriate explanation and follow-up. Practical tools are also available to guide return to participants’ relatives.39 Finally, how results are

35

36

CHAPTER 2  Oversight, governance, and policy for making decisions

returned implicates privacy and security. Where the data are coded, the individual must be reidentified to return the results.40 Obviously, there is no return if data are anonymized, that is, irreversibly delinked. Tracking samples and authenticating participants are also important to avoid breaches of researcher-participant confidentiality.

7 ­Children and adults who lack decisional capacity: Special considerations Return policies require certain modifications in contexts where participants lack the capacity to make research-related decisions, namely, minors and adults with limited or diminished capacity. One policy issue is whether or not the discretion of parents or legally authorized representatives (LARs) to refuse to receive individual findings should be limited. Indeed, these representatives have ethical and legal obligations to act in the individual’s best interests. Some professional genetics societies hold that parents should not be able to refuse clinically useful individual findings (for prevention or treatment) during childhood.32, 41 Refusing to receive or act on such information may in extreme cases amount to medical neglect under child or vulnerable adult protection laws. The matter is complicated for children in the case of individual findings relating to conditions that are not clinically useful or only become so in adulthood. There is a recognition of the need to preserve an “open future” for children, until they can make their own decisions to be tested or not as adults.27 Withholding individual findings about adult-onset conditions is problematic, however, as the individual may not have another opportunity to consider testing. Complicating matters further in the genetics context is that parents and LARs are often biologically related to the participant. On one hand, this may create a potential conflict of interest between the participant's and the parent’s (or LAR’s) preference to know (or not) about familial risk variants. On the other hand the parent’s (or LAR’s) general well-being may have welfare implications for their ward. Returning individual results to family members of adults unable to consent to disclosure may need to be guided by principles of substitute decision-making. When a researcher or LAR makes a decision on behalf of such an individual, they are generally required to consider the individual’s (previously expressed) wishes, values, and beliefs.42 As with a traditional duty to warn scenario, it may be appropriate to balance considerations of what the person would have wanted with concern for the well-being of the family member.30 Other consent and capacity issues must be considered in these contexts. First, ethical and legal principles of “assent” suggest that, where individuals lack the capacity to make a decision about return, they should still be provided with appropriate information and included in the decision-making process where possible. Second, in longitudinal research, if participants (i.e., children) develop capacity during the course of a study, should they be asked to reconsent to research participation when they reach the age of majority? Or to update their return preferences? There are also questions about how results should be returned in these contexts: only to the parent/ LAR or also to the child/individual?

9 ­ International data sharing and return

8 ­Return of secondary findings versus the right of access As we have seen, researchers have duties in some circumstances to return individual findings to participants. They also have a duty to share general research results with the public and participants.20 These duties should not be conflated or confused with the legal right of individuals to access their health record or personally identifying information upon request. The right of access is recognized in varying contexts in most countries, and is usually found in personal data protection legislation. There are sometimes exceptions in research contexts.43 Regardless, this right is rarely mentioned in research policies or governance documents. Since 1997, however, policies governing genetic data have articulated a right to know one’s health information (or not).44, 45 The right to know buttresses both the obligation to return individual results and the right of access. In the United States the Health Insurance Portability and Accountability Act (HIPAA) provides individuals a right to access their health record from covered healthcare entities, including genetic information.11, 28 Research participants in the United States do not normally have a right to access their individual research information, unless the research is conducted at an institution governed by HIPAA, and the information is considered part of their health record. In Europe, the General Data Protection Regulation (GDPR)12 has broad application to personal data and includes a general right to access, though member states may adopt exceptions for research. International projects that share patient data with European researchers may also find themselves subject to this right of access. While distinct from the researcher’s duty to return individual findings and the participant’s right of access raise overlapping ethical considerations. The quality of data, for example, is an important concern in both contexts. Growing emphasis on enabling individuals to control their genomic data constitutes a move away from protectionism of privacy interests to a reinforcement of personal autonomy.

9 ­International data sharing and return The international collaborative nature of genomic research complicates the handling of individual findings. Much of the genomic research enterprise, a form of infrastructure science, is dependent on shared biobanks and data repositories. Data sharing policies developed by funding agencies and journals often encourage or require genomic researchers to deposit individual-level genomic data in shared repositories, to serve as a resource for new studies, and also to improve scientific transparency by allowing for validation and refinement of results.46 Biobanks or databases may impose a return policy on users as a condition of access (e.g., in a material or data transfer agreement). This allows the return policy to “travel” with the samples or data. Denmark’s genomics guidelines, for example, explicitly require any external or foreign collaborators to abide by the same incidental findings policy applicable to Danish researchers.24 This can lead to confusion or even policy conflicts when sharing samples or data internationally, where the return

37

38

CHAPTER 2  Oversight, governance, and policy for making decisions

policy imposed by the access agreement disagrees with the policies applying to the foreign researcher.29

10 ­Conclusion Laws and policies have more and more to say about the return of individual findings in research contexts. IRBs are also turning more attention and scrutiny to how findings are handled. Unfortunately, expanding regulation does not appear to be resulting in greater certainty for researchers. This is partly because fundamental ethical and policy issues remain unresolved. An inherent difficulty in drafting policy is the diversity of research and resource contexts involving genomics. In addition, policy tends to focus on scientific, medical, and ethical considerations, namely, quality and utility. Policy often has much less to say about how to balance these considerations with competing concerns about limited expertise, time, and resources. It is increasingly clear that researchers and IRBs cannot be expected to handle individual findings alone. They need dedicated funding and institutional support to develop and review return plans, to assess the quality and usefulness of results, to effectively communicate results, and to ensure participants can act on the information they receive.18 Another trend influencing both research and research oversight generally is the rise of participant engagement. The return of individual findings is sometimes considered a form of engaging and showing respect for participants.18 Participants can also be engaged in designing research governance, including project-specific plans for governing individual findings (Ref. 21, Guideline 7). This amounts to a shift away from substantive rules for researchers and toward cocreation of governance. On one hand, this offers opportunities to ensure individual findings are handled in an effective and respectful manner. On the other hand, the replacement of substantive rules with engagement processes can decrease certainty and uniformity and equity of practices over time. Emphasis on the return of secondary findings also raises concern about a shift toward “routine genomic screening,” especially given the large size of modern research cohorts.37 Fortunately, there are research projects like Geisinger’s MyCode exploring the prevalence of secondary findings and clinical utility, as well as the costs, benefits, and risks of such routine screening.47 Ultimately, participants will only realize the value of secondary findings if there is a health system ready to accept a “hand off” from researchers. But health systems are awaiting evidence of clinical utility before adopting genomics. This remains one of the fundamental chicken-and-egg problems for genomics.

­Acknowledgments We would like to acknowledge the funding support of Genome Canada, Genome Quebec, and the Canadian Institutes of Health Research.

­References

­References 1. Campbell A, Glass KC. The legal status of clinical and ethics policies, codes, and guidelines in medical practice and research. McGill Law J. 2000;46:473. 2. World Health Organization. Standards and Operational Guidance for Ethics Review of Health-Related Research With Human Participants. https://www.who.int/ethics/research/en/; 2011. 3. Clayton EW, Zawati MH. Legal aspects of health applications in genomics. In: Genomics and Society. 2016:119–133. 4. Clayton EW, Haga S, Kuszler P, Bane E, Shutske K, Burke W. Managing incidental genomic findings: legal obligations of clinicians. Genet Med. 2013;15(8):624. 5. Mitchell C, Ploem C, Chico V, et al. Exploring the potential duty of care in clinical genomics under UK law. Med Law Int. 2017;17(3):158–182. 6. McGuire AL, Knoppers BM, Ma'n HZ, Clayton EW. Can I be sued for that? Liability risk and the disclosure of clinically significant genetic research findings. Genome Res. 2014. gr-170514. 7. Government of Quebec. Charter of Rights and Freedoms Chapter C-12. http://legisquebec.gouv.qc.ca/en/showdoc/cs/C-12; 2019. 8. de Paor  A. Genetic Risks and Doctors’ Disclosure Obligations—Revisiting the Duty of Confidentiality. 2018. 9. Zawati MNH, Thorogood A. The physician who knew too much: a comment on Watters V. White. Health Law J. 2014;21:1. 10. ABC v. St. George. ABC v. St. George's Healthcare NHS Trust & Ors. 2017. EWCA Civ 336. 11. United States Government. Health Insurance Portability and Accountability Act (“HIPAA”). 1996. 12. EU General Data Protection Regulation (GDPR). Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April 2016 on the protection of natural persons with regard to the processing of personal data and on the free movement of such data, and repealing Directive 95/46/EC (General Data Protection Regulation), OJ 2016 L 119/1 13. European Commission. European Convention on Human Rights and Biomedicine. 1997. 14. Laurie G. Recognizing the right not to know: conceptual, professional, and legal implications. J Law Med Ethics. 2014;42(1):53–63. 15. Rudnik-Schöneborn  S, Langanke  M, Erdmann  P, Robienski  J. Ethische und rechtliche Aspekte im Umgang mit genetischen Zufallsbefunden-Herausforderungen und Lösungsansätze. Ethik Med. 2014;26(2):105–119. 16. Multi-Regional Clinical Trials Center (MRCT). Return of Individual Results Workgroup, Return of Individual Results to Participants: Recommendations Document. 2017. 17. Canadian Tri-Council (Canadian Institutes of Health Research, Natural Sciences and Engineering Research Council of Canada and Social Sciences and Humanities Research Council of Canada). Tri-Council Policy Statement: Ethical Conduct for Research Involving Humans. December 2018. 18. National Academies of the Sciences, Engineering, and Medicine (NASEM). Return of Individual-Specific Research Results Generated in Research Laboratories. 10 July 2018. 19. Department of Health and Human Services. Federal Policy for the Protection of Human Subjects – Final Rule. United States 2017. 45 CFR §46 [US Common Rule] (23 August 2019).

39

40

CHAPTER 2  Oversight, governance, and policy for making decisions

20. World Medical Association. Declaration of Helsinki: Ethical Principles for Medical Research Involving Human Subjects. 2013. 21. Council for International Organizations of Medical Sciences/World Health Organization (CIOMS/WHO). International Ethical Guidelines for Health-related Research Involving Humans. 2016. 22. United States Department of Health and Human Services. Federal policy for the protection of human subjects (“Common Rule”). Fed Reg. 2015;80(173):53933–54061. 23. Spain. Law 14/2007 of 3 July on Biomedical Research. 2007. 24. Denmark. National Committee on Health Research Ethics, Guidelines on Genomics Research. 2018. 25. Australian Government. Australian Code for the Responsible Conduct of Research. 2018. 26. National Institutes of Health. How Can Consumers Be Sure a Genetic Test is Valid and Useful? 14 August https://ghr.nlm.nih.gov/primer/testing/validtest; 2018. 27. Knoppers BM, Zawati M, Sénécal K. Return of genetic testing results in the era of wholegenome sequencing. Nat Rev Genet. 2015;16(9):553. August 4. 28. Wolf  SM, Evans  BJ. Return of results and data to study participants. Science. 2018;362(6411):159–160. 29. Thorogood A, Dalpe G, Knoppers BM. Return of individual genomic research results: are laws and policies keeping step? Eur J Hum Genet. 2019;27:535–546. 30. Wolf SM, Branum R, Koenig BA, et al. Returning a research participant's genomic results to relatives: analysis and recommendations. J Law Med Ethics. 2015;43(3):440. 31. Burke  W, Beskow  LM, Trinidad  SB, Fullerton  SM, Brelsford  K. Informed consent in translational genomics: insufficient without trustworthy governance. J Law Med Ethics. 2018;46:79–86. 32. Boycott K, et al. The clinical application of genome-wide sequencing for monogenic diseases in Canada: position statement of the Canadian College of Medical Geneticists. J Med Genet. 2015;52(1):431–437. 33. van El  CG, et  al. Whole-genome sequencing in health care: recommendations of the European Society of Human Genetics. Eur J Hum Genet. 2013;21(6):S1. 34. ACMG. ACMG policy statement: updated recommendations regarding analysis and reporting of secondary findings in clinical genome-scale sequencing. Genet Med. 2015;17(1):68. 35. Genomics England. What Can Participants Find Out? https://www.genomicsengland. co.uk/information-for-participants/findings/; 2019. 36. Pujol  P, Perre  PV, Faivre  L, et  al. Guidelines for reporting secondary findings of genome sequencing in cancer genes: the SFMPP recommendations. Eur J Hum Genet. 2018;26(12):1732. 37. Murray  MF. The path to routine genomic screening in health care. Ann Intern Med. 2018;169(6):407–408. 38. Jarvik GP, Amendola LM, Berg JS, et al. Return of genomic results to research participants: the floor, the ceiling, and the choices in between. Am J Hum Genet. 2014;94(6):818–826. 39. Wolf  SM, Scholtes  E, Koenig  BA, et  al. Pragmatic tools for sharing genomic research results with the relatives of living and deceased research participants. J Law Med Ethics. 2018;46(1):87–109. 40. Thorogood A, Joly Y, Knoppers BM, et al. An implementation framework for the feedback of individual research results and incidental findings in research. BMC Med Ethics. 2014;15(1):88. 41. Botkin JR, et al. Points to consider: ethical, legal, and psychological implications of genetic testing in children and adolescents. Am J Hum Genet. 2015;97:6–21.

­References

42. Thorogood  A, Mäki-Petäjä-Leinonen  A, Brodaty  H, et  al. Consent recommendations for research and international data sharing involving persons with dementia. Alzheimers Dement. 2018;14(10):1334–1343. 43. Thorogood A, Bobe J, Prainsack B, et al. APPLaUD: access for patients and participants to individual level uninterpreted genomic data. Hum Genom. 2018;12(1):7. 44. Council of Europe. Convention for the protection of Human Rights and Dignity of the Human Being with regard to the Application of Biology and Medicine: Convention on Human Rights and Biomedicine. 1997. Oviedo, 4.IV.1997. 45. UNESCO. International Declaration on Human Genetic Data. 2003. 46. Taichman DB, Backus J, Baethge C, et al. Sharing clinical trial data—a proposal from the International Committee of Medical Journal Editors. N Engl J Med. 2016;374(4):384. 47. Schwartz ML, McCormick CZ, Lazzeri AL, et al. A model for genome-first care: returning secondary genomic findings to participants and their healthcare providers in a large research cohort. Am J Hum Genet. 2018;103(3):328–337.

41

CHAPTER

Selecting secondary findings to report: Creating a list that suits your study

3

Ann Katherine Major Foreman, Jonathan S. Berg Department of Genetics, School of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, NC, United States

With advances in multiplex genetic technology (MGT) making these techniques increasingly routine in human subjects research, it has become imperative for research teams to carefully consider their approach to the disclosure of results to research participants. This chapter focuses on the criteria that researchers should consider when making decisions about what types of results would qualify for disclosure as secondary findings—specifically those additional genomic findings that are not the main focus of the research. It is worth emphasizing at the outset that there is no single “right answer” and that each research protocol will have unique features that influence decisions about disclosure of secondary findings.1 Above all, we find that it is critical for research teams to invest time during the planning phase of the project and to make careful decisions about returning secondary findings in an a priori fashion. Making these choices prior to the generation of any sequencing data is beneficial both to research participants and investigators. For research participants the advantages largely relate to the enhancement of appropriate, meaningful informed consent, and avoidance of unnecessary surprise at the time of return of results. Investigators will find that, while proactively searching for secondary findings may seem burdensome to those without prior experience with genetic sequencing on this scale, a considered approach to analyzing and returning such variants minimizes that burden. Investigators can define the set of secondary findings that they will seek out, choosing a list that makes sense in the context of their research aims, their study population, and their resources. The desired set of secondary findings will inform the parameters of the investigators' bioinformatics pipeline, making the search for secondary findings as efficient as possible. This also minimizes the risk of identifying truly incidental findings and having to make post hoc choices about whether to return them. These topics will be covered at greater depth in the following chapters. In this chapter, we aim to assist investigators in making informed a priori choices about the secondary findings they wish to seek out and return to participants in their research.

Secondary Findings in Genomic Research. https://doi.org/10.1016/B978-0-12-816549-2.00003-5 © 2020 Elsevier Inc. All rights reserved.

43

44

CHAPTER 3  Selecting secondary findings to report

1 ­Types of genomic findings Any MGT will yield a number of genomic variants in each individual sequenced. For example, a typical exome sequence will yield 80,000–110,000 variants at which the individual's genomic sequence differs from the reference genome. A small fraction may be relevant to the reason for sequencing and, in the broadest definition of the term, the remainder are secondary findings. Given the current limitations of science in understanding the relevance of genomic variation, the majority of these will simply be uninterpretable. The implications of the remainder of these variants for the individual's health and well-being cover a spectrum from meaningless (e.g., rs72921001, a variant associated with whether one finds the taste of cilantro appealing or disgusting)2 to critical (e.g., rs80357906, a pathogenic BRCA1 gene variant associated with up to 80% lifetime breast cancer risk).3 Even among variants with clinical relevance, the ability to mitigate risk and thus the utility of disclosing these findings can be strikingly different. Here, we review several types of genomic information that may be said to have clinical relevance.

1.1 ­Multifactorial disease risks Most common health conditions and traits can be described as “complex” or “multifactorial” due to the many varying contributions of genetic variants, environmental or behavioral influences, and chance. Techniques such as genome-wide association studies (GWAS) have led to the identification of thousands of genomic variants that are statistically associated with different disease states. Often these will be single nucleotide polymorphisms (SNPs) that are found with different frequencies in cohorts of people with a disease of interest compared with unaffected cohorts. Using this type of data, an odds ratio may be calculated, which communicates the odds of having disease among people with the SNP relative to those without the SNP. Communicating multifactorial disease risk factors as secondary findings has some appeal. Multifactorial diseases include some of the most common problems impacting human health, such as heart disease, obesity, and diabetes, suggesting they may be of broad interest to many research participants. However, multifactorial disease risk SNPs have significant limitations when it comes to predicting disease risk for a given individual. For multifactorial disease, by definition, no single SNP is either necessary or sufficient for disease to manifest; the SNP is one of multiple factors that may include other SNPs, behaviors (e.g., diet and exercise habits), or other environmental influences. Even for diseases in which multiple risk SNPs have been identified and replicated, scientific understanding of how to interpret those SNPs in combination with one another remains limited. SNPs identified in cohorts of one race or ethnicity may not be generalizable to others. Even if a person appears to be at increased risk, there are often not useful actions to reduce the risk. If there are riskreducing behaviors, they would likely be recommended to those with a “low risk” genetic profile and those with apparently increased risk. For example, the odds ratio for age-related macular degeneration (ARMD) is high in homozygotes for a variant

1 ­ Types of genomic findings

in the ARMS2 gene called A69S or rs10490924 4. However, preventative actions one can take to limit the risk of ARMD include being a nonsmoker, maintaining a normal blood pressure, and limiting sun exposure by means such as wearing UV protective sunglasses. These are all recommendations that might be made to someone regardless of whether their genomic risk profile is suggestive of increased risk of ARMD. Overall, if returning risk SNPs for multifactorial disease is appealing to a research group, there are two main options. The first option would be to expend considerable energy defining what collection of SNPs to include and how to aggregate the risk information relative to those SNPs. Due to the nature of multifactorial disease, the presence or absence of any given SNP is essentially uninformative with respect to disease risk, and it is instead the collection of SNPs present in a given individual (sometimes reflected in a “polygenic risk score”) that is required to provide predictive information—and even in that case the predictive value of the information may still be quite limited. Because individuals who do not harbor any known risk SNPs may still develop disease, particularly if they have family history or nongenetic risk factors, a risk of returning this information as part of a portfolio of secondary findings is that those without increased genetic risk identified and reported may underestimate their risk. The main challenge with this option is that, to provide risk information responsibly, the research team would need to come up with a scientifically defensible plan to select SNPs for inclusion in the result disclosure. Further, if individualized risk is intended to be predicted and disclosed, the team should include expertise in calculating polygenic risk scores and communicating this information. The second option would be to provide research participants with raw sequencing data or variant calls that would enable them to pursue these analyses through third-party services that have become available. This option may appear to provide the most flexibility for research participants, but this may be somewhat misleading given the lack of clinically validated options and the potential for participants to be provided with incorrect information by third-party vendors.

1.2 ­Pharmacogenomics Pharmacogenomics involves the use of SNPs or haplotypes (combinations of SNPs in a genomic region) to predict something about a person's response to pharmaceuticals. That "something" may be the optimal dose for that individual, their likelihood to develop adverse drug reactions, or whether the medicine will be effective in them at all. Variations of the phrase “right patient, right drug, right dose, right time” have become a mantra of sorts for the field of pharmacogenomics, even being used as the name of a pharmacogenomics research protocol at the Mayo Clinic.5,6 The idea that by using genetics individual differences can be identified that help to choose the most appropriate drug and dose is both appealing and intuitive. Unfortunately, achieving the promise of pharmacogenomics has proven difficult. As of 2018, over 200 drugs include some pharmacogenomic information in their labeling; of these, only 10 are boxed warnings.7 Boxed warnings are made only when there is a serious risk or hazard, so not all pharmacogenomic gene-drug interactions

45

46

CHAPTER 3  Selecting secondary findings to report

would be expected to result in a boxed warning. Still, this suggests that pharmacogenomic information is thus far more useful to optimize benefit than to prevent serious harms. There has been controversy as to the true utility of pharmacogenomic information, even in cases where genetic findings have clear implications for drug metabolism and recommended clinical actions. For example, the anticoagulant warfarin was among the first drugs to have pharmacogenetic information included on its labeling, as variants in both CYP2C9 and VKORC1 have implications for its dosing. Certain genotypes require reduced dosage due to higher risk of bleeding complications. A randomized study of over 1000 patients assigned to clinical dosing with or without addition of pharmacogenomics concluded “genotype-guided dosing of warfarin did not improve anticoagulation control during the first 4 weeks of therapy.”8 A different randomized study of 455 participants found that patients with warfarin dosing based on pharmacogenomics reached optimal therapeutic dose quicker and stayed in a therapeutic range longer.9 Despite ongoing conflicting trial results, a boxed warning for warfarin dosing remains. For pharmacogenomic information to be useful, it must be available to the prescribing medical provider at the time a drug is needed. Before deciding to return pharmacogenomic secondary findings, a researcher should consider how likely it is that the findings will actually be useful at some point in the future should a participant have need of a drug. This includes the likelihood that the participant may need the drug in question at some point in the future and that their research result will be appropriately confirmed and accessible at that time. Additional information about pharmacogenetics, including genes and drugs for which there are clinical prescribing guidelines, can be found at PharmGKB, a web-based pharmacogenomics resource and database (www.pharmgkb.org). The Clinical Pharmacogenetics Implementation Consortium (CPIC) releases guidelines on the clinical use of some pharmacogenomics findings, including level of evidence supporting the recommendation. These guidelines are indexed on the CPIC website (www.cpicpgx.org/guidelines).

1.3 ­Mendelian disease Mendelian diseases, also known as monogenic or single-gene disorders, are those in which disruption of a single gene by a pathogenic variant or variants directly causes a disease or phenotype. These conditions are inherited in predictable patterns depending on whether mutation of one or two copies of a given gene is required to produce a phenotype and whether the gene is located on an autosome or sex chromosome. Inheritance patterns for Mendelian conditions include autosomal dominant, autosomal recessive, X-linked dominant, and X-linked recessive. There are also single-gene disorders caused by pathogenic variants in genes that are part of the mitochondrial DNA. Mendelian diseases are individually rare but common as a group. They are highly diverse, with varying age of onset, differing degrees of severity, expression of symptoms from isolated to syndromic, and affecting widely differing organ systems or disease states. They differ with respect to their penetrance, or the likelihood that a

1 ­ Types of genomic findings

person with a pathogenic variant or variants will exhibit symptoms. In some conditions, it is expected that everyone who inherits the relevant variant(s) will eventually develop disease, whereas in others the lifetime chance of developing disease could be lower than 10%. Identification of a pathogenic variant in a gene associated with a highly penetrant condition in an unaffected person can be considered predictive, while finding a pathogenic variant associated with a reduced penetrance condition may be called predispositional. Mendelian diseases also vary widely with respect to the extent that they can be intervened upon, from neurodegenerative conditions with no effective medical intervention whatsoever, to hereditary cancer predisposition enabling the opportunity to enhance surveillance and/or pursue prophylactic surgery, to conditions for which treatments are available with nearly curative potential. In clinical practice, presymptomatic testing for genetic conditions involves consideration of the possibility of medical intervention. Often extensive nondirective genetic counseling is performed in the context of an individual who is at risk due to a family history, and testing decisions are made in light of the available medical interventions. Clearly, this process may be inappropriate for many researchers to engage in, not to mention the substantial effort required to conduct professional genetic counseling for certain scenarios.

1.4 ­Carrier status Carrier status can be considered as a subcategory of Mendelian diseases for possible return as secondary findings. While having a Mendelian disease is relatively uncommon, virtually everyone is a carrier for one or more recessively inherited Mendelian diseases. Autosomal recessive diseases include sickle cell disease and cystic fibrosis. A recessive disease is caused by having pathogenic variants in both alleles of the same disease-associated gene. Carriers of recessive diseases have a pathogenic variant in just one allele, meaning they retain one functional copy of the gene. For most recessive diseases, carriers are clinically unaffected; however, they are at risk for having descendants with disease. Thus, while carrier status is expected to have no impact on the individual's personal health in most cases, it could be highly clinically relevant to an individual of reproductive age. When reproductive partners are both carriers of the same recessive disease, each naturally conceived child has a 25% risk of having the disease. Being a known carrier of a specific genetic condition puts one at markedly increased risk to have an affected child compared with the general population, but the absolute risk is still typically quite low as it requires one's reproductive partner to be a carrier for the same condition. Since any detailed discussion about the potential risk to have an affected child would depend on the prevalence of the condition(s) that the individual is found to carry and the actual risk depends on the partner's testing for the same condition, the utility of carrier status information is necessarily limited unless the analysis is conducted and interpreted in a specialized way that is likely to be far outside of the typical research protocol and more typically reserved for detailed reproductive genetic counseling.

47

48

CHAPTER 3  Selecting secondary findings to report

2 ­Criteria for a reportable secondary finding There will be a natural tension between some research participants' desire to know as much as possible from their research testing and the time and effort required by research teams to deliver these results in a responsible fashion. This is important to consider so as not to allow secondary findings reporting to overburden the research enterprise as a whole. For this reason, we propose that investigators planning to return secondary findings limit themselves to those with clear clinical relevance and possibly to a small subset of findings that are clinically actionable.

2.1 ­Confirmation and reporting The final interpretation of genetic variants implicated in Mendelian disorders should be done by a qualified molecular genetics professional who is experienced in clinical sequence variant interpretation. In addition, federal guidelines require that results being disclosed to research participants be confirmed in a clinical laboratory with CLIA/CAP certification. While this guideline may have been reasonable in the era when only one or a small number of results might be considered for disclosure, the possible wide range of findings that could be considered in the era of MGT makes this a much more daunting problem. That being said, if a result is going to be used in medical decision-making, it should be either generated or confirmed in a clinical laboratory with appropriate certifications and accreditations. A physician initiating a medical care plan on the basis of a secondary finding should require that the finding is clinically confirmed and interpreted. Research genomic findings should not be placed in patients' electronic medical records, unless those findings have been confirmed in a clinical lab. If research results are to be confirmed, it is worth considering who bears the cost for this confirmation. Assuming that the research sequencing is not being performed in a CLIA-certified and CAP-accredited laboratory, this means that secondary findings planned for return will need to be confirmed in such a lab, typically by Sanger sequencing. A research budget could quickly become overwhelmed by confirming results of various secondary findings, particularly for categories for which most if not all participants will have at least one finding. On the other hand, not all research participants will have equal resources to seek out clinical confirmation of any research findings of their own. The financial costs of confirmation are limited if secondary findings are restricted only to results expected to have a significant impact on a participant's medical care. The overall costs of learning about secondary findings—including financial cost of confirmation, professional genetic counseling, and possibly medical interventions that stem from the findings, as well as the risks of those interventions and psychological burdens—can be balanced by consideration of the harms that may be prevented by foreknowledge of a genetic disease and amelioration of the effects of that disease through appropriate intervention.

2 ­ Criteria for a reportable secondary finding

2.2 ­Genes Efforts to define recommended sets of reportable secondary findings generally start with lists of genes. Creating your own list of genes in which secondary findings will be reported is an important task, but it is only a first step. While a good foundation for formulating choices about secondary findings, the gene in which a variant is found is not the only important criterion to consider. Also important are the specific variant's likelihood of pathogenicity, participant factors (e.g., age), and research aims. It is impossible to predict in which genes you might identify variants within a given cohort of research participants, much less what those specific variants may be. Thus it is reasonable to generate a secondary finding list of genes prior to generation of sequence data but then individually consider any variants identified in that list of genes to determine if a variant is appropriate for return. Of course, there may be nuances in the way that gene lists behave in a bioinformatics algorithm. For example, some genes are associated with different disease entities depending on molecular mechanism (gain of function vs. loss of function) and/or inheritance pattern (dominant vs. recessive). These features of the genetic locus could lead to the unintentional discovery of variants that are not qualified to be returned, thus causing the research team to be aware of findings that the patient may not have consented to learn. Secondary finding gene lists are, in reality, lists of gene-disease pairs. A gene is only medically relevant as it relates to a human disease or phenotype, and thus the strength of evidence underlying the gene-disease association should be a fundamental concept governing the disclosure of secondary findings. There are a number of systems used to define the clinical validity of gene-disease associations, including a framework developed by the ClinGen resource.10 These types of approaches seek to define the genes that have sufficiently strong evidence supporting the gene-disease association to use that information for clinical purposes.

2.3 ­Actionability Another important concept that guides the generation of a secondary finding gene list for many research groups is the “actionability” of the finding. Unlike indicationbased testing, identification of secondary findings is an opportunistic venture aimed at making use of available sequence data to identify threats that can be prevented or ameliorated. That goal is only achieved if the finding is medically actionable, that is, if there are medical interventions that can be implemented that are effective in preventing the disease, detecting its onset at a more treatable stage, or otherwise altering the natural history in a beneficial way for the affected individual. Ideally, those interventions would be based on data and represent a medical consensus. The importance of having a strong, established association between the gene and the disease is linked to the concept of actionability. Almost no medical intervention is benign, and if a person is to submit to intervention on the basis of a secondary finding, it is critically important to have confidence that the person is at risk and in need of that intervention. If it is not clear that pathogenic variants in a given gene even cause a disease, it cannot be clear that medical action is appropriate.

49

50

CHAPTER 3  Selecting secondary findings to report

As one example of an approach to defining “actionability,” researchers at the University of North Carolina at Chapel Hill have developed a semiquantitative metric that scores five key parameters of actionability: the severity of the disease outcome of interest, the likelihood that the severe outcome will manifest in a person with a pathogenic variant in the gene, the efficacy of interventions aimed at ameliorating that severe outcome, the acceptability or burden of those interventions in an otherwise healthy individual, and the knowledge base regarding the gene-disease association and the parameters being scored.11 This metric was initially developed for determining “medically actionable secondary findings” for use in the North Carolina Clinical Genomic Evaluation by Next-generation Sequencing (NCGENES) study and later modified to take into account age-based actionability in the North Carolina Newborn Exome Sequencing for Universal Screening (NCNEXUS) study.12 In parallel the Clinical Genome Resource Actionability Working Group has adapted the semiquantitative metric,13 which is now being used for the American College of Medical Genetics and Genomics (ACMG) Secondary Findings Maintenance Working Group.14 Although the precise threshold for what might define “actionable” may vary from one setting to another, this method of scoring enables comparisons of the relative actionability between conditions and enables decision-making about what gene-disease pairs exceed a defined threshold in a transparent and evidencedriven manner. Some groups have opted to define gene lists for return of results using other methods, such as consensus expert opinion,15 while others have defined categories for return of results that do not hinge on actionability but incorporate more restricted features such as the severity and penetrance of the condition16 or other characteristics such as the organ systems involved.17

2.4 ­Participant and study factors There may be factors related to the research participants or the study that can further refine the creation of a gene list. The age of the individual undergoing sequencing has perhaps been the most discussed participant factor that can inform the search for secondary findings. If your research involves primarily adults from middle age onward, you may choose to exclude from secondary finding analysis even highly actionable conditions that would have high penetrance and clinical onset in childhood. Your population would have essentially outlived their risk for these conditions obviating the need to seek them out. While returning carrier status for autosomal recessive Mendelian diseases may not make sense for all studies, it may be a good fit for a study that enrolls mostly young adults of reproductive age, particularly if the study has anything to do with reproductive health making this kind of secondary finding potentially attractive or useful for participants. Considering participant age also brings into question whether it is appropriate to return results for adult-onset conditions in pediatric cohorts. In 2013 ACMG proposed a set of genes and conditions, with ages of onset ranging from infancy to adulthood, which should be proactively sought out and analyzed in the setting of clinical

2 ­ Criteria for a reportable secondary finding

exome or genome sequencing. In the recommendation the workgroup asserted that age of the individual undergoing sequencing should not be used to restrict this suggested list. Several groups pushed back against this recommendation, in part because it seemingly contradicted long-standing guidelines in the field of clinical genetics to delay predictive testing for adult-onset conditions until at-risk individuals reach adulthood and can participate in the decision of whether to have testing fully.18,19 Such guidelines were developed to address clinical settings in which there is knowledge of a specific adult-onset disorder in the family for which the at-risk child is at considerable (typically 50%) risk. The intent is to protect the future autonomy of the child to be able to make his or her own decisions about testing after reaching the age of majority and to avoid potential harm including psychological distress or inappropriate early interventions. While relevant scenarios may include actionable conditions, like inherited cancer risk, the guidelines also applied to situations in which the condition has no current disease-modifying intervention, such as Huntington disease. A representative statement against analysis of adult-onset secondary findings in minors comes from the Public Population Project in Genomics (P3G) Consortium. In their 2014 statement, they make the following recommendation: “Mutations that predispose the child to develop an adult-onset disorder, even if accidentally discovered in the research process, generally should not be returned. This allows the child to make his or her own decision about receiving the results as an adult.”20 Realistically, however, children who undergo sequencing do not have the opportunity to make decisions about receiving results as adults. It is unlikely that any research involving sequencing of minors will still have funding, staffing, and sufficient contact with participant families to make delayed consent practical. Because of the nature of secondary findings, these people may not have personal or family history indications for testing, making it less likely that there would be clinically available avenues for testing at the time of adulthood. If population-based screening for adult-onset genetic conditions becomes routine in the future, a stronger argument might be made for restricting secondary finding analysis in minors to childhood-onset conditions, but that is not the current status of genomics. It is also worth noting that intellectual disability is sometimes an enrollment criterion for research involving exome sequencing or other MGT and may also have implications for future autonomy. Depending on the degree of intellectual disability, it may be unlikely that the minor will achieve a level of autonomy enabling them to make future choices about genetic testing and that their parent will remain their medical decision-maker into adulthood. In such cases the main argument against return of adult-onset findings is moot. Another reason in favor of seeking adultonset actionable secondary findings in minors is that the nature of these Mendelian conditions is such that in most cases they will be inherited from an affected parent. Thus identification of minors with adult-onset conditions enables cascade testing in their relatives and may result in the avoidance of morbidity or mortality in their own parents. The benefit of possible cascade testing is particularly relevant when MGT is performed only for the minor and not as a duo or trio with joint sequencing and analysis of parental samples.

51

52

CHAPTER 3  Selecting secondary findings to report

Study factors may also inform the selection of an appropriate set of genes. If the research design essentially involves collection of samples up front with little ongoing interaction with participants, a more limited list of genes that are both actionable and generally familiar to most medical practitioners may be appropriate. If a study is able to include aims related to the return of secondary findings, for example, studying participant responses to such findings or their adherence to recommended interventions, utilizing a somewhat larger list may provide richer data.

2.5 ­Variants Predefining a list of genes in which to seek out secondary findings helps at the time of analysis by restricting secondary finding analysis only to variants in those genes. Still, depending on the size of the gene list, it is likely that the majority of research participants will have one or more variants in the secondary finding genes. The next step is determining whether any of the identified variants warrant reporting. We recommend that variants in Mendelian disease genes being assessed for possible disclosure as secondary findings be interpreted according to the consensus recommendations of the ACMG and the Association of Molecular Pathology (AMP).21 The ACMG and AMP guidelines specify that variants are classified into one of five categories: pathogenic, likely pathogenic, variant of uncertain significance (VUS), likely benign, or benign. Generally, there is consensus that only variants that are considered to be pathogenic should be eligible for return as secondary findings. This policy ensures that the results are sufficiently predictive of disease risk to be clinically useful. Of note the ACMG and AMP standards and guidelines are intended for use in interpreting variants implicated in rare single-gene disorders. They are not appropriate for use in interpreting or understanding risk SNPs or pharmacogenomic results. What constitutes a reportable variant may vary by gene and disease, but VUS should not be reported as secondary findings. Because secondary finding analysis is performed without a clinical or family history indication, the a priori risk that any particular research participant has a rare Mendelian disease is low. Most genetic variants are benign, and VUS by definition lack sufficient evidence of pathogenicity to justify altering medical care on the basis of their presence. It is standard of clinical care not to utilize VUS results as a justification for medical interventions, and it is not recommended to test unaffected family members for VUS for the purpose of risk assessment. The ACMG and AMP variant interpretation guidelines state that “a variant of uncertain significance (VUS) should not be used in clinical decision-­making.”21 This is consistent across other groups. For example, a statement of the IARC Unclassified Genetic Variants Working Group includes an assertion that “Class 3 [uncertain] variants carry sufficient uncertainty such that no clinical predictive testing should be done on other relatives and further data are needed before any action is undertaken based on the test result.”22 Although the label of VUS exists in the middle of the variant classification categories, it should not be inferred that a VUS is equally likely to be pathogenic or benign. On the contrary, recently reported evidence from clinics and laboratories

2 ­ Criteria for a reportable secondary finding

performing clinical genetic testing for hereditary cancer syndromes suggests that variants initially classified as VUS are much more likely to be downgraded to likely benign or benign than to be upgraded when reclassified. In one retrospective study of variant reclassifications over a 10-year time span, among VUS that received a reclassification, 91.2% were downgraded, and only 8.8% were reclassified as likely pathogenic.23 In a prospective study that included variant reclassifications from clinic patients who had hereditary cancer testing from various labs, only 7.5% of VUS reclassification resulted in an upgrade, with the remainder being downgrades.24 While similar data have not yet been widely reported outside of the hereditary cancer realm, because variant interpretation guidelines are similar or the same for noncancer Mendelian diseases, it is reasonable to assume that this preponderance of reclassifications of VUS to likely benign or benign categories would be observed for other conditions. In addition, these data come from populations of patients who underwent clinical genetic testing for hereditary cancer genes, presumably because of personal and/or family history suggestive of such a condition. Because secondary finding analysis is conducted in the absence of a specific clinical indication, it is possible that VUS identified in this unselected population are even less likely to be actually pathogenic than in a clinical cohort. Because VUS by definition lack sufficient evidence to be interpreted as being causally implicated in disease, their predictive value in the setting of secondary findings is extremely low. This is somewhat in contrast to the more routine disclosure of VUS results in clinical diagnostic testing, where there is a much higher prior probability that the individual being tested actually has the condition related to the gene. In such cases, disclosing VUS can be of value for further studies (e.g., segregation in other affected and unaffected family members, follow-up enzyme testing, and additional functional studies) that could ultimately clarify whether the variant is pathogenic or benign. In the setting of secondary findings, there is a greater risk that participants will experience psychological distress or even be subjected to inappropriate medical intervention if a clinician unaware of the proscription against acting on VUS recommends alterations to their medical management.

2.6 ­Balancing risks and benefits The opportunity to learn secondary findings may be viewed as valuable by some research participants. Research participants in studies involving MGT have entrusted researchers with their genomic information, often with little prospect of personal benefit from participation. Handled well, return of secondary findings can demonstrate a researcher's respect for and beneficence toward the participants in his or her study. Yet even in ideal circumstances, return of secondary findings is not without risks. While some participants with secondary findings may, in retrospect, have clinical or family histories that are consistent with the finding, the majority of secondary findings by their nature are likely to come as a surprise. Psychological adjustment to this unexpected information may pose a challenge. Some MGT studies enroll participants with diseases known or suspected to have a genetic component. A secondary

53

54

CHAPTER 3  Selecting secondary findings to report

finding may add additional burden onto participants who already have complex and costly medical needs. On the other hand a secondary finding in an otherwise healthy person transforms that participant into a patient who needs medical intervention in the form of surveillance or preventative intervention. Even for gene-disease pairs that are prime candidates for return as medically actionable secondary findings, there is burden associated with engaging in the medical action. Interventions could range from as innocuous as a blood draw once per year to lifelong surveillance with imaging (echocardiogram, CT, and MRI), or more invasive procedures or prophylactic surgeries. Ideally the medical benefit to the patient outweighs the burden they take on by engaging in the screening or preventative act. Burden is admittedly subjective, but perception of burden may be higher in individuals without personal experience with the disease they are avoiding. Similarly a person's willingness to submit to certain medical care and physicians' confidence in recommending a course of action are different depending on whether the patient is at risk for disease or diagnosed with disease. This highlights another difference between the types of secondary findings that can be discovered. In one case the risk for disease might be directly treated with prophylactic surgery in the absence of any extant disease per se, while in other cases the recommended surveillance may ultimately reveal evidence of underlying disease that then necessitates a more significant (often surgical) intervention. Thus research participants who receive certain types of secondary findings may be faced with decisions about engaging in an invasive management strategy for a disease they may never have heard of previously and that they have no signs of, with only the strength of evidence of the pathogenicity of their specific variant to inform them. One could reasonably consider different standards for defining a reportable variant depending on the risk/benefit ratio of the disease and intervention. For example, perhaps either pathogenic or likely pathogenic variants might be returned for genes with interventions where initial steps of surveillance represent low risk and the riskier surgical intervention is only applied when disease is measurable. For those other conditions in which a more invasive procedure is the norm, it may be more appropriate only to return known pathogenic variants. The population prevalence of each condition is also relevant, as the a priori risk that a variant in a gene associated with an ultrarare condition is pathogenic is less than in a gene associated with a more prevalent condition. To minimize false positives or overtreatment, a higher standard may be appropriate for genes associated with rare phenotypes. Some gene-disease pairs make it possible to adjudicate variants through a gold-standard biochemical test that can definitively diagnose or rule out the condition. When the ability to adjudicate a variant exists, it reduces some burdens of secondary findings, such as the worry about whether disease will manifest or concern about engaging in screening or treatment before evidence of disease develops. For gene-disease pairs amenable to adjudication of variants with gold-standard tests, it might be more acceptable to report likely pathogenic variants and known pathogenic variants, compared with genedisease pairs without a clear means of variant adjudication.

3 ­ Lists

3 ­Lists A goal of this chapter is to assist you in making a secondary finding gene list that can inform your consent process, bioinformatic processing, and other aspects of the research enterprise. However, it may be equally useful to consider what might be called “negative lists,” or lists of genes for which return of variants as secondary findings will not be undertaken. These might be situations in which the risks of disclosure are particularly high or the benefits particularly low. For example, highly penetrant neurodegenerative diseases for which there is no disease-modifying therapy would be generally considered to have low “clinical actionability” and therefore would be less useful to disclose as secondary findings. Risks, including psychological distress and risk of discrimination, are high in these conditions. Uptake of predictive testing in the at-risk community, those that have close relatives with one of these diseases, has historically been low. An increased risk of attempted and completed suicide has been observed in some of these populations, and the time point after presymptomatic testing occurs may be a particularly vulnerable period.25,26 The predictive genetic testing process for diseases like these has developed to reduce these risks and typically involves in-person pretest and posttest counseling, often over multiple sessions, and multidisciplinary input from genetics, psychology, and neurology. These conditions cannot and should not be replicated for all participants in a research study involving MGT, but neither should this type of result be returned cavalierly without any of the typical processes in place. Another type of result that could justifiably be put on a list of nonreturnable results is most risk SNP information, representing the low benefit end of the spectrum. As described previously in this chapter, risk SNPs have significant limitations when attempting to predict a specific individual's risk, and often any medical action that could be recommended would also be recommended for a person at average or even decreased risk (e.g., maintaining a healthy weight, engaging in regular exercise, and being a nonsmoker). It has been proposed that specific knowledge of one's own risk SNPs may provide a motivation for engaging in a generally recommended health behavior; however, research has failed to demonstrate this. For example, there was no association between a participant's knowledge of their genetic risk for obesity and either self-reported or objectively measured physical activity over the course of a 6-month study period.27 Overall, there is a lack of evidence supporting the idea that risk SNP information actually impacts behavior change (or at least sustained behavior change) for behaviors like increasing physical activity, diet modification, or smoking cessation.28–30 While there may still be value to personalized health coaching involving genetic risk information, it seems clear that SNP information alone would not be expected to meaningfully improve health. Research should continue to investigate programs aimed at health behavior change and the role of risk SNP information in these programs, but this type of information is not likely to be a benefit to participants if returned as a secondary finding.

55

56

CHAPTER 3  Selecting secondary findings to report

Despite the likely limited beneficial impact of risk SNP information, it seems that many find this desirable information. The existence of direct-to-consumer ­testing products that disclose risk SNP information shows that at least some people would pay for such information. In our own research, study participants undergoing exome sequencing were asked to make a choice about learning various types of nonmedically actionable secondary findings, and most (78%) intended to request at least some categories of this information.31 Some people may find personal utility in learning this type of information, even when clinical utility is lacking. This personal utility may vary from person to person and could involve satisfaction of curiosity, self-understanding, or motivation for behavioral change. Because of the individual variation in perceptions of personal utility, it cannot practically be used by researchers to inform decisions about what to return as a secondary finding. When defining the clinical utility of knowing a particular secondary finding, there may be a paucity of data, but experts will generally agree on the types of data that would be useful in determining actionability. Personal utility is completely individual and subjective, meaning it is not useful for defining a priori a set of secondary findings that makes sense for a particular research study and group of future participants. One strategy for defining a set of secondary findings to return is to utilize an existing published list. Among the first such list was that of the American College of Genetics and Genomics (ACMG). In 2013 ACMG proposed a list of 56 gene-disease pairs as a minimum standard for clinical labs performing “clinical germline (constitutional) exome and genome sequencing” to use to “seek and report” secondary findings.32 This list was often referred to in the literature as the “ACMG 56.” A 2016 update (published in 2017) resulted in a 59 gene list, with the new version of the ACMG list being called “ACMG SF v2.0.”14 Although this list was intended for use by clinical laboratories performing diagnostic exome or genome sequencing and not necessarily as a guide for researchers, this list has become somewhat of a standard, making it a reasonable default for researchers who do not wish to go through a process to define their own list of returnable genes. On the other hand, there may well be study-specific context or expertise that favors a smaller or larger list of secondary findings.

4 ­Conclusion Researchers who are generating large-scale sequencing data on their participants must consider what types of findings should be reported as secondary findings. The answer may depend to a great extent on the nature of study, the proximity between the participants and the research team, the institutional ethics oversight, ease of analysis and reporting, and resources available for result disclosure and genetic counseling. Such decisions should be made as part of the study design (and indeed will be required for an IRB protocol approval), budgeted for and conducted with care.

­References

­References 1. Jarvik GP, Amendola LM, Berg JS, et al. Return of genomic results to research participants: the floor, the ceiling, and the choices in between. Am J Hum Genet. 2014;94(6):818–826. 2. BMC. A Genetic Variant Near Olfactory Receptor Genes Influences Cilantro Preference. https://flavourjournal.biomedcentral.com/articles/10.1186/2044-7248-1-22; 2012. (Accessed May 21, 2019). 3. NCBI. ClinVar. https://www.ncbi.nlm.nih.gov/clinvar/variation/17677/; 2019. (Accessed May 5, 2019). 4. Rivera  A, Fisher  SA, Fritsche  LG, et  al. Hypothetical LOC387715 is a second major susceptibility gene for age-related macular degeneration, contributing independently of complement factor H to disease risk. Hum Mol Genet. 2005;14(21):3227–3236. 5. Bielinski SJ, Olson JE, Pathak J, et al. Preemptive genotyping for personalized medicine: design of the right drug, right dose, right time-using genomic data to individualize treatment protocol. Mayo Clin Proc. 2014;89(1):25–33. 6. Sadee W, Dai Z. Pharmacogenetics/genomics and personalized medicine. Hum Mol Genet. 2005;14. Spec No. 2:R207-214. 7. U.S. Food & Drug Administration. Table of Pharmacogenomic Biomarkers in Drug Labeling. https://www.fda.gov/Drugs/ScienceResearch/ucm572698.htm; 2019. (Accessed May 5, 2019). 8. Kimmel SE, French B, Kasner SE, et al. A pharmacogenetic versus a clinical algorithm for warfarin dosing. N Engl J Med. 2013;369(24):2283–2293. 9. Pirmohamed  M, Burnside  G, Eriksson  N, et  al. A randomized trial of genotype-guided dosing of warfarin. N Engl J Med. 2013;369(24):2294–2303. 10. Strande NT, Riggs ER, Buchanan AH, et al. Evaluating the clinical validity of gene-­disease associations: an evidence-based framework developed by the clinical genome resource. Am J Hum Genet. 2017;100(6):895–906. 11. Berg JS, Foreman AK, O'Daniel JM, et al. A semiquantitative metric for evaluating clinical actionability of incidental or secondary findings from genome-scale sequencing. Genet Med. 2016;18(5):467–475. 12. Milko  LV, O'Daniel  JM, DeCristo  DM, et  al. An age-based framework for evaluating ­genome-scale sequencing results in newborn screening. J Pediatr. 2019;209:68–76. 13. Webber EM, Hunter JE, Biesecker LG, et al. Evidence-based assessments of clinical actionability in the context of secondary findings: updates from ClinGen's Actionability Working Group. Hum Mutat. 2018;39(11):1677–1685. 14. Kalia SS, Adelman K, Bale SJ, et al. Recommendations for reporting of secondary findings in clinical exome and genome sequencing, 2016 update (ACMG SF v2.0): a policy statement of the American College of Medical Genetics and Genomics. Genet Med. 2017;19(2):249–255. 15. Dorschner MO, Amendola LM, Turner EH, et al. Actionable, pathogenic incidental findings in 1,000 participants' exomes. Am J Hum Genet. 2013;93(4):631–640. 16. Ceyhan-Birsoy O, Machini K, Lebo MS, et al. A curated gene list for reporting results of newborn genomic sequencing. Genet Med. 2017;19(7):809–818. 17. Brothers KB, East KM, Kelley WV, et al. Eliciting preferences on secondary findings: the preferences instrument for genomic secondary results. Genet Med. 2017;19(3):337–344. 18. Botkin  JR, Belmont  JW, Berg  JS, et  al. Points to consider: ethical, legal, and psychosocial implications of genetic testing in children and adolescents. Am J Hum Genet. 2015;97(1):6–21.

57

58

CHAPTER 3  Selecting secondary findings to report

19. Clarke A. The genetic testing of children. Working party of the clinical genetics society (UK). J Med Genet. 1994;31(10):785–797. 20. Knoppers  BM, Avard  D, Senecal  K, Zawati  MH. Return of whole-genome sequencing results in paediatric research: a statement of the P3G international paediatrics platform. Eur J Hum Genet: EJHG. 2014;22(1):3–5. 21. Richards S, Aziz N, Bale S, et al. Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology. Genet Med. 2015;17(5):405–424. 22. Plon SE, Eccles DM, Easton D, et al. Sequence variant classification and reporting: recommendations for improving the interpretation of cancer susceptibility genetic test results. Hum Mutat. 2008;29(11):1282–1291. 23. Mersch J, Brown N, Pirzadeh-Miller S, et al. Prevalence of variant reclassification following hereditary cancer genetic testing. JAMA. 2018;320(12):1266–1274. 24. Slavin  TP, Van Tongeren  LR, Behrendt  CE, et  al. Prospective study of cancer genetic variants: variation in rate of reclassification by ancestry. J Natl Cancer Inst. 2018;110(10):1059–1066. 25. Farrer LA. Suicide and attempted suicide in Huntington disease: implications for preclinical testing of persons at risk. Am J Med Genet. 1986;24(2):305–311. 26. Almqvist EW, Bloch M, Brinkman R, Craufurd D, Hayden MR. A worldwide assessment of the frequency of suicide, suicide attempts, or psychiatric hospitalization after predictive testing for Huntington disease. Am J Hum Genet. 1999;64(5):1293–1304. 27. Marsaux CF, Celis-Morales C, Livingstone KM, et al. Changes in physical activity following a genetic-based internet-delivered personalized intervention: randomized controlled trial (Food4Me). J Med Internet Res. 2016;18(2):e30. 28. Marteau  TM, French  DP, Griffin  SJ, et  al. Effects of communicating DNA-based disease risk estimates on risk-reducing behaviours. Cochrane Database Syst Rev. 2010;(10). Cd007275. 29. de Viron S, Van der Heyden J, Ambrosino E, Arbyn M, Brand A, Van Oyen H. Impact of genetic notification on smoking cessation: systematic review and pooled-analysis. PLoS One. 2012;7(7):e40230. 30. Hollands GJ, French DP, Griffin SJ, et al. The impact of communicating genetic risks of disease on risk-reducing health behaviour: systematic review with meta-analysis. BMJ. 2016;352:i1102. 31. Rini C, Khan CM, Moore E, et al. The who, what, and why of research participants' intentions to request a broad range of secondary findings in a diagnostic genomic sequencing study. Genet Med. 2018;20(7):760–769. 32. Green RC, Berg JS, Grody WW, et al. ACMG recommendations for reporting of incidental findings in clinical exome and genome sequencing. Genet Med. 2013;15(7):565–574.

CHAPTER

How secondary findings are made

4

Kevin M. Bowling, Michelle L. Thompson, Gregory M. Cooper HudsonAlpha Institute for Biotechnology, Huntsville, AL, United States

1 ­Secondary findings in clinical sequencing Technological developments have, over recent years, drastically improved the cost, speed, and quality of DNA sequencing methods. Owing to such improvements, protocols to sequence large subsets of the DNA present in any given individual in a single assay have become commonplace, including within a variety of clinical contexts. As a consequence of widespread implementation of large-scale sequencing in clinical molecular diagnostics, identification of genetic variation not associated with indication for testing has become prevalent.1–4 The presence of secondary or incidental findings poses not only challenges but also opportunities, for both clinicians and patients, and is an important consideration when doing large-scale DNA clinical sequencing. While the terms “secondary” and “incidental” are often used interchangeably, these types of findings can be quite different, and distinction of the two is important. Incidental findings are accidentally detected when conducting genetic variant analysis to identify variation associated with the primary indication for testing. In contrast, secondary findings represent variation that is not related to the indication for testing, but which is proactively pursued during analysis as a means of identifying variation that may be of medical relevance for the patient or study participant despite the absence of reported symptoms. For example, in an attempt to identify causal genetic variation in a child with a seizure disorder, one may search through all genetic variants that appear de novo (assuming parent samples are available for testing) and in the process of analysis detect a de novo variant in BRCA2 that puts the child at increased risk of cancer later in life, even though this was not relevant to the seizures, which prompted genetic testing. In contrast, genetic variants may be proactively sought out by clinical genetics or research laboratories in specific genes (e.g., ACMG SFv2.05, 6). In this case, genetic variation is not identified incidentally and is thus considered a secondary result. It is important to note that incidental and secondary findings are often not expected by patients or study participants, even in the case where pregenetic counseling occurs. As implementation of clinical sequencing expands, there is an increased concern among medical professionals regarding return of incidental and secondary findings. Secondary Findings in Genomic Research. https://doi.org/10.1016/B978-0-12-816549-2.00004-7 © 2020 Elsevier Inc. All rights reserved.

59

60

CHAPTER 4  How secondary findings are made

In response to a desire for standardized policy related to secondary finding identification and return, the American College of Medical Genetics and Genomics (ACMG) recommended that labs performing exome and genome sequencing report secondary findings to consented individuals who are found to harbor pathogenic findings in genes associated with clinically actionable disease, where actionability is defined on the basis of therapeutic and preventative options being available. As part of this recommendation, the ACMG published a specific gene list that focuses primarily, but not exclusively, on adult-onset cancer and heart disease.5, 6 With the launch of a number of large-scale genome projects (All of Us, NIH; Japan’s Initiative on Rare and Undiagnosed Diseases; United Kingdom’s 100,000 Genomes Project), the potential impact of identification and return of medically relevant secondary findings is on the rise. A large proportion of US clinical laboratories utilize the ACMG SFv2.0 gene list7; however, there are some US groups that use a modified gene list that may include carrier status (recessive disorders), may be a shortened version of the ACMG SFv2.0 gene list (i.e., 41 of the 59 ACMG SFv2.0 genes), and/or contain pharmacogenomic markers.7 The international genetics community is even more divided in regard to use of the ACMG SFv2.0 gene list primarily because of differences in culture and/or policy. Some international labs have adopted the ACMG SFv2.0 gene list and return secondary findings if patients have opted in, while others feel that returned results should be restricted to the patient’s primary indication for testing8, 9 (see Chapter 8 for more details on this topic). While consensus does not exist about which secondary results should be returned to patients, there is a consensus that patients and/or study participants should be informed of the possibility of incidental or secondary findings and given the ability to choose whether or not they receive such findings, should they be detected.10

2 ­Chapter overview In this chapter, secondary findings will be discussed in the context of exome and genome sequencing (ES/GS, respectively) being performed to identify germline variation in individuals suspected to have a genetic disorder; we note that detection of other types of variation, such as somatic mosaicism and tumor sequencing, are substantially different and are not discussed here. The workflows associated from taking a clinical blood draw to creating of a variant call format (VCF) file will be detailed (Fig.  1).11 Topics to be covered include the basic mechanics of genomics assays, Illumina sequencing technology (as the current predominant sequencing platform used in clinical genomic labs) and library preparation, sequence read alignment, variant calling, variant annotation, variant filtration, and variant analysis. Long-read sequencing technologies, which are not currently part of typical clinical testing but which have the potential to be beneficial to sequencing laboratories in the future, will not be reviewed. Potential secondary finding gene/disease categories will also be discussed, as will places in the workflow where discovery of secondary findings can be actively avoided, or alternatively, are unavoidable. Further, we will argue that variant validation via an alternative sequencing assay is needed to mitigate risk of

3 ­ Current methods for large-scale DNA sequencing in clinical laboratories

Blood draw from patient/ study participant

secondary findings

DNA isolation from whole blood

Sequencing library construction

Variant interpretation

Library capture if exome sequencing

Employ secondary findings gene list

Exome/genome sequencing

Variant filtration

Read alignment, variant calling, annotation

Quality control

FIG. 1 Workflow from sample intake to identification of secondary findings. Adapted from Fig. 1 of Neu MB, Bowling KM, Cooper GM. Clinical utility of genomic sequencing. Curr Opin Pediatr 2019;31(6):732-738. PMID 31693580.

false-positive return, although that may change in the future if certain technical accuracy and reliability standards can be met without such validation. There are a number of topics related to identification and return of secondary genetic variation that are relevant both up- and downstream of sequencing and analysis. Some of these may be mentioned in this chapter but are discussed in greater detail in other chapters (see Chapters 2 and 3). These topics include variant classification based on ACMG guidelines, clinical utility of secondary findings and how clinical usage of such findings is clinician-dependent (regardless of variant classification by the laboratory), the fact that secondary findings are typically unexpected by patients and study participants, policies and guidelines as they relate to secondary finding return, and finally patient rights to opt-in or opt-out of secondary finding disclosure.

3 ­Current methods for large-scale DNA sequencing in clinical laboratories The accessibility, affordability, and utility of modern sequencing methods have led to uptake of this technology by many molecular diagnostic labs. Illumina is the current leading provider of large-scale DNA sequencing platforms, which will be the focus

61

62

CHAPTER 4  How secondary findings are made

of this chapter. Over the past decade the use of ES, GS, and targeted gene panels have become prevalent in molecular diagnostic labs to detect genetic variation in patients with rare disease. GS, which is sometimes referred to as “whole genome sequencing” (WGS), aims to cover all bases in the genome (totality of DNA), including coding and noncoding regions, while ES targets only the protein-coding regions (gene exons) and their flanking DNA. Targeted gene panels capture and sequence the exons of only a specific set of genes, typically chosen based on their suspected relevance to the indications for testing. While the specific sequencing test ordered for genetic diagnosis is dependent upon a variety of factors (ordering physician, patient phenotype, disease-gene hypotheses, insurance reimbursement, etc.), genome-wide assays (ES and GS) are becoming more widely used as they become increasingly cost-effective and are far more comprehensive at detecting medically relevant genetic variation than gene panels. There is some debate within the clinical genetics community regarding the use of ES versus GS. The pros and cons of each assay are highlighted in Table 1. ES can be

Table 1  Differences between exome and genome sequencing. Pros

Cons

Exome sequencing (ES)

• Lower sequencing cost • Reduced data storage needs • Reduced analytical costs • Faster data processing

Genome sequencing (GS)

• Better coverage of the exome and coverage uniformity • Coverage of noncoding regulatory regions of the genome • No library capture required • Future sequencing may not be required, only reanalysis • Ability to call structural variation • Less sequencing depth required (compared with exome)

• Library capture required • Poor/inconsistent coverage of some exons and poor coverage uniformity • Unreliable detection of structural variation • Omits noncoding regulatory regions • More sequence depth required (compared with genome) • Higher cost • Increased data storage needs • Increased analytical costs • Slower data processing • Analysis and interpretation of noncoding regulatory regions is difficult

3 ­ Current methods for large-scale DNA sequencing in clinical laboratories

done at somewhat lower cost than GS, has lower data storage and analysis costs, and results in smaller datasets that can be processed and analyzed slightly more quickly. However, it also requires an extra experimental step (e.g., the targeted capture of coding exons), may result in poor or inconsistently covered exons (coverage refers to the number of times a specific base has been sequenced and correlates with accuracy in variant identification), and fails to assess some important functional regions of the genome, such as those that govern transcriptional regulation. Exome sequencing also substantially reduces the ability to detect structural variation like deletions and duplications. In contrast, GS results in better, more uniform coverage and in fact tends to more effectively capture the coding exons than does ES, includes gene regulatory regions, requires no experimental targeted capture step, and produces data that is far better for detecting structural variation. GS, however, has greater sequencing costs than ES, produces very large amounts of data that require increased storage, and imposes more time-consuming data processing and analysis steps. Further, interpretation of genomic regions outside of coding exons is difficult and has proven to have, thus far, minimal clinical value.12–14 It should be noted that the ratio of GSto-ES costs, including both sequencing and compute costs, has steadily declined over recent years. Additionally, both basic research efforts and the increased use of GS in large disease research have improved upon the ability to interpret gene regulatory regions. Combined, these factors make it likely that GS will steadily displace ES as a test for the diagnosis of rare disease. Conducting ES or GS begins with genomic DNA that is isolated from a patient sample. While ES/GS can be done from saliva, buccal swabs, or other sources of cells, the most typical source is patient blood leukocytes. Preparation of isolated genomic DNA for sequencing requires shearing of the large DNA molecules by sonication (i.e., fragmenting DNA molecules into small segments) and addition of adapter sequences to fragmented DNA molecules that will eventually facilitate attachment of those fragments to a glass slide. This slide is termed a “flowcell,” or “lane” within a flowcell. These terms are described in more details later, but we will use sequencing “lane” as a term to describe samples that are sequenced in the same space on a sequencing machine. For some, but not all library preparation protocols, a small number of PCR cycles may be used to amplify, or make multiple copies of, the fragments in the sequencing library. It is at the library preparation stage that samples may be barcoded to allow for sequencing multiple distinct samples within the same sequencing lane. Pooling of barcoded samples allows the lab to use sequencing lanes more efficiently, as machines often produce far more sequence “reads” (reads, defined in more detail later, are the stretches of DNA sequence that are identified by the sequencing machine) in a given lane than are needed to obtain the desired depth of coverage for a given sample. These barcodes are short, synthetically designed sequences that are uniquely assigned to a given sample. They are typically embedded within the adapters, described earlier, that are ligated to the fragmented genomic DNA. These barcodes can later be read on the sequencer and used to assign genomic sequence reads to the sample from which they originate. Further, barcoded pools of many samples can be

63

64

CHAPTER 4  How secondary findings are made

sequenced on multiple lanes, reducing the lane-specific variation (e.g., output levels, errors, and other factors) that might affect data quality. Following library preparation, if ES (or other large-scale targeted strategy) is being used, capture of the fraction of the library representing exonic regions of the genome is achieved using complementary oligonucleotide probes. Enrichment of the desired targets results from hybridization between the genomic DNA fragments and the provided oligonucleotide probes. The earliest implementations of ES involved hybridization to microarrays, so as to allow washing away of unhybridized/nontargeted genomic DNA. However, typical ES protocols now involve hybridization in solution15 with isolation of hybridized fragments achieved through use of magnetic beads and streptavidin/biotin; such protocols are far more amenable to high-throughput workflows than those that require array capture. After generation of ES or GS sequencing libraries, DNA molecules are put onto a flowcell. A flowcell is a glass slide containing lanes, or channels, across which the sequencing library will “flow.” This slide contains DNA sequences that are complementary to the previously described adapter sequences and will allow for the attachment of DNA molecules to be sequenced. After a DNA library is loaded onto a flowcell, the fragments attached to the flowcell are amplified via a specialized PCR process to generate “clusters” of identical copies of each attached fragment. After cluster generation the flowcell can then be loaded into a sequencing machine, such as the Illumina HiSeq X or NovaSeq. Sequencing is achieved by DNA synthesis using fluorescently labeled nucleotides, with differing colors for each of the four possible nucleotides (A, C, G, or T). During each “cycle,” a single-labeled nucleotide is incorporated into the nascent DNA strands growing at each cluster, and a camera captures images of each cluster after each cycle. Specialized software then analyzes these images to determine, by virtue of the intensities of each possible color, which nucleotide was incorporated into each cluster at each cycle. Over the course of many cycles, typically in the range of 100–300, a sequence “read” for each cluster is generated. Typically, a read is generated from both ends of the DNA fragments that were loaded onto the flowcell, and these are thus referred to as “paired-end reads.” Through this process, hundreds of millions or billions of DNA sequence reads can be generated simultaneously. These reads are then output into a “fastq” file for downstream analysis. The fastq files contain the raw sequence reads, along with the information relating to the quality of each base call within each read. Base quality scores are an important component of downstream analyses including alignment and variant calling.

4 ­Sequence read alignment, variant calling, and annotation A fastq file will typically contain many millions of reads for a particular sample. For example, over 300 million, 150 base pair paired-end reads (300 base pairs total for each library fragment) are needed to exceed 30 × mean coverage across the reference genome for GS, a depth considered to be a good target by many ­sequencing

4 ­ Sequence read alignment, variant calling, and annotation

l­aboratories. For ES, ~ 100 million, 100 base pair paired-end reads are needed to reach 75–100 × mean coverage over 64 megabases of targeted exonic content. However, average read depths are only one way to measure sequence coverage; metrics that capture uniformity of coverage, which depend on sequence alignment, are more informative (see in the following section). Before genetic variation within a sample that has been sequenced can be identified, the raw reads of the fastq file must be aligned to a reference genome (using an aligner software package such as Burrows-Wheeler Aligner (BWA16). It is at this analytical step that reads are assigned to specific locations within the genome, and this ultimately allows downstream detection of variation. During read alignment, quality scores are calculated to determine how well the read maps are to a specific place within the genome (MAPQ score). This metric can be used, along with other metrics, to determine confidence in genetic variant calls. Once reads are mapped to a reference assembly, a variety of quality control (QC) metrics can be computed to determine the effectiveness and accuracy of the sequencing data. For example, one is often interested in assessing how well the exome or genome was covered as a means to assess overall potential variant sensitivity. While mean coverage levels are often used, which can be inferred simply from the total number of reads generated divided by the total target space (as discussed earlier), uniformity of coverage is also important to consider. High average coverage is of less value if reads are distributed such that a few regions receive extremely high coverage while many receive little to no coverage. Indeed, variant calling accuracy (in the context of germline variant detection) at a given position depends on the number of reads at that position. In turn, variant calling accuracy rises dramatically as read depth increases past a few reads but tends to plateau after 20 such that increases beyond this level lead to minimal accuracy gains. As such, metrics for sequencing coverage that describe the percentage of bases covered at some minimal depth level are essential. For example, it is often desirable in GS to have > 80% of all bases covered by > 20 reads, with > 90% at 20 × often achieved, and ES targets similarly may be > 80% (or > 90%) of exonic bases at > 20 reads. It is important to keep in mind that coverage at a given position/region can vary randomly from sample to sample, and there are also regions in both ES and GS that tend to be consistently poorly covered. These regions may contain genes that are known to be clinically relevant, and these coverage limitations may interfere with identification of secondary or incidental findings.17 Coverage variation and its consequences on variant calling sensitivity must be taken into consideration, especially when returning negative findings to patients or study participants. Further, if there are specific genes that are of particular interest given the indication for testing, it may be necessary to assess coverage levels specific to that gene for each sequenced sample to determine whether or not the individual may harbor variation in that gene that was simply missed due to coverage limitations. After sequence alignment and associated QC are performed, variant calling is done. Variant calling is the process by which genomic locations are identified where the assayed individual harbors a sequence that does not match that of the reference

65

66

CHAPTER 4  How secondary findings are made

sequence (Genome Analysis Toolkit [GATK]18). Within the context of detecting germline variation in humans, which are diploid (i.e., harbor two allelic copies at any given genetic location, one maternal and one paternal), the goal is to identify sites where an individual is heterozygous for the reference sequence and an “alternative” (i.e., nonreference) sequence or sites where an individual is homozygous for an “alternative” sequence. For example, if a particular base is covered by 38 reads and 19 of those reads match the reference “T,” while 19 match the alternate call “A,” the sample is likely to be flagged as being a T/A heterozygote at that particular position. As one can imagine, reads are often not perfectly balanced for heterozygous variation, and there will often be sequencing errors such that one or more reads incorrectly appear to have an alternative sequence at a particular position. Therefore, making accurate variant calls can be challenging. This is particularly true for insertion or deletion variants (“indels”), for which alignment is more challenging and which tend to exhibit much higher error rates than do single-nucleotide variants (SNVs). Current variant calling methods, including the predominant standard known as “GATK,”18 tend to be relatively complex pipelines that account for many features. Such features include base and base quality scores of each read, alignment quality/confidence, the presence/absence of variation at a given location in population variant databases, local sequence context, and the position of a base within a read or set of reads. Detailed descriptions of these approaches are well outside the scope of this chapter, but see19, 20 for additional information. Most approaches, including GATK, produce variant calls that include the sequence(s) observed at a given location and the genotype of those sequences (i.e., homozygous alternate and heterozygous). In addition, a variety of useful features for each call are also generated, including quality metrics related to the estimated confidence of the call and the number of reads that support each inferred allele. Such data are typically output into a “VCF” file, which is a format specifically designed to capture variant calling results from a sample or set of samples. One general property that is relevant to variant calling is the distinction between single-sample and “batch” variant calling; for the former, each sample is analyzed independently of all other samples, while in the latter the data derived from a batch of samples is used in aggregate to identify variants in each sample. While singlesample calling is logistically simpler, batch calling can provide some advantages. For example, in the case that an individual within a batch has a small minority of reads that suggest the presence of a specific alternate variant, the observation of the same specific alternative variant in other samples within the same batch makes it more plausible, and the variant can be called within the sample in question with more confidence. While this works quite effectively for common variants (since common variants are more likely to be seen in multiple individuals within a batch), such a strategy can also be effective for rare variants, especially when multiple members of the same family are sequenced together. Once variants have been called, they must be annotated before moving forward with variant filtration and analysis. The raw VCF contains information related to chromosome, position, reference sequence, alternate sequence, quality, batch allele counts, and other products of the variant calling process. The information within this

5 ­ Use of variant filtration to detect or avoid secondary findings

file must be annotated with biological information to understand the genes affected by variation and to understand potential disease consequences. There is a plethora of different annotations used by clinical and research laboratories, and the specifics of this process can vary considerably between labs. However, in general, annotation features include, but are not limited to, protein effect, population frequency data (e.g., gnomAD21), conservation scores, computational predictions of deleteriousness, and presence in clinical disease databases (e.g., ClinVar22 and HGMD23), among many other possibilities. For a more complete assessment of how annotations can be used to predict disease relevance, see Refs.24–26. After annotations have been added to identified genetic variation, calculation of VCF statistics aides with quality control. An exome or genome should, for example, contain a predictable number of variants globally and per chromosome, and these variants should be distributed in certain percentages across different categories of annotation (missense, nonsense, splice, synonymous, etc.). If a particular sample is an outlier for one or more of these measures, that may be an indication of problems in the sequencing, alignment, or variant calling process for that sample.

5 ­Use of variant filtration to detect or avoid secondary findings After variants have been called and annotated, they can be filtered based on their annotations and pattern of inheritance to reduce the number of variants that requires curation. Such steps are, in fact, essential, as ES typically yields tens of thousands of variants in each person and GS yields millions, very few of which put individuals at high risk for disease. One of the most critical annotations for filtering is variant frequency information from population genetic databases.21 Frequency of a variant in the general population, as a general rule, cannot exceed the prevalence of that disease, assuming it is fully penetrant. As such, when one is searching for highly penetrant variants that contribute to the risk of a rare disease, removing any variant seen at appreciable frequencies in the general population is a highly effective step at removing benign variants, or at least those not relevant to the disease in question. While specific frequency thresholds depend on a variety of factors including the prevalence of the disease, mode of transmission, expected penetrance, ancestry of the particular individual, and ancestries of individuals in the variant database, typical thresholds tend to be much lower than 1% (often even below 0.1%), especially for rare, dominant conditions, and such thresholds can remove the vast majority, typically > 90%, of the variation in any particular individual. When sequencing an affected individual along with other affected or unaffected family members, patterns of familial inheritance can also be useful. For example, when analyzing a severely affected child of two healthy parents, one effective filtering strategy is to identify genetic variants that are de novo, that is, present in the child but absent from his/her parents. Moreover, filters can also be designed to identify, for example, compound heterozygous, recessive, or maternally inherited hemizygous

67

68

CHAPTER 4  How secondary findings are made

variation (i.e., X-linked disease variants in an affected male proband). Inheritance pattern alone can dramatically reduce lists of identified genetic variation. Aside from inheritance pattern, annotations that are useful for variant filtration include effect on gene protein product, algorithmic predictions of variant deleteriousness, conservation scores, and information contained within clinical disease databases. This list of annotations is not comprehensive, and annotations applied and utilized in variant filtration vary considerably among clinical and research laboratories. It is also important to note that batch (when variants are called in a sample batch) and internal (variants within a clinical sequencing lab) allele frequencies may be informative for detecting false positives that result from site-specific errors, whether they be generated via sequencing or are an artifact of the analytical pipeline. Identified variants that are observed multiple times within a given batch, for example, or have been seen numerous times in samples sequenced at a specific site, but appear to be very rare in or are absent from external population frequency databases, are suspect and should be treated as such. It is during variant filtration that secondary findings may be selected for or avoided and when incidental findings may be detected (also see Chapter 3). Based on recommendations by the ACMG, many clinical laboratories in the US actively search for pathogenic genetic variation in the list of 59 genes (ACMG SFv2.05, 6) that have been defined as clinically actionable. This gene list, along with the aforementioned annotations, may serve as a filter for identification of secondary findings. To the converse, one might use a gene “blacklist” to reduce the chance of identification of incidental genetic findings; however, this is rarely done in practice. Moreover, this can be problematic depending on the phenotype of the patient or patient cohort, as some genes may have implications for both primary and secondary findings. For example, PTEN is a gene included on the ACMG SFv2.0 list of 596 given its association with cancer risk, but PTEN is also associated with a neurodevelopmental disorder.27 It would thus be a problem to filter out all genetic variation in PTEN, as the effort to avoid detecting secondary or incidental genetic variation in this gene may also result in missing a primary finding. Even when attempting to avoid secondary findings, it is plausible that a finding not associated with the indication for testing will be identified (i.e., incidental finding). A filter designed to detect de novo variation in a disease-affected individual that has been sequenced as part of a proband-parent trio may, for example, expose a variant in a well-established cancer or heart-disease risk gene. In this case the variant is truly an incidental finding, and its discovery is unavoidable. Finally, interpretation of secondary findings in a patient or patient cohort can be challenging given that such findings often relate to diseases that are not manifest in the sequenced individual, and searching for secondary variants can be time consuming and likely requires more manual curation than findings linked to the patient’s primary phenotype. Detailed personal and family history of disease can be useful when conducting analysis to detect secondary genetic findings. It may also be valuable to consider the preferences, if available to the analysis lab, of the patient and/or family when conducting secondary variant analysis, as time spent proactively looking for

6 ­ Potential categories of secondary findings

variation linked to nonprimary disease genes can obviously be avoided if the patient/ family has opted out.

6 ­Potential categories of secondary findings In the context of secondary genetic variation, there are many categories of findings that a clinical or research laboratory may decide to return. In short, these categories may include variation in genes on the ACMF SFv2.0 list,6 variants listed as pathogenic or likely pathogenic in a clinical disease database (e.g., ClinVar28), variants associated with cancer risk variation either inside or outside the ACMG SFv2.0 list, variants linked to treatable and untreatable adult-onset disorders, variants in ­pediatric-onset disorder genes, carrier status for variation in recessive disease genes, and pharmacogenomic markers. We note that these listed categories are neither comprehensive nor mutually exclusive, but rather represent different types of approaches that are often used to organize secondary findings and/or within which one might make decisions about filtration, interpretation, return, etc. Consensus regarding which categories of secondary findings should or should not be returned does not exist among clinical genetics laboratories; however, numerous studies have investigated patient preferences for return of secondary findings organized into different groupings.29–32 The ACMG prioritizes variation in genes for conditions that pose a major threat to health and for which standard treatment or preventative measures are available.5 In contrast, studies have shown that patients and study participants prefer report of variation based on the expected severity and likelihood of disease.29 Moreover, many studies suggest that, when it comes to receipt of secondary findings, most individuals want to know “everything.”3, 32–36 From the perspective of the clinical laboratory, analysis and curation of secondary genetic findings can be a large task, depending on the desired categories of return. Several studies have summarized the reporting practices of genome sequencing laboratories in the United States.7, 37 In practice, return of secondary results to participants may range from less than 2% of participants when returning only pathogenic or likely pathogenic variation in ACMG SFv2.0 genes3, 38–40 to greater than 90% when returning comprehensive carrier results.1 As more clinical and research laboratories have converged on the principle of searching for and providing at least some sets of secondary findings to patients and study participants, development of standard practices related to categories of return has begun, as has development of standardized consent documentation and opt-in/opt-out options.37, 41 One category of secondary findings that may have a large impact on personalized medicine is pharmacogenomics. Adverse drug reactions and ineffective drug choice or dosing greatly impact morbidity and mortality rates and healthcare costs.42–48 Because of genetics research the understanding of patient drug response variability has advanced greatly over the last decade. Pharmacogenomic variants may allow physicians to make more effective decisions about patient medication, both in terms of which drugs to use or avoid and what dosage to prescribe. Newfound knowl-

69

70

CHAPTER 4  How secondary findings are made

edge about drug efficacy, dosage, and variable response, in light of association with specific genotypes, has decreased detrimental effects.49 Studies suggest that, of all individuals tested for variation that alter drug response, most harbor at least one allele that could alter response to medication.49 In fact, everyone has common variant genotypes that are relevant to predicting drug response; for example, common alleles near VKORC1 are correlated with effective dosage levels of the blood-thinner warfarin.50–55 However, pharmacogenomics secondary findings pose special challenges to clinical laboratories and are not typically sought or returned by clinical labs, at least those focused on diagnosis of rare genetic disease. Beyond the fact that returning common pharmacogenetically relevant variants would lead to a massive increase in rates of variant interpretation and return, specialized filters must be established to detect such variation given that clinical labs focused on rare diseases typically filter out all variants that are at appreciable frequencies in the general population (as discussed earlier). As such, there are currently a wide variety of approaches by clinical labs to secondary pharmacogenomic results within ES or GS data, with many labs eschewing such analysis altogether.

7 ­Variant validation via alternative sequencing assay False positives do exist in ES/GS data. While some may be flagged by quality metrics (e.g., base quality scores, read mapping qualities, variant call quality scores, and coverage levels), filtration on quality metrics alone may not completely remove all false calls as there will always be some that meet quality control metric cutoffs. Further, as QC stringency increases to reduce false positives, the rate at which true positives fail QC metrics will increase and lead to reduced variant sensitivity. Moreover, certain classes of biologically interesting but rare variants are at higher intrinsic risk of being false discoveries owing to the fact that their true-positive rates are low.56 This is particularly true for variants that are, for example, apparently de novo in a given proband or are rare nonsense variants, which are examples of findings that are more likely to be medically relevant (if real). There can be great risk with return of false positives, and the associated risk varies from gene to gene, person to person, and condition to condition. For example, there is the possibility that returning a false-positive variant in BRCA2 could result in unnecessary surgery. In contrast, returning a false-positive variant in LDLR, associated with hypercholesterolemia, is likely to be less risky, as it may simply lead to more frequent cholesterol testing. Most clinical and research laboratories choose to validate medically relevant genetic variation with a different sequencing assay prior to issuance of a variant report that is returned to a patient or inserted into the medical record. Sanger dideoxy terminator sequencing has long been considered the gold standard in DNA sequencing and is widely used to confirm the presence and genotype of an ES- or GS-detected variant. Although Sanger confirmation can increase costs and delay issuance of a report, there are important benefits. Beyond eliminating false positives, for example,

­References

Sanger sequencing may be used to determine the correct genomic location and/or the specific insertion or deletion event of ES-/GS-identified indels, which are often inaccurately mapped even if real. Sanger may also serve as qualitative confirmation of unbalanced ES/GS reference/alternate reads indicative of mosaicism. Thus, while some studies have suggested that Sanger verification of ES- or GS-identified variants is unnecessary,57, 58 because costs are modest and it remains the safest approach, Sanger validation of ES/GS variation remains standard.

8 ­Conclusion In this chapter, we summarize how secondary findings are discovered in clinical large-scale DNA sequencing. We describe the sequencing assays (ES/GS) that are currently being employed by clinical and research laboratories and the downstream analytical pipelines that are utilized to detect potentially medically relevant genetic variants among the thousands to millions of benign (or at least not highly penetrant) DNA changes. We also discuss places within this analytical pipeline where secondary genetic findings can be proactively sought out or, conversely, avoided. Also stressed in this chapter is the importance of quality control metrics (throughout the analytical process) and validation via an alternative sequencing assay to reduce the risk of false positives. With the decreasing cost and increasing uptake of DNA sequencing in clinical and research settings, identification of incidental and secondary genetic findings becomes increasingly more common. These findings are often detected in individuals who exhibit no obvious clinical phenotype and to whom the results are completely unexpected. Return of incidental and/or secondary genetic findings is a recent phenomenon resulting from adoption of large-scale DNA sequencing by clinical laboratories. Research is underway to understand how return of this genetic information affects patients, study participants, outcomes, and healthcare systems, but the consequences associated with incidental and/or secondary genetic findings remain unknown. Further, no consensus exists among members of the clinical genetics community regarding the types of secondary findings that should be returned to patients and study participants; therefore, it is up to individual laboratories to adopt procedures that are deemed effective, sustainable, and appropriate.

­References 1. Porter KM, Kauffman TL, Koenig BA, et al. Approaches to carrier testing and results disclosure in translational genomics research: the clinical sequencing exploratory research consortium experience. Mol Genet Genomic Med. 2018; https://doi.org/10.1002/mgg3.453. 30133189. 2. Hart MR, Biesecker BB, Blout CL, et al. Secondary findings from clinical genomic sequencing: prevalence, patient perspectives, family history assessment, and health-care costs from a multisite study. Genet Med. 2018; https://doi.org/10.1038/s41436-018-0308-x. 30287922.

71

72

CHAPTER 4  How secondary findings are made

3. Thompson ML, Finnila CR, Bowling KM, et al. Genomic sequencing identifies secondary findings in a cohort of parent study participants. Genet Med. 2018; https://doi.org/10.1038/ gim.2018.53. 29790872. PubMed Central PMCID: PMC6185813. 4. Johnston  JJ, Rubinstein  WS, Facio  FM, et  al. Secondary variants in individuals undergoing exome sequencing: screening of 572 individuals identifies high-penetrance mutations in cancer-susceptibility genes. Am J Hum Genet. 2012;91(1):97–108. https://doi. org/10.1016/j.ajhg.2012.05.021. 22703879. PubMed Central PMCID: PMC3397257. 5. Green RC, Berg JS, Grody WW, et al. ACMG recommendations for reporting of incidental findings in clinical exome and genome sequencing. Genet Med. 2013;15(7):565–574. https://doi.org/10.1038/gim.2013.73. 23788249. PubMed Central PMCID: PMC3727274. 6. Kalia SS, Adelman K, Bale SJ, et al. Recommendations for reporting of secondary findings in clinical exome and genome sequencing, 2016 update (ACMG SF v2.0): a policy statement of the American College of Medical Genetics and Genomics. Genet Med. 2017;19(2):249–255. https://doi.org/10.1038/gim.2016.190. 27854360. 7. Ackerman SL, Koenig BA. Understanding variations in secondary findings reporting practices across U.S. genome sequencing laboratories. AJOB Empir Bioeth. 2018;9(1):48–57. https://doi.org/10.1080/23294515.2017.1405095. 29131714. 8. van El  CG, Cornel  MC, Borry  P, et  al. Whole-genome sequencing in health care: recommendations of the European Society of Human Genetics. Eur J Hum Genet. 2013;21(6):580–584. https://doi.org/10.1038/ejhg.2013.46. 23676617. PubMed Central PMCID: PMC3658192. 9. Boycott K, Hartley T, Adam S, et al. The clinical application of genome-wide sequencing for monogenic diseases in Canada: Position Statement of the Canadian College of Medical Geneticists. J Med Genet. 2015;52(7):431–437. https://doi.org/10.1136/­ jmedgenet-2015-103144. 25951830. PubMed Central PMCID: PMC4501167. 10. Scheuner  MT, Peredo  J, Benkendorf  J, et  al. Reporting genomic secondary findings: ACMG members weigh in. Genet Med. 2015;17(1):27–35. https://doi.org/10.1038/ gim.2014.165. 25394173. 11. Neu MB, Bowling KM, Cooper GM. Clinical utility of genomic sequencing. Curr Opin Pediatr. 2019;31(6):732–738. PMID 31693580. 12. Zhu Y, Tazearslan C, Suh Y. Challenges and progress in interpretation of non-coding genetic variants associated with human disease. Exp Biol Med (Maywood). 2017;242(13):1325– 1334. https://doi.org/10.1177/1535370217713750. 28581336. PubMed Central PMCID: PMC5529005. 13. Oetting WS, Beroud C, Brenner SE, et al. Non-coding variation: the 2016 annual scientific meeting of the human genome variation society. Hum Mutat. 2017;38(4):460–463. https:// doi.org/10.1002/humu.23169. 28054414. 14. Cooper GM, Shendure J. Needles in stacks of needles: finding disease-causal variants in a wealth of genomic data. Nat Rev Genet. 2011;12(9):628–640. https://doi.org/10.1038/ nrg3046. 21850043. 15. Fisher S, Barry A, Abreu J, et al. A scalable, fully automated process for construction of ­sequence-ready human exome targeted capture libraries. Genome Biol. 2011;12(1):R1. https:// doi.org/10.1186/gb-2011-12-1-r1. 21205303. PubMed Central PMCID: PMC3091298. 16. Li H, Durbin R. Fast and accurate long-read alignment with Burrows-Wheeler transform. Bioinformatics. 2010;26(5):589–595. https://doi.org/10.1093/bioinformatics/btp698. 20080505. PubMed Central PMCID: PMC2828108. 17. Sanghvi  RV, Buhay  CJ, Powell  BC, et  al. Characterizing reduced coverage regions through comparison of exome and genome sequencing data across 10 centers. Genet Med. 2018;20(8):855–866. https://doi.org/10.1038/gim.2017.192. 29144510.

­References

18. McKenna A, Hanna M, Banks E, et al. The Genome Analysis Toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data. Genome Res. 2010;20(9):1297– 1303. https://doi.org/10.1101/gr.107524.110. 20644199. PubMed Central PMCID: PMC2928508. 19. DePristo MA, Banks E, Poplin R, et al. A framework for variation discovery and genotyping using next-generation DNA sequencing data. Nat Genet. 2011;43(5):491–498. https:// doi.org/10.1038/ng.806. 21478889. PubMed Central PMCID: PMC3083463. 20. Van der Auwera GA, Carneiro MO, Hartl C, et al. From FastQ data to high confidence variant calls: the Genome Analysis Toolkit best practices pipeline. Curr Protoc Bioinform. 2013;43:https://doi.org/10.1002/0471250953.bi1110s43. 11.10.1-33 25431634. PubMed Central PMCID: PMC4243306. 21. Lek M, Karczewski KJ, Minikel EV, et al. Analysis of protein-coding genetic variation in 60,706 humans. Nature. 2016;536(7616):285–291. https://doi.org/10.1038/nature19057. 27535533. PubMed Central PMCID: PMC5018207. 22. Landrum MJ, Lee JM, Benson M, et al. ClinVar: improving access to variant interpretations and supporting evidence. Nucleic Acids Res. 2018;46(D1):D1062–D1067. https:// doi.org/10.1093/nar/gkx1153. 29165669. PubMed Central PMCID: PMC5753237. 23. Stenson PD, Mort M, Ball EV, et al. The Human Gene Mutation Database: towards a comprehensive repository of inherited mutation data for medical research, genetic diagnosis and next-generation sequencing studies. Hum Genet. 2017;136(6):665–677. https://doi. org/10.1007/s00439-017-1779-6. 28349240. PubMed Central PMCID: PMC5429360. 24. Steward CA, Parker APJ, Minassian BA, Sisodiya SM, Frankish A, Harrow J. Genome annotation for clinical genomic diagnostics: strengths and weaknesses. Genome Med. 2017;9(1):49. https://doi.org/10.1186/s13073-017-0441-1. 28558813. PubMed Central PMCID: PMC5448149. 25. Yang  H, Wang  K. Genomic variant annotation and prioritization with ANNOVAR and wANNOVAR. Nat Protoc. 2015;10(10):1556–1566. https://doi.org/10.1038/ nprot.2015.105. 26379229. PubMed Central PMCID: PMC4718734. 26. McLaren W, Gil L, Hunt SE, et al. The ensemble variant effect predictor. Genome Biol. 2016;17(1):122. https://doi.org/10.1186/s13059-016-0974-4. 27268795. PubMed Central PMCID: PMC4893825. 27. Orrico  A, Galli  L, Buoni  S, Orsi  A, Vonella  G, Sorrentino  V. Novel PTEN mutations in neurodevelopmental disorders and macrocephaly. Clin Genet. 2009;75(2):195–198. https://doi.org/10.1111/j.1399-0004.2008.01074.x. 18759867. 28. Landrum  MJ, Lee  JM, Benson  M, et  al. ClinVar: public archive of interpretations of clinically relevant variants. Nucleic Acids Res. 2016;44(D1):D862–D868. https://doi. org/10.1093/nar/gkv1222. 26582918. PubMed Central PMCID: PMC4702865. 29. Ploug T, Holm S. Clinical genome sequencing and population preferences for information about ‘incidental’ findings—from medically actionable genes (MAGs) to patient actionable genes (PAGs). PLoS One. 2017;12(7):e0179935https://doi.org/10.1371/journal. pone.0179935. 28671958. PubMed Central PMCID: PMC5495206. 30. Murphy Bollinger  J, Bridges  JF, Mohamed  A, Kaufman  D. Public preferences for the return of research results in genetic research: a conjoint analysis. Genet Med. 2014;16(12):932–939. https://doi.org/10.1038/gim.2014.50. 24854226. PubMed Central PMCID: PMC4241188. 31. Daack-Hirsch  S, Driessnack  M, Hanish  A, et  al. ‘Information is information’: a public perspective on incidental findings in clinical and research genome-based testing. Clin Genet. 2013;84(1):11–18. https://doi.org/10.1111/cge.12167. 23590238. PubMed Central PMCID: PMC4334458.

73

74

CHAPTER 4  How secondary findings are made

32. Brothers  KB, East  KM, Kelley  WV, et  al. Eliciting preferences on secondary findings: the preferences instrument for genomic secondary results. Genet Med. 2017;19(3):337– 344. https://doi.org/10.1038/gim.2016.110. 27561086. PubMed Central PMCID: PMC5326612. 33. Shahmirzadi L, Chao EC, Palmaer E, Parra MC, Tang S, Gonzalez KD. Patient decisions for disclosure of secondary findings among the first 200 individuals undergoing clinical diagnostic exome sequencing. Genet Med. 2014;16(5):395–399. https://doi.org/10.1038/ gim.2013.153. 24113345. PubMed Central PMCID: PMC4018499. 34. Sanderson SC, Linderman MD, Suckiel SA, et al. Motivations, concerns and preferences of personal genome sequencing research participants: Baseline findings from the HealthSeq project. Eur J Hum Genet. 2016;24(1):153. https://doi.org/10.1038/ejhg.2015.179. 26508568. PubMed Central PMCID: PMC4795219. 35. Loud  JT, Bremer  RC, Mai  PL, et  al. Research participant interest in primary, secondary, and incidental genomic findings. Genet Med. 2016;18(12):1218–1225. https://doi. org/10.1038/gim.2016.36. 27101135. PubMed Central PMCID: PMC5074919. 36. Bishop CL, Strong KA, Dimmock DP. Choices of incidental findings of individuals undergoing genome wide sequencing, a single center's experience. Clin Genet. 2017;91(1):137– 140. https://doi.org/10.1111/cge.12829. 27392285. 37. O'Daniel  JM, McLaughlin  HM, Amendola  LM, et  al. A survey of current practices for genomic sequencing test interpretation and reporting processes in US laboratories. Genet Med. 2017;19(5):575–582. https://doi.org/10.1038/gim.2016.152. 27811861. PubMed Central PMCID: PMC5415437. 38. Dorschner MO, Amendola LM, Turner EH, et al. Actionable, pathogenic incidental findings in 1,000 participants’ exomes. Am J Hum Genet. 2013;93(4):631–640. https://doi. org/10.1016/j.ajhg.2013.08.006. 24055113. PubMed Central PMCID: PMC3791261. 39. Gambin T, Jhangiani SN, Below JE, et al. Secondary findings and carrier test frequencies in a large multiethnic sample. Genome Med. 2015;7(1):54https://doi.org/10.1186/s13073015-0171-1. 26195989. PubMed Central PMCID: PMC4507324. 40. Amendola LM, Dorschner MO, Robertson PD, et al. Actionable exomic incidental findings in 6503 participants: challenges of variant classification. Genome Res. 2015;25(3):305– 315. https://doi.org/10.1101/gr.183483.114. 25637381. PubMed Central PMCID: PMC4352885. 41. Jamal SM, Yu JH, Chong JX, et al. Practices and policies of clinical exome sequencing providers: analysis and implications. Am J Med Genet A. 2013;161A(5):935–950. https:// doi.org/10.1002/ajmg.a.35942. 23610049. PubMed Central PMCID: PMC3708985. 42. Nuckols TK, Paddock SM, Bower AG, et al. Costs of intravenous adverse drug events in academic and nonacademic intensive care units. Med Care. 2008;46(1):17–24. https://doi. org/10.1097/MLR.0b013e3181589bed. 18162851. 43. Gurwitz JH, Field TS, Judge J, et al. The incidence of adverse drug events in two large academic long-term care facilities. Am J Med. 2005;118(3):251–258. https://doi.org/10.1016/j. amjmed.2004.09.018. 15745723. 44. Vargas  E, Terleira  A, Hernando  F, et  al. Effect of adverse drug reactions on length of stay in surgical intensive care units. Crit Care Med. 2003;31(3):694–698. https://doi. org/10.1097/01.CCM.0000049947.80131.ED. 12626971. 45. Lazarou  J, Pomeranz  BH, Corey  PN. Incidence of adverse drug reactions in hospitalized patients: a meta-analysis of prospective studies. JAMA. 1998;279(15):1200–1205. 9555760.

­References

46. Field TS, Gilman BH, Subramanian S, Fuller JC, Bates DW, Gurwitz JH. The costs associated with adverse drug events among older adults in the ambulatory setting. Med Care. 2005;43(12):1171–1176. 16299427. 47. Classen DC, Pestotnik SL, Evans RS, Lloyd JF, Burke JP. Adverse drug events in hospitalized patients. Excess length of stay, extra costs, and attributable mortality. JAMA. 1997;277(4):301–306. 9002492. 48. Ernst FR, Grizzle AJ. Drug-related morbidity and mortality: updating the cost-of-illness model. J Am Pharm Assoc (Wash). 2001;41(2):192–199. 11297331. 49. Epstein  RS, Moyer  TP, Aubert  RE, et  al. Warfarin genotyping reduces hospitalization rates results from the MM-WES (Medco-Mayo Warfarin Effectiveness study). J Am Coll Cardiol. 2010;55(25):2804–2812. https://doi.org/10.1016/j.jacc.2010.03.009. 20381283. 50. Limdi NA, Wadelius M, Cavallari L, et al. Warfarin pharmacogenetics: a single VKORC1 polymorphism is predictive of dose across 3 racial groups. Blood. 2010;115(18):3827– 3834. https://doi.org/10.1182/blood-2009-12-255992. 20203262. PubMed Central PMCID: PMC2865873. 51. International Warfarin Pharmacogenetics C, Klein TE, Altman RB, et al. Estimation of the warfarin dose with clinical and pharmacogenetic data. N Engl J Med. 2009;360(8):753– 764. https://doi.org/10.1056/NEJMoa0809329. 19228618. PubMed Central PMCID: PMC2722908. 52. Yin T, Miyata T. Warfarin dose and the pharmacogenomics of CYP2C9 and VKORC1— rationale and perspectives. Thromb Res. 2007;120(1):1–10. https://doi.org/10.1016/j. thromres.2006.10.021. 17161452. 53. Cooper GM, Johnson JA, Langaee TY, et al. A genome-wide scan for common genetic variants with a large influence on warfarin maintenance dose. Blood. 2008;112(4):1022–1027. https://doi.org/10.1182/blood-2008-01-134247. 18535201. PubMed Central PMCID: PMC2515139. 54. Wadelius  M, Chen  LY, Lindh  JD, et  al. The largest prospective warfarin-treated cohort supports genetic forecasting. Blood. 2009;113(4):784–792. https://doi.org/10.1182/ blood-2008-04-149070. 18574025. PubMed Central PMCID: PMC2630264. 55. Wen MS, Lee M, Chen JJ, et al. Prospective study of warfarin dosage requirements based on CYP2C9 and VKORC1 genotypes. Clin Pharmacol Ther. 2008;84(1):83–89. https:// doi.org/10.1038/sj.clpt.6100453. 18183038. 56. MacArthur DG, Tyler-Smith C. Loss-of-function variants in the genomes of healthy humans. Hum Mol Genet. 2010;19(R2):R125–R130. https://doi.org/10.1093/hmg/ddq365. 20805107. PubMed Central PMCID: PMC2953739. 57. Baudhuin LM, Lagerstedt SA, Klee EW, Fadra N, Oglesbee D, Ferber MJ. Confirming variants in next-generation sequencing panel testing by sanger sequencing. J Mol Diagn. 2015;17(4):456–461. https://doi.org/10.1016/j.jmoldx.2015.03.004. 25960255. 58. Beck  TF, Mullikin  JC, Program  NCS, Biesecker  LG. Systematic evaluation of sanger validation of next-generation sequencing variants. Clin Chem. 2016;62(4):647–654. https://doi.org/10.1373/clinchem.2015.249623. 26847218. PubMed Central PMCID: PMC4878677.

75

CHAPTER

Informed consent and decision-making

5

Sebastian Schleidgena, Kyle B. Brothersb a

Institute of Philosophy, Faculty of Humanities and Social Sciences, FernUniversität in Hagen, Hagen, Germany b Endowed Chair of Pediatric Clinical and Translational Research, University of Louisville, Louisville, KY, United States

1 ­Introduction The application of whole exome and whole-genome sequencing techniques (WES/ WGS) in research and clinical contexts facilitates, among other things, the discovery and identification of disease-causing genetic variants. At the same time, WES and WGS have the potential to reveal unexpected genetic information associated with health conditions of probands and patients that were not the focus of the respective investigation. The possibility of such genetic secondary findings has long raised ethically relevant questions, in particular: “How much does the subject actually wish to know about his or her genetic constitution? […] What influence could this information exert over the way the subject views and plans his or her life?” The normative significance of the first question stems from the right to informational self-determination. This in turn is justified by appropriate consideration of patient and participant autonomy, which is a fundamental principle of biomedical ethics. The right to informational self-determination implies both a right to know and a right not to know. According to the former an autonomous individual may decide what they want to know about their genetic constitution. According to the latter, “[n]o individual may be forced to (want to) know all that can be discovered about his or her genetic constitution […]”.1 In principle, this holds true for the contexts of WES/ WGS and any secondary findings that may occur as well as in other medical contexts. When speaking of a patients’ or research participants’ autonomous decision about what information they want to know (or do not want to know) about their genetic constitution, it may be important to assess the potential effects of this information on his or her life plans. In other words, “What influence could this information exert over the way the subject views and plans his or her life?”1 While this question emphasizes the effect on the patient or research participant, it also draws attention Secondary Findings in Genomic Research. https://doi.org/10.1016/B978-0-12-816549-2.00005-9 © 2020 Elsevier Inc. All rights reserved.

77

78

CHAPTER 5  Informed consent and decision-making

to the duties of physicians with respect to their patients, and a potentially distinct set of duties that researchers owe the participants in their research. In both contexts the physician or research is obligated to maximize benefits and minimize harms. This set of duties, therefore, refers to additional principles of biomedical ethics: beneficence and nonmaleficence. This means that physicians and researchers have some obligation to consider what genetic information may or should be disclosed, while at the same time giving due consideration to patients’ and participants’ autonomous decisions regarding disclosure of their genetic information. Obviously, this is easier said than done. In practice the obligation of healthcare providers or researchers to maximize benefits and minimize harms can collide with, or last create tension with, the duty to give appropriate consideration to the patients’ or research participants’ autonomy. Patients or research participants may want more information despite its risks, even if professionals would not anticipate that this information would provide benefits. This gives rise to the particularly difficult question of which secondary findings may or should be disclosed, to which patients or research participants, and under which conditions. As discussed in Chapter 1, debates on the disclosure of genetic secondary findings originated in the research context but have also been discussed for some time with reference to clinical contexts.2 Given the scope of this book, we will focus in this chapter on the handling of genetic secondary findings in research contexts utilizing WES/WGS. Specifically, we will explore how researchers can approach informed consent, decision-making, and disclosure of results in ways that maximize both the autonomy of research participants and the researcher’s obligation to provide benefits and avoid harms.

2 ­Ethical issues 2.1 ­Respect for autonomy and informed consent Before digging deeper into how respect for autonomy can be operationalized through the informed consent process, it will be helpful to briefly consider what autonomy is and why researchers are expected to demonstrate respect for it. This is a surprisingly difficult topic. The most popular way of thinking about autonomy in the context of clinical and research ethics is that it is a midlevel principle. That is to say, it is not the sort of “high-level” moral theory that has been thoroughly specified through philosophical analysis. And it is also not the sort of low-level moral axiom—such as not to lie or not to kill—that constitutes “common” morality that most people are taught as children. It falls somewhere in the middle: Most people will agree that people should be able to make certain decisions for themselves, even without a great degree of philosophical specification about when this right should apply, under what circumstances, based on which first principles, etc.3 There are certainly other ways of looking at autonomy, however. Philosophers have laid out more comprehensive accounts of the conceptual basis for autonomy and what its implications might be.4 Others have taken a more consequentialist approach,

2 ­ Ethical issues

thinking of autonomy not as a goal that has intrinsic value, but rather as a means to achieving well-being. In other words, because having the ability to makes one’s own decisions is a necessary (and possibly sufficient) means for ensuring individual wellbeing, it is an important value to protect.5 For the most part we will think of autonomy in the midlevel sense proposed by Beauchamp and Childress in their classic book Principles of Biomedical Ethics.3 This is a particularly helpful framing, given that Beauchamp and Childress emphasize the need to balance various midlevel principles, including beneficence and nonmaleficence, to generate appropriate outcomes to difficult cases or to generate ethically appropriate policies. This framing of autonomy is similar to the one provided in the Belmont Report, except perhaps in the sense that the Belmont Report implicitly emphasizes the primacy of autonomy in the context of research.6 Regardless of other considerations, the Belmont Report seems to say, the ability of competent adults to accept or decline research participation needs to be respected. We will keep both views in tension: accepting that the decision to participate or not participate should always be available to potential research participants, while examining how other dimensions of the research process should balance the preferences of participants with the obligation of researchers to maximize benefits and minimize harms. As a starting place, however, we will examine the dimensions of the research process that take place prior to a genomic test being performed on a participant. The informed consent process certainly plays a central role in this stage of research, but genomic research also frequently incorporates formal pretest genetic counseling. This type of counseling might be offered to participants prior to their decision to participate or after informed consent has been obtained. As an example, some genomic research studies focus primarily on getting a genetic test; if you do not want to undergo the genomic test, that means you do not want to be in the research. In this case, it might make sense to provide genetic counseling before the informed consent process or as a part of the informed consent process. In other types of genomic research, undergoing a genomic test might not be required. A common example here would be a study examining outcomes from different approaches to genetic counseling. In this type of study the informed consent process would come first, and then genetic counseling would typically be provided afterward. For the purposes of this discussion, we will first consider the informed consent process and will then turn to possible approaches to genetic counseling and shared decision-making. In discussions of informed consent in medical or research contexts, four conditions of an effective informed consent are typically mentioned3,7: (1) Appropriate education of the patient or research participant by the healthcare provider or researcher. (2) Understanding of the educational content by the patient or research participant. (3) Voluntariness of consent, in particular by ensuring that decisions are made without undue influence, intimidation, or manipulation. (4) Explicit expression of consent, typically by signing a document.

79

80

CHAPTER 5  Informed consent and decision-making

The first two of these conditions highlight that the adjective “informed” is a critical piece of the phrase “informed consent.” In order for a potential research participant to realize the ability to make autonomous decisions, the research staff members conducting the informed consent process need to provide appropriate information. Although it is arguable whether the opportunity to make autonomous decisions should be thought of as a right, it is useful in this context to make an analogy with the idea of negative and positive rights. Negative rights are those that can be realized as long as others “stay out of the way.” In this way, negative rights obligate others not to obstruct the ability of individuals to realize these rights. Positive rights, on the other hand, can only be realized if others take action. In this analogy an individual’s interest in making autonomous decisions about research participation are similar to a positive right. To make an autonomous decision, potential research participants depend upon researchers to provide information that is both understandable and sufficient to make a good decision. In other words, one of the implications of the principle of autonomy is that researchers have a duty to provide potential research participants with quality information.7 While this general framework may seem somewhat intuitive, intense debate has surrounded its details.8–11 What is the scope of the obligation for researchers to inform research participants about research? Beyond the decision to participate or not participate in research, what other aspects of participant autonomy need to be accounted for in the design of research studies? In particular, to what extent do the principle of autonomy and the duty to inform carry over to the management of secondary findings in genetics research?12–14 Since the completion of the Human Genome Project in 2003, there has been a remarkable increase in clinically oriented research involving genome and exome sequencing.15 Some of this research focuses on identifying the cause of rare genetic conditions, while other research focuses on identifying risk factors for common conditions or on the development of novel, targeted therapeutics.16–20 All of these types of research have the potential to generate genetic results that are not directly related to the aim of the study, but could, nonetheless, carry clinical relevance for the research participant.14,21 Depending on the study, this may involve, for example, the identification of a genetic predisposition for passing on sickle cell disease or cystic fibrosis to offspring, or to develop Huntington’s disease or Alzheimer’s disease. An increased risk for developing breast or intestinal cancer, for instance, might also be identified.22–25 As discussed in Chapter 1, secondary findings are not fundamentally new phenomena in research. However, they have gained new significance in the context of genetics research. On the one hand, this is because the probability of secondary findings has increased dramatically in the context of genetics due to technological and scientific innovation.26 When a whole-genome sequence is generated for research purposes, the range of possible secondary findings that might be generated is virtually unlimited in terms of the potential quantity of types of results that could be generated, as well as the qualitative dimensions of the likelihood of the disease, its implications for different patients, etc. Many have also observed that secondary

2 ­ Ethical issues

r­ esults generated in genomics research can lead to scenarios that are uncommon in other types of research, such as the likelihood that a result could carry implications not only for the research participant but also for his or her relatives. The potential for identifying future disease risk is also often highlighted as an implication that is distinctive to genetics, a scenario that is complicated by the fact that many genetic risks are not modifiable through actions like taking medications or making changes in health behaviors. Timing is also a potential complicating factor, as secondary findings generated in genomics research might not be uncovered until long after a participant enters a study, such as when samples are first stored as a part of a biorepository.22,27 Although none of these features is unique to genetics or genomics research, they do highlight why secondary findings generated in this type of research tend to create complicated questions about research participants’ interest in making autonomous decisions and the duty of researchers to provide appropriate information. In particular, we will consider the following issues that are raised by this set of challenges12,26–29: I. Providing information: What information on possible secondary findings must an appropriate disclosure contain? In view of the complexity of molecular genetic studies, what information is likely to be understandable for potential research participants? II. Offering options and eliciting preferences: In the context of informed consent for genetics and genomics research, what options for receiving secondary findings should research participants be offered? III. Deciding what to disclose: Which secondary findings should research participants or their relatives receive and under which conditions?

2.2 ­Providing information As we discussed earlier the obligation to provide potential research participants with information about a research study, and specifically about the potential for receiving secondary findings, might take place during a conventional research informed consent encounter, during a genetic counseling session, or both. Regardless of how researchers will deliver this information to participants, it is necessary to consider what information about secondary findings should be provided. One conventional answer to this question is that research participants should be informed of “everything” and that they should receive information that is as comprehensive as possible.21 In addition to informing potential participants about the purpose and significance of the planned study, this might include information about the scope of possible secondary findings for probands and their relatives and how these secondary findings will be handled if they are identified. Although it is intuitive to assume that researchers should not withhold information from potential research participants, there are reasons to question whether it is even feasible to provide information about possible secondary findings that is truly comprehensive. Even if a study were to limit potential secondary findings to those

81

82

CHAPTER 5  Informed consent and decision-making

conditions and genes listed in the actionable secondary findings list published by the American College of Genetics and Genomics,30 this would still require a discussion of about 34 distinct conditions (as of early 2020), about which most research participants will know almost nothing. In reality, it is not clear that more information is always better, unless of course the goal of informed consent is merely to protect researchers and their institutions from legal exposure.31 From an ethical and moral perspective, the goal of an informed consent process in research is to ensure that potential participants are able to make decisions that reflect their values and priorities, in short, to make “good” decisions. When this is accepted as the goal, it is clear that by providing comprehensive information, researchers might actually be inhibiting the ability of a potential research participant to make a “good” decision. Too much information can create confusion and can make it difficult for the potential research participant to identify and consider the factors that are most important to making a good decision.31 In light of this dynamic, there are a range of viewpoints about how much and what kinds of information are sufficient to ensure that potential research participants are able to make good decisions. At one end of this spectrum are suggestions that the information provided to research participants should be simplified but should still cover all of the relevant topics. At the other end of this spectrum is the suggestion that autonomous decision-making is actually best supported when researchers provide potential participants with only basic information about the planned study.32,33

2.3 ­Offering options and eliciting preferences The various frameworks for classifying secondary findings are discussed in detail in Chapter  3 of this book, while strategies for eliciting preferences on secondary findings are explored in Chapter 6. For the discussion in this chapter, it is important simply to highlight that procedures for offering options to potential research participants about which research results they would like to receive are unavoidably linked with the approach to informed consent. In some contexts, researchers designing a genomic research study may decide that they wish to provide research participants with several options about what kinds of results they would like to receive. For example, they might offer participants options to receive all secondary findings, certain secondary findings, or no secondary findings. When this kind of strategy is utilized, it is important for either the informed consent process or a separate genetic counseling session to explain these options. We have found in our own research that participants sometimes see options like these and say something like, “Why wouldn’t I want all the results you can give me?” If the genetic counselor is then able to explain briefly why some people choose not to receive results, this can be quite helpful to participants to begin to evaluate their own values. Regardless of whether they end up deciding to receive all results, some results, or no results, the opportunity to think through some of the implications that were not immediately apparent can help them make a “good” decision.

2 ­ Ethical issues

In other types of genomics research, enrollment in a study can be contingent on the participant’s willingness to receive secondary findings. If, for example, a researcher wanted to examine how participants respond to secondary findings, then it would make sense to include willingness to receive secondary findings among the inclusion criteria for the study. In this case, of course, it remains important to explain the possible implications of receiving secondary findings so that potential research participants can decide whether participation is right for them.21 Regardless of whether a study offers participants with options about which secondary findings they would like to receive or makes receipt of secondary findings a requirement of research participation, an explanation of each and every possible secondary finding is typically unnecessary. Consistent with our earlier observations about the level of detail needed to make “good” decisions, participants typically just need information about the types of secondary findings that might be generated, combined with some examples of the implications (including both benefits and harms) that these results my carry.

2.4 ­Deciding what to disclose In addition to deciding whether receipt of secondary findings will be optional or mandatory for participants, researchers also need to decide which types of results they will work to identify and subsequently disclose to participants. This issue has been an important source of debate and controversy for over a decade. As discussed in Chapter 3, there are a wide array of strategies for organizing potential secondary findings, and many of these are organized around the idea that researchers’ obligations to search for and return a particular type of secondary finding vary depending on features of that finding.21,34–37 For example, a researcher’s obligation to search for and return a finding is higher when the scientific understanding of that finding is high (scientific validity or analytical validity), when understanding of the link between the genetic change and a clinical condition is strong (clinical validity), and when the result could lead to actions on the part of the participant and/or a healthcare provider that could improve the participant’s health (clinical utility).38 When one or more of these features is absent, however, the researcher’s obligation to identify and disclose findings would be lower. In fact, it could be argued that a genetic change is not really a “result” until the evidence for scientific validity and clinical validity is quite strong. Practical challenges also need to be considered when deciding which secondary findings will be identified and returned (and even whether secondary findings will be identified at all). To disclose findings to participants in a way that maximizes benefits and minimizes risks, a range of resources are needed: laboratories with enough time, resources, and experience to effectively identify secondary findings of potential value; clinical research staff and/or healthcare providers who have the time and experience needed to interpret secondary findings and communicate them to participants; etc.39 Over the past decade, research studies performing genomic sequencing on human biosamples have increasingly planned and been funded to have these resources available and have even been designed to examine how the receipt of secondary findings affects research participants.40

83

84

CHAPTER 5  Informed consent and decision-making

However, there are likely still many research studies that are located at sites or have funding sources that simply cannot support the return of secondary findings to participants using high-quality practices.41 As the old saying goes: “If you can’t afford to do it right, you can’t afford to do it.” In these settings, it might be necessary to plan not to return secondary findings and simply make this plan clear during the informed consent process.42 Once the researcher had decided on a plan for identifying and disclosing (or not identifying or disclosing) secondary findings, the question sometimes arises whether that plan should be broken. For example, a researcher might have included in the informed consent document that secondary findings will not be returned, but then later unexpectedly discovers a finding that seems potentially important for a participant to be aware of. A similar event could arise when a participant is given options for receiving secondary findings and chooses to decline receipt of these results. The question in either case is: Should earlier promises be kept in light of new information, such as unexpected findings that seem very important to a participant’s health? Best practices can help avoid these types of difficult decisions. For example, it is generally preferable that researchers design laboratory procedures for analyzing genomics data so as not to generate secondary findings when the study has promised in the consent document that secondary findings will not be returned, or when the participant has declined receipt of secondary findings. It is better not to identify secondary findings at all if they will not be returned, since this discrepancy can create moral distress for the research team. If these situations would be unavoidable and that limitation is foreseeable, then the researchers should strongly consider from the beginning not adopting a plan that would prevent them from returning results that might be important. When unforeseeable circumstances arise, however, this can create a difficult ethical challenge: the researcher must choose whether to respect the participant’s stated preferences or disclose the information anyway. This is a very difficult decision and in fact represents a classic type of dilemma in bioethics: autonomy versus beneficence. There is significant controversy about how best to solve this type of dilemma, with much of the debate centering around a claimed “right not to know.”43 That is, many argue that respect for a person’s autonomy also means respecting that person’s wishes if they decided that they do not want certain types of information. In general terms, we believe that competent adults can and should be able to decline the receipt of information that they choose not to receive. Children and others who have limited autonomy, however, pose a more difficult challenge.44 In practice, however, there are ways to approach informed consent in these situations such that participants retain the opportunity to accept or decline secondary findings. For example, if the informed consent document promises that secondary findings will not be returned, and researchers later change their mind about that plan, then it is possible to reconsent all participants (or everyone who can be reached) so that the new plan is made clear. This gives participants the opportunity to reconsider whether they would like to receive secondary findings, without incidentally disclosing that a secondary finding is already available for disclosure to them.

3 ­ Counseling challenges

3 ­Counseling challenges In the previous section we briefly discussed the tension between providing too much and too little information in the course of the informed consent process, as well as any pretest genetic counseling that may also take place. This is undoubtedly an important issue and one that plays into nearly every research study involving genomic sequencing. However, there are a number of other important challenges that can arise in efforts to help research participants understand the implications of secondary findings, including potential benefits and potential risks. In the sections that follow, we will consider three of these challenges: barriers to understanding, changing information, and the therapeutic (or diagnostic) misconception.

3.1 ­Addressing barriers to understanding If there is a single characteristic that distinguishes genomics from other domains of science and medicine, it is probably complexity. Changes in the human genome interact with one another and with other factors in the environment to create changes in human cells. These changes then sometimes accumulate to create observable changes in an individual’s health. Some of these changes are preventable, some are not preventable, and some are so uncertain that it is difficult even to discern whether prevention efforts ultimately changed an outcome. All of these interactions and dynamics are extremely difficult to understand, even for experts in science and medicine.45 It should not be surprising, then, that members of the public, including potential research participants, demonstrate limited knowledge of genetics and genomics.46–49 These challenges are further complicated in the context of research consent, where potential research participants might also encounter unfamiliar topics that are nonetheless relevant to their decision about whether to participate, including issues like the purpose of the study, the procedures that will be carried out, and the potential risks and benefits of research participation. Frequently, education is proposed as the solution to this challenge. “If only we could provide participants with relevant information in an understandable format,” this way of thinking goes, “they would understand and could make a good decision.” However, there are reasons to believe that this solution oversimplifies the problem. In reality the barriers to making good decisions related to research participation and receipt of secondary findings extend far beyond the participant’s lack of familiarity with these topics and limited background in scientific knowledge about genetics. Making a “good” decision about whether to participate in a research study and whether to subsequently receive secondary findings depends as much on an individual’s values, priorities, and past experiences as it does on understanding the underlying knowledge. Despite this critical connection, at the time potential research participants are approached about research participation, most have not previously had the opportunity to clarify for themselves what their values might be around genetic information and its use in medical care, or how their past experience with conditions like Alzheimer’s

85

86

CHAPTER 5  Informed consent and decision-making

disease or breast cancer might impact their decisions in the future. These are not topics that come up every day, and the vast majority of potential research participants simply have not thought much about it. This means that “understanding” is not just about understanding the details of a research study and the background information about genetics; understanding also requires having the chance to work through one’s own values and priorities for the first time. This connection is, in some ways, the basis for the entire profession of genetic counseling: the factors that go into making decisions around genetic information are so complex that it can be important to support decision-making by enlisting the help of a professional who not only has expertise in genetics but also has special skills in helping clients explore their values and priorities. This is not to say that formal genetic counseling is required in every genomic research study that might return secondary findings but that it can be an important strategy to reduce risks and maximize research participants’ opportunity to make decisions that are consistent with their values and preferences.

3.2 ­Accounting for changing information Another important challenge for effective genetic counseling in genomics research is the changing nature of scientific knowledge about genetics. Although there are a number of ways that scientific knowledge is advancing, the key evolution for this discussion has to do with the evolving understanding of the association between specific genetic variants and the causation of medical conditions. Recall that except in the case of identical twins, humans differ from one another at millions of locations in our genomes. Many differences have no observable effect on us, and others cause only benign changes, such as changes in our appearance. When a genomic test on a research participant uncovers a genetic variant that has never been seen before, scientists typically are not able to tell immediately whether that change contributes to risk for a disease, or whether it is simply a “silent” change. One of the most important ways to draw conclusions about this is to test enough individuals that scientists can begin to get an idea of whether a particular change seems to occur more frequently in people with a specific medical condition. These types of insights accumulate over time, so new information is always emerging. A genetic change that was thought to be benign when it was previously discovered in a research participant might a year later be recognized as a risk factor for a medical condition. This changing state of knowledge creates at least two challenges for researchers who want to inform potential participants of what to expect in a research study. First the likelihood of future changes in knowledge makes it very difficult to give participants a conclusive explanation of the possible secondary findings they might receive. This is one of the important challenges raised by “positive lists,” which set out a list of conditions or genetic variants that will be returned to participants if they are uncovered. Because such lists are likely to change over time (with conditions both added and subtracted as knowledge evolves), providing the current list to potential participants may cause misunderstandings. A common alternative is simply to notify participants of the types of results that will be returned.42 For example, participants

3 ­ Counseling challenges

might be informed that they will receive any secondary findings that could affect their health, or any findings that indicate the patient or their healthcare provider should take specific preventive actions (i.e., “actionable” results). The challenge with this approach is that most potential research participants do not understand medical conditions well enough to infer what conditions might be included in those categories. This means that if a secondary finding is uncovered, they might still be surprised that they agreed to receive such a result. Changing knowledge also creates challenges related to longitudinal reanalysis. When a research participant enrolls in a genomics study, the first few steps of that participation are typically straightforward: The participant gives consent, and a biosample is collected. The sample is analyzed, and any secondary findings are returned to the participant. It is after these initial steps that things become more difficult to anticipate. Since scientific understanding of genomics is evolving, the initial secondary findings disclosed to a participant could change over time. If a participant’s sample is reanalyzed a year later, new secondary results could be uncovered, even if the participant was initially told no secondary results were found.50 A secondary finding that was at first classified as likely pathogenic might later be reclassified as a variant of unknown significance (VUS).51 This means that the investigators need to plan for whether and when they will conduct reanalyses of secondary findings that were returned previously.50 Furthermore, they need to decide how to explain to participants that results disclosed to them during the initial phase of a study might need to be revised or updated at a later time. It turns out that this collection of uncertainties—uncertainty about which results will be considered returnable as time passes, and whether results returned early in the study will require revision later in the study—is extremely difficult to explain to potential research participants. Even if researchers are able to explain this clearly and participants are able to understand it, there remain important questions about how best to respect participants’ autonomy as time passes. If the list of potentially returnable results changes, do participants need to be offered the opportunity to reconsent to the study? If they receive a result that needs to be revised later, should their previous decision to withdraw from the study be respected?

3.3 ­Managing therapeutic and diagnostic misconceptions Research ethicists have long emphasized the importance of maintaining a distinction between research and clinical care. This concern is rooted in the principle of autonomy: Research participants should have the opportunity to make an autonomous decision about whether to voluntarily assume the potential risks of research participation. To ensure that these decisions remain voluntary, researchers should avoid creating situations in which individuals are coerced into research participation. This is why excessive payments for research participation are not allowed, and it is also one important reason that there is an emphasis on separating clinical care and research. Autonomy is threatened if research participants are induced to participate in research to receive either money or medical care that they desperately need.

87

88

CHAPTER 5  Informed consent and decision-making

For many types of medical research, however, it can be very difficult for research participants to tell where clinical care ends and research begins. In many cases both take place in the same location, a patient’s physician may also be a researcher, and the interventions that are required for research purposes may be indistinguishable from clinical interventions. There are numerous examples throughout the history of medical research where research participants believed that they were taking part in interventions for their own care, whereas the truth was that the interventions were being done exclusively for research purposes. The most famous and most influential of these cases was the so-called Tuskegee study, in which African-American men with syphilis believed that they were receiving treatment for their condition. An effective antibiotic treatment had become available during the course of their participation in the study, but it was being withheld by the researchers. Had the participants realized that the “care” they were receiving was in fact just research procedures, they would have had the opportunity to seek appropriate medical care elsewhere. Preventing this kind of therapeutic misconception has been an important goal for research ethicists almost as long as research ethics has existed as a field. It is also a key concern of government bodies that regulate research participation and institutional review boards (IRBs) and research ethics committees (RECs) that review research protocols. Genomics research is one domain, however, where the distinction between clinical care and research has been increasingly blurred in recent years.52 Many of the studies that are being conducted to understand how genomics might improve clinical care are intentionally designed to incorporate research observations into settings where clinical care is being provided to patients.53 In many of these studies, parents choose to participate in research because this is the only way to access genomic sequencing to uncover a diagnosis for their child’s rare disease. This type of research has inspired a relatively new term—the diagnostic misconception—that points to the possibility that patients might be induced to participate in research to receive a diagnosis, not just to receive medical therapy.54 In fact the return of secondary findings to research participants is itself a practice that blurs the boundary between clinical care and research and can motivate individuals to participate in research to receive medical information about themselves.55 Genomic research is not the only domain of medical research where clinical care and research are increasingly becoming blurred. Another important context is the so-called Learning Healthcare System strategy, where analysis of clinical data is used to iteratively modify (and hopefully improve) medical care.56 In these types of settings, it is important from an ethical perspective to ask whether participants’ perception that they are receiving medical therapy or diagnosis should be classified as a “misconception,” since these studies are intentionally designed to blur these boundaries. For researchers conducting genomic research, the key question is how studies can best be designed to generate the scientific knowledge needed while at the same time ensuring that research participation remains, to the extent possible, a decision that is made without coercion. Although this remains an active area of controversy, it is virtually certain that any solutions that arise to address these issues will retain

4 ­ Approaches for addressing ethical issues and counseling challenges

the informed consent process as an important stage in the research process, since it provides a critical opportunity to ensure that participants know what they are getting themselves into and understand the nonresearch alternatives available to them.

4 ­Approaches for addressing ethical issues and counseling challenges 4.1 ­Genetic counseling As we have discussed, pretest genetic counseling is often offered (or even required) as a part of enrolling participants in genomic studies that plan to return secondary findings to participants. Genetic counseling may take place prior to the informed consent process or after, and in some cases the informed consent process may itself involve a conversation with a genetic counselor. Fundamentally, genetic counselors are professionals trained to help individuals and families with decision-making about whether to use genetic testing technologies and what to do with the results. Although knowledge about genetics is certainly an important prerequisite for becoming an effective genetic counselor, the skills required for this profession extend far beyond scientific knowledge. Genetic counselors are, above all else, counselors. They are trained to help individuals and families explore their values and preferences, think through the possible outcomes of genetic testing, and explore how they might want to utilize those results in ways consistent with their values and preferences. Genetic counselors are trained to be effective communicators, to listen actively, and to empower individuals and families to make decisions consistent with their values. Given this background, it is easy to understand why pretest genetic counseling is considered the gold standard for helping patients or research participants decide whether to undergo genetic testing. There is an extensive literature demonstrating that genetic counseling can help reduce adverse effects like anxiety that people may experience before and after a decision to receive genetic information. Genetic counseling can also support the adoption of adaptive strategies, like feelings of self-­ control, that can help individuals and families make good decisions for themselves and effectively deal with any results they may receive.57

4.2 ­Decision-making without formal counseling The only problem with genetic counselors, perhaps, is that there might not be enough of them. According to an analysis commissioned by the American Board of Genetic Counseling (ASGC), the National Society of Genetic Counselors (NSGC), and other key professional organizations, there were 4242 Certified Genetic Counselors (CGC) in the United States in 2017. However, only 2476 of those were involved in direct patient care. Conservatively the analysis estimated that 3266 genetic counselors were needed at that time to address patient demand, indicating a deficit of 791 genetic counselors in the United States alone.58

89

90

CHAPTER 5  Informed consent and decision-making

The analysis also indicates that recent increases in the number of training programs would lead to a supply of genetic counselors that would meet estimated demand by 2026.58 However, many scientists and clinicians who are involved in genomics research would like to see the demand for genetics services increase at a far faster rate than accounted for in this analysis. Frameworks like precision medicine and public health genomics envision a time in the near future when healthy patients will undergo genomic testing to identify their genetic risks, and then work with their primary care provider to target preventive measures to address these risks.59 In this vision the demand for genetic services like genetic counseling would quickly outstrip growth in the supply of genetic counselors. This is one reason that genomics researchers have begun to focus on alternatives to genetic counseling. Research consortia like the Clinical Sequencing EvidenceGenerating Research (CSER) Consortium53 and Electronic Medical Records and Genomics (eMERGE) Network60 are testing whether some of the positive outcomes created by genetic counseling can be achieved using alternative strategies. For example, SouthSeq, one of the research studies that comprise the CSER Consortium,61 is randomizing families to receive their infant baby’s genomic results from a genetic counselor or a nongenetic healthcare professional, such as a neonatologist. Other studies have examined whether disclosure of results online might be another reasonable alternative.62 There are a number of other factors that have driven interest in alternatives to genetic counseling. One of the important concerns that have motivated the use of genetic counselors is the risk that individuals who have received a high-risk genetic result will experience adverse psychological effects like depression or anxiety. However, accumulating evidence indicates that although some individuals do experience modest psychological symptoms after receiving a high-risk genetic result, most deal with this information quite well. The timing and dynamics created by secondary findings is another factor that has driven interest in alternatives to genetic counseling. As discussed earlier, technologies like exome and genome sequencing have the potential to produce many different results. It is difficult to imagine how any amount of pretest counseling could prepare a patient or research participant for all of the possible results they might receive. This is one reason that genomics researchers are increasingly exploring whether formal pretest counseling can be eliminated. In this approach, substantive discussions are delayed until secondary findings have been generated, so that counseling can focus on the findings that a patient or research participant may actually receive. The discussion prior to testing would just focus on providing enough general information for the patient or research participant to decide whether to proceed with testing. These types of research goals have created somewhat of a dilemma for genomics researchers. In one respect, pretest genetic counseling is considered a desirable part of the informed consent process for genomic research studies. After all, it is considered the gold standard for preparing research participants for potentially receiving secondary findings and is thus attractive to researchers, IRBs, and RECs who want to reduce the risks associated with research participation. Researchers are

5 ­Consent models

interested, however, in conducting clinical research to find out whether alternatives to genetic counseling can be used effectively in preventive health settings where genetic counselors are unavailable. In order test these kinds of questions, it is necessary to enroll participants in genomic research studies without providing pretest genetic counseling. This is, of course, an inherent problem in medical research. Although it is important to design research studies to include safeguards, at some point, it becomes necessary to test interventions in “real-world” settings where conventional research safeguards are not available. We consider it reasonable to conduct genomic research without pretest genetic counseling, as long as the informed consent process addresses potential risks and participants are monitored for psychological symptoms and other adverse effects.

5 ­Consent models Having considered a number of issues related to genetic counseling as a part of ­decision-making about research participation and the receipt of secondary findings, it will now be helpful to consider a few of the approaches to informed consent that can be used for this type of research. Although we will discuss a number of consent models in the sections that follow, it is important to note that these models are not necessarily mutually exclusive with one another. It is possible, for example, for a modular consent model and a family consent model to be used together in developing a single informed consent process for a single study. In fact, we will recommend that researchers consider using several of these models together in most genomic research studies. It is expedient for discussion, however, to separate out each of the models so that their advantages and disadvantages can be highlighted.

5.1 ­Staged consent Just as the field of genetics is typified by its complexity, genomic research studies stand out for their multiple facets and “moving pieces.” The studies that comprise the CSER Consortium, for example, typically begin with the collection of demographic and clinical information, and at least one biosample. At a later point, when sequencing data are available, participants will receive not only a primary result (related to the clinical reason for testing) but also one or more secondary findings. Participants will receive these results during a clinical encounter or research visit, which is typically recorded and analyzed. Researchers will then want participants to complete several questionnaires and semistructured interviews. In addition, many studies involve permission for accessing and analyzing medical records and other data. Most of the procedures involved with these steps are unfamiliar to potential participants, and each poses its own set of risks and benefits, which typically need to be explained. Other types of studies that involve genomic testing, such as biorepositories and cohort studies, involve their own complex combination of steps, with a distinct set of risks and benefits.

91

92

CHAPTER 5  Informed consent and decision-making

Due to the complexity of these types of studies, it is often necessary to adopt an approach to organize and simplify the consent process. A staged consent addresses this challenge by spreading the consent process across several encounters. This approach is not unique to genomics studies; researchers have long helped potential research participants prepare for the informed consent process by providing them with consent documents ahead of time. As research has gotten more complex, however, many researchers have explored ways to break down the informed consent process into even more stages, with the goal of making each stage less overwhelming to potential participants. From the perspective of secondary findings, a staged consent approach typically means separating discussion of secondary findings from the initial informed consent process.63 In this approach, research participants will typically receive an initial notification that secondary findings could be generated and then as results are discovered participants are given separate opportunities to decide whether to receive these results. Although this approach provides a number of advantages, there are some drawbacks. Depending on how researchers approach research participants once secondary findings are available, these subsequent conversations may themselves reveal information to research participants that they would not have wanted to know. For example, if a researcher recontacts a research participant to find out whether she wants information about her risk for developing breast cancer, the participant may infer that a high-risk variant was uncovered. Even if a participant is not able to infer which secondary finding has been discovered in their genomic data, the knowledge that “someone else knows something about me that I don’t know” can create a dynamic in which participants feel trapped into receiving information they would not otherwise have wanted. These dynamics are not fatal flaws of the staged consent approach, however, as they can be avoided through careful design of study procedures.

5.2 ­Modular consent Modular consent is another approach that can be used to manage the complexity of genomic research studies that involve the disclosure of secondary findings. Fundamentally a modular consent is simply a consent process that utilizes distinct elements that can be rearranged and reused. Most informed consent documents involve some degree of modularity, as IRBs and RECs often require that investigators use template language to address specific issues. However, this particular use of modular elements can create more problems than they solve, as IRB-supplied consent templates are often complex and include exculpatory language.64,65 Ideally, modular consent approaches can be used to simplify the informed consent process. For example, multimedia tools like explanatory videos and animations can make certain elements of the informed consent process more understandable and accessible.66 Multimedia elements can also create a welcome mental break and transition during a process that can be monotonous and overwhelming. However, researchers getting individual research studies off the ground typically do not have

5 ­Consent models

sufficient time or funding to generate their own bespoke multimedia resources that are study specific. Modular consent designs make it possible to reuse multimedia resources that explain specific elements and intermix them with written or verbal explanations that are specific to the study. Staged consents are also typically modular, since separating a consent into several stages necessitates dividing the consent into distinct elements.

5.3 ­Family consent Since genetic variants can be heritable, genetic test results, including secondary findings, often create implications for family members. If a genetic variant conferring risk for cancer is found in a research participant, for example, other family members might want to undergo testing to identify whether they, too, inherited this risk. Typically, when genetic results potentially relevant to family members are uncovered in clinical or research settings, a reactive approach called cascade testing is utilized. In this approach, family members who might also have inherited a finding are contacted and encouraged to seek testing for themselves. While this approach is used widely, some research ethicists have proposed a more proactive approach. That is, instead of waiting until results are uncovered, family members who might be affected by secondary findings can be linked into the research from the beginning. This approach allows for family members to make their own prospective decisions about whether they would like to receive secondary findings and potentially establish ongoing contact with researchers through mobile or other digital platforms.67

5.4 ­Broad consent The broad consent model is not, strictly speaking, a consent model designed to address secondary findings. Rather broad consent refers to an approach used primarily in biorepository and cohort studies in which researchers ask research participants for permission to utilize data and biosamples for unspecified future research. Broad consent is often distinguished from categorical consent, in which research participants agree for their data and biosamples to be used for certain categories of future research.68 Frequently, categorical consent is used in research studies focused on a particular disease. For example, investigators collecting data and biosamples from children with asthma may request permission to reuse the same data and biosamples in future research studies focused specifically on asthma. Tiered consent refers to a strategy where research participants are given options about how widely their data and samples can be reused, such as giving participants a choice between broad research uses and categorical research uses. Although these distinctions focus on future research uses, they can carry important implications for management of secondary findings. In research studies designed to utilize study data and biosamples from each participant only once, it is relatively easy to anticipate, explain, and control the types of secondary findings

93

94

CHAPTER 5  Informed consent and decision-making

that might be generated. When studies are designed to return secondary findings to participants and to make data and biosamples available for future research uses, the potential for generating secondary findings becomes much more open-ended. At the time participants grant broad consent for the use of their data and biosamples, it is typically unknown which future research studies will take place. This is the central reason that broad consent remains controversial: how can a participants’ agreement to open-ended use of their data and biosamples be considered “informed” when no one, including the researcher, know what those future uses might be? Because these future uses are unspecified at the time of consent, the future secondary findings that might be generated are also unknown. Studies that combine broad consent with disclosure of secondary findings therefore create particularly thorny ethical issues related to informed consent. A number of solutions have been suggested to address this problem. In fact, staged consent, as discussed earlier, provides a relatively straightforward strategy for managing secondary findings that are generated in the course of future research but were not anticipated at the time of the initial consent. This approach can also be incorporated into an online platform in which participants are able to access new results overtime, while at the same time maintaining control over how their samples are used and which secondary findings are reported.69

6 ­Conclusion In this chapter we have examined the important role a robust informed consent process can play in ensuring that secondary findings are managed in ways that meet research participants’ needs and preferences. We have also explored why pretest genetic counseling is considered the gold standard for helping patients and research participants make good decisions and why a great deal of contemporary genomics research is focused on replacing genetic counselors with alternative approaches. As with many of the chapters in this book, there is still a great deal more that could be said. As we have pointed out repeatedly in this chapter, the complexity of genomics and genomics research has created novel challenges related to informed consent. In fact the complexity of genomics has created a fascinating dynamic over the past decade in which the conventional principles and strategies of research ethics have needed to be revisited in light of new genomic study designs, including challenges related to secondary findings, that had not previously been envisioned.70 This dynamic relationship between innovation in genomic research and innovation in research ethics has proven to be extremely generative to both. In the coming years, it is likely that genomics researchers and research ethicists will continue to learn from one other, creating new opportunities in both fields to uncover new insights that had previously been obscured. We hope this chapter will provide a place for those unfamiliar with these fields to gain their footing, and then subsequently to join this important work of finding new and better ways to address participants’ needs in increasingly complex forms of human research.

­References

­References 1. Lanzerath D. Incidental findings in genetic research: ethical issues. In: Incidental Findings. Scientific, Legal and Ethical Issues. Köln: Deutscher Ärzte-Verlag; 2014:73–81. 2. Lohn Z, Adam S, Birch PH, Friedman JM. Incidental findings from clinical genome-wide sequencing: a review. J Genet Couns. 2014;23(4):463–473. 3. Beauchamp TL, Childress JF. Principles of Biomedical Ethics. USA: Oxford University Press; 2001. 4. Richardson  HS. Autonomy’s many normative presuppositions. Am Philos Quart. 2001;38(3):287–303. 5. Brandt R. A Theory of the Right and the Good. Oxford: Clarendon; 1979. 6. National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research. The Belmont Report: Ethical Principles and Guidelines for the Protection of Human Subjects of Research. Bethesda, MD: National Institutes of Health; 1979. 7. Kollek R. Informed Consent. Comment on article 6. Universal Declaration on Bioethics and Human Rights. Background, Principles and Application. Paris: UNESCO Publishing; 2009.123–138. 8. Manson  NC, O’Neill  O. Rethinking Informed Consent in Bioethics. Cambridge: Cambridge University Press; 2007. 9. Brownsword R. The cult of consent: fixation and fallacy. King’s Law J. 2004;15(2):223–251. 10. Christman J. The Politics of Persons: Individual Autonomy and Socio-Historical Selves. Cambridge: Cambridge University Press; 2009. 11. Donchin  A. Autonomy and interdependence: quandaries in genetic decision-making. In: Relational Autonomy. Feminist Perspectives on Autonomy. Oxford: Oxford University Press; 2000.236–258. 12. Wolf SM, Lawrenz FP, Nelson CA, et al. Managing incidental findings in human subjects research: analysis and recommendations. J Law Med Ethics. 2008;36(2):219–248. 211. 13. Wolf SM, Annas GJ, Elias S. Patient autonomy and incidental findings in clinical genomics. Science. 2013;340(6136):1049–1050. 14. Dal-Ré  R, Katsanis  N, Katsanis  S, Parker  LS, Ayuso  C. Managing incidental genomic findings in clinical trials: fulfillment of the principle of justice. PLoS medicine. 2014;11(1):e1001584. 15. Mestan KK, Ilkhanoff L, Mouli S, Lin S. Genomic sequencing in clinical trials. J Transl Med. 2011;9(1):222. 16. Mardis ER, Ding L, Dooling DJ, et al. Recurring mutations found by sequencing an acute myeloid leukemia genome. N Engl J Med. 2009;361(11):1058–1066. 17. O’Roak BJ, Deriziotis P, Lee C, et al. Exome sequencing in sporadic autism spectrum disorders identifies severe de novo mutations. Nat Genet. 2011;43(6):585. 18. Vilariño-Güell C, Wider C, Ross OA, et al. VPS35 mutations in Parkinson disease. Am J Hum Genet. 2011;89(1):162–167. 19. Völzke H, Alte D, Schmidt CO, et al. Cohort profile: the study of health in Pomerania. Int J Epidemiol. 2010;40(2):294–307. 20. Greulich H. The genomics of lung adenocarcinoma: opportunities for targeted therapies. Genes Cancer. 2010;1(12):1200–1210. 21. EURAT P. des menschlichen Genoms T. Eckpunkte für eine Heidelberger Praxis der Ganzgenomsequenzierung. Heidelberg: Marsilius-Kolleg der Universität Heidelberg; 2013.

95

96

CHAPTER 5  Informed consent and decision-making

22. Van Ness  B. Genomic research and incidental findings. J Law Med Ethics. 2008;36(2):292–297. 23. Craddock  NJ, Jones  IR. Wellcome Trust Case Control Consortium. Genome-wide association study of 14,000 cases of seven common diseases and 3,000 shared controls. Nature. 2007;447:661–678. 24. Worthey EA, Mayer AN, Syverson GD, et al. Making a definitive diagnosis: successful clinical application of whole exome sequencing in a child with intractable inflammatory bowel disease. Genet Med. 2011;13(3):255. 25. Nicolas  G, Wallon  D, Charbonnier  C, et  al. Screening of dementia genes by wholeexome sequencing in early-onset Alzheimer disease: input and lessons. Eur J Hum Genet. 2016;24(5):710. 26. Clayton EW. Incidental findings in genetics research using archived DNA. J Law Med Ethics. 2008;36(2):286–291. 27. Meacham MC, Starks H, Burke W, Edwards K. Researcher perspectives on disclosure of incidental findings in genetic research. J Empir Res Hum Res Ethics. 2010;5(3):31–41. 28. Rotimi CN, Marshall PA. Tailoring the process of informed consent in genetic and genomic research. Genome Med. 2010;2(3):20. 29. Lucassen A, Houlston RS. The challenges of genome analysis in the health care setting. Genes. 2014;5(3):576–585. 30. Kalia SS, Adelman K, Bale SJ, et al. Recommendations for reporting of secondary findings in clinical exome and genome sequencing, 2016 update (ACMG SF v2.0): a policy statement of the American College of Medical Genetics and Genomics. Genet Med. 2016;17. 31. Lantos  J. Informed consent. The whole truth for patients. Cancer. 1993;72(9 Suppl):2811–2815. 32. Beskow LM, Friedman J, Hardy NC, Lin L, Weinfurt KP. Simplifying informed consent for biorepositories: stakeholder perspectives. Genet Med. 2010;12(9):567. 33. McGuire AL, Beskow LM. Informed consent in genomics and genetic research. Annu Rev Genomics Hum Genet. 2010;11:361–381. 34. Berg JS, Khoury MJ, Evans JP. Deploying whole genome sequencing in clinical practice and public health: meeting the challenge one bin at a time. Genet Med. 2011;13(6):499. 35. Bredenoord AL, NC O‐M, Van Delden JJ. Feedback of individual genetic results to research participants: in favor of a qualified disclosure policy. Hum Mutat. 2011;32(8):861–867. 36. Green RC, Berg JS, Grody WW, et al. ACMG recommendations for reporting of incidental findings in clinical exome and genome sequencing. Genet Med. 2013;15(7):565. 37. Boyd S, Galli S, Schrijver I, Zehnder J, Ashley E, Merker J. A balanced look at the implications of genomic (and other “omics”) testing for disease diagnosis and clinical care. Genes. 2014;5(3):748–766. 38. Knoppers  BM, Dam  A. Return of results: towards a lexicon? J Law Med Ethics. 2011;39(4):577–582. 39. Appelbaum PS, Waldman CR, Fyer A, et al. Informed consent for return of incidental findings in genomic research. Genet Med. 2014;16(5):367. 40. Hart MR, Biesecker BB, Blout CL, et al. Secondary findings from clinical genomic sequencing: prevalence, patient perspectives, family history assessment, and health-care costs from a multisite study. Genet Med. 2018;. 41. Cho MK. Understanding incidental findings in the context of genetics and genomics. J Law Med Ethics. 2008;36(2):280–285.

­References

42. Brothers KB, Lynch JA, Aufox SA, et al. Practical guidance on informed consent for pediatric participants in a biorepository. Mayo Clin Proc. November 2014;89(11):1471–1480. 43. Berkman BE, Hull SC, Biesecker LG. Scrutinizing the right not to know. Am J Bioeth. 2015;15(7):17–19. 44. Knoppers BM, Avard D, Senecal K, Zawati MH. Return of whole-genome sequencing results in paediatric research: a statement of the P3G international paediatrics platform. Eur J Hum Genet. 2014;22(1):3–5. 45. Selkirk CG, Weissman SM, Anderson A, Hulick PJ. Physicians’ preparedness for integration of genomic and pharmacogenetic testing into practice within a major healthcare system. Genet Test Mol Biomark. 2013;17(3):219–225. 46. Molster C, Charles T, Samanek A, O’Leary P. Australian study on public knowledge of human genetics and health. Publ Health Genom. 2009;12(2):84–91. 47. Lanie AD, Jayaratne TE, Sheldon JP, et al. Exploring the public understanding of basic genetic concepts. J Genet Couns. 2004;13(4):305–320. 48. Henneman L, Timmermans DR, van der Wal G. Public experiences, knowledge and expectations about medical genetics and the use of genetic information. Commun Genet. 2004;7(1):33–43. 49. Carver RB, Castera J, Gericke N, Evangelista NA, El-Hani CN. Young adults’ belief in genetic determinism, and knowledge and attitudes towards modern genetics and genomics: the PUGGS questionnaire. PLoS One. 2017;12(1):e0169808. 50. Bombard  Y, Brothers  KB, Fitzgerald-Butt  S, et  al. The responsibility to recontact research participants after reinterpretation of genetic and genomic research results. Am J Hum Genet. 2019;104(4):578–595. 51. Harrison SM, Rehm HL. Is ‘likely pathogenic’ really 90% likely? Reclassification data in ClinVar. Genome Med. Nov 21 2019;11(1):72. 52. Childerhose JE, Finnila CR, Yu JH, et al. Participant engagement in translational genomics research: respect for persons-and then some. Ethics Hum Res. 2019;41(5):2–15. 53. Green RC, Goddard KA, Jarvik GP, et al. Clinical sequencing exploratory research consortium: accelerating evidence-based practice of genomic medicine. Am J Hum Genet. 2016;99(2):246. 54. Nobile H, Vermeulen E, Thys K, Bergmann MM, Borry P. Why do participants enroll in population biobank studies? A systematic literature review. Expert Rev Mol Diagnost. 2013;13(1):35–47. 55. Kauffman  TL, Irving  SA, Leo  MC, et  al. The NextGen Study: patient motivation for participation in genome sequencing for carrier status. Mol Genet Genomic Med. 2017;5(5):508–515. 56. Minari J, Brothers KB, Morrison M. Tensions in ethics and policy created by National Precision Medicine Programs. Hum Genom. 2018;12(1):22. 57. Bernhardt  BA, Biesecker  BB, Mastromarino  CL. Goals, benefits, and outcomes of genetic counseling: client and genetic counselor assessment. Am J Med Genet. 2000;94(3):189–197. 58. Hoskovec JM, Bennett RL, Carey ME, et al. Projecting the supply and demand for certified genetic counselors: a workforce study. J Genet Couns. 2018;27(1):16–20. 59. Stoll  K, Kubendran  S, Cohen  SA. The past, present and future of service delivery in genetic counseling: keeping up in the era of precision medicine. Am J Med Genet C. 2018;178(1):24–37. 60. Gottesman O, Kuivaniemi H, Tromp G, et al. The electronic medical records and genomics (eMERGE) network: past, present, and future. Genet Med. 2013;15(10):761–771.

97

98

CHAPTER 5  Informed consent and decision-making

61. Amendola LM, Berg JS, Horowitz CR, et al. The clinical sequencing evidence-generating research consortium: integrating genomic sequencing in diverse and medically underserved populations. Am J Hum Genet. 2018;103(3):319–327. 62. Tabor HK, Jamal SM, Yu JH, et al. My46: a Web-based tool for self-guided management of genomic test results in research and clinical settings. Genet Med. 2017;19(4):467–475. 63. Appelbaum PS, Parens E, Waldman CR, et al. Models of consent to return of incidental findings in genomic research. Hastings Cent Rep. 2014;44(4):22–32. 64. Paasche-Orlow MK, Brancati FL, Taylor HA, Jain S, Pandit A, Wolf MS. Readability of consent form templates: a second look. IRB. 2013;35(4):12–19. 65. Paasche-Orlow  MK, Brancati  FL. Assessment of medical school institutional review board policies regarding compensation of subjects for research-related injury. Am J Med. 2005;118(2):175–180. 66. Anderson  EE, Newman  SB, Matthews  AK. Improving informed consent: stakeholder views. AJOB Empir Bioeth. 2017;8(3):178–188. 67. Minari J, Teare H, Mitchell C, Kaye J, Kato K. The emerging need for family-centric initiatives for obtaining consent in personal genome research. Genome Med. 2014;6(12):118. 68. Garrison NA, Sathe NA, Antommaria AH, et al. A systematic literature review of individuals’ perspectives on broad consent and data sharing in the United States. Genet Med. 2016;18(7):663–671. 69. Kohane IS, Mandl KD, Taylor PL, Holm IA, Nigrin DJ, Kunkel LM. Reestablishing the researcher-patient compact. Science. 2007;316(5826):836–837. 70. Brothers KB, Rivera SM, Cadigan RJ, Sharp RR, Goldenberg AJ. A Belmont Reboot: building a normative foundation for human research in the 21st century. J Law Med Ethics. 2019;47(1):165–172.

CHAPTER

Reporting of secondary findings in genomic research: Stakeholders’ attitudes and preferences

6

Gesine Richtera,b, Eva De Clercqc, Marcel Mertzd, Alena Buyxe a

Institute of Experimental Medicine, Division of Biomedical Ethics, Kiel University, University Hospital Schleswig-Holstein, Kiel, Germany b Institute of Epidemiology, Kiel University, University Hospital Schleswig-Holstein, Kiel, Germany c Institute for Biomedical Ethics, University of Basel, Basel, Switzerland d Working Group Research/Public Health Ethics & Methodology, Institute for History, Ethics and Philosophy of Medicine, Hannover Medical School, Hannover, Germany e Institute for History and Ethics of Medicine, Technical University of Munich (TUM), Munich, Germany

1 ­Introduction and background When bioethics began to emerge as a distinct field in the 1960s and 1970s, it was a predominantly theoretical, normative endeavor. It was interdisciplinary from its very beginning, with academic disciplines such as medicine, law, philosophy, and theology presiding over “the birth of bioethics”.1, 2 However, for a field concerned with very practical matters in medical research and clinical decision-making, it did take a surprisingly long time to incorporate empirical methods and findings into its discussions in a systematic way. Developing sound ways of integrating the views on bioethical issues of patients, health professionals, and the lay public into decisionmaking and policy was certainly not an early priority.1 When, however, this did start to happen increasingly (also against the background of a rise of the evidence-based medical paradigm), an intense debate ensued. True to bioethics’ nature as a discipline asking whether, why and how something should be done, for roughly two decades in the 1990s and the 2000s, it was hotly contested if “the empirical turn” in bioethics was justified, problematic, or long overdue (to name just a few contributions from an extensive debate, see Refs. 1, 3–13). Among many things, protagonists in these debates examined the reasons for including empirical data on stakeholder views and preferences; they discussed the quality of these initiatives within bioethics; they

Secondary Findings in Genomic Research. https://doi.org/10.1016/B978-0-12-816549-2.00006-0 © 2020 Elsevier Inc. All rights reserved.

99

100

CHAPTER 6  Reporting of secondary findings in genomic research

touched on potential methodological conflicts between bioethics and disciplines that had long pursued empirical research, such as the social sciences; and most importantly, they inquired about the proper place and function of empirical findings within the process of normative reasoning. How could simplistic “is/ought fallacies”a be avoided? Was there a danger that referring to empirical data could replace normative analysis and argument? Conversely, was policymaking without hearing the views of those affected by a particular policy intrinsically paternalistic? For the most part, these discussions have quietened somewhat. Obviously, there remains tension between an empirical and a normative approach to bioethics; and there are still many open questions, particularly around the best way to integrate diverse methodologies and fields,14 and on how to make sure the quality of bioethical empirical studies is as high as possible.15, 16 However, at this point it is fair to state that at the very least, empirical evidence has an important place in bioethics.17 If done in a careful and proportionate way, it can support and enhance clinical decisionmaking and policy development in many ways, and it can lend greater legitimacy to policies and governance approaches. In the following chapter, we briefly summarize the main points from the debate on whether and why empirical evidence should be integrated into decision-making and policy in bioethics (Section 2). We then turn to the case of feedback on secondary findings in genomic research and provide an overview of tools and instruments on how to elicit stakeholder preferences (Section 3), before we offer a summary of the rich evidence that is available on the views of participants, health professionals, and the lay public regarding secondary findings (Section 4). The example of feedback on secondary findings in genomics research, particularly in large-scale, biobank-based research, is highly suitable to illuminate not only the opportunities but also the difficulties of including empirical evidence in decision-­making and policy (Refs. 18, 19, Chapter 5). Important ethical questions are involved—How much genetic information should be disclosed? Why? When? And for whom? Is there a right to that knowledge? These questions can likely be answered only incompletely without a good understanding of the implications different policy options have for the lives of those involved and how these options and implications are indeed assessed by those involved. Not surprisingly, there is now a wealth of studies aiming to answer exactly these questions. We hope that this chapter provides those with an interest in the complex question on how to integrate empirical methods and data into a normative field with enough material to appreciate how much bioethics has developed in this regard. We also hope that those who are simply looking for an overview on stakeholder views on secondary findings come away from reading this chapter with an appreciation that while eliciting stakeholder preferences might be important, it is certainly not simple.

a

No direct normative ethical inference is possible from empirical data alone, as it is logically fallacious to infer a normative conclusion exclusively out of empirical premises. At least one normative premise besides the empirical premises is needed for a valid inference.

2 ­ Why measure preferences? Normative ethical questions

2 ­Why measure preferences? Normative ethical questions If empirical research increasingly becomes a part of the methods used in bioethics (called by some “empirical bioethics”), it is important to consider the following crucial questionb: Should empirical data about stakeholders’ perspectives (preferences, attitudes, opinions, etc.) have an influence on the normative ethical stance we have toward this topic, that is, on what we think should morally be done (or not be done) or on what is morally right or wrong? After all, even following the aforementioned empirical turn, bioethics remains a predominantly normative discipline. How, then, should empirical data on preferences be included as part of a normative enterprise? Should we change our stance because of the preferences of the stakeholders we measured? If yes, under what circumstances, and why exactly? Additionally, is it, from an epistemological point of view, reasonable or legitimate to change a stance on these grounds at all? Questions such as these are not answered by just doing an empirical study of stakeholders’ preferences; they require normative ethical and metaethical considerations.c In the following, a number of typical methodological (Section 2.1) and epistemological (Section 2.2) answers to these questions will be considered.

2.1 ­Methodological answer: It depends on the validity of the empirical research Whether empirical data should have any legitimate bearing on our normative ethical stance partly depends on the quality of the empirical study and on the subsequent validity of the results. Changing our stance based on questionable data would not be reasonable. If the method that was used introduced multiple biases that confounded the results (internal validity), or if broad and general conclusions were drawn based on data that lacked the potential for generalizability (external validity) (e.g., Ref. 20), then data should not be used to convince us of a specific stance. Additionally, it might be questionable if the results are valid in the sense of construct validity, that is, whether what was actually measured by items in a questionnaire is also what was intended to be measured in the first place. If we have doubts about validity, we should refrain from basing too much of our normative ethical stance on that data.

b

“Empirical bioethics” refers here to a methodological position that proposes to explicitly combine (socio-) empirical data gathering and analysis with normative ethical arguments or analysis (see Ref. 15), which relies substantially on interdisciplinary research practices. This concept differs from an understanding of the term which defines it as empirical studies done in bioethics or about bioethical topics without (explicit) reference to any kind of normative ethical argumentation or analysis. c Metaethics is concerned with the metaphysical, epistemological, logical, and linguistic basis of ethics (e.g., do ethical values exist in nature, can ethical norms be known or are they “constructed,” are there ethical norms that can claim universal validity). Normative ethics is concerned with determining what is morally right or wrong (or good or bad) and why it is so (e.g., the basis for reason, values, norms, rules, and criteria).

101

102

CHAPTER 6  Reporting of secondary findings in genomic research

2.2 ­Epistemological answer: It depends on the intended use of the empirical data The methodological answer alone falls short. Even if we had “ideal” data about preferences, whether and how the data should rationally influence normative ethical stances would not be answered entirely. As no direct normative ethical inference is possible from empirical data alone (see footnote a), it must also to be demonstrated how the empirical and the normative ethical parts can convincingly be “bridged,” that is, by clarifying what exact epistemological use (function) or relevance empirical statements have for ethical arguments or ethical decision-making. This bridging depends heavily on metaethical and normative ethical background premises. Obviously, starting from different background premises can lead to different results regarding the use and relevance of stakeholders’ preferences. Such premises are part of ethical theories or normative ethical frameworks (see Ref. 21). Some of these theories will deny relevance of preferences entirely, which is the case with strong deontological positions. These positions claim, essentially, that if an action is fundamentally wrong a priori, no preferences of stakeholders whatsoever will ever be able to change that.d In the following discussion, however, it will be presumed that empirically measured preferences can have some use or relevance in normative ethical decisions. At least the following six uses of empirically measured preferences are possible (e.g., Refs. 21, 22):

2.2.1 ­Gaining insights into current moral stances or into moral behavior

Preferences can be a means for gaining insights into current moral stances in a population or social group, or they can be a means for getting a deeper understanding of why people do, or do not, adhere to ethical norms, rules, principles, etc. These uses are arguably the most descriptive ones for empirically measured preferences—which also, at least prima facie, renders unclear how they might be relevant for normative goals. This occurs because, in themselves, such insights do not automatically carry any normative implications. However, such insights may help us to better understand the social practice under consideration, which might result in an ethical analysis or judgment that is more sensitive. Nevertheless, there is no immediate relevance for normative ethical arguments or analysis because such insights are often more likely to be a means for other uses, as described in the following sections.

d

Deontology is an ethical theory in which an action is morally right when it corresponds to a moral duty or rule that is defined, for the most part, a priori. Often, a more or less clear hierarchy of moral duties or rules can be ascertained for such theories. The most prominent example is Kantian ethics with its categorical imperative as the central moral principle; any action contradicting the categorical imperative is morally wrong a priori, independent of any preferences.

2 ­ Why measure preferences? Normative ethical questions

2.2.2 ­Identifying ethical problems or aspects

As Salloch et al.21 suggest, another use of preferences can be to identify ethical problems or aspects that ethical theory or “ethics experts” have not (yet) recognized. By assessing preferences that are measured by questionnaires or interpreted in interviews, ethicists might discover (additional) ethical conflicts that enrich the analysis of an ethical problem or shed light on new aspects that are ethically relevant. Preferences can also indicate whether accepted ethical principles might be violated in practice. If, for example, many patients want to be informed about secondary findings and secondary findings are regularly not shared with them, this could be a hint that an ethical problem (e.g., lack of respect for patient autonomy or the value of self-determination) is not being sufficiently considered in practice.

2.2.3 ­Solidifying or contextualizing (specifying) accepted norms or principles

Another use of preferences is solidifying or contextualizing ethical norms or principles. In the widely used principlism approache (e.g., Ref. 23), this is also called “specifying.” The norms or principles themselves are regarded as morally right or ethically justified, but for determining what concrete action is to be taken, it is necessary to enrich a principle with empirical content. What action is exactly implied by “respecting patient autonomy” in regard to sharing or not sharing secondary findings? Also, if secondary findings should be shared, how should this be done? Here, preferences can help to specify when, how, and under what circumstances secondary finding should be shared or not shared as a result of respecting a patient’s right to self-determination. In principlist approaches, however, all principles have to be considered, and when these principles are in conflict, they are weighed against each other. For this weighing, in turn, preferences can also be relevant if they clarify which principle a patient would prioritize. It must be borne in mind, however, that the results of empirical studies, at least when they come in the form of summary statistical statements, do not automatically hold true for single cases. While the particular preferences of a particular patient are relevant for specifying and weighing, the aggregated data of a study empirically measuring preferences of many patients might not be relevant (although such a study could provide some “reference points”).

2.2.4 ­Justifying norms: Information regarding acceptability/practicality

Preferences can be used for adapting norms (or principles) because of acceptability or practicability. This is a weak approach to the justification of norms. The rationale behind this use is that if preferences and norms are in stark contrast, people will e

Principlism is a bioethical approach in which, instead of a full-blown ethical theory, several ­midlevel principles are defined, such as respect for (patient) autonomy, nonmaleficence, beneficence, and justice, which in turn allow concrete actions corresponding to each principle to be derived (e.g., when following nonmaleficence, X should be done, and when following beneficence, Y should be done). No principle has a higher a priori relevance than another one; therefore, principles have to be weighed and balanced against each other a posteriori, that is, after a close analysis of the specifics of the case at hand.

103

104

CHAPTER 6  Reporting of secondary findings in genomic research

p­ robably not adhere to the norm. For this reason, it is not reasonable to implement such an (unmodified) norm into practice. In this scenario, preferences only have an instrumental role. They are used for anticipating if a norm that is deemed morally right or ethically justified in its own right would be accepted and/or effective if implemented, and for determining how the norm might be changed to achieve better compliance in practice.24 It is important to note, however, that this framework involves a set of assumptions. It assumes, for example, that the successful implementation of a norm in social practice is itself important—that it carries normative weight. Not every ethical approach would accept this assumption; one might argue, for example, that doing the right thing is critical even if it is difficult to implement in practice. For this reason, this justification for the use of empirical data in normative deliberation requires a set of metaethical premises that prioritize the practical application of norms in “the real world.”

2.2.5 ­Justifying norms/single actions: As a component in coherentism

Preferences can also be used as a possible component—among many others—for the justification of norms or single actions. This takes place especially in coherentistf approaches that aim to find a coherent final statement via ethical reflection, that is, a statement that “fits” the rest of all the other statements that are considered. In such an approach, preferences can be used as “well-considered judgments” that are starting points for this reflection. However, it is unclear whether all measured preferences can be deemed “well considered.” The answers to questionnaires or interviews “[…] may be expressions of intense moral reflection, as well as products of self-interest, self-deception or historical or cultural accidents” (Ref. 21, p. 598).

2.2.6 ­Justifying norms/single actions: As (more or less) direct justification

The strongest approach to the justification of norms or single actions is when preferences—more or less—directly justify the rightness (or wrongness) of a norm (principle) or single action. As established earlier, this is not possible as an immediate inference from empirical data alone. However, there are metaethical positions that would claim that moral truths—at least very basic or general ones—can be “grasped” without inferencing them, for example, by moral intuitions (intuitionismg). In these cases, justification is considered prereflective or pretheoretical and is akin to sense experience (e.g., vision). Nevertheless, even when assuming a position such as intuitionism, the measurement of preferences would not truly lead to a pretheoretical justification of an ethical norm or a single action. Although the measurement of preferences could be understood as a way of carving out moral intuitions of stakeholders, this premise is dependent on f

Coherentism is an approach to the justification of moral norms or single actions in which different normative and empirical statements have to be considered to find a (final) statement that is the most coherent, that is, consistent with, or supportive of all these considered statements. g Intuitionism is a metaethical theory that makes moral intuitions the basis of morality (i.e., instead of rationality or discourse), which thus are also the (final) instances for justifying norms or single actions.

2 ­ Why measure preferences? Normative ethical questions

additional methodological premises, such as the aforementioned questions regarding construct validity. In other words, it is not clear whether preferences measured empirically even may “count” as the moral intuitions actually sought by intuitionism. Other more successful strategies for justification involve inferences, although they add a normative premise and are therefore not immediate inferences from empirical data alone. One strategy is to argue that preferences are highly relevant because of the associated participatory importance of the stakeholders involved (see Ref. 25) or due to strong contextualist conceptions of (bio)ethics (see Ref. 26). According to these views, broadly speaking, preferences are relevant because the “moral authority” (determining what is right/wrong or good/bad) is located in social practice itself,26 and not in ethical theories or “ethics experts” that paternalistically decide what, for example, patients should know. Hence, it is relevant to know what professionals or patients actually prefer. If, for example, patients prefer to be informed about secondary findings, then, given participatory/contextualist background premises, it is—prima facie—morally right to do so. There is little doubt that the argument for including more participatory approaches in biomedicine, for example, to counter entrenched paternalism, is strong.27, 28 However, some have accused this stance as amounting to “ethics by opinion poll”,21 which could risk ad populum or “democratic” fallacies.h In particular, this account risks prioritizing ad hoc, unconsidered preferences as if they are preferences that evolved due to long and informed ethical reflections. That said, obviously the results of ethical reflection can still include bias or rely on wrong premises, which might be uncovered by careful elicitation of stakeholder preferences. Other strategies argue that preferences indicate what could be (or will be) positive consequences. For example, for a position such as preference utilitarianismi (e.g., Ref. 29), preferences have special relevance because they are exactly the information that is needed for calculating utility. For such calculations, aggregated data of an empirical study could be quite suitable. Additionally, there is a critical, argumentative element that stands between the measurement of preferences (empirical part) and the ethical judgment (normative part). Nevertheless, such an approach also faces problems, such as the existence of preferences that are stated but do not honestly reflect the preferences held by people versus “true” preferences that people conceal from others (Ref. 29, p. 102). This again raises relevant questions about construct validity. In sum, preferences (empirical data) by themselves are not able to ethically justify norms, principles, or single courses of action. To do that, normative and methodological premises are necessary. However, when they are sought with awareness h

Justifying something as true—or morally right—by just referring to a majority opinion is not valid because all this can show is what people think is true (or right); it cannot establish the truth or rightness of something by itself. Furthermore, the majority view could also be wrong, and a minority view could be true or right. i Utilitarianism is a normative ethical theory in which the single action is morally right that maximizes “utility.” Utility is measured by the consequences for all involved and affected by the action, such as how the consequences increase luck and reduce pain, or how the preferences of those involved are realized or affected (= preference utilitarianism).

105

106

CHAPTER 6  Reporting of secondary findings in genomic research

of their limitation regarding the formulation of normative stances and courses of action, they can be very useful. As described, they can enhance the formulation of norms and rules, they can help uncover hidden ethical conflicts or issues, and they can ensure that the voices of those who are affected by a normative stance are heard in the long run.

2.3 ­Relevance of background premises for guideline development In addition, preferences are not just elicited to determine normative rules. More importantly for our context, empirical research is especially valuable for developing policy. In 6.4, we present a rich array of studies on preferences for secondary findings in genetic biobanking; it is hard to imagine that the complexity of this topic could have been anticipated well enough for proportionate and fair policymaking based on normative, theoretical reflection alone. Preferences help us to understand the ethical dilemmas and challenges at hand. In addition, eliciting preferences can also help better legitimize a policy by involving (different) stakeholders (also see Section 3 in the succeeding text). This refers back to uses such as acceptability and coherentism (Sections 2.2.4 and 2.2.5). Often, it is not so much a problem whether the preferences that are measured are in principle able to justify a norm, but rather how exactly one should balance and weigh the preferences that were measured. In reality, of course, preferences are most often mixed, and it will be quite rare that all stakeholders express the same preferences. Even while there can be clear statistical tendencies, it is still questionable whether a result such as “80% of stakeholders that were interviewed have preference Y” should be sufficient to provide a recommendation to act accordingly in a guideline or policy. Additionally, it could be problematic if stakeholders have different (professional and social) backgrounds, interests, and knowledge bases that represent different groups (e.g., professionals, regulators, patients, and citizens in general). Which preferences from which group “weigh” more in such cases, and what if they have conflicting preferences? Would it be the preferences of the patients (those affected), for example, or the preferences of the professionals or regulators? In order to answer these questions, it is necessary to rely again on the normative ethical premises that determine whether, for example, preferences of a specific group should be weighed more (or less). However, it is possible to help inform these considerations by eliciting stakeholder views on exactly this question. Obviously, methodological premises about how to balance conflicting preferences or how to adequately report and include them in a policy or recommendation become relevant here. In sum, an answer to the question of whether empirical data about preferences regarding secondary findings should have bearing on our normative ethical stance, especially in the context of developing policy, can only be given if the following conditions obtain: (i) The general use of empirical data (the preferences) has been clearly determined (which, in turn, is most often based on normative ethical premises, or in other words what makes a norm or action morally right or wrong).

2 ­ Why measure preferences? Normative ethical questions

(ii) It is clear how the descriptive and the normative parts are related to each other in a given use. (iii) There is some discussion about how conflicting preferences within a stakeholder group and across stakeholder groups might be balanced and weighed against each other (relying on premises from methodology in ethics). (iv) The empirical results are actually sufficiently valid, especially regarding construct validity (which involves premises from empirical methodology). If these conditions obtain, empirical findings can be very illuminating and serve as an important corrective to policymaking that is potentially one-sided. In the following section, examples of these various premises are discussed in further detail for developing policies about secondary findings.

2.4 ­Eliciting preferences for developing policies An instructive case study in this regard is the normative concept of biological citizenship. This notion was introduced by anthropologist Adriana Petryna,30 who describes it as a typically “postsocialist” phenomenon in which citizens demand “access to a form of social welfare based on medical, scientific, and legal criteria that both acknowledge biological injury and compensate for it” (p. 6). Western authors, while expanding on the work of Petryna,31, 32 often use the concept of biological citizenship in a different sense. Rather than referring to demands for financial support from state authorities, here it encompasses not only individuals’ right to know genomic findings but also their responsibility to know and share them with kin and the community at large.33–36 Where does this right and responsibility come from? Although altruism is often featured as an underlying force for population biobanks and individual benefit serves as a determining factor for participation in disease-specific biobanks,37 a recent systematic literature review on participants’ motivations to enroll in biobank studies38 indicates that in most cases the decision to participate in genomic research is driven by a combination of self-interest and altruism, as well as by prosocial motivations such as solidarity.39, 40 This finding might explain the growing consensus about returning results to patients, including those not related to the primary reason for testing.41 The democratization of genetic information can be understood against the background of the overall tendency within the medical landscape to move from top-down paternalism to a focus on the individual choices of the patient or participant, who should be empowered to take part in a shared decision-making process. Nevertheless, although respect for autonomy is a key ethical principle, the question remains about exactly how much freedom participants and patients can and should be granted.42 Biobank studies pose important ethical and practical challenges to participants’ individual acts of choice. First, although most ethicists defend both a person’s right to know and their right not to know,43, 44 participation in biobank research not only is never purely an individual choice but also implies a kind of responsibility to know, since genomic findings could be related to family members.45 This possibility raises the question

107

108

CHAPTER 6  Reporting of secondary findings in genomic research

of whether participants have the right to decline the disclosure of findings that could prevent harm to them or their kin. According to the American College of Medical Genetics and Genomics’ (ACGM) recommendations of 2013,46 they clearly do not. Highly actionable secondary findings should be returned without considering participant preferences. In reaction to a heated debate on how a loss of control over decision-making might deter people from participating in research, the ACMG subsequently altered this stance to grant participants the right to opt out of the analysis for medically actionable genes. Second, since genomic sequencing produces a great number of secondary results with varying degrees of significance and actionability, the question arises whether participants have the right to the maximum number of options.42 The revised ACMG47 recommendations, for example, do not seem to consider the possibility that participants might want to receive findings that are not limited to so-called actionable results.48 Critics maintain that although empirical data on preferences are not enough to guide policy (Ref. 49; also see in the preceding text), recommendations around the return of incidental findings should take into consideration the preferences of all relevant stakeholders, including nonprofessionals, to render the decision-making process more transparent, foster greater public trust, and comply with the ideal of participants and investigators as partners in research.50 Of course, different stakeholders might have different views, and thus some preferences might be fulfilled while others are overruled. Nevertheless, as was stressed in the previous section, policy implementation is likely to benefit from a fair deliberation process, and policymakers will be better placed to inform the public about their decisions if they are aware of their preferences. Third, although eliciting and including participants’ preferences on the return of secondary findings might be the best approach for policy development, this approach still raises the following concerns: (a) Given the large number of potential future secondary results, how can individuals’ preferences be elicited effectively? (b) Given the lay public’s limited knowledge about the production and analysis of secondary results, how can their preferences be elicited faithfully, without extensive counseling (see Section 2.1, concerns regarding validity)? Some authors argue that fully individualized feedback would be too time-­ consuming, lead to unjustified costs, and actually undermine people’s autonomy due to information overload.48 Hence, they suggest developing a general return policy based on a combination of population (rather than individual patient/participant) preferences and professional standards.48 Nevertheless, they emphasize that this approach is not without shortcomings because it bypasses the fact that population preferences might change over time and subgroups within a population might have different preferences based on their family history, personal experiences, or individual group characteristics (e.g., age, gender, ethnicity, or educational background). Hence, the policy should be fine-tuned over time to find equilibrium between individual preferences and practicability.48

3 ­How to measure preferences?

3 ­How to measure preferences? Strategies and tools for eliciting preferences on secondary findings Given the exponential growth of biobank research, the question of whether, when, and how to provide secondary findings to participants continues to merit urgent policy attention. In the following sections, we aim to discuss whose preferences merit being taken into account when developing a policy for returning secondary findings, as well as which (individual or disease-related) factors might influence these preferences and which tools or instruments currently exist to elicit stakeholders’ preferences. We conclude with a brief discussion of when to elicit these preferences.

3.1 ­Whose preferences need to be elicited? The notion of stakeholder is a key concept in organization and project management theory.51 It denotes a “person, group, or organization that affects or can be affected by the planned action(s)”52 of an organization. In the literature on the return of results, relevant stakeholders are generally divided into three broad groups50: professionals, participants, and the general public. The group of professionals has generally been considered a key interlocutor in the development of an adequate return policy for secondary results because of their presumed knowledge in the field. Nevertheless, this group is far from being homogeneous; the attitudes of healthcare professionals, researchers, IRB or ethics committee members, and representatives of funding organizations could differ due to potential differences in interests and expertise. Even among health professionals, opinions might be divided based on the division of skills and responsibilities. Despite their lack of knowledge in genetics, for example, family physicians, are important stakeholders since they are often the ones with whom families get in touch when a positive result is returned. Unlike clinicians, researchers are not involved in the management of clinically relevant results but are concerned with the feasibility of the study and with legal liability. Biobank funders, on the other hand, will consider the extra costs involved in verifying and transferring secondary findings. The primary role of IRB and ethics committee members is to ensure the protection of research participants. Nevertheless, many of them might be unfamiliar with the complexities of biobanks,53 and this might cause “mission creep” based on the real or perceived need to assume responsibility for all potential risks, including responsibility to the researcher and the institutions.54 The participant group includes patients with specific diseases and in different disease stages, “healthy” participants, and fellow group members, such as relatives (with parents of minor participants being a specific subcategory). Although patients lack scientific knowledge, their illness experience is considered valuable by an increasing number of researchers because it might enhance the quality and relevance of the research process and thus complement the professionals’ opinion of the debate.55 The involvement of “healthy” participants is important as well given that to be respectful, biobanks are expected to be responsive to the values and beliefs of

109

110

CHAPTER 6  Reporting of secondary findings in genomic research

research participants.56 Additionally, the inclusion of the family perspective is relevant because secondary findings might have a huge impact on the lives of biological relatives. Given the increase in genomic research in pediatric research settings, the preference of parents or guardians regarding the return of secondary results should be assessed, too. Finally, although parents are expected to act in the child’s best interest, children’s and especially adolescents’ perspectives should be given consideration as well since they are the ones who will bear the consequences of the decision to return secondary results (or not).36 Furthermore, the involvement of minor participants resonates with various ethical guidelines, such as The United Nations Convention on the Rights of Children,57 which increasingly emphasize the importance of involving children and adolescents in the healthcare decision-making process at a level that is appropriate for their development. To date, the academic literature on pediatric biobanking has mainly focused on the parental perspective when addressing the return of secondary findings.36 Not all studies include the public on the list of relevant stakeholders due to their apparent lack of direct involvement.58 At first sight, it might seem unreasonable to elicit the preferences of individuals about a situation which they are not “in”.48 Still, critics point out that the preferences of this group should be considered because its members are all potential participants and potential beneficiaries of biobank research. And although they might be unaware of it, all people have genetic variants.48, 59 As scholars have already pointed out, that does not mean, however, that this group is homogeneous; different subgroups might have different preferences depending on both contextual and personal factors.

3.2 ­Factors influencing preferences: Personal and disease related Several personal characteristics might influence stakeholder preferences regarding the return of secondary results. Older participants, for example, might want to know more results than young people because of their overall greater concern with their health. The perspective of patients affected by a genetic or rare disorder might differ from those with a common multifactorial disease because the impact of results for relatives might be bigger. Social class and educational background might influence stakeholders’ understanding of the potential psychosocial impact of the knowledge of secondary results. Ethnic minorities might have a different viewpoint than the overall population, as their personal life story might make them more susceptible to fear of discrimination. Other background characteristics that could be considered include gender, religion, nationality, and sexuality. It is important to keep in mind, however, that all these characteristics might interact and operate at the same time. Hence, the concept of intersectionality should be employed when analyzing these influencing factors. Next, to stakeholders’ personal characteristics, the attributes of secondary results might also have an impact on their preferences, such as severity of disease, risk of developing disease, possibility of prevention, time to first manifestation, scientific reliability, possibility of treatment, and reproductive relevance (carrier status).48

3 ­How to measure preferences?

3.3 ­Tools or instruments to elicit stakeholders’ preferences To inform the development of adequate preference-setting tools, scholars have conducted online surveys, focus groups, and interviews with stakeholders to better understand their preferences for the return of secondary research results. Given the complexity of the debate and the need to faithfully elicit stakeholders’ preferences, these channels have often been complemented with other strategies, such as (1) workshops in which professionals can openly discuss their viewpoints and deepen the existing ethical, legal, financial, and organizational challenges60 or (2) the deliberative democracy approach in which members of the public are educated by experts on secondary findings, then engage with each other in small groups and express their preferences by voting.61 Both information sessions (e.g., presentations by experts and educational videos) and the possibility to engage with others are believed to contribute to a faithful elicitation process. The findings of these qualitative and quantitative studies have led to the development of different preference-setting models that can be used in biobank research.62 Most models focus on specific disease characteristics (e.g., preventability and/or severity of a condition) or offer the possibility to opt out of certain sensitive disease results (e.g., mental illness and psychological conditions, developmental disorders and learning disabilities, and childhood-onset degenerative conditions). In other models, the results are organized based on a combination of different categories,63 such as disease risk, treatability, carrier status, genetic syndrome, metabolic disorder, and affected organ system. In one study, a preference instrument for secondary results was developed that considers participants’ personal and family experiences with specific medical conditions, such as autism, chronic fatigue, diabetes, Down syndrome, and fibromyalgia.63 The assumption behind this model is that it will be feasible to infer preferences for future secondary genomic results from this small number of preference items. All these tools have their limitations as they all leave out options that certain stakeholders might find important. However, given the large number of potential secondary findings, such trade-offs, is somehow inevitable if one wants to keep the process of eliciting preferences feasible and effective. Despite the increasing involvement of minor patients in biobank research, tools aimed at recording the preferences of children and adolescents with regard to the return of secondary findings are still largely missing. More empirical research is needed to explore minors’ view on the kind of secondary results they want to have returned, as well as whether and how they conceptualize genetic risks. These data can then be used to develop adequate guidelines and tools for eliciting the preferences of children and adolescents.36 Furthermore, differences among children need to be considered as well because they might have an impact on their preferences and some minors might be in good health, whereas others might be disabled or chronically ill.64

3.4 ­When to elicit stakeholders’ preferences? The question regarding when to elicit stakeholders’ preferences primarily concerns biobank participants. There are generally two moments in which participants can

111

112

CHAPTER 6  Reporting of secondary findings in genomic research

be asked about which results they would like to receive: (1) during the informed consent process and (2) when an actual secondary finding is made.65 The difference between these two stages is important as studies have shown that although some participants might want to oppose any return of secondary findings, the mere existence of a finding (once the result is out there) often leads participants to wanting to know it because they do not want others to know things they do not know themselves.63 A computer-based analysis that only selects the results that participants want and filters out those they do not want might overcome this knowledge gap.63 Nevertheless, preferences are not static but often alter due to changing circumstances, such as parenthood and illness.65 A one-time elicitation process during the recruitment process might not be sufficient to capture these changes.50 Finally, within the context of pediatric biobanks, professional guidelines generally recommend restricting the return of secondary findings to results that are treatable during childhood and to defer findings on carrier status with reproductive implications, as well as adult-onset conditions without treatments, until adolescents and children are old enough to express their own preferences and thus decide for themselves. The reasoning behind this approach is that minors should have the right to an open future and parents should not compromise this right by wanting to learn everything about their children’s genetic makeup (Refs. 36, 66). What if the knowledge of genetic risks affects children’s psychosocial well-being? What if parents have access to the findings but shield their children from them? Do minors have the right to genomic privacy (Refs. 36, 67)? Can they express their preferences on the return of results, and what if their choices conflict with those of parents and/or professionals? To tackle these questions, more research is needed on the decisionmaking capacity and the preferences of minor participants with regard to the return of secondary findings.

4 ­What has been found? Stakeholders’ attitudes and perspectives As mentioned in the preceding text, the reporting of secondary findings involves the dilemma of balancing actionability or nonactionability of findings from a medical point of view with the subjective and contextualized attitudes, expectations, misconceptions, and preferences of all stakeholders involved. In order to shed light on the attitudes and preferences of different stakeholders, it is essential to distinguish between actionable findings on the one side and nonactionable findings on the other side. The literature on stakeholders’ attitudes and preferences has to be interpreted with caution, and the results cannot easily be compared since the empirical studies are based on different presumptions concerning the differentiation between actionable and nonactionable items.68 Apart from this issue, it is essential to be aware of the individual situations of stakeholders. Stakeholders include patients with different kinds and different levels of diseases, patients with rare diseases, patients with genetic disorders, and their

4 ­ What has been found? Stakeholders’ attitudes and perspectives

relatives, as well as healthcare professionals in genetics or other specialties, primary care physicians, clinicians of different specialties, genetics and other researchers, and institutional review boards. When analyzing studies to learn about stakeholders’ attitudes and preferences, it has to be kept in mind that these studies have been conducted on different populations, in different countries with different policies and cultural backgrounds; most of the studies were carried out in North America.68 What is common in most stakeholders is that their attitudes are influenced by a variety of constant and temporary factors, such as general values, social and cultural influences, information and education, and the actual (health) situation of their lives.69 Similarities between the attitudes of all stakeholders have been observed concerning the key aspect that it is ethical to disclose secondary findings because it grants certain health and emotional benefits to participants.70 In general, when studies distinguished between actionable and nonactionable findings, the majority of all stakeholders who were interviewed were in favor of receiving actionable secondary findings, but there are still a remarkable number of all stakeholders who wished to receive nonactionable secondary findings.68 However, a closer look reveals that stakeholder groups differ in their attitude toward the following50: – Preferences for secondary findings – Impacts and implications of secondary findings for the participant – Impact and implications of secondary findings for the family and the relatives of the participant – Understanding and literacy – The information process before consent to genetic research – The process after disclosure of secondary findings – Rights and responsibilities – Policies and practices By recognizing how the attitudes and preferences of all stakeholders differ or coincide, it becomes more feasible to develop recommendations or strategies that are capable of respecting all concerns during the process of healthcare decision-making.

4.1 ­Professionals’ attitudes toward disclosing or not disclosing: Clinicians, researchers, and IRB members 4.1.1 ­Preferences for secondary findings

In general, the majority of clinicians and researchers have agreed that patients and participants should have the choice to receive at least some secondary findings (e.g., Ref. 71). However, in regard to the question concerning the amount of information that should be disclosed, there are not only varying positions within groups of professionals but also a disconnect between the views of clinicians and researchers on the one side and participants and patients on the other side.68, 72 The professionals’ positive attitude to support disclosure of actionable secondary findings is about as high as the participants’ preference.68 Be that as it may,

113

114

CHAPTER 6  Reporting of secondary findings in genomic research

this ­support differs according to experience and field of specialization. Studies have shown that genetic healthcare professionals were slightly less supportive than nongenetic healthcare professionals.73 Additionally, researchers with no clinical training and no direct interaction with patients had a slightly more positive attitude toward the return of secondary findings, and this trend also applied to nonactionable findings in which no intervention is possible or known.74 Studies concerning professionals’ preferences for secondary findings mainly deal with the questions of what to disclose as well as which criteria to use. It has been observed that professionals were in favor of creating guidelines but only if they were flexible enough to be adapted to individual situations.75 Aspects such as clinical and laboratory best practices, patients’ preferences, the guidance for practice, and the meaning of the informed consent seem to guide professionals in their decision to disclose or not to disclose secondary findings.76 More practical aspects might influence the decision to return secondary findings, such as problems in terms of time, costs, and logistics for follow-up, or sometimes the limited availability of genetic counselors in certain centers,75 which is something institutional review boards are well aware of.77 In general, professionals agree that individual preferences should guide the return of secondary findings and that the choice of patients has to be respected (Refs. 78, 79). Studies have shown that institutional review boards share the position that actionable secondary findings in a research context must be returned, but only if the patients want to know the result.80 A special case involves secondary findings for children in which parents sign the consent form for research purposes. In this case, professionals face the question of not only what to disclose but also to whom they should disclose.81 It has been observed that professionals think that children and their parents have the right to choose to be informed or not to be informed about secondary findings. A majority of professionals believe that clinically actionable variants before adulthood should be reported.71 Yet, some professionals think parents should not have access to data on adult-onset diseases.82 Only a few studies have touched on the question of the position of professionals concerning the right of the parents to know or not to know versus the right of the child to know or not know the results. Some studies concluded that there is a broad consensus among professionals that parents should have the right to opt out of receiving secondary findings and that there is no consensus about the ethics of justifying disclosure based on parental benefit.82 There is hardly any literature on the question of how professionals would act in a case in which parents refuse to receive knowledge of any secondary findings and professionals uncover an actionable secondary finding. Concerning the disclosure of nonactionable findings, professionals take a much more negative attitude than participants and patients. Studies have shown that for the professionals’ decisions, the criteria of actionability or nonactionability are supplemented by further criteria that influence the decision to return data or not return data, such as questions of quality and validity of information, respect for rules and the welfare of the patient.83, 84

4 ­ What has been found? Stakeholders’ attitudes and perspectives

4.1.2 ­Impacts and implications of secondary findings and the question of understanding and literacy

Findings from previous studies also showed that the impact of potential psychological harms of secondary findings were a frequent concern for professionals (e.g., Refs. 78, 85–88). Several studies have shown that professionals were often concerned that participants and patients do not sufficiently appreciate the impacts and implications of secondary findings (e.g., Refs. 71, 89). They often felt that participants’ and patients’ enthusiasm for a wide range of secondary findings that could be reported may result from a lack or an incomplete understanding of the implications.90 This issue leads to the problem of whether professionals can adequately inform participants and patients about their decision on what should be reported back. They raised concerns about overwhelming participants and patients with too much information about the return of secondary findings during the informed consent process, which includes information such as the possibilities of stigmatization and discrimination in insurance and employment (e.g., Ref. 72). Nevertheless, studies have shown that professionals feel unprepared to interpret secondary findings and are concerned about their general genomic knowledge (e.g., Refs. 85, 91). Physicians reported that they had specialty-specific concerns about their preparedness, a lack of laboratory guidance, time pressures, and a lack of standards contributed to them feeling unprepared.77, 91 Moreover, professionals point to problems arising from the rapid development of genetic knowledge. One consequence that professionals are aware of is that there might be deficits in the qualifications of healthcare professionals to manage or interpret secondary findings.71 In addition, professionals raise the concern that knowledge about the clinical importance and relevance of genetic information is currently limited and still developing.92 This limitation could mean that current secondary research findings of unknown relevance could gain relevance in the future. A reporting of current findings could thus lead to misconceptions. It is an open issue how health professionals (and medicine overall) should deal with this challenge in the future.93

4.1.3 ­Information process before consent to genetic research and process after disclosure of secondary findings

Studies observed a general consensus that disclosure should be guided by decisions made during the consent process82, 89, 94 and they attach crucial importance to the pretest process in the form of written information and pretest discussions. Concerning the question of the pretest information process and content, professionals highlighted the often unrealistic expectations and the enthusiasm of participants and patients to know about all secondary findings and the need to counteract this trend.86, 95 Studies have shown that professionals would recommend pretest information that should be flexible and tailored to the recipient.95 Professionals share a general opinion that pretest discussion should include the possibility of secondary findings, but they differ in regard to the details to include, such as details of the problem of false-positive/false-negative results, the possibility of changes in the interpretation of variants, the question of what the results mean for relatives, and the

115

116

CHAPTER 6  Reporting of secondary findings in genomic research

potential for anxiety (e.g., Refs. 72, 95). Although they are aware of the importance of pretest discussions, for professionals the length of time required to achieve informed consent was seen as burdensome.95 The same issue has been observed in their opinions about the time of disclosure. It goes without saying for the majority of professionals that genetic counseling at the time of disclosing the results, face-to-face meetings with a professional who has knowledge about genetics, and the implications of the findings were seen as essential (e.g., Refs. 78, 88). However, when asked about the time for disclosure, a study among US genetic researchers found out that the majority did not want to spend more than 30 min returning these results. At the same time, they expressed concerns that the disclosed information would overwhelm participants.72 Due to an awareness of the psychological implications of disclosure, some healthcare practitioners felt that they would not be able to provide psychological support themselves and that they would require support from specialists specific to the secondary findings that were disclosed.82, 86

4.1.4 ­Rights, responsibilities, policies, and practices

The question of pretest procedures for genetic research is preceded by the crucial aspect of feeling responsible for producing secondary findings and the subsequent question of how to deal with them. There is not much literature about this aspect. A study among Canadian researchers found that only a minority of researchers feel a strong responsibility to look for meaningful secondary results. However, medical geneticists were significantly more likely than genomics researchers to report a feeling of responsibility to examine the data for clinically relevant secondary findings. Yet, when such a result has been uncovered, the majority of researchers strongly think they have an obligation to share the result.96 In general, clinicians and researchers agree that the autonomy of the participant has to be respected, and most of them believe that consent decisions concerning secondary findings have to be respected and should guide the return of findings (e.g., Ref. 78). Questions regarding their professional responsibility are a special concern for clinicians and researchers. Professionals express a strong sense of responsibility for the participants, and they are aware that fulfilling these responsibilities demands careful consideration about the utility of the secondary finding. Therefore, professionals have proposed a multidisciplinary approach to reach a comprehensively informed and responsible decision for disclosure.89 Some professionals anticipate a dilemma between their professional responsibility in their duty to provide warnings and respecting the participant’s decision and autonomy in cases in which the participant has exercised their right not to know any secondary findings (e.g., Refs. 85, 89). In such cases, professionals take into consideration whether they should paternalistically override the participant’s decision if a clinically significant secondary finding has been discovered.71, 72, 85 The fact that genetic knowledge is developing rapidly raises the issue of duration of responsibility,50 which has been mentioned in some studies. Most professionals

4 ­ What has been found? Stakeholders’ attitudes and perspectives

are in favor of obtaining consent to recontact participants, as new interpretations of secondary findings become known.72, 89 However, there are different opinions about the duration of responsibility, which depends on the nature of the project and whether it has a research or a clinical focus. A Canadian study found that researchers felt that their responsibility is restricted to the funding period.96

4.2 ­Participants’ and lay attitudes and preferences Biobank participants have individual interests and needs.97 It has become state of the art to keep their expectations and preferences in mind when thinking about a policy of reporting secondary findings, in order to balance the attitudes and preferences of all stakeholders involved. In this sense, the inclusion of the participants’ perspective is postulated in the literature to offset the dominance of professional viewpoints in the current debate. A policy for reporting secondary findings that exclusively takes into account researchers’ and healthcare professionals’ attitudes, as well as official standards, is widely regarded as paternalistic. It is recommended that a general reporting policy be based on a combination of population and professional preferences, as well as official standards.48 Several studies have tried to identify the “perspectives of actual end users”98 to work out the vast spectrum of attitudes, expectations, preferences, misunderstandings, and misconceptions of a variety of participants in different situations of their lives or disease stages and to obtain insights into the factors that have contributed to their formulation. In the studies cited in this section, different groups of participants in genetic research studies were observed, such as patients with specific diseases, patients with rare diseases, patients with inherited genetic disorders, parents and families of pediatric patients, and patients with different educational, cultural, or religious backgrounds. Studies have tried to work out participants’ understanding of the nature of secondary findings resulting from genetic research and their awareness of the consequences of secondary findings. They have evaluated their attitudes and tried to categorize what participants want to know about secondary findings, what kind of impact they expect on their future lives, what kind of information and support they demand before and after disclosure, and how they assess the implications for their relatives. However, the ethical conduct of clinical research does not end when informed consent is obtained. Consequently, knowledge about expectations and preferences of participants who have been affected by and have received secondary findings is needed to complete the debate about a strategy for reporting secondary findings. Until now, studies examining the participants’ experiences after disclosure are rather uncommon. A study that explored the experiences and preferences of populationbased research participants to whom an incidental finding was communicated observed that participants were grateful for the disclosure of the secondary findings and that they did not regret their consent to be notified about incidental findings. The study showed that disclosure of the finding had great impact on the lives of most participants.99

117

118

CHAPTER 6  Reporting of secondary findings in genomic research

Impacts on patients, their lives, their families, and the management of disease after disclosure have hardly been explored100 and need to be further assessed.50

4.2.1 ­Preferences for secondary findings

Empirical research on population and patient preferences, as well as criteria for reporting of secondary findings, has mainly revealed that a majority thinks that some form of secondary findings should be returned.68 It has been reported that participants expressed concerns about their providers deciding which results should be reported. They have referenced their autonomy in relation to their genes in the strongest terms (e.g., Ref. 101). Most of all, they have a significant desire to receive clinically actionable findings (e.g., Refs. 101–105). Since “actionability” is one of the key aspects not only for professionals and official standards but also for the lay population and for patients, a number of studies have tried to specify what participants mean by expecting reporting of actionable secondary findings and to further evaluate what kind of influence on their lives participants expect from the knowledge about such findings. Personal utility was determined to be an important reason for disclosure. Several studies have shown that participants wish to receive information about treatable or preventable conditions (e.g., Ref. 106). Parents especially had an interest in learning about findings related to treatable or preventable disorders in childhood (e.g., Refs. 107, 108) since they felt responsible for transmitting this information to their child, irrespective of the disease severity.109 Parents expressed uneasiness about learning about predispositions for untreatable adult-onset conditions and carrier status for recessive conditions.108 Some studies found that the majority of parents of children at risk for genetic abnormalities wished to be informed about incidental findings even if there were no preventive or therapeutic options available in childhood.107 However, after education by experts, studies discovered that parents changed their preferences to a more restrictive disclosure of secondary results related to adultonset conditions. It has been observed that their awareness for the child’s future autonomy and appreciation of potential harms to the child was raised by increasing their knowledge and awareness.110 Nevertheless, parents demonstrated intentions to filter the information they tell their children, even when they are no longer minors.110 Since opportunities for pediatric populations to participate in genomic research projects and in pediatric biobanking are increasing—which means there will be storage and use of pediatric data and biomaterial for a longer time or even until the child reaches the age of majority—empirical research on adolescents’ views regarding the disclosure of incidental findings is gaining importance. Currently, there are only a few studies that ask for adolescents’ preferences. Some have found that adolescents reported they would hypothetically want to receive all results from whole exome sequencing, including carrier status and untreatable adult-onset conditions.111, 112 These participants most frequently endorsed future planning as the reason for disclosure.111 The lay public agrees that parents should receive all information since they are seen as informational gatekeepers.81 Despite this agreement, there is an ongoing discussion in the literature about the ethical aspects of whether parents should have

4 ­ What has been found? Stakeholders’ attitudes and perspectives

the right to know all kinds of secondary findings from genetic research, especially those that have an adult onset. Attention has been drawn to the right of the minor to an open future and the minors’ rights not to know, which means that the minor can decide whether certain genetic information is given to him or her.113 There is an ethical argument that revealing information on treatable or preventable conditions that have adult onset should be postponed until the young person can make their own decisions.114 In asking adolescents this question, a majority thought that parents of a child less than 12 years old should have the right to have access to the information about adult-onset findings.111 As another reason for disclosure of actionable findings, participants mentioned issues with the ability to plan or alter their lifestyle (e.g., Refs. 115, 116) or to influence their reproductive decision-making while learning about secondary findings concerning their carrier status for recessive disorders.117 Only a few studies have tried to determine how the lay population would weigh relevant factors for the reporting of secondary findings.48, 118, 119 The general aspect of “actionability” is further differentiated in relation to the severity of disease and in relation to the degree of treatability, such as a given hypothetical situation of a finding with low scientific certainty concerning a very serious disease combined with a high risk of developing the disease and displaying the first manifestation in more than 20 years. This scenario points to hypothetical situations that can weigh the possible aspects of secondary findings such as the following48: – – – – – – –

Scientific certainty of a relationship between genetic findings and disease Risk of developing a disease given genetic findings Severity of the disease Possibility of treatment Possibility of prevention Time between finding and first clinical manifestation of a disease Reproductive relevance

A study performed in the general Danish population observed a ranking of these attributes influencing preferences for feedback. This ranking rated severity of disease and risk of development highest, followed by the possibility of prevention, the time of first manifestation, and scientific reliability. The possibility of treatment and the reproductive relevance were found to be less important for this sample.48 The aspect that “the population prefers reporting based on the severity and likelihood of the resulting disease and gives less weight to the existence of preventive measures and treatment” (Ref. 48, p. 9) points to a wider understanding of “actionability.” This understanding differs slightly from the understanding of professionals and official standards. Existing studies have found that the lay population has a “strong interest in gaining information about genetic factors influencing future health” (Ref. 48, p. 9). Apart from the attitudes on receiving information on actionable secondary findings, it has been observed that some participants wish to be informed about all secondary findings irrespective of any kind of expected actionability (e.g., Refs. 50, 120).

119

120

CHAPTER 6  Reporting of secondary findings in genomic research

Their motivation is mainly based on the attitude that it is their right to gain the knowledge obtained from their biomaterial and that access to secondary findings is a kind of compensation for participation in the research.121 The attitude of the majority of participants to learn about all secondary findings, or at least all actionable secondary findings, and the occasionally observed enthusiasm for a wide range of secondary findings, leads to the question of whether participants understand such findings and whether they appreciate the implications for their future lives, their families, and relatives.

4.2.2 ­Impacts and implications of secondary findings

In several studies, participants have discussed the potential burden of knowing about secondary findings. Findings about disease risks may have a severe effect on themselves and their families, especially findings that may potentially come into effect in some years after disclosure. There is also potential for a secondary finding to change their family as a whole (e.g., Refs. 90, 115). It has been noted that patients were well aware of the potential psychological harms that knowledge about secondary findings could bring to their lives (e.g., Ref. 88). In general, the consequences of any potential anxiety and uncertainty did not override the strong desire for most participants in studies to receive secondary findings (e.g., Ref. 106). However, studies have underlined the importance of individuals to carefully consider the personal implications if they elect to receive results with no effective therapy. When participants were confronted with examples of preference options for the return of secondary findings and became aware of the respective accompanying psychological implications, they acknowledged in a number of studies that their wish to learn about all secondary findings diminished (e.g., Refs. 90, 122). Whereas participants with inherited genetic diseases reported that they are paying little attention to the potential implications of incidental findings, they admitted that their experience living with a genetic condition prepared them to adjust to incidental findings.123 These findings are similar to the attitude of parents of children affected by various rare diseases who generally perceive a favorable riskbenefit ratio in receiving all incidental findings despite potential negative consequences.109 They mentioned that their experiences enduring their child’s disease would help them to cope with the disclosure of secondary findings. Studies have shown that it is of utmost importance in this context to emphasize that parents’ preferences and perspectives are specific to their child’s disease and the needs of the family as a whole,124 which is an aspect that needs careful consideration for any policy of disclosure of secondary findings. Nevertheless, further research is needed to evaluate the impact of disclosure on parents themselves, their child, and the whole family.109 In general, relatives play an important role concerning the management of disclosure of secondary findings (e.g., Ref. 123). Studies have shown that participants are willing in principal to share their personal genetic results. Moreover, some participants mentioned a willingness to provide genetic information about inheritable genetic effects to relatives and future generations as a motivating factor to receive

4 ­ What has been found? Stakeholders’ attitudes and perspectives

secondary findings. This attitude has been observed within studies with healthy participants and with participants who have a medical condition (e.g., Refs. 101, 116). Nevertheless, it has been recognized that participants realize the challenges that could arise from communicating genetic results with their family. Despite the fact that participants wanted to learn about secondary findings on their own, studies have found that they have a tendency to filter information about secondary findings concerning their families. Some would only share actionable findings (e.g., Ref. 107); others would make a decision to tell individual relatives based on whether the relative is currently ill (e.g., Ref. 101). When it came to sharing information about conditions that are expected to have adverse psychological effects, such as late-onset or untreatable diseases, participants are particularly protective of their children.109 In general, participants have a clear opinion about telling family members that they have potentially impactful genetic information, and they respect that each individual should ultimately decide what has to be shared (e.g., Ref. 101).

4.2.3 ­Information process before consent to genetic research and process after disclosure of secondary findings

In general, studies have demonstrated that members of the public generally have a have a view on genetic testing and the informational value of genetic results.125, 126 This appears to have changed over time, with respondents in 2010 expressing a greater expectation for the use of genetic information compared with respondents in 2002. Respondents in 2010 were more likely to believe that knowledge about the genetic background of disease would help them live longer, and they were more interested in their genetic make-up compared with respondents from 2002.127 This trend might indicate that participants’ attitude that all results should be disclosed to them reflect an inflated perception of the clinical utility of this information.68 This finding correlates with the observation that the public wants more or less unrestricted access to secondary genetic results (e.g., Refs. 73, 88, 105–108, 116, 128). Although providing individual research results is a strong motivation to participate in genetic research,129 studies have shown that participants’ views may shift after learning more about the possible consequences of learning incidental findings.73, 130 Therefore, all stakeholders felt strongly that attention should be paid to informing participants before they provide consent for genetic research, to returning results to participants using a careful disclosure process, and to disclosing results in a way consistent with the participants’ decisions made during the consent process.87, 106 In the consent process, many studies confirmed that participants want to be informed that incidental findings are possible and be given a choice to learn about them.73, 88, 105–108, 115, 116, 128 It has been observed that participants expect detailed information in the consent form about a number of issues, including that the findings may be wrong, that there might be errors in the research interpretations, that certain results could have a negative psychological impact, and about the possibility that results could evolve along with advances in scientific knowledge.72

121

122

CHAPTER 6  Reporting of secondary findings in genomic research

Hence, participants generally agree with policies on the return of secondary findings that include the flexibility to choose to receive or not receive information, and they strongly disagree with policies that provide no information at all and no flexibility to choose.110 Generally, participants want to make their choice about which results to receive (Refs. 40, 121, 131). They see it as imperative to include participants in the decision about which results to return rather than having predetermined rules about which results to disclose.115 Consequently, studies have identified a need to improve the pretest communication of incidental findings by giving clear information about which findings will be disclosed.99 To optimize the process of informed decision-making, participants felt that pretest discussions should last as long as possible, up to 2 h,72 and be conducted along with written information (e.g., Refs. 88, 106) that is culturally sensitive and accessible to all educational levels.120 Participants suggested safeguards against the risks of choice, such as strategies to ensure an informed decision by giving patients the opportunity to ask questions and to consider the potential consequences, as well as to meet trained specialists who can help them understand complex genetic information, before consenting.110 It has been observed that genetic counseling and increased knowledge shifts participants’ preferences toward a desire for less, rather than more, genetic findings.73, 130 Findings demonstrate the benefits of providing examples of preference options concerning the scope of secondary findings that should be reported. It has been noted that participants’ tendency to want all results disclosed decreased.122 Priority was placed on findings that participants understood well.128 The extent of support for disclosure varied when participants learned more about the potential for negative consequences, such as psychological distress or social discrimination. Studies observed that the preference for full disclosure dropped when participants were confronted with features of hypothetical diagnostic scenarios during the pretest discussions that provided a more detailed description of the possible findings and the degree of controllability of disease.128, 132 For the disclosure process and the time after disclosure of secondary findings, participants stressed the importance of genetic counseling. For participants, the face-to-face discussion with a professional with expert knowledge about genetics and the implications of the genetic secondary findings is seen to be essential.101, 112 Participants have a strong demand for professional support for managing secondary findings. They believe that disclosure of secondary findings should be followed by a discussion and a clear arrangement with medical specialists to guarantee a follow-up for the finding.99, 106, 109 The management of secondary findings is seen as a shared endeavor with the professional obliged to disclose along with options for a follow-up plan and with the participant bearing the responsibility for their respective choice concerning the reporting of secondary findings.106

4.2.4 ­Rights, responsibilities, policies, and practices

The attitudes of participants and patients on rights and responsibilities have been widely examined in studies. Participants in research studies believe that researchers in general are obliged to disclose any suspected findings (e.g., Ref. 133).

5 ­ Summary

It has been observed that participants stress the importance of their autonomy, which refers to their right to be given a choice about receiving or not receiving secondary findings and which secondary findings to receive at all.88, 90, 101, 106, 107, 109, 112, 115, 132 These studies indicated that participants are driven by their strong feeling of ownership and their demand to maintain control over their genetic information. Moreover, studies have shown that participants extend this feeling of ownership even to the ­genetic data of family members when they indicated that they would want to be informed about results concerning family members and expressed their wish that these results would be shared with family members.107 However, little is still known about the experiences participants in genetic research have concerning the disclosure of secondary findings affecting family members and the way participants cope with this burden of knowledge, as well as the perceptions about the rights of family members to know about these findings. Out of their strong desire to gain control over their child’s health, parents especially had strong feelings that it is their right to know all secondary findings concerning their children.107–109 Some questions concerning the potential negative impacts arising from full disclosure to parents still need further discussion, such as the loss of long-term autonomy through decision-making for the child—the aforementioned open future—or a potential change in family dynamics if the child is found to have a serious condition.134 It has been observed that participants have strong feelings that it is their right to choose to know everything about their genetic dispositions and everything that is incidentally found (e.g., Refs. 128, 132). When asked about their right not to know, there were strong objections to paternalism. Participants strongly objected to the idea that their physician should know everything about genetic findings and decide what knowledge and which information they would disclose. For these participants, their strong feelings of respect for their autonomy seem to outweigh the medical duty to care.132 Moreover, this study has shown that participants objected to the idea that their physician should have the right to ignore the patient’s wish not to know. However, some participants in this study would tolerate the idea that the physician has the right to ignore the participants’ ignorance in exceptional cases, such as cases involving life, death, and consequences for others. However, only a few studies have examined the situation of disclosure of secondary findings to family members in cases in which the research participant has died prior to the genetic results being reported. A number of questions remain, such as whether and how to share deceased participant results with relatives. This points to the open question of weighing the participants’ privacy against family members’ interest in their genetic information.

5 ­Summary This chapter summarizes the main points from the debate on whether and why empirical evidence should be integrated into decision-making and policy in bioethics. There is an obvious conceptual tension between empirical and normative approaches

123

124

CHAPTER 6  Reporting of secondary findings in genomic research

to bioethics. There are also many questions around the best way to integrate diverse methodologies and fields and on how to make sure the quality of bioethical empirical studies is as high as possible. This chapter presents these issues and concludes that if done in a careful and proportionate way, empirical data in bioethics can support and enhance decision-making and policy development in many ways and contribute toward greater legitimacy of policy and governance. The chapter focuses on the examples of feedback on secondary findings in genomic research and provides an overview of tools and instruments for eliciting stakeholder preferences. The presentation is structured around a set of questions: whose preferences need to be elicited, which personal and disease-related factors are influencing attitudes and preferences, what are the available tools to elicit stakeholders’ preferences, and when stakeholders’ preferences should be elicited. Although measuring and subsequently responding to participants’ preferences on the return of secondary findings is a preferred approach for policy development, concerns remain given the high number of potential future secondary results. Open questions around the way individuals’ preferences can be elicited effectively given the lay public’s limited knowledge about the production and analysis of secondary results and how preferences can be elicited faithfully without extensive counseling. The final part of the chapter presents an overview of recent empirical studies into stakeholder preferences on secondary findings in genomic research. Stakeholders include patients with different types and different levels of diseases; patients with rare diseases; patients with genetic disorders, as well as relatives; healthcare professionals in genetics or other specialties; primary care physicians, genetics and other researchers; and institutional review boards. All these groups differ in their attitudes toward secondary findings, for example, regarding impacts and implications of secondary findings for the participant, for their family and relatives, etc. Stakeholders vary in their understanding and literacy, in their preferences concerning the information process before consenting to genetic research and the process after disclosure of secondary findings and, finally, in their attitudes concerning participant rights and responsibilities. The literature on stakeholders’ attitudes and preferences has to be interpreted with caution, and the results cannot easily be compared since studies are frequently based on different presumptions. It also has to be kept in mind that these studies have been conducted on different populations and in different countries with different legislation and cultural backgrounds. Notwithstanding, by recognizing how the attitudes and preferences of stakeholders differ or coincide, it becomes more feasible to develop recommendations or strategies that are capable of respecting all concerns during the process of healthcare decision-making in a particular context. Policy implementation is likely to benefit from a fair deliberation process, and policymakers will be better placed to inform the public about their decisions if they are aware of public preferences. It should be kept in mind that population preferences might change over time and subgroups within a population might have different preferences based on their family history, personal experiences, or individual group characteristics. Hence, any policies should be reassessed over time to deliver equilibrium between individual preferences and practicability.

­References

­References 1. Borry P, Schotsman P, Dierickx K. The birth of the empirical turn in bioethics. Bioethics. 2005;19(1):49–71. 2. Jonsen AR. The birth of bioethics. Oxford: Oxford University Press; 1998. 3. Alvarez  AA. How rational should bioethics be? The value of empirical approaches. Bioethics. 2001;15(5–6):501–519. 4. Ashcroft  RE. Constructing empirical bioethics: Foucauldian reflections on the empirical turn in bioethics research. Health Care Anal. 2003;11:3. https://doi.org/10.102 3/A:1025329811812. 5. De Vries  R, Subedi  J, eds. Bioethics and society. Constructing the ethical enterprise. New Jersey: Prentice Hall; 1998. 6. Goldenberg  MJ. Evidence-based ethics? On evidence-based practice and the “empirical turn” from normative bioethics. BMC Med Ethics. 2005;6:11. https://doi. org/10.1186/1472-6939-6-11. 7. Ten Have H, Lelie A. Medical ethics research between theory and practice. Theor Med Bioeth. 1998;19:263–276. 8. Herrera  C. Is it time for bioethics to go empirical? Bioethics. 2008;22(3):137–146. https://doi.org/10.1111/j.1467-8519.2007.00621.x. 9. Hurst S. What ‘empirical turn in bioethics’? Bioethics. 2010;24(8):439–444. https://doi. org/10.1111/j.1467-8519.2009.01720.x. 10. Singer P, Pellegrino E, Siegler M. Research in clinical ethics. J Clin Ethics. 1990;1:5. 11. Solomon  MZ. Realizing bioethics’ goals in practice: ten ways “is” can help “ought”. Hastings Cent Rep. 2005;35(4):40–47. 12. Strech D. Evidence-based ethics—what it should be and what it shouldn’t. BMC Med Ethics. 2008;9:16. https://doi.org/10.1186/1472-6939-9-16. 13. Ives  J, Draper  H. Appropriate methodologies for empirical bioethics: it’s all relative. Bioethics. 2009;23(4):249–258. 14. Mertz M, Schildmann J. Beyond integrating social sciences: reflecting on the place of life sciences in empirical bioethics methodologies. Med Health Care Philos. 2018;21(2):207– 214. https://doi.org/10.1007/s11019-017-9792-z. 15. Ives J, Dunn M, Molewijk B, et al. Standards of practice in empirical bioethics research: towards a consensus. BMC Med Ethics. 2018;19:68. 16. Mertz M, Inthorn J, Renz G, et al. Research across the disciplines: a road map for quality criteria in empirical ethics research. BMC Med Ethics. 2014;15(1):17. https://doi. org/10.1186/1472-6939-15-17. 17. Wangmo T, Hauri S, Gennet E, Anane-Sarpong E, Provoost V, Elger BS. An update on the “empirical turn” in bioethics: analysis of empirical research in nine bioethics journals. BMC Med Ethics. 2018;19:6. https://doi.org/10.1186/s12910-018-0246-9. 18. Prainsack B, Buyx A. A solidarity-based approach to the governance of biobanks. Med Law Rev. 2013;21(1):71–91. 19. Prainsack  B, Buyx  A. Solidarity in biomedicine and beyond. Cambridge University Press; 2017. 20. Parker M. Two concepts of empirical ethics. Bioethics. 2009;23(4):202–213. 21. Salloch  S, Vollmann  J, Schildmann  J. Ethics by opinion poll? The functions of attitudes research for normative deliberations in medical ethics. J Med Ethics. 2014;40:597–602. 22. Düwell M. Wofür braucht Medizinethik empirische Methoden? Eine normativ-ethische Untersuchung. Ethik Med. 2009;21(3):201–211.

125

126

CHAPTER 6  Reporting of secondary findings in genomic research

23. Beauchamp TL, Childress JF. Principles of biomedical ethics. 6th ed. New York: Oxford University Press; 2009. 24. Birnbacher  D. Where and when ethics needs empirical facts. In: Brand  C, ed. Dualprocess theories in moral psychology. Interdisciplinary approaches to theoretical, empirical and practical considerations. Wiesbaden: VS Verlag für Sozialwissenschaften/ Springer Fachmedien; 2016:41–55. 25. Schicktanz  S. Zum Stellenwert von Betroffenheit, Öffentlichkeit und Deliberation im empirical turn der Medizinethik. Ethik Med. 2009;21(3):223–234. 26. Musschenga  B. Empirical ethics and the special status of practitioner’s judgements. Perspectives. 2010;17(2):203–230. 27. Buyx  A, Del Savio  L, Prainsack  B, Völzke  H. Every participant is a PI. Citizen science and participatory governance in population studies. Int J Epidemiol. 2017;46(2):377–384. 28. Del Savio  L, Prainsack  B, Buyx  A. Crowdsourcing the human gut. J Sci Commun. 2016;15(3):A03. 1–16. 29. Birnbacher  D. Utilitarismus/Ethischer Egoismus. In: Düwell  M, Hübenthal  C, Werner MH, eds. Handbuch Ethik. Stuttgart/Weimar: J.B. Metzler; 2011:95–107. 30. Petryna A. Life exposed: biological citizens after Chernobyl. Princeton University Press; 2004/2013. 31. Heath D, Rapp R, Taussig K-S. Genetic citizenship. In: Nugent D, Vincent J, eds. A companion to the anthropology of politics. Malden, MA: Blackwell; 2007:152–167. 32. Rose N, Novas C. Biological citizenship. In: Ong A, Collier S, eds. Global assemblages: technology, politics and ethics as anthropological problems. Malden, MA: Blackwell; 2005:439–463. 33. Cooter R. Biocitizenship. Lancet. 2008;372(9651):1725. 34. Heineman T. Biological citizenship. In: ten Have H, ed. Encyclopedia of global bioethics. Springer; 2015. 35. Mulligan S. Biological citizenship. Oxford Bibliographies in Anthropology; 2017. https:// doi.org/10.1093/obo/9780199766567-0164. 36. Sabatello  M, Appelbaum  PS. Raising genomic citizens: adolescents and the return of secondary genomic findings. J Law Med Ethics. 2016;44:292–308. 37. Zawati L, Lang M. Biobank donors and the concept of benefit: time for reciprocity. J Law Biosci. 2017;371–376. 38. Nobile H, et al. Why do participants enroll in population biobank studies? A systematic literature review. Expert Rev Mol Diagn. 2013;13(1):35–47. 39. Prainsack  B, Buyx  A. Thinking ethical and regulatory frameworks in medicine from the perspective of solidarity on both sides of the Atlantic. Theor Med Bioeth. 2016;37(6):489–501. 40. Richter  G, Krawczak  M, Lieb  W, Wolff  L, Buyx  A. Broad consent for health care-­ embedded biobanking: understanding and reasons to donate in a large patient sample. Genet Med. 2018;20(1):76–82. 41. Faucett WA, Davis FD. How Geisinger made the case for an institutional duty to return genomic results to biobank participants. Appl Transl Genom. 2016;8:33–35. 42. Elger BS, De Clercq E. Returning results: let’s be honest! Genet Test Mol Biomarkers. 2017;21(3):134–139. 43. Domaradzki  J. Patient rights, risk, and responsibilities in the genetic era—a right to know, a right not to know, or a duty to know? Ann Agric Environ Med. 2015;22:156–162. 44. Sheehan M. The right to know and genetic testing. J Med Ethics. 2015;41:287–288.

­References

45. Barazetti G, Cavalli S, Benaroyo L, Kaufmann A. “Still rather hazy at present”: citizens’ and physicians’ views on returning results from biobank research using broad consent. Genet Test Mol Biomarkers. 2017;21(3):159–165. 46. Green RC, et al. ACMG recommendations for reporting of incidental findings in clinical exome and genome sequencing. Genet Med. 2013;15(7):565–574. 47. ACMG Board of Directors. ACMG policy statement: updated recommendations regarding analysis and reporting of secondary findings in clinical genome-scale sequencing. Genet Med. 2015;17(1):68–69. 48. Ploug T, Holm S. Clinical genome sequencing and population preferences for information about ‘incidental’ findings—from medically actionable genes (MAGs) to patient actionable genes (PAGs). PLoS One. 2017;12(7):e0179935. 49. Strong KA, Zusevics KL, Bick D, Veith R. Views of primary care providers regarding the return of genome sequencing incidental findings. Clin Genet. 2014;86:461–468. 50. Mackley MP, Fletcher B, Parker M, Watkins H, Ormondroyd E. Stakeholder views on secondary findings in whole-genome and whole-exome sequencing: a systematic review of quantitative and qualitative studies. Genet Med. 2017;19(3):283–293. 51. Bjugn R, Casati B. Stakeholder analysis: a useful tool for biobank planning. Biopreserv Biobank. 2012;10(3). https://doi.org/10.1089/bio.2011.0047. 52. Friedman AL, Miles S. Stakeholders. Theory and Practice. Oxford: Oxford University Press; 2006. 53. Brothers KB, Lynch JA, Aufox SA, et al. Practical guidance on informed consent for pediatric participants in a biorepository. Mayo Clin Proc. 2014;89(11):1471–1480. https:// doi.org/10.1016/j.mayocp.2014.07.006. 54. Gunsalus CK, Bruner EM, Burbules NC, et al. Mission creep in the IRB world. Science. 2006;312(5779):1441. https://doi.org/10.1126/science.1121479. 55. Mitchell  D, et  al. Biobanking from the patient perspective. Res Involv Engagem. 2015;1(1):4. 56. Lemke AA, Halverson C, Ross LF. Biobank participation and returning research results: perspectives from a deliberative engagement in South Side Chicago. Am J Med Genet A. 2012;158A:1029–1037. 57. United Nations Convention on the Rights of the Child. 1989. 58. MacLean S, Burgess MM. In the public interest: assessing expert and stakeholder influence in public deliberation about biobanks. Public Underst Sci. 2009;19(4):486–496. https://doi.org/10.1177/0963662509335410. 59. Hunter  KG, Laurie  GT. Involving publics in biobank governance. In: Widdows  H, Mullen  C, eds. The governance of genetic information: who decides? Cambridge: Cambridge University Press; 2014. 60. Budin-Ljøsne I, Mascalzoni D, Soini S, et al. Feedback of individual genetic results to research participants: is it feasible in Europe? Biopreserv Biobank. 2016;14(3). https:// doi.org/10.1089/bio.2015.0115. 61. Gornick MC, Scherer AM, Sutton EJ, et al. Effect of public deliberation on attitudes toward return of secondary results in genomic sequencing. J Genet Couns. 2017;26(1):122–132. 62. Bacon PL, Harris ED, Ziniel SI, et al. The development of a preference-setting model for the return of individual genomic research results. J Empir Res Hum Res Ethics. 2015;10(2):107–120. 63. Brothers  KB, East  KM, Kelley  WV, et  al. Eliciting preferences on secondary findings: the Preferences Instrument for Genomic Secondary Results (PIGSR). Genet Med. 2017;19(3):337–344.

127

128

CHAPTER 6  Reporting of secondary findings in genomic research

64. Liebel M. From evolving capacities to evolving capabilities: contextualizing children’s rights. In: Stoecklin D, Bonvin JM, eds. Children’s rights and the capability approach. Dordrecht, the Netherlands: Springer; 2014:67–84. 65. Parker L. Returning individual research results: what role should people’s preferences play? Minn J Law Sci Technol. 2012;13(2):449–484. 66. Hens K, Nys H, Cassiman J, et al. The return of individual research findings in paediatric genetic research. J Med Ethics. 2011;37:179–183. 67. Levenseller BL, Soucier DJ, Miller VA, Harris D, Conway L, Bernhardt BA. Stakeholders’ opinions on the implementation of pediatric whole exome sequencing: implications for informed consent. J Genet Couns. 2014;23:552–565. https://doi.org/10.1007/ s10897-013-9626-y. 68. Delanne  J, Nambot  S, Chassagne  A, et  al. Secondary findings from whole-exome/genome sequencing evaluating stakeholder perspectives. A review of the literature. Eur J Med Genet. 2018. https://doi.org/10.1016/j.ejmg.2018.08.010. 69. Richter G, Borzikowski C, Lieb W, Schreiber S, Krawczak M, Buyx A. Patient views on research use of clinical data without consent: legal, but also acceptable? Eur J Hum Genet. 2019;25(01):2019. 70. Cole  C, Petree  LE, Phillips  JP, Shoemaker  JM, Holdsworth  M, Helitzer  DL. ‘Ethical responsibility’ or ‘a whole can of worms’: differences in opinion on incidental finding review and disclosure in neuroimaging research from focus group discussions with participants, parents, IRB members, investigators, physicians and community members. J Med Ethics. 2015;41(10):841–847. 71. Klitzman  R, Appelbaum  PS, Fyer  A, et  al. Researchers’ views on return of incidental genomic research results: qualitative and quantitative findings. Genet Med. 2013;15(11):888–895. 72. Appelbaum PS, Waldman CR, Fyer A, et al. Informed consent for return of incidental findings in genomic research. Genet Med. 2014;16(5):367–373. https://doi.org/10.1038/ gim.2013.145. 73. Middleton A, Morley KI, Bragin E, et al. Attitudes of nearly 7000 health professionals, genomic researchers and publics toward the return of incidental results from sequencing research. Eur J Hum Genet. 2016;24:21–29. 74. Wynn  J, Martinez  J, Duong  J, et  al. Association of researcher characteristics with views on return of incidental findings from genomic research. J Genet Couns. 2015;24:833–841. 75. Keogh  LA, Fisher  D, Gorin  SS, et  al. How do researchers manage genetic results in practice? The experience of the multinational colon cancer family registry. J Community Genet. 2014;5(2):99–108. https://doi.org/10.1007/s12687-013-0148-y. 76. Scheuner  MT, Peredo  J, Benkendorf  J, et  al. Reporting genomic secondary findings: ACMG members weigh in. Genet Med. 2015;17(1):27–35. 77. Gliwa C, Yurkiewicz IR, Lehmann LS, Hull SC, Jones N, Berkman BE. Institutional review board perspectives on obligations to disclose genetic incidental findings to research participants. Genet Med. 2016;18(7):705–711. https://doi.org/10.1038/gim.2015.149. 78. Yu J-H, Harrell TM, Jamal SM, Tabor HK, Bamshad MJ. Attitudes of genetics professionals toward the return of incidental results from exome and whole-genome sequencing. Am J Hum Genet. 2014;95(1):77–84. https://doi.org/10.1016/j.ajhg.2014.06.004. 79. Kleiderman  E, Avard  D, Besso  A, Ali-Khan  S, Sauvageau  G, Hebert  J. Disclosure of incidental findings in cancer genomic research: investigators’ perceptions on obligations and barriers. Clin Genet. 2015;88:320–326.

­References

80. Dressler LG, Smolek S, Ponsaran R, et al. IRB perspectives on the return of individual results from genomic research. Genet Med. 2012;14(2):215–222. https://doi.org/10.1038/ gim.2011.10. 81. Driessnack  M, Daack-Hirsch  S, Downing  N, et  al. The disclosure of incidental genomic findings: an “ethically important moment” in pediatric research and practice. J Community Genet. 2013;4(4):435–444. 82. Barajas  M, Ross  LF. Pediatric professionals’ attitudes about secondary findings in genomic sequencing of children. J Pediatr. 2015;166(5):1276–1282.e7. https://doi. org/10.1016/j.jpeds.2015.01.032. 83. Meacham MC, Starks H, Burke W, Edwards K. Researcher perspectives on disclosure of incidental findings in genetic research. J Empir Res Hum Res Ethics. 2010;5(3):31–41. https://doi.org/10.1525/jer.2010.5.3.31. 84. Turbitt  E, Wiest  MM, Halliday  JL, Amor  DJ, Metcalfe  SA. Availability of treatment drives decisions of genetic health professionals about disclosure of incidental findings. Eur J Hum Genet. 2014;22(10):1225–1228. https://doi.org/10.1038/ejhg.2014.11. 85. Downing NR, Williams JK, Daack-Hirsch S, Driessnack M, Simon CM. Genetics specialists’ perspectives on disclosure of genomic incidental findings in the clinical setting. Patient Educ Couns. 2013;90:133–138. 86. Gray SW, Park ER, Najita J, et al. Oncologists’ and cancer patients’ views on wholeexome sequencing and incidental findings: results from the CanSeq study. Genet Med. 2016; [Epub ahead of print]. 87. Miller FA, Hayeems RZ, Bytautas JP, et al. Testing personalized medicine: patient and physician expectations of next-generation genomic sequencing in latestage cancer care. Eur J Hum Genet. 2014;22:391–395. 88. Townsend A, Adam S, Birch PH, Lohn Z, Rousseau F, Friedman JM. I want to know what’s in Pandora’s box: comparing stakeholder perspectives on incidental findings in clinical whole genomic sequencing. Am J Med Genet A. 2012;158A(10):2519–2525. 89. Grove ME, Wolpert MN, Cho MK, Lee SS, Ormond KE. Views of genetics health professionals on the return of genomic results. J Genet Couns. 2014;23:531–538. 90. Christenhusz GM, Devriendt K, Van Esch H, Dierickx K. Focus group discussions on secondary variants and next-generation sequencing technologies. Eur J Med Genet. 2015;58:249–257. 91. Christensen KD, Vassy JL, Jamal L, et al. MedSeq Project Team. Are physicians prepared for whole genome sequencing? a qualitative analysis. Clin Genet. 2016;89:228–234. 92. Klitzman R, Buquez B, Appelbaum PS, Fyer A, Chung WK. Processes and factors involved in decisions regarding return of incidental genomic findings in research. Genet Med. 2014;16:311–317. 93. Fiske A, Buyx B, Prainsack B. Health information counselors: a new profession for the age of big data? Acad Med. 2018. https://doi.org/10.1097/ACM.0000000000002395. 94. Simon  CM, Williams  JK, Shinkunas  L, Brandt  D, Daack-Hirsch  S, Driessnack  M. Informed consent and genomic incidental findings: IRB chair perspectives. J Empir Res Hum Res Ethics. 2011;6(4):53–67. https://doi.org/10.1525/jer.2011.6.4.53. 95. Bernhardt BA, Roche MI, Perry DL, Scollon SR, Tomlinson AN, Skinner D. Experiences with obtaining informed consent for genomic sequencing. Am J Med Genet A. 2015;167A:2635–2646. 96. Fernandez CV, Strahlendorf C, Avard D, et al. Attitudes of Canadian researchers toward the return to participants of incidental and targeted genomic findings obtained in a pediatric research setting. Genet Med. 2013;15(7):558–564.

129

130

CHAPTER 6  Reporting of secondary findings in genomic research

97. Wolf SM. The past, present, and future of the debate over return of research results and incidental findings. Genet Med. 2012;14(4):355–357. 98. Saelaert M, Mertes H, De Baere E, Devisch I. Incidental or secondary findings: an integrative and patient-inclusive approach to the current debate. Eur J Hum Genet. 2018;2018. https://doi.org/10.1038/s41431-018-0200-9. 99. de Boer AW, Drewes YM, de Mutsert R, et al. Incidental findings in research: a focus group study about the perspective of the research participant. J Magn Reson Imaging. 2017. https://doi.org/10.1002/jmri.25739. 100. Hart MR, Biesecker BB, Blout CL, et al. Secondary findings from clinical genomic sequencing: prevalence, patient perspectives, family history assessment, and health-care costs from a multisite study. Genet Med. 2018; [Online published]. 101. Hitch K, Joseph G, Guiltinan J, Kianmahd J, Youngblom J, Blanco A. Lynch syndrome patients’ views of and preferences for return of results following whole exome sequencing. J Genet Couns. 2014;23:539–551. 102. Allen NL, Karlson EW, Malspeis S, et al. Biobank participants’ preferences for disclosure of genetic research results: perspectives from the OurGenes, OurHealth, OurCommunity project. Mayo Clin Proc. 2014;89:738. 103. Jelsig AM, Qvist N, Brusgaard K, Ousager LB. Research participants in NGS studies want to know about incidental findings. Eur J Hum Genet. 2015;23:1423–1426. 104. Kaphingst KA, Ivanovich J, Biesecker BB, et al. Preferences for return of incidental findings from genome sequencing among women diagnosed with breast cancer at a young age. Clin Genet. 2016;89:378–384. 105. Sanderson SC, Linderman MD, Suckiel SA, et al. Motivations, concerns and preferences of personal genome sequencing research participants: baseline findings from the Health Seq project. Eur J Hum Genet. 2016;24(1):14–20. 106. Daack-Hirsch S, Driessnack M, Hanish A, et al. ‘Information is information’: a public perspective on incidental findings in clinical and research genome-based testing. Clin Genet. 2013;84:11–18. 107. Fernandez CV, Bouffet E, Malkin D, et al. Attitudes of parents toward the return of targeted and incidental genomic research findings in children. Genet Med. 2014;16:633–640. 108. Sapp JC, Dong D, Stark C, et al. Parental attitudes, values, and beliefs toward the return of results from exome sequencing in children. Clin Genet. 2014;85:120–126. 109. Kleiderman E, Knoppers BM, Fernandez CV, et al. Returning incidental findings from genetic research to children: views of parents of children affected by rare diseases. J Med Ethics. 2014;40(10):691–696. 110. Ryan KA, De Vries RG, Uhlmann WR, Roberts JS, Gornick MC. Public’s views toward return of secondary results in genomic sequencing: it’s (almost) all about the choice. J Genet Couns. 2017;26(6):1197–1212. 111. Hufnagel SB, Martin LJ, Cassedy A, Hopkin RJ, Antommaria AH. Adolescents’ preferences regarding disclosure of incidental findings in genomic sequencing that are not medically actionable in childhood. Am J Med Genet A. 2016;170(8):2083–2088. 112. Levenseller  BL, et  al. ‘Stakeholders’ opinions on the implementation of pediatric whole exome sequencing: implications for informed consent. J Genet Couns. 2014;23(4):552–565. 113. Borry P, Shabani M, Howard HC. Is there a right time to know? The right not to know and genetic testing in children. J Law Med Ethics. 2014;42(1):19–27. https://doi.org/10.1111/ jlme.12115.

­References

114. Hens K, Van El CE, Borry P, et al. Developing a policy for paediatric biobanks: principles for good practice. Eur J Hum Genet. 2013;21:2e7. 115. Clift KE, Halverson CM, Fiksdal AS, Kumbamu A, Sharp RR, McCormick JB. Patients’ views on incidental findings from clinical exome sequencing. Appl Transl Genom. 2015;4:38–43. 116. Facio  FM, Eidem  H, Fisher  T, et  al. Intentions to receive individual results from whole-genome sequencing among participants in the ClinSeq study. Eur J Hum Genet. 2013;21:261–265. 117. Shahmirzadi L, Chao EC, Palmaer E, Parra MC, Tang S, Gonzalez KD. Patient decisions for disclosure of secondary findings among the first 200 individuals undergoing clinical diagnostic exome sequencing. Genet Med. 2014;16:395–399. 118. Bennette CS, Trinidad SB, Fullerton SM, et al. Return of incidental findings in genomic medicine: measuring what patients value—development of an instrument to measure preferences for information from next-generation testing (IMPRINT). Genet Med. 2013;15(11):873–881. 119. Murphy Bollinger  J, Bridges  JFP, Mohamed  A, Kaufman  D. Public preferences for the return of research results in genetic research: a conjoint analysis. Genet Med. 2014;16(12):932–939. https://doi.org/10.1038/gim.2014.50. 24854226. 120. Wynn J, Martinez J, Duong J, et al. Research participants’ preferences for hypothetical secondary results from genomic research. J Genet Couns. 2016;26(4):841–851. 121. O’Daniel J, Haga SB. Public perspectives on returning genetics and genomics research results. Public Health Genomics. 2011;14:346–355. 122. Christensen KD, Savage SK, Huntington NL, et al. Preferences for the return of individual results from research on pediatric biobank samples. J Empir Res Hum Res Ethics. 2017;12(2):97–106. 123. Bergner  AL, Bollinger  J, Raraigh  KS, et  al. Informed consent for exome sequencing research in families with genetic disease: the emerging issue of incidental findings. Am J Med Genet A. 2014;164A:2745–2752. 124. Tabor  HK, Brazg  T, Crouch  J, et  al. Parent perspectives on pediatric genetic research and implications for genotype-driven research recruitment. J Empir Res Hum Res Ethics. 2011;6(4):41–52. 125. Condit CM. Public attitudes and beliefs about genetics. Annu Rev Genomics Hum Genet. 2010;11:339:359. 126. Etchegary  H, Cappelli  M, Potter  B, et  al. Attitude and knowledge about genetics and genetic testing. Public Health Genomics. 2010;13(2):80–88. 127. Henneman  L, Vermeulen  E, van El  CG, Claassen  L, Timmermans  DRM, Cornel  MC. Public attitudes towards genetic testing revisited: comparing opinions between 2002 and 2010. Eur J Hum Genet. 2013;21(8):793–799. https://doi.org/10.1038/ejhg.2012.271. PMID: 23249955. 128. Murphy Bollinger J, Scott J, Dvoskin R, Kaufman D. Public preferences regarding the return of individual genetic research results: findings from a qualitative focus group study. Genet Med. 2012;2012(14):451–457. 129. Kaufman D, Murphy J, Scott J, Hudson K. Subjects matter: a survey of public opinions about a large genetic cohort study. Genet Med. 2008;10(11):831–839. 130. Bradbury AR, Patrick-Miller LJ, Egleston BL, et al. Patient feedback and early outcome data with a novel tiered-binned model for multiplex breast cancer susceptibility testing. Genet Med. 2015;18:25–33.

131

132

CHAPTER 6  Reporting of secondary findings in genomic research

131. Murphy J, Scott J, Kaufman D, Geller G, LeRoy L, Hudson K. Public expectations for return of results from large-cohort genetic research. Am J Bioethics. 2008;8:36–43. 132. Flatau L, Reitt M, Duttge G, et al. Genomic information and a person’s right not to know: a closer look at variations in hypothetical informational preferences in a German sample. PLoS One. 2018;13(6):e0198249. https://doi.org/10.1371/journal.pone.0198249. 133. Erdmann P. Handling incidental findings from imaging within IM related research. In: Fischer T, Langanke M, Marschall P, Michl S, eds. Individualized medicine. Ethical, economical and historical perspectives. Heidelberg: Springer; 2015:231–250. 134. Byrjalsen A, Stoltze U, Wadt K, et al. Pediatric cancer families’ participation in whole‐ genome sequencing research in Denmark: parent perspectives. Eur J Cancer. 2018. https://doi.org/10.1111/ecc.12877.

­Further reading Dresser  R. Public preferences and the challenge to genetic research policy. J Law Biosci. 2014;1(1):52–67. Lemke AA, Bick D, Dimmock D, Simpson P, Veith R. Perspectives of clinical genetics professionals toward genome sequencing and incidental findings: a survey study. Clin Genet. 2013;84(3):230–236. Radecki Breitkopf C, Wolf SM, Chaffee KG, et al. Attitudes toward return of genetic research results to relatives, including after death: comparison of cancer probands, blood relatives, and spouse/partners. J Empir Res Hum Res Ethics. 2018;13(3):295–304.

CHAPTER

Disclosing genomic sequencing results

7

Janet L. Williams Senior Research Genetic Counselor, Genomic Medicine Institute, Geisinger, Danville, PA, United States

This chapter will focus on important elements to consider in the communication of the various types of research genomic findings to participants. While many articles have been written about determining participant’s desire to learn results, whether to return results, and which results to return, there has been considerably less written about the actual process of disclosing research findings to research participants. Studies evaluating the implementation of the American College of Medical Genetics and Genomics (ACMG) recommendations1, 2 on returning incidental findings share some of the processes used in the clinical setting.3–5 While there are no practice guidelines regarding communication of research findings for either the clinical or the research settings, there are lessons learned3, 4, 6 and published frameworks7–11 to help guide development of a disclosure process within a research study. An important disclaimer is necessary to address the language choices for this review. As indicated in the first chapter, terminology for genomic sequence results varies widely with published references to individual research findings, secondary findings, incidental findings, or additional findings to name a few. Inasmuch as these terms have been used somewhat interchangeably in published studies, at times, it is difficult to distinguish which type of finding the study referenced. In many studies, disclosure efforts and lessons learned are applicable across the various types of findings to return. It seems worthwhile to clearly describe how the terms will be used in this chapter (see Table 1). A key message in the literature recommends planning for the need to return research results within the study protocol and consent process.5, 10, 11 For most researchers the study protocol and consent form will supply the context for disclosure of research results of any kind. According to Thorogood et al.9 the research protocol should address three steps: evaluation of the research finding with regard to suitability for return, use of a data broker for reidentification of the research participant, and development of a process for disclosing the research results. Previous chapters have outlined the process for evaluation of research findings, deciding which results to return and establishing a data broker for reidentification of participants for whom disclosure is indicated. This chapter will propose options for the infrastructure needed Secondary Findings in Genomic Research. https://doi.org/10.1016/B978-0-12-816549-2.00007-2 © 2020 Elsevier Inc. All rights reserved.

133

134

CHAPTER 7  Disclosing genomic sequencing results

Table 1  Genomic sequence research findings: terminology definitions. Primary finding Secondary finding

Incidental finding Research findings Research results

Genomic sequencing result that directly relates to a participant’s indication for research sequencing Genomic sequencing result that is identified based on intentional interrogation of the research sequence, often based on specified criteria Genomic sequencing result that is found unexpectedly upon review of the research sequence of interest Genomic sequence results that are revealed in the context of population sequencing A generic term used to encompass all types of research results

to complete the process of communicating research results to participants. Elements of the infrastructure include context of the research, characteristics of the study participants, notification of study participants of a research result to be disclosed, disclosure modalities, formal laboratory result reports, and counseling lessons learned in communicating research results.

1 ­Context of the research: Managing participant expectations The best description of the role that research context plays in the development of a disclosure plan is reported in the paper by Jarvik et al. on behalf of the NIH-funded CSER and eMERGE research consortia in which areas of agreement with genomics research findings were proposed. The researchers propose five guiding principles to inform disclosing research results to participants in genomics research, including disclosure of incidental findings (see Table 2). The first principle reflects the differences between the goals and objectives of research versus clinical care. Returning results in the clinical setting requires attention to clinical standards and an ethical obligation to ensure appropriate clinical treatment based on the results returned. While researchers have ethical responsibilities to participants relative to disclosure, the relationship of the researcher and legal Table 2  Guiding principles to inform disclosure of research findings.12 1. 2. 3. 4. 5.

Differences between the goals and objectives of research versus clinical care Research does not require analysis for secondary or incidental findings Research is needed to define methods, outcomes, and value related to disclosure of research results Research results that are analytically and clinically valid and have actionable health implications for participants should be disclosed when possible Participants should be given the opportunity to decline any of their research findings

1 ­Context of the research: Managing participant expectations

r­ esponsibilities for disclosure are distinct from those of a clinical provider. The context suggests that participants be made aware of the potential to return results, but that the clinical responsibility to act on the result may or may not be included as part of their consent for participation in the research. The second principle refers to the difference that searching for incidental findings is not expected in the way that such effort has been recommended when clinical ­genomic sequencing is offered.1 Focusing resources on the identification of incidental findings and the need for clinical confirmation divert resources from the research goal. Research teams may not have the requisite expertise. Participants may have been recruited within a specific disease cohort with which the researcher is well acquainted. Therefore the group consensus supports determining the “floor” in developing a list of the anticipated actionable findings to return or that might be found in the course of the research.12, 13 This is particularly important in establishing the context for disclosing results because it begins to set the expectations for participants of the types of results that they may learn.9, 14 Developing informational content for the disclosure will be directed by the expectations of the participants relative to the types of results that they are told that could be revealed. Soliciting, understanding, and responding to participant expectations are crucial to the actual return process. The third principle recognizes that research is needed to define the outcomes and value related to returning research results. Khoury et al.15 discussed the value of including research into the process of disclosure of research results to build evidence related to the complete translational pathway.15, 16 Anticipating the need to add research efforts to evaluate the clinical and personal value of research results to participants, when possible, furthers the translational aspect of the research endeavor. In this principle, there is a decision about the types of results participants may choose, ranging from the “floor” of restricted findings to the “ceiling” of results, described in this paper as the return of the entire genome.12 This principle relates again to setting appropriate expectations for participants. The fourth principle encourages researchers to return results that are known to impact an individual’s health. In studies in which the researcher may not have the ability to reidentify participants and all data have been deidentified or anonymized, there may be no option to return individual research results. However, more recently, this principle is more and more frequently quoted in qualitative and quantitative studies involving participants, genomics researchers, institutional review boards, and ­healthcare providers.17 In the article by Jarvik et al.,12 the assumption is made that return of results will be outlined in the research consent form. Klitzman et al.18 surveyed genomics researchers regarding their return practices and noted that some researchers reported difficulty in discerning the right balance between a participant’s right to know versus the potential for harm. In this study the majority of researchers recommended that involvement of clinically trained personnel such as genetic counselors should be available to disclose results. These same researchers indicated their concern that participants might misinterpret the return of incidental findings as having more significant implications than their medical or family history might justify. Whether or not the research consent

135

136

CHAPTER 7  Disclosing genomic sequencing results

has addressed explicit consent to learn genomic findings, more and more ­researchers ­advocate for disclosing research results that will affect the participant’s health. In this context the infrastructure for disclosure will rely on the characteristics of the study cohort and engagement of the institutional review board and clinical experts.10, 19 Managing expectations in this scenario is different and requires different content in developing the process to reach participants and disclose results. The final principle described by Jarvik et al.12 is now universally embraced, that is, individual participants should be able to decline to receive any research results. If the research study includes disclosing research results, then participants should have the option to decline all participation in the research, which reflects an accepted standard for research. This would occur at the time of consent. When research results are found in the course of review of genomic findings and the option to learn results was not presented at the time of enrollment, then many recommend reconsent that includes the return of results at the time when the potential for returnable results becomes recognized.20, 21 At any point in the process to discuss research results and incidental findings, the participant should be given the option to decline.12 The context reflected in principle five is important in aligning participant choice in the development of the disclosure process. In summary, context matters for managing participant expectations that will in turn directly inform the process designed for returning results in any given research study. Clearly establishing the “floor and ceiling” of potential types and methods of result disclosure will facilitate the actual disclosure process once results become available to researchers.12 The optimal time to set the overall context for participation is in the consent process. Participants who clearly understand who will contact them, how they may be contacted, and what is expected of them if a research finding is uncovered are likely to be better equipped to learn of their finding and move forward with next steps relative to the genomic finding reported. Finally, anticipating that life events change and challenge participants leads to the suggested practice of reconfirming the choices regarding learning genomic findings and incidental findings at the outset of any return process. Each of these will be explored in the chapter.

2 ­Characteristics of the study participants Designing a disclosure process will require recognition of demographic information, the medical circumstances, and an estimate of the prior probability for genetic ­conditions of the study participants.3, 22, 23 Returning incidental finding results within a research protocol targeted to a specific medical condition will have different implications for the participants than returning results in the setting of a healthy population that is receiving genomic sequencing as part of a general population health study. In addition, disclosing results to participants who are less than age 18 and their parents requires careful consideration of the types of results that will be disclosed. As discussed in the chapter on ethics, there is disagreement about

2 ­ Characteristics of the study participants

opting to return to minors those genomic findings associated with conditions that have onset in adulthood. Many individuals who consent to participate in research do so in part, because they hope to learn something new about their medical condition.3, 4, 23, 24 Participant explanations of their rationale for participating in genomics research include the expectation that genomic information will increase understanding about the causes and treatment for their diagnosis. This sentiment has the potential to influence participant receptivity to incidental findings. For example, participants who are recruited into a study to evaluate genomic findings associated with colorectal cancer may be very interested in the colon c­ ancer-related genomic findings, but not interested in the incidental findings unrelated to their medical diagnosis. In such scenarios the participant has been reported to express disbelief that no finding related to their diagnosis was found, but rather an entirely different and unrelated finding was disclosed.4, 6 Finally in a setting of general population genomic sequencing, the terms primary or secondary findings are meaningless as there are only genomic findings. Some participants may have a personal or family history that gives context to the genomic finding; however, for many, any result will be entirely unexpected. For others, their participation comes with the expectation of some type of genomic result, so negative findings are a disappointment.24–26 In summary, attention to the characteristics of the study cohort will shape the approach and content for disclosure sessions.

2.1 ­Notification to participants of a sequence result In surveys of researcher’s attitudes about returning research results, the biggest conundrum often cited is “how to notify participants that there is a result.”23, 27, 28 There are two overarching scenarios to consider in planning to notify participants. The most straightforward scenario is when the possibility for results to return has been anticipated and covered in the consent form, and participants are aware and prepared that notification will happen if certain results are identified. The other more ethically complex scenario happens when the study consent did not anticipate return of results, but upon consultation with the IRB and other local experts, findings now compel the study investigators to decide to return study results.9, 28 Once it is decided that there are results to return by following the recommendations outlined in previous chapters, many suggest that any notification process should give the participant the option of deciding to learn results at this time. As mentioned earlier, circumstances and life events change, which may lead participants to choose differently than they did at the time of consent. In addition, this serves the purpose to allow participants who did not know that they might receive a research result the option to decline learning any research results. In a study by Meacham et al.,27 qualitative responses from researchers suggested an approach that involves sending a generic letter to participants that allows for the participant to choose whether to contact the researcher to learn about research findings. The researchers interviewed also suggested that all participants receive this notification and perhaps the option to reconsent to receive findings, rather than just

137

138

CHAPTER 7  Disclosing genomic sequencing results

those with an actual research result. Researchers with the MyCode Community Health Initiative utilized this approach to reach participants who were consented to recontact but had enrolled prior to the option to learn of research results.29 The approach to contact everyone was suggested as a means to address the potential for recognition of those with positive research findings, if they are the only participants to receive letters. The generic letter approach is less controversial for those studies in which participants have given consent to recontact even if they did not give consent for return of results. For those studies in which no consent for recontact exists, the letter should be crafted with input from the IRB and local healthcare providers with knowledge of the population and the medical conditions implicated by the study results.16, 18, 30 For studies in which the return of results process is anticipated and participants expect that they could notified of a research result, many still opt for a more generic notification that allows participants to choose whether or not to learn their result.12, 31 More recently, wider variability in methods of notification have been described, including within large research consortia such as the Electronic Medical Records and Genomics Network, eMERGE, and the Clinical Sequencing Exploratory Research (CSER) group (statement of purpose of eMERGE). Personal participation in the eMERGE network has facilitated access to a collection of the methods utilized by the different sites in the consortia. Most are using a combination of methods to ensure contact and consent of participants. Meacham et  al.27 describes researchers’ concern about how much effort needs to go into trying to contact participants who do not respond. This has also been an issue encountered by the researchers in the eMERGE network. Many sites report using a combination of phone calls, letters, notification through the patient portal of electronic health records, and reliance on the participant’s healthcare provider to relay the fact that a result is now available. Based on personal experience at Geisinger in the large genomic sequencing project, the MyCode Community Health Initiative, in which participants are consented to receive actionable research results, initial notification of participants utilized phone conversations consisting of condition-specific phone scripts that research coordinators delivered to participants.20, 29, 31 The script anticipates frequently asked questions and refers more complex issues to genetic counselors or the participant’s primary care provider. Based on feedback from participants and healthcare providers and the recognition that urgent medical implications might be discovered when the participant is contacted, the research protocol is changing to have genetic counselors notify participants of their research result to recognize and address potential complex medical presentations that may have more immediate healthcare implications for the participant (Box 1). Recognizing the clinical implications of a research result is a concern mentioned frequently in qualitative interviews with researchers.9, 18, 19, 23 Ravitsky and Wilfond describe that researchers have a responsibility to assess their “capability” to provide appropriate communication of the result. In situations in which the researcher is uncomfortable providing full communication of the result, notification may include referral to appropriate clinical support to fully disclose the result and discuss implications of findings.32

2 ­ Characteristics of the study participants

Box 1  Urgent medical implications of genomic finding. Kadesha didn’t really remember signing up for participation in the biobank research. Weeks from delivery of her first child, it didn’t even enter her thinking about things to come. Out of the blue, she got a phone call from the biobank research coordinator that her analysis of her exome sequence revealed a result that she needed to learn. She told them she was pregnant and asked if she couldn’t learn it in a few months? The coordinator indicated that she would be hearing from a genetic counselor to set up an appointment and the result might have implications for her pregnancy and delivery. The research team, including a medical geneticist and a genetic counselor, met with her the next day to review her result, a pathogenic variant in the gene COL3A1 associated with Vascular Ehlers-Danlos syndrome, which carried significant risk for serious pregnancy complications in the third trimester and at delivery. The disclosure session included physical exam for findings associated with EDSIV, solicitation of medical and family history, discussion of specific actions relative to the pregnancy and delivery, referral to a maternal and fetal medicine specialist, and the invitation to meet again to review this information and implications at a later point when the timing allowed for a more overarching review of the condition. The psychosocial issues involved included attention to shock, distress, and anxiety about the implications for her safety and the well-being of the pregnancy.

Some sites within the eMERGE research network are returning negative sequence results. Most rely on a letter to disclose the negative genomic result. As a reminder, in the eMERGE network, many sites do not have participants enrolled within an indication-based cohort, so no results are considered “primary” versus “secondary.” Whatever the method of notification, sites have chosen to list a contact person for the study who is to be available to answer questions about the research, potentially reconsent for return of results, and withdraw participants who wish to discontinue participation. In addition, a contact for the IRB has generally been given to provide an avenue for complaint regarding the conduct of the study and/or study personnel. An innovative online model for result notification (www.my46.org)33 has been developed that facilitates participant choice about which results to learn, educational materials regarding results and actual return of genomic results.34, 35 The developers caution that it will not provide clinical interpretation for the participant, but it does offer access to genetic counselors and offers a means to store, track, and share results with other providers and family members. An additional advantage is the option for a participant to change one’s choices regarding which results to view over time. This can be seen as a tool for notification and useful in preparing participants for in-person visits during which the time can be focused to answer specific questions and review specific health implications of a result and plan for next steps in health management. In summary, many researchers have suggested that notification takes generic form that gives participants the option to choose whether or not to learn of genomic sequence results including primary or secondary findings. The method of notification is changing as overall methods of communication with patients/participants change. Ideally the method of notification was outlined in the study consent process so that

139

140

CHAPTER 7  Disclosing genomic sequencing results

participants are prepared for the types of results that could be returned and of the process by which they will learn of.

3 ­Disclosure modality and content Many studies have focused on the process of whether to return results and which results to return. Fewer studies report on the actual disclosure process, that is, an in-person visit, telephone or telemedicine visit, or by written communication only. What has been published consists mainly of research evaluating the disclosure process. The modality is often gleaned in the methodology section of qualitative research studies,4, 6, 36 with the mode most frequently described as an in-person visit. In addition, with practice guidelines lacking, it appears that the default mode was in-person visits, often supported by the notion that, because too little is known about how participants will react, it may be better to err on the side of being able to meet the participant to discuss concerns and reveal results. There appears to be consensus that disclosure of research results, including incidental findings, is supported when one can define that the results are meaningful to participants, that the results are valid, that implications can be explained in clear and understandable language, and that comprehensive next step guidance can be offered to participants.10 However, there is no consensus on the best method by which to disclose the findings. Early studies based on hypothetical scenarios reported that respondents preferred that the disclosure visit be divided into two in-person visits separating the primary result disclosure and the incidental findings, thereby potentially avoiding conflation of the findings and their implications.37 Research by Wynn et  al.37 in follow-up with participants of their results’ return process noted that participants wanted the ability to make individual choices about the sequence of the return process and that they did not necessarily want more than one visit. More recent studies noted participants were overwhelmingly interested in findings that apply to their condition and that there was a tendency to “tune out” disclosure of other findings. Wynn suggested that visits be structured in such a way to begin by asking participants which order they prefer to learn results and then stage the disclosure visit according to participant choice.4, 36 In genetic counseling structure, this early discussion is called contracting. It is the opportunity to review expectations for results, reconfirm choices regarding types of results the participant wishes to learn, and identify specific questions and concerns at the outset of the visit. This allows for disclosure visits to be shaped by the participant and enables attention to the specific contextual issues of the participant. The question of who is responsible to disclose results in a given research project is important to consider in the development of an overall disclosure process. Qualitative interviews of researchers and participants alike point to the need to consider the qualifications of the agent returning results.10, 18, 21, 23 Timely, transparent, and effective communication are valuable attributes of a disclosure process that

3 ­ Disclosure modality and content

involves incidental findings. Many published studies of the return of incidental findings have relied on genetic counselors and/or clinical geneticists who conducted in-person disclosure visits. In part, this design was utilized because of the novelty of returning research findings particularly returning incidental findings. The experience of these experts, who are comfortable with discussing complex genetic concepts with patients from all walks of life, can provide templates for the elements of a return process. As pointed out earlier, these experts agree that within the disclosure process there is need to recognize the context of the visit including characteristics and age of the study population, the content of the consent form, and the types of findings to be communicated before structuring the content of an in-person disclosure visit.4, 6, 19, 21 Two studies that have published case scenarios with lessons learned4, 6 suggest several recurrent themes worthy of consideration in the result disclosure visit. The first theme involved managing expectations related to the types of results that could be found and the likelihood that a relevant result would be identified. This was most important in the context of sequencing done to provide genotype information specific to a participant’s personal or family medical history. Participants may elevate the likelihood of a definitive genomic finding once genomic sequencing is discussed and while waiting for the result to be returned. Participants also reported that a comprehensive genomic sequencing approach should be expected to find a diagnosis for their child’s problems.18, 38 In this setting, also reported by Amendola et al.,6 when negative findings were returned, a few participants misinterpreted this to mean that the condition must not be genetic.18 An outline of potential counseling strategies to address this concern published by Brett et al.4 recommended reviewing the participant’s expectations and choices for return at the outset of the disclosure visit to shape the order of the disclosure content. The disclosure process is also recommended to include assessing the psychosocial state of the participant by soliciting past experience with the condition associated with the finding and watching for participant cues regarding the significance of past experiences. Communicating the primary and secondary findings, within an evidence-based discussion of what is known and what may be uncertain, is also recommended.4 Disclosure of research findings in the context of general population studies, which resemble screening programs rather than diagnostic studies, is shaped by the fact that no particular finding is anticipated. All results are in some sense unexpected. There are several reasons why participants may be less prepared to learn about their research finding in this setting. For example, the consent process may not have addressed specifics regarding research results, or pretest counseling may not have been part of the research protocol. Other participants report greater uncertainty related to learning of a research finding that does not fit with their lived experience. To aid in understanding, participants who have taken part in a disclosure process suggest that disclosure visits should allow ample opportunity to discuss their reactions to learning their research results, encourage communication with healthcare providers and family members, provide referral to specialists, and outline supportive information relative to anticipatory guidance.4, 6, 19, 38

141

142

CHAPTER 7  Disclosing genomic sequencing results

Box 2  Dual results. William had signed up to participate in the MyCode biobank a few years ago. When he got the notice that a medically actionable result had been found, he was not surprised. His mother, maternal aunt, and maternal uncle had all been diagnosed with colorectal cancer, and he had genetic testing at the recommendation of his primary care provider and GI provider. He was curious to know for sure, whether the same finding he already knew that he had, a change in the MSH2 gene, was detected. He went in to talk with the genetic counselor associated with the MyCode genetic counseling program and was very surprised to learn that, in addition to the MSH2 variant (it was the same), a variant in the BRCA2 gene had also been detected. He reported being totally overwhelmed. As he and the genetic counselor discussed his family history, he realized that his father’s family history did include an uncle with prostate cancer and his grandmother who had died at an older age of ovarian cancer. With all the focus on his mother’s family and colon cancer, they hadn’t really discussed his father’s cancer history. So, what did this mean now? In some respects, for William, this new finding did not change his cancer screening significantly. He would need to be aware of any changes in his breast tissue, but he was already being watched for prostate cancer risk.

A second theme discussed by Amendola et al.6 involves preparing for multiple findings for the same participant. When more than one finding is noted, which of the findings is considered primary and which is secondary may or may not be relevant to the disclosure process but may impact how participants view the findings. Several instances of dual genomic findings in the Geisinger MyCode Community Health Initiative shed light on the disclosure issues encountered (Box 2). In each of the reported cases, one of the research findings represented a result that could be explained based on known participant personal history or family history. However, often the second result was a surprise. Although Amendola et  al.6 suggest that individual appointments might be necessary to discuss multiple findings, participants in the MyCode disclosure process preferred to learn both results in the same visit.31 For the genetic counselors in the study by Amendola et al.,6 the order of the disclosure of multiple findings was decided after discussion with the participant. In the disclosure visit, soliciting participant’s views of what results they expected to hear often led to discussion of one of the findings with the disclosure of the unanticipated finding after the initial disclosure. Being prepared with supporting information and resources for each of the results will help the participant to begin to adjust to the implications of the results. Research personnel within the MyCode study have chosen to contact participants with more than one result soon after the disclosure visit, anecdotally at about 1–2 weeks, to check in to see if the participant has new questions or concerns. MyCode Community Health Initiative participants are encouraged to invite other close individuals to attend the visit, who can support the participant during the visit.31 This is frequently recommended in traditional genetic counseling visits, to provide support for the individual when “bad news” or complex information will be communicated. This support person is available to also hear the information presented and act as a resource for the individual receiving the genetic counseling.

3 ­ Disclosure modality and content

The third theme acknowledged that disclosure visits in which results are communicated with participants who have no personal or family history of the a­ ssociated condition are likely to take longer than when the results match a participant’s personal or family history.6, 26 Traditional genetic testing has been based on an individual who presents with a history, signs, or symptoms that lead to the choice of a specific genetic test, the result of which hopefully confirms the suspected diagnosis.26 Research participants who learn that they have a genetic finding for which they have no personal context benefit from a careful review of what is known and what is uncertain about the implications of the finding. Current wisdom regarding future risks and appropriate follow-up have been based on data from individuals who have symptoms and/or families with strong evidence of the condition. The overarching theme of uncertainty related to the participant’s personal meaning of the result, interpretation of the risk for adverse health in the future, and appropriate healthcare management decision-making contribute to challenges in formulating the disclosure visit content. Michael Murray presents a model in a commentary, “DNA is not your diagnosis: getting diagnoses right following secondary genomic findings,” that may be useful in shaping the disclosure content and diagnostic process for the variety of contexts encountered among participants particularly those whom he described as having the “genotype without phenotype”38 (see Fig. 1). The model offers clarification in ways to think about what it means to have a genomic finding considering the participant’s personal and family medical history. When researchers have access to participant medical records, the model can be used to illustrate the various implications based on prior review of the participant's history. For those without medical record access, the model can continue to serve to match the participant’s lived experience with the identified genomic finding based on information solicited from the individual that is analogous to that in the medical record. Two main scenarios are possible in the model. The genomic finding is associated with the medical condition that is manifest at the time of or appears after disclosure or the medical condition never appears. The two groups of individuals are divided into useful subgroups. In group 1, individuals with the finding have an existing diagnosis of the medical condition that includes the genetic cause identified by traditional clinical testing. In group 2, previous characteristics of the condition have been noted, with the genomic finding linked to those characteristics providing a unifying diagnosis. Group 3 consists of diagnosis after recommended evaluations based on the genomic finding revealing subclinical manifestations of the medical condition. In group 4, individuals have no diagnosis of the medical condition either clinical or subclinical at the time of disclosure but develop the associated condition over time. Group 5 consists of individuals who have the genotype but are never diagnosed with the medical condition. Groups 1–4 eventually manifest the phenotype expected based on the genomic finding—that is, the variant is penetrant. Group 5 demonstrates lack of penetrance of the genomic finding. Murray makes the point that the genomic finding does not in and of itself imply a diagnosis of the medical condition associated with that finding. It is only after careful review of medical and

143

144

CHAPTER 7  Disclosing genomic sequencing results

Secondary or incidental finding of a PATHOGENIC/LIKELY PATHOGENIC VARIANT

GENE SPECIFIC EVALUATION Including history, exam, testing, consultation

DIAGNOSIS OF GENOMIC SYNDROME WITH TESTING AND INITIAL EVALUATION Both Genotype and Phenotype Present

NO DIAGNOSIS OF GENOMIC SYNDROME WHEN TESTED Genotype without Phenotype

GROUP 1

GROUP 2

GROUP 3

GROUP 4

GROUP 5

Existing Genomic Syndrome Diagnosis Confirmed

Unifying Genomic Syndrome Diagnosis

New Genomic Syndrome Diagnosis Achieved

No Genomic Syndrome Diagnosis Achieved Initially

No Genomic Syndrome Diagnosis Achieved Initially

Previous genotype and phenotype documented

Previously documented phenotype and new genotype

Subclinical phenotype revealed thru evaluation

Phenotype emerges over time

Phenotype does not emerge

GENOMIC SYNDROME DIAGNOSED Both genotype and phenotype

No Genomic Syndrome

FIG. 1 A model for the classification of patients with incidental or secondary findings on genomic sequencing.38 Current strategies for genotype-phenotype correlation cannot distinguish group 4 from group 5. The increasing identification of patients who fit into “genotype without phenotype” (i.e., groups 4 and 5) creates a need for new management plans and healthcare policies.

family history together with physical examination as appropriate that a diagnosis can be made relevant to the participant. Use of the groups described in the model in the disclosure visit when explaining a finding that appears inconsistent may well help participants to understand the nuance behind a genomic finding in a context with no known personal or family medical history. Amendola et al.6 emphasize the need to encourage participants to stay connected with the healthcare system and to take advantage of surveillance options provided. In addition, they recommend that soliciting evidence of participant understanding together with discussion of their feelings and reactions to the information may support healthy adjustment to the results.6 Finally, enlisting the participant’s primary care provider and equipping them to provide ­ongoing information about the result may support appropriate follow-up of genomic sequencing findings.

3 ­ Disclosure modality and content

Another common theme that many researchers did not anticipate involved planning for the possibility that a research participant may be deceased by the time a genomic finding is identified.6 Anecdotally, in the eMERGE network, several sites have had to confront this issue. Most have approached this dilemma in an ad hoc manner, with researchers meeting with representatives of their IRB and institutional ethicists. A few sites have instituted consent changes that specify how and to whom such results will be disclosed. Based on personal communication, one site changed the consent form to request that participants list on the consent form the names of individuals to whom they release results. At Geisinger after consultation with patient advisory groups, the MyCode governing board, members of the IRB, and members of the external ethics advisory board, a process was developed to amend the consent form to include standard language that, if necessary, researchers will notify the emergency contact/next of kin listed in the electronic health record.29 There is no evidence to say one approach is superior to the other; however, these are examples of informing participants of the process to communicate genomic findings should the participant be unable to receive results.29, 39 Recognizing the family setting is the final theme discussed relative to disclosure of genomic sequence results. For study populations that include minor children and adolescents, the disclosure plan may benefit from added scrutiny relative to the types of results that are planned to be disclosed.40 In this setting of sequencing of children, Wilfond et al.40 recommend that parents be offered the option to learn their child’s results for adult-onset conditions. There are acceptable reasons such as significant illness of the child that may lead parents to decline learning results that don’t have a childhood impact on the health of their child. The discussion with parents about deciding whether to receive their child’s secondary findings is recommended to include benefits and risks of learning such findings. It is helpful to share with parents that there are situations in which parents and/or their children may decline learning such results (Box 3).

Box 3  Adolescent scenario. In addition to the search for gene variants linked to her rare cancer diagnosis, Helen, age 16, and her parents were informed that genomic sequencing could identify other gene variants associated with known conditions such as cardiomyopathy, anesthesia risk, or heart arrhythmias for which there are well-formulated guidelines around medical management for her should a pathogenic gene variant be identified. Two other conditions involve cancer syndromes that would not develop until Helen was older with no immediate relevance to her current healthcare management. These conditions, hereditary breast and ovarian cancer and Lynch syndrome, were explained, and Helen’s questions were addressed. The questions and concerns of her parents were also discussed. Helen was interested in knowing about all information available from the testing. She referenced her current high school biology class where they have discussed recent news stories about miraculous cures resulting from whole genome sequencing and gene therapy. Her parents, while interested in the information, did not want to disrupt the focus on her current treatment choice and process. Helen and her parents discussed the options and ultimately coconsented (i.e., assented and gave parental permission) to have genome sequencing and to learn first of the results that pertain to her cancer. At a time when Helen was well and in remission, they would revisit the opportunity to learn of other results available from her sequence information.

145

146

CHAPTER 7  Disclosing genomic sequencing results

3.1 ­Laboratory reports of genomic findings Several studies have evaluated and made recommendations regarding the structure and format for laboratory reporting of genetic testing and exome sequencing.40–42 Most reviews reference clinical testing reports as several agencies regulate reporting clinical test results.43 In addition to well-documented CLIA report requirements (42 CFR 493.1291),43 other agencies including the Centers for Disease Control, American College of Genetics and Genomics, and College of American Pathologists supplement CLIA requirements with the recommendation to include reference to the indication for testing, result interpretation and follow-up, referral for genetic counseling, supporting evidence with references included, and race/ethnicity when appropriate to the result. When asked for their feedback on result templates, providers requested that the clinically relevant information be most prominent in the report.44, 45 Most healthcare providers asked that reports include less information, a concise summary of the findings, active clinical guidance, referral information for relevant specialists, and supporting information that could help them communicate the result with their patient. In a qualitative study by Pet et  al.,46 in which physicians were asked about their reactions to unsolicited genomic findings for their patients, physicians desired access to resources and information with clear evidence-based guidance for any recommended follow-up. A project funded through the Patient-Centered Outcomes Research Institute (PCORI) developed online genomic sequencing report templates. Patients and providers were interviewed regarding preferences for report content to address the types of results available following genomic sequencing. Both patients and providers approved the reports that included patient demographic information, primary care provider contact information, prominent placement of any primary result and the clinical rationale supporting the interpretation, less prominently placed secondary findings and their clinical rationale, the presence or absence of clinical confirmation (when needed), resources for the provider, resources for patients and families, access to research and/or clinical trials, a summary of the clinical features of the patient utilized in evaluating the result, anticipatory guidance for the result, inheritance patterns, the expected next steps for management of the results, and technical documentation of the sequencing methods. In this study, providers indicated that the content of the report could serve as the basis for communication about the genomic sequencing results and management implications with their patients by reminding them of the information to be discussed when returning a result or set of results.45, 47 The need for reports designed specifically for patients is a more recent endeavor38, 48 impacted by “meaningful use” components of the Centers for Medicare and Medicaid Services electronic health record mandate that patients have access to all results from laboratory testing. Researchers formulated guidance for genomic reports that recommended the inclusion of an interpretive summary section, a summary letter to the patient, development of a user guide, and a completely revised patient-centered laboratory report.44, 48 Qualitative interviews with participants regarding genomic sequencing result reports identified three themes regarding these

3 ­ Disclosure modality and content

reports. The first is that participants report that they are always looking for valid and reliable information and resources about the genomic finding. The second theme was similar to that mentioned earlier with regard to enhanced communication with healthcare providers and other professionals who interact in their care. The third theme involved specific elements of a report that were judged critical in a genomic result report: simple language, logical flow of information, interesting and engaging presentation of information, specific next steps in the care and management expected based on the finding (which could be extensive), what to expect in the future, resources for families, and a glossary of the medical terms used in the report.38 A report designed with these elements may be used in place of the traditional summary letter following a disclosure visit. In this research setting the report was available online within the electronic health record that presented the added benefit of allowing participants to go back to the information as questions arose and allowed them to share the report with other professionals and family members.47 The opportunity to revisit the information has been recognized in other studies.22, 49 Participants and providers appreciated the fact that the genomic result information was equally available to both creating a sense of shared understanding, reported as empowering participants and providers alike.47

3.2 ­Who will return results Publications describing important elements to consider in the disclosure process often list the need to identify knowledgeable agents who can respond to questions regarding the result.21, 38, 50–52 Participants also preferred someone they trust such as their primary care provider. As previously pointed out, many published accounts of disclosing secondary research findings have involved genetic counselors and/or medical geneticists. In accounts describing important elements to consider in the disclosure process, a recurrent element involves enlisting the support of local experts in the disclosure process to help ensure appropriate education around the result and provide genetic counseling to support the participant and their family and for medical care required in follow-up of the result. Darnell et al.53 describe a structured clinical service that could be acknowledged to reside as a bridge between research obligations and clinical communication and support of incidental findings. It is anticipated that not all genomic findings will require an in-person visit or genetic counseling expertise. The Consent and Disclosure Recommendations (CADRe) workgroup of the Clinical Genome Resource (www.clinicalgenome.org)52 recently published a rubric (see Fig. 2) for nongenetic providers that includes salient issues to use in content formulation for consent and disclosure processes. The rubric allows for the consideration of many of contextual issues discussed in this chapter.54 The workgroup proposed several attributes of patients who may benefit when the disclosure visit is scheduled to occur with genetic professionals. While the summary reflects clinical testing, there are important parallels that can be made to the research

147

148

CHAPTER 7  Disclosing genomic sequencing results

FIG. 2 CADRe levels of communication.54

setting. The attributes described by Ormond et al. included patients who request genetic counseling, are highly anxious, will learn of positive predictive results, have low health literacy or limited scientific knowledge, and are adolescents. They also recommended disclosure by genetic professionals when a provider is uncertain or uncomfortable with any aspect of the disclosure visit.

4 ­Summary Early in the design of a genomic sequencing research protocol, researchers are encouraged, and some argue are obligated, to consider whether or not the research will result in reportable findings, primary or secondary, with significant health implications for participants, and whether the option to receive results should be provided

­References

to participants. The protocol should address the range and types of such findings, anticipate which results will be identified for return, and formulate a plan for returning results to participants. The disclosure program should consist of the who, what, how, and when of their disclosure process. The consent process should outline the program, consent participants regarding disclosure choice, and begin to set expectations about the return of research results. Participants will need to be notified when a research finding is available for them to learn. Most advocate for a process that allows for participant choice whether to learn that a result is available or to decline the information. The disclosure process for primary or secondary findings related to genomic sequencing has most frequently involved in-person visit(s) to provide education about the findings, set the clinical context for the result, outline medical management implications relative to risk, discuss psychosocial issues, respond to questions, and encourage sharing of results with family members. Reports of research results are recommended to include clear and concise language, the primary genomic finding and relationship to the current presentation of the participant, any incidental/secondary findings and clinical rationale, specific medical management actions, referral to appropriate specialists, and resources for providers and participants. Proper planning will facilitate a robust program for disclosure of genomic findings in research settings.

­References 1. Green RC, Berg JS, Grody WW, et al. ACMG recommendations for reporting of incidental findings in clinical exome and genome sequencing. Genet Med. 2013;15(7):565–574. https://doi.org/10.1038/gim.2013.73. [Erratum in: Genet Med 2017;19(5):606]. 2. ACMG Board of Directors. ACMG policy statement: updated recommendations regarding analysis and reporting of secondary findings in clinical genome-scale sequencing. Genet Med. 2014;17(1):68–69. https://doi.org/10.1038/gim.2014.151. 3. Smith  LA, Douglas  J, Braxton  AA, Kramer  K. Reporting incidental findings in clinical whole exome sequencing: incorporation of the 2013 ACMG recommendations into current practices of genetic counseling. J Genet Couns. 2014;24(4):654–662. https://doi. org/10.1007/s10897-014-9794-4. 4. Brett GR, Wilkins EJ, Creed ET, et aI. Genetic counseling in the era of genomics: what's all the fuss about? J Genet Couns. 2018. https://doi.org/10.1007/s10897-018-0216-x. 5. Lohn  Z, Adam  S, Birch  PH, Friedman  JM. Incidental findings from clinical genomewide sequencing: a review. J Genet Couns. 2013;23(4):463–473. https://doi.org/10.1007/ s10897-013-9604-4. 6. Amendola  LM, Lautenbach  D, Scollon  S, et  al. Illustrative case studies in the return of exome and genome sequencing results. Pers Med. 2015;12(3):283–295. https://doi. org/10.2217/pme14.89. 7. Jamal SM, Yu J, Chong JX, et al. Practices and policies of clinical exome sequencing providers: analysis and implications. Am J Med Genet A. 2013;161A:935–950. https://doi. org/10.1002/j.1552-4833.2013.35942.x.

149

150

CHAPTER 7  Disclosing genomic sequencing results

8. Levy  KF, Blake  K, Fletcher-Hoppe  C, et  al. Opportunities to implement a sustainable genomic medicine program: lessons learned from the IGNITE Network. Genet Med. 2018;https://doi.org/10.1038/s41436-018-0080-y. 9. Thorogood A, Joly Y, Knoppers BM, et al. An implementation framework for the feedback of individual research results and incidental findings in research. BMC Med Ethics. 2014;15(1):https://doi.org/10.1186/1472-6939-15-88. 10. Wolf SM, Lawrenz FP, Nelson CA, et al. Managing incidental findings in human subjects research: analysis and recommendations. J Law Med Ethics. 2008;36(2):219–248. https:// doi.org/10.1111/j.1748-720X.2008.00266.x. 11. Parens  E, Appelbaum  P, Chung  W. Incidental findings in the era of whole genome sequencing? Hastings Cent Rep. 2013;43(4):16–99. https://doi.org/10.1002/hast.189. 12. Jarvik GP, Amendola LM, Berg JS, et al. Return of genomic results to research participants: the floor, the ceiling, and the choices in between. Am J Hum Genet. 2014;94(6):818–826. https://doi.org/10.1016/j.ajhg.2014.04.009. 13. Klitzman R, Buquez B, Appelbaum PS, et al. Processes and factors involved in decisions regarding return of incidental genomic findings in research. Genet Med. 2014;16(4):311–317. https://doi.org/10.1038/gim.2013.140. 14. Wolf SM. Return of individual research results and incidental findings: facing the challenges of translational science. Annu Rev Genom Hum Genet. 2013;14:557–577. https:// doi.org/10.1146/annurev-genom-091212-153506. 15. Khoury MJ, Gwinn M, Bowen MS, Dotson WD. Beyond base pairs to bedside: a population perspective on how genomics can improve health. Am J Publ Health. 2012;102(1):34– 37. https://doi.org/10.2105/AJPH.2011.300299. 16. Wolf  SM, Burke  W, Koenig  BA. Mapping the ethics of translational genomics: situating return of results and navigating the research-clinical divide. J Law Med Ethics. 2015;43(3):486–501. https://doi.org/10.1111/jlme.12291. 17. Bennette CS, Trinidad SB, Fullerton SM, et al. Return of incidental findings in genomic medicine: measuring what patients value. Genet Med. 2013;15(11):873–881. https://doi. org/10.1038/gim.2013.63. 18. Klitzman R, Applebaum PS, Fryer A, et al. Researcher views on return of incidental genomic research results: qualitative and quantitative findings. Genet Med. 2013;15(11):888– 895. https://doi.org/10.1038/gim.2013.87. 19. Brandt DS, Shinkunas L, Hillis SL, et al. A closer look at the recommended criteria for disclosing genetic results: perspectives of medical genetic specialists, genomic researchers, and institutional review board chairs. J Genet Couns. 2013;22(4):544–553. https://doi. org/10.1007/s10897-013-9583-5. 20. Carey DJ, Fetterolf SN, Davis FD, et al. The Geisinger MyCode community health initiative: an electronic health record-linked biobank for precision medicine research. Genet Med. 2016;18(9):906–913. https://doi.org/10.1038/gim.2015.187. 21. Berrios  C, James  CA, Raraigh  K. Enrolling genomics research participants through a clinical setting: the impact of existing clinical relationships on informed consent and expectations for return of research results. J Genet Couns. 2018;27(1):263–273. https://doi. org/10.1007/s10897-017-0143-2. 22. Krier  JB, Green  RC. Management of incidental findings in clinical genomic sequencing. Curr Protocols Hum Genet. 2013. https://doi.org/10.1002/0471142905.hg0923s77. [Chapter 9: Unit 9.23]. 23. Shalowitz DI, Miller FG. Communicating the results of clinical research to participants: attitudes, practices, and future directions. PLoS Med. 2008;5(5):e91, https://doi.org/10.1371/ journal.pmed.0050091.

­References

24. Facio FM, Lee K, O’Daniel JM. A genetic counselor’s guide to using next generation sequencing in clinical practice. J Genet Couns. 2014;23(4):455–462. https://doi.org/10.1007/ s10897-013-9662-7. 25. Wynn J, Martinez J, Duong J, et al. Association of researcher characteristics with views on return of incidental findings from genomic research. J Genet Couns. 2015;24(5):833–841. https://doi.org/10.1007/s10897-014-9817-1. 26. Biesecker LG. Opportunities and challenges for the integration of massively parallel genomic sequencing into clinical practice: lessons from the ClinSeq™ project. Genet Med. 2012;14(4):393–398. https://doi.org/10.1038/gim.2011.78. 27. Meacham  MC, Starks  H, Burke  W, Edwards  K. Researcher perspectives on disclosure of incidental findings in genetic research. J Emp Res Hum Res Ethics. 2010;5(3):31–41. https://doi.org/10.1525/jer.2010.5.3.31. 28. Beskow LM, Burke W. Offering individual genetic research results: context matters. Sci Transl Med. 2010;2(23):38cm20, https://doi.org/10.1126/scitanslmed.3000952. 29. Williams  MS, Buchanan  AH, Davis  FD, et  al. Patient-centered precision health in a learning health care system: Geisinger’s genomic medicine experience. Health Aff. 2018;37(5):757–764. https://doi.org/10.1377/hlthaff.2017.1557. 30. Wolf  SM, Paradise  J, Caga-anan  C. The law of incidental findings in human subjects research: establishing researchers’ duties. J Law Med Ethics. 2008;36(2):361–383. 214, https://doi.org/10.1111/j.1748-720X.2008.00281.x. 31. Schwartz MLB, McCormick CZ, Lazzeri AL, et al. A model for genome-first care: returning secondary genomic findings to participants and their healthcare providers in a large research cohort. Am J Hum Genet. 2018;103(3):328–337. https://doi.org/10.1016/j. ajhg.2018.07.009. 32. Ravitsky V, Wilfond BS. Disclosing individual genetic results to research participants. Am J Bioethics. 2006;6(6):8–17. https://doi.org/10.1080/15265160600934772. 33. https://www.my46.org; (Accessed September 1, 2018). 34. Yu  JH, Jamal  SM, Tabor  HK, Bamshad  MJ. Self-guided management of exome and whole-genome sequencing results: changing the results return model. Genet Med. 2013;15(9):684–690. https://doi.org/10.1038/gim.2013.35. 35. Tabor HK, Jamal SM, Yu JH, et al. My46: a Web-based tool for self-guided management of genomic test results in research and clinical settings. Genet Med. 2016;19(4):467–475. https://doi.org/10.1038/gim.2016.133. 36. Wynn J, Martinez J, Bulafka J, et al. Impact of receiving secondary findings from genomic research: A 12-month longitudinal study. J Genet Couns. 2018;27(3):709–722. https://doi. org/10.1007/s10897-017-0172-x. 37. Levenseller BL, Soucier DJ, Miller VA, et al. Stakeholders' opinions on the implementation of pediatric whole exome sequencing: implications for informed consent. J Genet Couns. 2014;23(4):552–565. https://doi.org/10.1007/s10897-013-9626-y. 38. Stuckey H, Williams JL, Fan AL, et al. Enhancing genomic laboratory reports from the patients' view: a qualitative analysis. Am J Med Genet A. 2015;167A(10):2238–2243. https:// doi.org/10.1002/ajmg.a.37174. 39. Wilfond BS, Fernandez CV, Green RC. Disclosing secondary findings from sequencing to families: considering the “benefit to families”. J Law Med Ethics. 2015;43(3):552–558. https://doi.org/10.1111/jlme.12298. 40. Brownstein CA, Beggs AH, Homer N, et al. An international effort towards developing standards for best practices in analysis, interpretation and reporting of clinic genome sequencing results in the CLARITY challenge. Genome Biol. 2014;15(3):R53. https://doi. org/10.1186/gb-2014-15-3-r53.

151

152

CHAPTER 7  Disclosing genomic sequencing results

41. Lubin IM, Caggana M, Constantin C, et al. Ordering molecular genetic tests and reporting results: practices in laboratory and clinical settings. J Mol Diagn. 2008;10(5):459–468. https://doi.org/10.2353/jmoldx.2008.080050. 42. Scheuner MT, Hilborne L, Brown J, Lubin IM, Members of the RAND Molecular Genetic Test Report Advisory Board. A report template for molecular genetic tests designed to improve communication between the clinician and laboratory. Genet Test Mol Biomark. 2012;16(7):761–769. https://doi.org/10.1089/gtmb.2011.0328. 43. 42 CFR 493.1291. https://www.law.cornell.edu/cfr/text/42/493.1291; (Accessed August 25, 2018). 44. Dorschner  MO, Amendola  LM, Shirts  BH, et  al. Refining the structure and content of clinical genomic reports. Am J Med Genet C Semin Med Genet. 2014;166C(1):85–92. https://doi.org/10.1002/ajmg.c.31395. 45. Williams JL, Rahm AK, Stuckey H, et al. Enhancing genomic laboratory reports: A qualitative analysis of provider review. Am J Med Genet A. 2016;170A(5):1134–1141. https:// doi.org/10.1002/ajmg.a.37573. 46. Pet DB, Holm IA, Williams JL, et al. Physicians’ perspectives on receiving unsolicited genomic results. Genet Med. 2018. https://doi.org/10.1038/s41436-018-0047-z. 47. Goehringer  JM, Bonhag  MA, Jones  LK, et  al. Generation and implementation of a ­patient-centered and patient-facing genomic test report in the EHR. EGEMS (Wash DC). 2018;6(1):14. https://doi.org/10.5334/egems.256. 48. Haga  SB, Mills  R, Pollak  KI, et  al. Developing patient-friendly genetic and genomic test reports: formats to promote patient engagement and understanding. Genome Med. 2014;6(7):58. https://doi.org/10.1186/s13073-014-0058-6. 49. Tabor HK, Stock J, Brazg T, et al. Informed consent for whole genome sequencing: a qualitative analysis of participant expectations and perceptions of risks, benefits, and harms. Am J Med Genet A. 2012;158A:1310–1319. https://doi.org/10.1002/ajmg.a.35328. 50. Kleiderman E, Avard D, Besso A, et al. Disclosure of incidental findings in cancer genomic research: investigators’ perceptions on obligations and barriers. Clin Genet. 2015;88(4):320–326. https://doi.org/10.1111/cge. 12540. 51. Jones LK, Rahm AK, Gionfriddo MR, et al. Developing pharmacogenomic reports: insights from patients and clinicians. Clin Transl Sci. 2018;11(3):289–295. https://doi. org/10.1111/cts.12534. 52. Clinical Genome Resource. https://www.clinicalgenome.org; (Accessed August 24, 2018). 53. Darnell AJ, Austin H, Bluemke DA, et al. A clinical service to support the return of secondary genomic findings in human research. Am J Hum Genet. 2016;98(3):435–441. https://doi.org/10.1016/j.ajhg.2016.01.010. 54. Ormond KE, Hallquist MLG, Buchanan AH, et al. Developing a conceptual, reproducible, rubric-based approach to consent and result disclosure for genetic testing by clinicians with minimal genetics background. Genet Med. 2018. https://doi.org/10.1038/ s41436-018-0093-6.

­Further reading Bunnik EM, Bodegom LV, Pinxten W, Beaufort IDD, Vernoojj MW. Ethical framework for the detection, management and communication of incidental findings in imaging studies, building on an interview study of researchers’ practices and perspectives. BMC Med Ethics. 2017;18(1):https://doi.org/10.1186/s12910-017-0168-y.

­Further reading

Chan  B, Facio  FM, Eidem  H, et  al. Genomic inheritances: disclosing individual research results from whole-exome sequencing to deceased participants' relatives. Am J Bioeth. 2012;12(10):1–8. https://doi.org/10.1080/15265161.2012.699138. Murray  MF. Your DNA is not your diagnosis: getting diagnoses right following secondary genomic findings. Genet Med. 2016;18(8):765–767. https://doi.org/10.1038/gim.2015.134.

153

CHAPTER

Implications of secondary findings for clinical contexts

8

Michael Morrisona, Harriet Tearea,b, Gabrielle Bertierc,*, James Buchanand,*, Yasmin Bylstrae,*, Clara Gafff,g,*, Leigh Jacksonh,*, Kazuto Katoi,*, Elke Kaufmannj,*, Susan Kellyk,*, Gabriel Lázaro-Muñozl,*, Liis Leitsalum,*, Lili Milanim,*, G. Owen Schaefern,*, Christoph Schickhardtj,*, Mahsa Shabanio,*, Erin Tuttyp,*, Eva C. Winklerj,*, Sarah Wordsworthd,* a

Centre for Health, Law and Emerging Technologies (HeLEX), Faculty of Law, University of Oxford, Oxford, United Kingdom b HeLEX Melbourne, Melbourne Law School, The University of Melbourne, Carlton, VIC, Australia c Center for Genomic Health, Charles R. Bronfman Institute for Personalized Medicine, Icahn School of Medicine at Mount Sinai, Manhattan, NY, United States d Health Economics Research Centre (HERC), Nuffield Department of Population Health, University of Oxford, Oxford, United Kingdom e SingHealth Duke-NUS Institute of Precision Medicine (PRISM), Singapore, Singapore f Melbourne Genomics Health Alliance, Parkville, VIC, Australia g Faculty of Medicine, Dentistry and Health Sciences, The University of Melbourne, Melbourne, VIC, Australia h University of Exeter Medical School, Royal Devon & Exeter Hospital, Exeter, United Kingdom i Department of Biomedical Ethics and Public Policy, Graduate School of Medicine, Osaka University, Suita, Japan j National Center for Tumour Diseases (NCT), German Cancer Research Center (DKFZ), Heidelberg, Germany k Egenis Centre for the Study of the Life Sciences, University of Exeter, Exeter, United Kingdom l Center for Medical Ethics and Health Policy, Baylor College of Medicine, Houston, TX, United States m Estonian Genome Center, University of Tartu, Tartu, Estonia n Centre for Biomedical Ethics, Yong Loo Lin School of Medicine, National University of Singapore, Singapore, Singapore o Centre for Biomedical Ethics and Law, Katholieke Universiteit Leuven, Leuven, Belgium p Victorian Clinical Genetics Services, Murdoch Children’s Research Institute, Royal Children's Hospital, Parkville, VIC, Australia

*

Each of these authors should be recognized for an equal contribution to the chapter.

Secondary Findings in Genomic Research. https://doi.org/10.1016/B978-0-12-816549-2.00008-4 © 2020 Elsevier Inc. All rights reserved.

155

156

CHAPTER 8  Implications of secondary findings for clinical contexts

1 ­Introduction In recent years, several countries have initiated research projects that will set the stage for rollout of next-generation genome sequencing (NGS) in the clinic. For example, the 100,000 Genomes Project in the United Kingdom (announced in 2013), the Precision Medicine Initiative in the United States (announced in 2015), and the Australian Genomics Health Alliance (AGHA) (announced in 2015) all examine the opportunities for introducing genomics into routine patient care. Implementation, whether of whole genome sequencing (WGS), whole exome sequencing (WES), or multiple panel tests, will not occur overnight; there will be a continuing need to gather data at a population level to progress our understanding of genomics, necessarily bringing clinical care into close contact with research, as research participants are recruited through the clinic and research findings are relevant for immediate patient care. This blurring of lines between research and clinical care raises distinct ethical challenges, as the limits of our understanding of the implications of genetic variation are quickly reached. Developing clear policies to manage this potential friction is crucial to support researchers and clinicians working in this area, managing the findings from genomic sequencing and supporting individuals, as both patients and research participants, in their understanding of the implications of these data. The results arising from genomic sequencing are often separated into two distinct categories: (1) health-related findings, or findings that are “pertinent” to the question intended to be answered by sequencing and (2) variants that are not immediately relevant to the question, described as incidental, accidental, secondary, unsolicited, unexpected, unrelated, nonpertinent, ancillary, and additional. A third category could perhaps be (3) variants of unknown significance (VUS), as, given their significance has not yet been determined, it is difficult to tell whether they might eventually be pertinent or not. VUS are usually included in category 2, however, and treated as an incidental or secondary finding (the term used herein). The categorization of findings is usually drawn along the lines of pertinence, rather than distinguishing between somatic or germ line findings, as this is determined by the initial question. The likelihood of secondary findings may differ in clinical care and research settings, given that genomic sequencing within the clinic is more usually initiated for diagnostic purposes, which may give rise to a more defined question being asked, and a clearer understanding of the regions within the genome that will be of interest. However, there is still significant risk that, in these explorations, other information will be unearthed that could be relevant to the patient. Understanding how to broach these occurrences in both the clinic and research is vital, if the information has health implications. Given the clear need for guidance for healthcare practitioners and researchers working in this area, the first challenge is to define secondary findings. The American College of Medical Genetics and Genomics (ACMG) define “incidental or secondary findings” as “the results of a deliberate search for pathogenic or likely pathogenic alterations in genes that are not apparently relevant to a diagnostic indication for which the sequencing test was ordered.” In the United Kingdom a report commissioned by the Medical Research Council (MRC) and Wellcome Trust

1 ­ Introduction

(WT) (referred to herein as the “MRC/WT report”) defines them as a finding “that is discovered in the course of conducting research but is beyond the aims of the study.” Interestingly the ACMG definition alludes to a deliberate search for such findings, while the WT/MRC report, in its more general description, would relate to findings both sought and those “stumbled across” without specific prior intent. As the previous chapters in this volume describe, secondary findings are intensely debated within the genomics community, and there is uncertainty about how best to manage them, whether they should be fed back to research participants and, if so, in what circumstance. In the clinic, opinion is divided on how best to proceed and what information the patient should receive. Secondary findings are not a challenge solely contained within genomics; in any medical specialism, there might be risk of finding something about the patient’s health that is not related to the initial question, for example, the startling frequency with which cancer diagnoses are made in the course of accident and emergency department care or tumors being detected in imaging scans that were requested for something completely different. The difference with genomics is the relative infancy of the field, the scale of uncertainty about what different findings mean, and the implications they have for a patient’s health (and that of their family). The time scale of influence is also relevant—often secondary findings in other specialisms have immediate consequences for a patient—while in genomics the information could be related to a risk of disease where symptoms could lie decades into the future or may never materialize. As focus on genomics continues and more countries conduct population-level research projects to set the foundations for rollout of sequencing in clinical care, the issue of secondary findings is increasing in prominence. Researchers are universally encouraged to consider their policies on feedback of results and are looking to their professional bodies for guidance. In response to the clear need for a position on secondary findings to support healthcare practitioners, the ACMG ignited discussion on this topic with their announcement in March 2013 of a list of 57 genes deemed medically actionable. These were mutations that could lead to severe outcomes, including inherited cancers, inherited cardiac diseases, connective tissue diseases affecting the cardiovascular system, familial hypercholesterolemia, and malignant hyperthermia susceptibility. The ACMG recommended that known pathogenic (or in some cases expected pathogenic) variants discovered in any of these genes should be reported regardless of an individual’s age. The international community was largely critical of the two elements of this recommendation, one that these variants should be actively sought and two that they should be fed back to the individual whether they wanted to receive them or not. For many, this undermined the ethical concept that a patient had a “right not to know” while creating significant implications for project resource. The European Society of Human Genetics (ESHG) took a much narrower approach, advising that WGS be restricted to regions of the genome that were most likely relevant to the patient’s potential diagnosis, with wider testing needing specific justification. Actively searching for additional findings would therefore go against this premise. In recognition that unsolicited findings would occasionally occur, the ESHG made clear that the

157

158

CHAPTER 8  Implications of secondary findings for clinical contexts

patient’s right not to know was important, although, in instances where the patient is not in a position to fully understand the implications of not knowing a result, the physician may still have a moral duty to inform relatives. The general advice was to adopt a cautious approach. In response to this criticism, the ACMG amended their recommendations in 2016 stating that “Adherence to these recommendations is completely voluntary and does not necessarily assure a successful medical outcome.” The approach taken by the MRC/WT report was more to guide researchers in how to think about the problem, rather than prescribing a solution, acknowledging that policy would be highly project dependent and further research was needed to fully access risks and benefits to patients, and concluding that above all projects needed to think about secondary findings and draft a clear policy in advance. Several high-profile examples from the United Kingdom demonstrate the difficulty encountered when preemptively designing policy. The UK Biobank recruited 500,000 participants aged between 40 and 69 (2006–10) to gather health data and samples for future research. The policy upfront was that participants would not receive any results, as stated in the consent form: “I understand that none of my results will be given to me (except for some measurements during this visit).” When UK Biobank introduced an imaging project, this policy changed, to allow for the possibility that secondary findings might be returned, specifically relating to the findings from scans. While this is not a genomics example, it demonstrates the difficulty different projects may have in anticipating issues ahead of time and balancing practical measures with the concerns of personnel involved in analyzing results, who might feel obliged to feedback information. Part of the challenge is that patients have mixed views of how to broach secondary findings. Patients involved in focus groups that fed into the MRC/WT report offered a broad spectrum of opinion, ranging from the supportive You’ve got the choice then to respond…and for your family to be aware also, if maybe there’s a possibility it's hereditary, or there’s maybe a female carrier or a male carrier with an illness (Focus group participant, Cardiff) to the doubtful It might create unnecessary worry which potentially could lead on to psychosomatic illnesses. You think you’re ill, you’re going to be ill (Focus group participant, Cardiff). A 2016 paper1 questioned different stakeholders about feeding back research findings to participants that may not be related to primary project aims. This crosssectional survey gathered views from 6944 individuals (75 countries), including members of the public (4961), genetic health professionals (533), nongenetic health professionals (843), and genomic researchers (607). The vast majority of respondents (98%) considered treatability of the condition/disease to be paramount, but most did not expect researchers to actively search for secondary findings within a research setting. The study demonstrated that, on many issues, the genetic health professionals held more conservative views than other stakeholders, which might suggest the need for further exploration of this area to better understand the disconnect between the views of the professionals, who arguably understand the implications of the data more clearly, and those of the patient.

1 ­ Introduction

The differences of opinion across the genomics community extend beyond just the policy approach to feeding back results. One of the major challenges within this complex discussion lies with the science itself. Clinical actionability is put forth as a useful criterion for feeding back results or not—and understandably so—if there is something that could immediately be done to improve the health of the patient, rationally, and indeed ethically, this should be promoted. Defining clinical actionability, however, is difficult. There is disagreement among genomics experts about which variants might be acted upon and how and crucially what this might mean for immediate patient care. Furthermore, clinical actionability suggests relevance to the immediate patient; however, some of the results that could have major significance are related not to the health of the patient, but instead to future generations, relating to the carrier status of individuals, which may have implications for future reproductive decisions. Pediatric care similarly raises further challenges, given the time frames for actionability—should variants relating to adult-onset conditions be disclosed to children (via their parents) and what implications might that have for their childhood and future autonomy? This level of disagreement across different aspects of secondary findings is the reason for the need for careful consideration of policies, further discussion with patients, and better understanding of the implications of this information. The language used to describe this group of findings may influence how they are addressed by the community and received by patients. As alluded earlier, “incidental” is commonly used and clearly refers to findings that were not immediately relevant to the question in hand; however, the term has received criticism from patients, for the connotation of nonimportance. Given these findings could refer to risk of life-threatening disease, including cancer and heart disease, they may be of huge significance to the patient, regardless of the initial clinical question.2 It is helpful to draw a distinction between findings that are not pertinent to the primary clinical purpose for which NGS analysis is being carried out but that are actively looked for and those that are accidentally discovered, as these two groupings would require different procedures from the clinical teams involved. It would be relatively straightforward to have a policy that stipulates that findings will not be fed back and thus states that the clinical team will not actively look for these variants. However, if a variant that could have clinically significance is genuinely unexpectedly uncovered, it may then fall to the conscience of the bioinformatician about whether this should be fed back. Anecdotally, this is said to happen reasonably often, where a researcher has gone against an explicit policy because a result could have such dramatic and immediate consequence for a patient. By encouraging teams to devise policies in advance, it is hoped this scenario could be avoided and the policy could be guided directly by the wishes of the patient. This chapter will explore initial experiences in different nation states to understand how return of secondary and incidental findings in clinical contexts has played out so far and what can be learnt from these approaches (Table 1).

159

160

CHAPTER 8  Implications of secondary findings for clinical contexts

Table 1  Definitions provided in each case study. Country

Term

Definition

The United States The United Kingdom

Secondary findings Incidental findings

Australia

Incidental/secondary

Germany

Additional findings

France

Secondary or incidental

Actively looked for, clinically relevant Secondary findings—that are actively sought after but are not related to the condition in question Result from analysis of the data/ interrogation of genes that are not indicated by the patient’s clinical presentation Findings unrelated to the initial investigation/question but relevant for the health and/or reproductive plans of the person and/or relatives Will not hunt for specific variants, would only return actionable results

Canada

Secondary findings Incidental findings Secondary findings or incidental findings (of potential health or reproductive importance) Incidental findings (or secondary) Incidental/secondary findings

Singapore

Estonia Japan

Incidental findings are broadly defined in the law to include what are known internationally as “secondary findings” Clinically significant unrelated to the indication of testing Nonprimary results; germ line variants of inherited diseases that could be found during analysis of cancer tissues and blood samples using gene panels

2 ­International approaches to genomics in clinical care and translational medicine Genomics is a global phenomenon, although it is not developing at the same pace or in the same way in all territories and jurisdictions. This section presents a snapshot of the way genomics is being implemented in clinical practice in a range of countries and regions. This is intended to provide a sense of how the issue of secondary or additional findings is being dealt with in different countries with different healthcare systems and different governance regimes. Since genomics is not yet a routine part of clinical care for most patients in most states, these country profiles also review current translational genomics efforts and, where relevant, the influence of research governance on the process of the translation of genomic sequencing into the clinic.

2.1 ­United States Medical practices, such as how to manage secondary findings, in the United States are most often determined by the standard of care or the type of care that is expected

2 ­ International approaches to genomics

from a minimally competent physician in the same field, with similar experience, and under similar circumstances.3 The standard of care is often shaped by the way physicians typically manage a particular situation in the clinic. However, many states also require that clinicians do not just do what others ordinarily do but that their actions are also those of a reasonably prudent clinician.3 A standard of care may also be influenced by hospital policies, state or federal regulations, and statements from influential professional groups. When novel technologies, like genomic testing, enter the clinical context, the standard of care may be unclear because it takes time for a standard to develop. When it comes to managing secondary findings in the clinical setting, there are no US state or federal regulations that directly address this issue. Thus, their management depends mainly on the standard of care that develops among clinicians handling secondary findings. In 2013 The American College of Medical Genetics and Genomics (ACMG) published recommendations regarding how to manage some aspects of secondary findings in the clinical setting. The ACMG recommended that “whenever clinical sequencing is ordered, the ordering clinician should discuss with the patient the possibility of [secondary] findings and that laboratories should seek and report [57 secondary findings]… described in [these recommendations] without reference to patient preferences.”4 These recommendations were influential in the United States for various reasons, including the standing of the ACMG as a professional organization, and that they published at a time in which there was little guidance about how to manage these findings. Thus, these recommendations helped fill a gap in the standard of care. As described in more detail in the succeeding text, medical professionals and institutions in the United States are, to a large extent, following ACMG’s current recommendations regarding secondary findings. After the ACMG published its original recommendations, there was significant backlash because many interpreted the ACMG recommendation not to ask patients whether they want secondary findings analyzed, as a violation of the tradition and legal obligation of respect for patient autonomy in the United States.5,6 The ACMG responded by modifying their recommendations to offer patients the opportunity to opt out of the analysis of secondary findings.7 More recently the ACMG published an updated list of the now 59 “medically actionable genes” that the organization recommends should be analyzed as secondary targets, whenever clinical sequencing is performed.8 The Secondary Findings Maintenance Group of the ACMG will periodically curate and update their list (available at: www.acmg.net) of medically actionable genes recommended for analysis as secondary targets when clinical sequencing is performed. Studies have examined the degree to which clinical sequencing laboratories in the United States are following the ACMG guidelines. Fowler and colleagues found that 94% of consent forms examined from these clinical laboratories contained language that stipulates whether the laboratory will examine and report findings regarding the ACMG list of 59 secondary target genes.9 Approximately 80% of clinical laboratories offered patients the opportunity to opt out or opt in of the report of some or all the 59 ACMG genes. This suggests that the ACMG guidelines have been influential in practice and that, based on their consent forms, clinical genomics laboratories in

161

162

CHAPTER 8  Implications of secondary findings for clinical contexts

the United States are following the ACMG recommendations regarding secondary findings. As the practice of offering the analysis of ACMG-recommended secondary targets in the clinical setting becomes more prevalent, a more concrete standard of practice and standard of care emerge for clinical sequencing laboratories and clinicians, respectively. The ACMG list of medically actionable genes is likely the most influential and widely used in the United States when it comes to offering secondary findings. However, it is important to note that there is no clear consensus regarding what constitutes a medically actionable gene. Geisinger Health System’s MyCode Community Health Initiative encourages patients to submit samples for genomic sequencing, and it analyzes and reports back findings from 77 medically actionable genes associated with 25 conditions.10,11 Notably, MyCode is a screening program, and therefore, strictly speaking, these would not be considered secondary findings because examining these medically actionable genes is the primary goal of this program. To further evidence the differences in what is considered a medically actionable gene that should be offered to individuals, in the research context, there are groups that have identified up to 168 medically actionable genes.12 The way certain issues are managed in a research context often influences how they are later managed in clinical practice. Thus, it will be important to be attentive to how projects such as the All of Us Research Program manage secondary findings. All of Us plans to recruit 1 million individuals in the United States and to offer to return medically actionable findings, but the program has not determined which findings it will analyze and make available to participants; what kind of opt out alternatives, if any; or how it will return findings.13,14 Furthermore the US National Academies of Science, Engineering, and Medicine (NASEM) recently published a report in which it recommends that researchers carefully consider returning findings, particularly medically actionable findings, to individual research participants.15 Thus, in the United States, the trend toward offering to return medically actionable findings as secondary or primary targets is on the rise in both the research and clinical settings. On the other hand, there is much more uncertainty about what is considered a medically actionable gene, what criteria are used to determine this, and how these criteria should be applied. Since secondary findings offered in clinical care are generally medically actionable genes, it will be important to follow how this debate progresses and for groups that offer secondary findings to be transparent about their selection criteria.

2.2 ­United Kingdom The United Kingdom’s major national genomic sequencing program is the 100,000 Genomes Project. The project was launched in 2013 by then Prime Minister David Cameron as a way of operationalizing recommendations made in a previous UK government white paper on genetics.16,17 Genomics England, a wholly owned company founded by the UK Department of Health, was charged with running the project, which began recruiting through Genomic Medicine Centres located in the

2 ­ International approaches to genomics

United  Kingdom’s National Health Service (NHS) in 2015. This heralded a concerted focus on genomic medicine within the NHS, seeking to exploit existing academic expertise and establish the United Kingdom at the forefront of the field and to boost the country’s commercial science industry. The project aimed to sequence 100,000 genomes by WGS from around 70,000 NHS patients and their families with a rare disease or adults who have cancer and the less widely discussed or transparent infectious disease arm led by Public Health England. Rare disease was chosen in an attempt to reduce the diagnostic odyssey that many patients with rare disease experience, and cancer was picked to stimulate the development of patient-specific medications by members of the Genetics Expert Network for Enterprises (GENE) consortium. This group of commercial partners have contributed financially toward the project and will have access to patient data to facilitate drug development. The project is described as “clinical transformation,” and the varied components situate it somewhere between research and routine clinical practice. While this has the benefit of catalyzing the use of this technology in the health service, it also poses challenges in encouraging clinician and patient uptake. Primary findings are fed back to patients through standard NHS pathways and were initially based on a gene-panel (via PanelApp, a crowdsourced tool allowing curation and review of condition-specific gene panels) approach, with pathogenic variants in known disease-causing genes reported. Where this is not successful, the relevant Genomics England Clinical Interpretation Partnership (GeCIP) will be tasked with finding an answer to the clinical question. GeCIPs provide access to sequence data for over 2500 academic and public-sector researchers and clinicians organized around particular disease categories or crosscutting topics such as functional genomics and machine learning). The project literature18 makes no mention of incidental findings; however, it does explain the meaning of secondary findings (referred to in the project as additional findings) defined as findings that are actively sought but not related to the condition in question. These are available on an opt-in basis and relate to a small list of serious and clinically actionable genes containing variants responsible for inherited cancer syndromes or cardiovascular disease. This list is not static however, and the patient consents to receive information on all the conditions and genes on the list at the time of reporting, not those present when consenting. In children, this will be limited to information concerning childhood-onset conditions. Carrier status will be reported on a small list of conditions (currently only cystic fibrosis) providing both members of a couple are recruited into the study and both consent to receiving this additional information. X-linked carrier status may be returned if appropriate. These secondary findings are returned to patients separately and at a later date compared with the primary clinical findings relating to their reason for testing. Outside the 100,000 Genomes Project, within the NHS, there are no other examples of secondary findings being routinely sought or returned. Policies concerning truly incidental findings are often ad hoc and vary by laboratory. In this sense the 100,000 Genomes Project policy of returning defined (if not fixed) secondary findings could be seen to provide clarification throughout the NHS, but the lack of

163

164

CHAPTER 8  Implications of secondary findings for clinical contexts

any guidance regarding incidental findings still leaves laboratories, clinicians, and patients unsure as to what constitutes best practice. As genomic medicine becomes more mainstream, a clinician’s obligation to disclose genomic findings becomes complicated. There is no legislation at the current time surrounding the duties of care in genomic sequencing and no legal duty to look for secondary findings or return incidental findings.19 The onus therefore reverts to a duty of care or candor and what is deemed (probably retrospectively if challenged) to be the appropriate standard of care. Both the ESGH20 and the Public Health Genomics (PHG) Foundation, a UKbased health policy think tank in genomics,21 oppose the ACMG position on secondary findings and advocate a targeted approach to sequencing. As well as the difficulties surrounding legislation and best clinical practice, there are also concerns regarding clinical utility of secondary findings in a population context. The majority of knowledge regarding the pathogenicity of variants is derived from the traditional clinical genetics phenotype-to-genotype model where patients with symptoms or a family history of a condition present to the health service. Variants found in these patients can be interpreted in the context of the clinical picture of the proband or their affected family members. This is not the case in a secondary finding context. Patients will usually have no symptoms relating to the variants identified and may have no family history. Population genomic databases such as ExAC and gnomAD have demonstrated that it is possible to carry a pathogenic variant and not exhibit serious disease symptoms, leading to reclassification of certain variants. This could be due to incomplete or age-dependent penetrance or other genetic cofactors.22 Without further data and knowledge of the effects of secondary findings in the wider UK population, it is difficult to predict what the pathogenicity of variants is even in well-characterized disease genes.23 Studies have shown that the disease burden of carrying a “pathogenic” variant in an asymptomatic person is unclear but could be significantly lower than previously suspected. 23,24 This could lead to overdiagnosis of future disease, and those patients may endure unnecessary stress and worry dealing with variants they cannot comprehend and additional and potential risky confirmatory tests or prophylactic treatments they may choose to undertake.6 For this reason a number of commentators, including a Genomic Medicine Multidisciplinary Team (GM-MDT) responsible for local review of genomic sequence data in Oxford, have advocated a “not pathogenic until proven otherwise” approach, where only variants definitively shown to be pathogenic should be fed back as secondary findings.25 However, there is then a risk of a two-tier variant classification for primary and secondary findings, which could lead to confusion for both patients and interpreting clinicians. In summary, within the United Kingdom, the 100,000 Genomes Project is leading the way regarding clinical secondary findings in genomics and is likely to form a template for the ongoing NHS genomic medicine service. The political and medical importance of the project was reinforced in 2016 as the Chief Medical Officer chose to base her annual report on “Generation Genome.” The report highlighted the potential of genomic medicine to aid the NHS in areas such as diagnostics, personalized medicine, drug discovery, and disease prevention.26 More recently, in spring

2 ­ International approaches to genomics

2018, the UK Parliament outlined a strategic approach to embed a genomic medicine service in the NHS, realizing the 2012 vision of transforming patient care through genetic and genomic services.27 Seven national genomic hubs will provide genome testing for various conditions and were planned to be operational by January 2019.28 However, more population-level data associated with clinical outcomes (including from the 100,00 Genomes Project itself) will be needed to help appropriately interpret genomic variants and guide counseling to patients.

2.3 ­Australia In Australia, genomic sequencing tests are making a rapid transition from research to clinical practice. Australia’s component states and territories are responsible for the delivery of public hospital services, which include genetic services. There is therefore a risk of divergent clinical practices arising in potentially contentious areas such as the return of secondary or incidental findings. National bodies, however, have not yet set policy, which directly addresses secondary results. For the purposes of this section, the term “secondary findings” is used to mean result from an analysis of the data/interrogation of genes that are not indicated by the patient’s clinical presentation (e.g., ACMG list genes in a patient with intellectual disability). In 2015 the Australian National Health and Medical Research Council produced a “decision tree” for the management of findings from research and healthcare within their Principles for the translation of omics-based tests from discovery to health care.29 Although recognizing the potential for the identification and/or return of incidental and secondary findings, it simply advises following policy, patient preferences, or national genomics guidelines (which do not yet exist on this issue). The Human Genetics Society of Australasia has produced a commentary on the ACMG guidelines,30 stating that they represent a significant change from previously accepted guidelines on genetic testing and advising exercising caution in adoption. The Royal College of Pathologists of Australasia recommends targeted analysis, that is, only analyzing those genes relevant to the clinical indication for testing, noting that it is a pragmatic approach to minimize the ethical dilemmas arising from “incidental” findings.31 They also observe that, while debate and guideline development are in progress, different practices are emerging and each laboratory should provide a clear verbal and written communication of their policy. In the Australian State of Victoria, the Melbourne Genomics Health Alliance was formed to integrate genomics into clinical practice across its member organizations.32 This currently encompasses five hospitals, each with a genetic service, and five accredited laboratory services. Initial consultations in 2013–14 indicated that secondary findings were a concern to some medical specialists.33 Stakeholder interviews conducted in 2015 showed that no consensus on secondary findings had emerged.33 The community advisory group, clinical and laboratory stakeholders, and ethics committees supported an approach whereby analysis and interpretation are restricted to the genes indicated by the patient’s clinical presentation. Analysis and interpretation of the ACMG genes are deliberately blocked, unless a specific gene is clinically indicated.

165

166

CHAPTER 8  Implications of secondary findings for clinical contexts

The rationale for excluding the ACMG gene list during the early implementation of genomics in practice was as follows: Firstly, patients needed only consider the implications of their diagnostic test in pretest counseling. Secondly, clinicians who were not expert in genetics or with little experience in genomics did not have the additional burden of discussing secondary findings prior to testing or managing any results afterward. In addition, laboratory workload was minimized and prioritized results of most immediate clinical use. Publicly funded healthcare systems, such as Australia’s, are constrained. Any increased use of resources—be they personnel time, infrastructure, or consumables— has an opportunity cost. That is, those resources are not then available to provide another health service. The ratio of the cost of health intervention to the benefit it produces must be shown to warrant the allocation of resources. In Australia, national government reimbursement of new medical services and tests is largely dependent on a health technology assessment (HTA) of this nature.34 American laboratories report that the ACMG guidance has meant that they have had to incorporate a new workflow35, but the costs and the benefits of testing have not yet been measured. Australian evidence of both the costs and benefits is also lacking. The Melbourne Genomics Health Alliance is now assessing a sequential model for offering patient secondary findings.36 Adults are offered a secondary finding analysis after the clinically indicated WE/GS results are available. This is possible since the current guidelines suggest that genomic data arising from clinical testing should be stored31 meaning the data are available for reanalysis. The Melbourne Genomics Health Alliance refers to these as “additional findings,” in concordance with patient preferences2 and to emphasize that it is “extra” information arising from an additional analysis. Testing is offered with pre- and posttest counseling. If the patient accepts, their stored genomic data are reanalyzed for genes predictive of actionable adult-onset conditions. These genes all have defined clinical management pathways, which are publicly funded in the Victorian health system. Evaluation of this model is designed to inform HTA and clinical service delivery decisions by capturing costs and assessing the process of service provision.36 This evaluation does not provide the information needed to determine if secondary findings should be available but is essential to inform decisions regarding how they can be offered and returned. There are at least two anticipated advantages of a sequential model for secondary findings. Firstly, if people wish to learn secondary findings, they can choose the point in their life that best suits them. This may not be when they are seeking a diagnosis or making clinical management decisions. Secondly, they have an opportunity to fully consider the implications of the secondary findings before proceeding. Genetic counselors and patients have reported that scant attention is paid to these when diagnostic testing and secondary findings are offered concurrently.37,38 The concurrent model of testing—whereby consent for both diagnostic testing and secondary findings are sought at the same time—is also being tested with a cohort of infants diagnosed with congenital deafness through the Victorian Infant Hearing Screening Program. Parents who consent to genomic sequencing of their infant are offered three alternatives: to restrict analysis to genes known to cause ­deafness, to

2 ­ International approaches to genomics

also include analysis of genes known to cause childhood-onset conditions with treatment pathways, or to include all genes known to cause childhood-onset conditions. Parents’ choice and experience are being captured through evaluation surveys and interviews.39 An international study reporting laboratory practices found that none of the participating Australasian laboratories in any state searched for secondary findings at that time.40 There were, however, some differences in practice relating to the return of unsolicited findings, that is, potentially disease-causing variants inadvertently identified in genes unrelated to the original rationale for testing. In May 2018 the Australian Government announced a AUD$500M investment over 10 years in genomics research to set a robust foundation for genomic healthcare. With funding from the National Health and Medical Research Council, a collaboration of more than 80 organizations, Australian Genomics, is conducting a program of work to develop tools (such as a national clinical consent form) and provide evidence for the equitable, effective, and sustainable delivery of genomic medicine in healthcare. Australian Genomics aims to harmonize approaches nationally by strengthening networks between clinicians, researchers, and diagnostic geneticists from its participating organizations across all Australian States.41 At the level of government, national consistency will be achieved through implementation of the National Health Genomics Policy Framework by the state and federal jurisdictions.42

2.4 ­Germany The landscape of genomics in Germany is rather heterogeneous. Whereas panel sequencing is part of routine clinical practice, WGS and WES do not yet form part of routine clinical diagnostics and are mainly employed in basic and translational research settings. This is reflected by the status of WGS and WES within the system of reimbursement for clinical care services in German public healthcare. The public reimbursement system does not encompass WGS and WES as reimbursable standard healthcare diagnostics. NGS panels up to 25 kb or designated as basic diagnostics are covered by the statutory health insurance system. Reimbursement of larger panels and extended diagnostics using WGS or WES is available only within special programs; in general, reimbursement has to be applied for individually through statutory health insurance.43 Clinical use of WGS and WES is usually driven by research-­ oriented interests and initiatives or used within clinical contexts, which are traditionally close to research such as pediatric genetics (e.g., diagnosis of rare or unknown developmental diseases). A growing number of interdisciplinary molecular tumor boards are also being established in academic oncology centers to apply NGS analyses to the diagnosis of certain cancer patients.44 There is no national program to promote genomic research and its implementation into the clinic comparable with the UK 100,000 Genomes Project. However, several smaller public research programs fund genomic approaches. One example is the e:Med program sponsored by German Ministry of Education and Research (BMBF), which aims to establish systems medicine in Germany. Another, the German Medical

167

168

CHAPTER 8  Implications of secondary findings for clinical contexts

Informatics Initiative (MII), aims to establish medical data integration centers to allow for integrated analyses and research uses of clinical data from all German university hospitals, including genetic/genomic data from routine care and, ultimately, genomic data from research contexts as well.45 One of the institutional hubs in Germany for the paradigmatic “omics” field of cancer research is Heidelberg, with three institutions (German Cancer Research Centre, the European Molecular Biology Laboratory, and the University Hospital) engaging in joint genomic activities. It is no coincidence that one of the most detailed guidelines for genomic sequencing was issued by the Heidelberg-based EURAT Group (see in the succeeding text). The German Gene Diagnostics Act (Gendiagnostikgesetz, 2009) does not explicitly mention or address secondary or incidental findings. However, the parliament’s official motivation for the act refers to “unexpected results” from genetic analyses (“unerwartete Untersuchungsergebnisse” 27) and to “excess (or surplus) information” (“Überschussinformationen”).46 Such unexpected results are considered to result from methods that generate information on genetic characteristics of patients beyond the medical or clinical scope of the genetic analysis. The official motivation document of the act states that during the consent process patients have to be informed about the possibility of “excess information” and “unexpected results” and should be able to decide whether they want to have “unexpected” results returned to them or not. The act itself provides that the return of any genetic results to patients must be performed by physicians. The act thus effectively determines legal provisions within the clinical context that apply to incidental findings and that would also apply to secondary findings. It does not apply to the research context, leaving many unresolved questions concerning incidental and secondary findings in research genomics. Several prominent German scientific organizations have published statements or recommendations concerning genetics and genomics. In 2013 the BerlinBrandenburgische Akademie der Wissenschaften issued an “ad hoc statement on the consequences of new sequencing technologies for genetics in the clinic.”47 The statement uses the term “excess information” (“Überschussinformation”) but does not discuss secondary findings or incidental findings. The same year the German Ethics Council (DER) published a statement on “the future of genetic diagnostics.”48 As indicated by the title, the statement mainly refers to the clinical context. The DER statement does not explicitly refer to secondary or incidental findings, but refers to “excess information” and “additional result” (“Nebenbefund,” literally “beside result”). The statement defines “excess information” as information generated through genetic analyses, which is not needed for answering the clinical question of the analysis or which occurs unexpectedly or undesirably. “Additional results” are defined as results generated from “excess information,” which are beyond the medical goal of a directed genetic analysis. As to practical handling the statement mainly states that excess information and additional findings are likely to increase in the future and that pertinent ethical and legal challenges, relating, for instance, to the information and consent process, need to be addressed especially with respect to WGS and WES in the research context.

2 ­ International approaches to genomics

The German Society of Human Genetics (GfH) issued a “statement on genetic additional findings in diagnostics and research” in 2013.49 As indicated by the title, the statement focuses on “additional findings” and chooses a differentiated approach, considering the clinical context and research context separately. The authors declare the English term “incidental finding” does not completely suit their conceptual needs, and they prefer the German term “Zusatzbefund” (which might be best translated as “additional finding”). They define “additional findings” as findings unrelated to the initial investigation or question, but which nonetheless have relevance for the health and/or reproductive plans of the person herself and/or her relatives. The term “secondary finding” is not mentioned. In the clinical context the authors state (in accordance with paragraph 9 of the Gene Diagnostics Acts) that consent must include information on the possible occurrence of additional findings and whether additional findings are to be communicated to the patient. Data security, the patient’s right not to know, and the protection of people incapable of giving informed consent must also be considered. In the research context, by contrast, the report finds neither a duty to establish a diagnosis nor an obligation to report additional findings. If additional findings are likely to occur, the possibility and handling of additional findings must be clearly addressed during the consent process. If there is an agreement with the patient/participant to return additional findings, it is necessary to define a time period in which the information will be given. Where patients and participants agree to the return of incidental findings, it is necessary to clarify which kind of results will be reported. Results have to be reliable and scientifically validated and are categorized as follows: 1. Genetic characteristics bearing a significant risk to develop a specific disease, which can be treated effectively or preventively. 2. Genetic characteristics bearing a significant risk to develop a specific disease, which cannot currently be treated. 3. Genetic characteristics bearing a slightly modified risk to develop a certain disease. 4. Genetic characteristics bearing no health-related risk for the patient/ participant—but which are heritable and may have consequences for reproduction. Following the statement, information on significant health-related risks that can be treated effectively or preventively (category 1) should be passed to the patient. Context-related decisions have to be made in case of other findings. The statement summarizes that communicating risks of diseases lacking treatments (category 2) and characteristics bearing a slightly modified risk to develop a particular disease (category 3) are of limited value. A 2016 statement by the German Research Foundation (DFG) on “human genome sequencing challenges for a responsible application in the sciences” recognized the value of this categorization of additional findings and the attendant recommendations. The Heidelberg-based interdisciplinary EURAT Group (Ethical and Legal Aspects of WGS) published a “position paper: cornerstones for an ethically and

169

170

CHAPTER 8  Implications of secondary findings for clinical contexts

legally informed practice of whole genome sequencing,” initially in 2013 with a revised version in 2016.50 The position paper contains a code of conduct for researchers (as distinct from treating physicians) involved in genomic research, as well as two participant consent templates. The position paper distinguishes two sorts of research findings: findings of individual health relevance for the patient that pertain directly to the sequencing request and are within the scope of actual investigation (“primary findings”) and “additional findings” that are findings of health relevance that are nonintended and are beyond the scope of the original investigation. Echoing previous debates the authors find the term “additional finding” more appropriate than “incidental finding” since, even though it refers to unsought findings, researchers can generally expect to encounter findings of individual health relevance for patients when engaging in NGS analyses, so that the findings do not truly occur “incidentally” even though they are nonintended. The EURAT position paper makes the following recommendations: 1. As NGS techniques used in research may not be approved for clinical care, all findings from research (primary findings and additional findings) need to undergo clinical validation before being reported back to the patient via a physician. 2. Primary findings should always be returned if they can be combined with existing knowledge and methods to provide treatment and care measures tailored to the patient’s specific condition and needs and if the patient has consented. 3. Reporting of additional findings should be part of the informed consent process. The patient has the right and burden of deciding whether or not additional findings are to be reported and, if so, which findings they want to know. 4. The physician in charge must decide whether routine laboratory diagnostics will be performed to validate the findings and, depending on the results, communicate them to the patient. 5. The researcher has a duty of care to notify the responsible physician of all primary and additional findings that have been recognized as medically relevant for the patient, if and only if the researcher has the awareness that, in the absence of this knowledge, the patient would be subject to additional harm or increased suffering and if the patient’s statement of consent does not rule out such reporting. 6. The researcher is not obligated to engage in the active or deliberate search for findings beyond the specified context of a sequencing request. The last element (6) implicitly refers to (the concept of) secondary findings as defined in this handbook (results identified based upon intentional interrogation and beyond the original scope/the patient’s treated clinical condition). The EURAT statement rejects the idea of a moral or legal duty of researchers to actively search for findings of individual health relevance beyond the scope of investigation or disease. The EURAT Group also explicitly rejected two elements proposed by the 2013 ACMG recommendation: (i) That presymptomatic and untreatable findings in pediatric patients be returned and (ii) that prespecified findings be reported back to adult patients with no option to opt out.

2 ­ International approaches to genomics

2.5 ­France In France, genomic sequencing technologies have been used in the context of cancer and rare diseases for a number of years but primarily in the context of clinical trials, research, or clinical research projects. Only a handful of teams are using WES and WGS tests as opposed to gene panels, which are more widespread. Indeed, neither WES nor WGS is listed in the National Biology Table, which lists approved and priced acts to be reimbursed by the publicly funded, universal social security system. Until recently, there were no official French guidelines framing the return of secondary findings. Teams that had been pioneering the clinical use of these tests therefore adapted European51,52 or American guidelines8 to establish return of results protocols that they thought would respect the best interest of patients.53 Striking the right balance between these guidelines is challenging, knowing that the ESHG recommends that measures should be taken to limit risks of secondary or incidental findings as much as possible. Without hunting for a specific list of variants, they would then decide to return only actionable individual findings on a case-by-case basis, after collecting specific consent from patients and families and through a consensual, collegial decision involving biologists, clinical geneticists, medical doctors familiar with the case, and bioinformaticians. However, the situation has recently changed considerably. In 2016 the Agency for Life and Health Sciences (agence nationale pour les sciences de la vie et la santé, Aviesan) published a report establishing a 10-year plan for the realization of genomic medicine in France, entitled “genomic medicine France 2025.”54 The final recommendations stemming from the plan mandate the establishment of 12 sequencing platforms throughout the territory of France, the first two having already been identified in early 2018. The plan sets a target of sequencing over 200,000 genomes by 2025, including cancer and rare disease patients. As part of the plan, a specific WES/ WGS consent form will be established, and a decision will be made on the best way to handle secondary findings at the national level. In addition, the French Society for personalized and precision medicine published recommendations on the return of actionable secondary findings in cancer genes.55 This guideline, which specifically excludes pediatric-onset conditions, lists 36 genes in which variants are recommended to be returned to patients, including quite a few more than the ACMG recommendations. It also recommends a two-step consent, where patients are consented for the return of secondary findings not only at the pretest counseling session but also once the primary results are returned to them, to ensure they have a chance to revise their preferences at this point. It will be instructive to observe the uptake of these guidelines in practice.

2.6 ­Canada (Quebec) In Canada, there are two guidelines that include recommendations on the return of secondary findings to patients or research participants. In 2015 the Canadian College of Medical Geneticists (CCMG) published a position statement intended to “provide recommendations for Canadian medical geneticists, clinical laboratory geneticists,

171

172

CHAPTER 8  Implications of secondary findings for clinical contexts

genetic counselors, and other physicians regarding the use of genome-wide sequencing of germline DNA in the context of clinical genetic diagnosis.”56 While each province and territory in Canada is responsible for determining clinical test reimbursement, this position statement aimed to provide nonbinding guidance to increase consistency in clinical genomic testing offered to patients across Canada. Although the CCMG is aligned with the ESHG in stating that incidental findings should be avoided as much as possible,56 they also value individual laboratories’ autonomy and suggest ways to frame their offer to report such findings. The fundamental document regulating research ethics is the Tri-Council Policy statement on the Ethical Conduct for Research Involving Humans, which stems from the Canadian Institutes of Health Research, the Natural Sciences and Engineering Research Council of Canada, and the Social Sciences and Humanities Research Council of Canada. (TCPS2). This document is not legally binding in Quebec, but it is the gold standard followed by all ethics committees that oversee research in all provinces and territories in Canada. This guideline provides details on the requirements for researchers to establish a clear plan “for managing information revealed through genetic research.” Researchers are free to decide whether to share individual findings with participants or to exclusively disclose “nonidentifiable research results.” In case researchers do decide to return individual results to participants, measures should be taken so that they can (a) make informed choices about whether they wish to receive information about themselves and (b) express preferences about whether information will be shared with biological relatives or others with whom the participants have a family, community, or group relationship. The “right not to know” is specifically mentioned, and researchers are required to detail how the return of results will be organized, whether directly or through a healthcare provider, as long as appropriate genetic counseling options are given to patients when necessary. In 2014 the Quebec Ministry of Health and Social Services (MoHSS) specifically excluded the use of WES and large gene panels in patients, by adding them to the “list of analyses not covered by Quebec medical insurance and not reimbursed in the framework of the authorization and reimbursement mechanism for medical biology analyses not available in Quebec.”57 Teams who wanted to offer those tests to patients were effectively funding them through research projects and observed the TCPS2 guidelines regarding the return of secondary findings. However, in August 2018, the Minister of Health and Social Services Gaetan Barrette announced the creation of the Quebec Center for Clinical Genomics, located at the Ste-Justine hospital in Montreal. The decision for the sequencing platform to be located in Ste-Justine is mainly based on the fact that, in October 2014, they initiated an Integrated Clinical Genomic Centre in Pediatrics58 at the same site. That platform, based on the Illumina 2500 technology, started functioning in the summer of 2015, offering sequencing services to researchers in Ste-Justine and other institutions in the province, and is

2 ­ International approaches to genomics

in the process of obtaining CLIA and ISO certification to be able to offer clinically validated tests. Teams throughout the province will shortly be able to request WES or WGS tests and a standard consent form, including guidance on how to handle secondary findings will be established, most probably using the CCMG guidelines as a starting point.

2.7 ­Singapore During the rise of genomics, individual research centers and clinical units in Singapore pursued their own genomic programs. As a result, practices concerning secondary findings varied depending on institutional capacity, expertise, and ethical judgment. More recently, Singapore’s Ministry of Health has begun exploring the potential of large-scale genomics programs for healthcare, and the National Precision Medicine Alliance has been formed as a ground-up coalition to harmonize existing genomics activities. The development of national regulations and guidelines concerning secondary findings is also providing further standardization, depending on whether they occur in the research or clinical context. Translational genomics research in Singapore is regulated by the 2015 Human Biomedical Research Act.59 Among other provisions the Act requires the informed consent process to disclose whether incidental findings will be returned. “Incidental findings” are broadly defined in the law to include what are known internationally as “secondary findings.” In this section the terms are used interchangeably. Regulations further specify that it is the responsibility of research institutions to establish institutional policies concerning the return of such findings. As of this writing, local institutions and their IRBs are in the process of developing reporting policies. Some further guidance can be found in the Singapore Bioethics Advisory Committee’s 2015 report, “Ethics Guidance for Human Biomedical Research,” which states that there is “some duty” for researchers to report clinically significant incidental findings.60 Nonclinically significant findings are not discussed, nor does the report go into detail concerning appropriate reporting mechanisms. It does emphasize, however, that a reporting plan should be decided prior to commencing a study and that the plan should be disclosed to participants—with the option of not receiving any findings, should researchers choose to make these available. Clinical genomics is regulated separately. Beyond general professional ethics standards governing clinical practice, in June 2018, Singapore’s Ministry of Health issued a code of practice on the standards for clinical and laboratory genetic and genomic testing services.61 The code defines three tiers of practice corresponding to the risk and complexity of the genetic test, with germ line testing (excluding pharmacogenomic associations), occupying the highest tier. At that level, genetic counseling must be provided both before a genetic test is conducted and at results disclosure. The pretest counseling must include discussion of potential incidental findings, detail the particular laboratory’s policies on the matter, and explore the ramifications (clinical, social, and psychological) on the return of results. In addition, the consent process must allow patients to opt in or opt out of receiving such findings. Notably

173

174

CHAPTER 8  Implications of secondary findings for clinical contexts

the guidelines do allow space for a laboratory to decline to report any incidental findings, as long as that policy is made clear to patients during pretest counseling and adhered to later on. No definition of “incidental findings” is given, nor is a standard for generating or reporting such findings (reproductive relevance, clinical relevance, and clinical actionability) provided. Instead, laboratories are advised to consider the clinical actionability and patient consent preferences and document the management of incidental findings in their standard operating procedures. Given that there are limited accredited genetic counselors and clinical geneticists practicing in Singapore, the clinical guidelines stipulate that the pre- and postgenetic counseling can also be performed by a medical practitioner with several years of experience in clinical genetic counseling. One particular risk that any pretest counseling session should discuss is genetic discrimination. Unlike other countries, such as the United States, Canada or the United Kingdom, Singapore currently lacks a legal prohibition or moratorium on genetic discrimination in insurance or employment. While there is a universal health insurance program through MediShield Life, many Singaporeans “top up” with private health insurance along with life or disability insurance policies. Secondary findings, then, pose theoretical but nontrivial risks to Singaporeans: when signing up for an insurance program after having a genetic test, there is no legal prohibition on asking whether the purchaser knows they are a carrier of certain traits affecting disease risk. Likewise, similar issues could be experienced when seeking employment. Unless a moratorium on genetic discrimination is issued in future, the risk of discrimination needs to be disclosed and discussed in the genetic counseling process. Despite the diverging regulatory regimes, there are substantial similarities in the regulations for returning incidental findings in research and clinical settings. In both cases a plan must be in place prior to the test, disclosed to the volunteer/patient, and the option given to opt in or opt out. The type of findings returned, or whether there is a duty to generate secondary findings in the first place is at the discretion of the researcher or clinical laboratory. A substantial point of difference, though, is the pre- and posttest genetic counseling requirements present in clinical testing, but not for research. It may indeed be impracticable to require individual pretest counseling for large-scale genomics research, though, in situations where clinically relevant findings could be generated, compromise solutions like standardized decision aids (videos, interactive apps, and brochures) are worth exploring. Clinical care and research activities are not always clearly delineated, however. For example, the SingHealth Duke-NUS Institute of Precision Medicine (PRISM) was established to transition research participants with significant genomics findings into the healthcare system for ongoing care. PRISM analyses genomic data from volunteers recruited for genomics research. Participants are then given relevant information from genetics specialists and, after counseling and consent, receive clinically validated secondary findings that are transferred to the clinical setting for follow-up care. Integrating genetic research units into clinical practice in this way requires substantial scientific, organizational, and infrastructural investment but may become more feasible as local capacity in clinical genetics and genomic research grows.

2 ­ International approaches to genomics

2.8 ­Estonia In Estonia, research in the field of genomics and its transfer into the healthcare system has been closely intertwined with the developments of the Estonian Biobank, which is held by the Estonian Genome Centre, University of Tartu. The population biobank was founded in early 2000 with the long-term goal of benefiting the health of the public. Between 2002 and 2011, close to 52,000 participants were recruited, and by now, all of the subjects in the biobank have been genotyped. In addition, the genomes of 2600 and exomes of 2500 participants have been sequenced. Although the primary goal of the biobank was research,62 the tremendous amount of genetic data that has been produced raises the question of incidental findings. As there are no national guidelines concerning secondary or incidental findings in population biobank settings, the practice of the Estonian Biobank is based on the Estonian Human Genes Research Act, which regulates its work. This legislation states that the biobank participants have the right to know results and the right to genetic counseling.63 Currently, results are only offered on a project-based manner through the biobank.64 For example, participants detected to carry a pathogenic finding associated with hereditary breast and ovarian cancer or familial hypercholesterolemia have been recontacted and invited to participate in a project where results are returned upon consent.65 A second blood sample is taken to confirm the results, and validated findings are returned to the participants during a face-to-face genetic counseling session. Carriers are referred to clinical specialist for follow-up and encouraged to contact their first- and second-degree relatives to introduce the option of cascade screening. The response, emotions, and general opinion of the participants are surveyed prior to and after the counseling session to inform future projects. As of September 2018, over a thousand individuals have received individual results and counseling at the biobank. Clinical exome sequencing was first implemented in 2011 through a research collaboration between the Estonian Genome Centre and the children’s hospitals of Tallinn and Tartu. The service was recognized as a diagnostic analysis by the Estonian Health Insurance Fund in 2014. The indication for exome sequencing covered by the health insurance fund is “an undiagnosed disease of suspected genetic etiology, where more specific genetic analysis is not available and difficult to specify due to genetic heterogeneity or has been negative.” The informed consent form for clinical exome sequencing mentions a small possibility of clinically significant findings unrelated to the indication for testing. When consenting for exome sequencing, one can decide to opt in or opt out of being informed of incidental findings. The consent form refers to the ACMG recommendations for reporting incidental findings and mentioning the option of other genetic findings that could be of significant importance to the patient or relatives.66 Based on these guidelines the expert board of the Department of Clinical Genetics of Tartu University Hospital decide which findings are reported. In 2015 a national personalized medicine pilot project was initiated based on the Estonian Biobank.67 The initiative foresees incorporating the genotype data of biobank participants into the national health information system to allow better targeted methods for healthcare and disease prevention. The approach involves targeting rare

175

176

CHAPTER 8  Implications of secondary findings for clinical contexts

disease-causing mutations (e.g., BRCA1/2, LDLR, APOB, and PCSK9), polygenic risk scores for common diseases (coronary artery disease and type 2 diabetes), and pharmacogenetics. As part of the pilot phase of the initiative, clinical flagship projects have been launched involving personalized risk prediction and treatment of breast cancer and cardiovascular disease in clinical settings. In 2018 the government allocated funding for the recruitment of an additional 100,000 biobank participants.68 Within the first 5 months, over 52,000 individuals signed up and consented to participate in the biobank project.

2.9 ­Japan In the late 1990s formal discussions about ethical issues regarding human genome research at the government level started in Japan. At that time the Human Genome Project, of which Japan was a member, was making a rapid progress. At the same time, large genomics research programs were being promoted by the government in Japan. With this background the Japanese government set up the Bioethics Committee in the Council for Science and Technology to lay out national frameworks to deal with ethical and social issue of genomics research.69 In 2000 the committee issued a report on “The Fundamental Principles of Research on the Human Genome.” In this document, both the right to be informed and right not to be informed were stated. Based on this document, new government guidelines for genomics research, “Ethical Guidelines for Human Genome/Gene Analysis Research (Ethical Guidelines)” were established in 2001.70 Although the Ethical Guidelines are not legally binding, they have been serving as guiding regulations across the country for more than 15 years. The Ethical Guidelines not only have a section describing the right of research participants to be informed of the research results but also have stipulations on the exemptions from being informed of results. The latter includes when disclosing the results is judged likely to harm the life, body, properties, and other rights and interests of research participants. Until around 2015 many of the basic genomics research projects utilizing human samples did not adopt a policy of disclosing even primary results from genomic analysis. The main reason was because the results of the genomic analyses did not produce clinically significant data. This situation began to change rapidly around 2015. There were several important events behind the change. First, new laws were enacted to promote and revitalize the Japanese policy for medical research in the face of an aging society. These were the Act on the Promotion of Healthcare Policy and the Act on the Independent Administrative Agency of Japan Agency for Medical Research and Development (AMED). As a result the Headquarters for Healthcare Policy (HHP) was established in the Cabinet in 2014.71 One of the committees created within the HHP, the Council for Realization of Genomic Medicine (CRGM), began to steer the national policy for genomic medicine. A key report on the genomic medicine published by the CRGM in July 2015 contained a section on hurdles for implementation of genomic medicine in clinical settings. Within the section, it was stated that incidental findings were a challenging issue and requested specific government funded projects to work on it.

2 ­ International approaches to genomics

The second important event was the establishment of a new funding agency AMED in 2015 that coordinates funding activities for medical and translational research across the ministries. It began to promote genomics research projects that were more relevant for clinical medicine. Even projects aiming at basic understanding of mechanisms of diseases, researchers were required to explain potential clinical outcomes of the research. The first large project that worked on the issue of incidental/secondary findings was funded by the AMED and led by Dr. Hitoshi Nakagama, the director of the Research Institute of the National Cancer Center. It aimed to carry out large-scale clinical sequencing in several disease areas including cancer, neuromuscular diseases, and cardiovascular diseases, although in the strict sense the sequencing was still research stage. The project contained a research group for ethical issues with the author of this section, Professor Kato of Osaka University, as one of its members. The group conducted literature surveys and held meetings with genomic researchers who were trying to implement clinical sequencing. It became clear that, in some of the cancer areas, additional findings were being returned to sample donors, while researchers were not returning those findings in noncancer areas. The final report published in March 2017 included five key areas of recommendations. They include the importance of providing genetic counseling; long-term follow-up of patients and families; and capacity building of relevant specialists such as data scientists, genetic counselors, and ethicists. Establishment of measures to prevent genetic discrimination was also mentioned.72 Another research project is a large national genome cohort, the Tohoku Medical Megabank, which was set up after the Tohoku earthquake.73 It was set up by the government as one of the projects for restoration of the area and has collected samples and data from about 150,000 healthy people. The Megabank also completed whole genome analysis of several thousand samples. Utilizing the results, they have started to return the genomic analysis results of familial hypercholesterolemia. The two latest projects that are working on the issue of additional findings are also funded by the AMED. One of the projects, led by Professor Shinji Kosugi, is working on the establishment of a practical policy for returning genomic results in the clinical setting. In this project the term “secondary finding” is used for germ line variants of inherited diseases (mostly cancer) that could be found during the analysis of cancer tissues and blood samples using gene panels. The first draft was published in March 2018 and describes necessary issues for consideration at the stages of pretesting explanations to participants, informed consent, and disclosure of results. For example, the draft report recommends that, at the time of informed consent, patients are given an explanation of the possibility of secondary findings and the opportunity to state their preferences of receiving the results. Another new project, known as the Leader Project (leading the way for genomic medicine) is looking into the requirements when researchers have to deal with the secondary findings in the research setting. Also funded by AMED the project team plans to release a final report by March 2019. One further recent activity is worth mentioning. Japan began to perform “cancer genomic medicine,” as a part of the government initiatives to implement genomic

177

178

CHAPTER 8  Implications of secondary findings for clinical contexts

medicine begun in 2017. Using gene panels to analyze cancer tissues, 11 designated hospitals and approximately 100 associated hospitals will perform cancer treatment. It is of note that none of the above activities has published a list of genes that are the targets of secondary findings. In conclusion a dramatic change has happened in terms of the policy concerning secondary findings in genomic medicine in Japan over the last several years. A remaining challenge is to further elaborate the policy, particularly for the clinical settings. Deciding which genes (variants) need to be considered as targets of returning as secondary findings is an issue that still requires further work. It will also be necessary to work with patients to incorporate their perspectives. Effective communication and collaborations among key stakeholders including medical professionals, policy makers, and patients will be the key to maximizing the benefit for the society.

2.10 ­Summary The aforementioned accounts demonstrate there is no single, straightforward method for implementing genome sequencing as a routine clinical procedure. As seen in almost all the national case studies, the initial development of medical genomics has taken place in research settings and subsequently moved into clinical practice. There are important differences between the research and clinic settings. Medical research is understood as experimentation designed to produce generalizable knowledge. Medical ethicists over the years have emphasized the importance of ensuring that patients taking part in medical research do so on the clear understanding that they should not expect any therapeutic or diagnostic benefit from participation (socalled therapeutic or diagnostic misconceptions). Return of findings to participants is rarely expected or even considered suitable as their clinical significance is often unknown. The regulation of medical research, through national and international laws and mechanisms such as Institutional Review Boards or Hospital Research Ethics Committees, tends to focus on protection of human participants from undue harm by managing the risks to which they are exposed. Research projects also tend to have a fixed time limit and budget. In clinical care, by contrast, the injunction to “first do no harm” is only a minimum basic requirement, and the oversight of care is orientated toward securing the best possible outcome for the individual patient, which usually means securing evidence-based diagnosis and treatment. This is established in legal requirements such as the physician’s duty of care to the patient, the rules for establishing medical negligence, and the concept of an established “standard of care” that physicians and healthcare institutions are obligated to provide. Healthcare systems also operate in a different financial setting where costs must be met on an ongoing basis, either from a constrained budget in public healthcare settings or to the reimbursement policies of different payers (mainly insurance companies and employers) in a private system. In either case, there are well-established processes of HTA designed to assess which new processes and interventions represent value for money. The very idea of “translational” research was coined to recognize that the transition from research to clinical practice is complicated and requires work, resources,

2 ­ International approaches to genomics

and time.74 Translation is about more than simply developing a new machine or procedure. It requires finding ways to make that technology work in the wider context of a particular healthcare system so that it can move from a successful prototype to a widely used routine process.75 Genomics presents an “excess” of data, beyond what is needed for most diagnostic applications. Although unsought findings are not unique to WGS, the phenomenon meant that there was uncertainty about how to manage this “disruptive” feature; for example, in the United Kingdom, it remains unclear whether the established legal duty of care extends to returning additional or incidental findings in genomics to patients (and indeed potentially to their families), and if so, what should be returned. In the early stages of translation, individual clinics and hospitals must often make decisions about what should be done under conditions of uncertainty and a lack of centralized guidance. This means there is a risk of divergent standards of care emerging within jurisdictions, something widely seen as undesirable. When a new technology like WGS enters the clinical realm, there is rarely economic evidence for benefit or a clear existing standard of care and seldom time for new legislation to be passed to stipulate exactly how the technology should be used. Instead, there are more likely to be “inherited regulations”76 devised for earlier technologies, in this case genetic testing, and existing institutions and professional bodies that act as an authoritative reference point. In this context the ACMG recommendations that (only) a specific subset of genomic variants should be examined for all clinical sequencing of patients can be understood as an attempt to generate a standard of care that would deal with the problem of unsolicited findings and allow the clinical implementation of genomics to proceed. At the same time, it is clear from the studies of different territories presented earlier that the ACMG list of secondary findings has not been universally accepted (see Table 2 for a summary). This should not be surprising. The ACMG approach of a defined, if flexible, list of specific genomic variants to be designated “secondary findings” was developed in the specific setting of US healthcare. Different territories have different legal and regulatory systems, different institutions and economic underpinnings of their healthcare systems, and different cultural preferences and attitudes to topics such as risk and even defining clinical benefit. It is therefore hardly surprising that different countries and different regions within federated territories took divergent stances on secondary findings as a solution to the problem of unsolicited genomic results. Although the ACMG list of secondary findings does not act as a universal standard of care, even in the United States, they did succeed in driving forward the international debate, by prompting responses about how best to proceed, in each jurisdiction. Even though the outcomes are distinctive, there is a remarkable similarity in the mode of responses in different jurisdictions. Only Germany has a statutory provision that relates to secondary or incidental findings, and the 2009 German Gene Diagnostics Act does not actually use those terms at all, but introduces its own lexicon of “unexpected results” arising from “surplus information.” However, most territories deployed some form of nonbinding but centralized guidelines on what best practice might look like for clinicians. These were mainly issued by existing professional bodies such as the ACMG, the ESHG, Canadian College of

179

Table 2  Overview of stances on secondary and additional findings in national policies, guidelines, and translational genomic projects. Legislation concerning secondary or additional findings

Guidelines on secondary findings for clinical practice

The United States

No

ACMG guidelines

The United Kingdom

No

ESHG guidelines Public Health Genomics Foundation recommendation

Australia

No

Germany

German Gene Diagnostics Act (Gendiagnostikgesetz, 2009) states that potential for “unexpected findings” must be discussed with patients No

Territory

France

Stance on secondary findings

Major translational project

Stance on secondary findings

Recommend return of ACMG list of secondary findings Advise against return of secondary or incidental findings

All of Us precision medicine cohort

Policy in development

100,000 Genomes Project

NHMRC “Principles” HGSA commentary on ACMG Guidelines Royal College of Pathologists of Australasia recommendations

Advise caution and advocate targeted use of genomics

National: Australian Genomics State specific: Melbourne Genomics Health Alliance for the state of Victoria

German Ethics Council statement on “the future of genetic diagnostics”

Patients must have right to refuse additional findings. Only additional findings indicating significant risk of treatable or preventable disease should be offered for return

Closest equivalent: BFMB e:Med program; German Medical Informatics Initiative (MII)

French Society for Personalized and Precision Medicine recommendations

Recommend return of expanded list of secondary findings (more variants than on ACMG list)

Aviesan network of 12 regional sequencing platforms

Return project specific list of additional findings on an opt-in basis Pilot return of project specific list of additional findings on an opt-in basis Australian Genomics policy is in development Not clear, but see Section 2.4 for guidelines from German Society of Human Genetics (GfH), German Ethics Council and EURAT group Policy in development

Canada (Quebec)

No

CCMG recommendations

Singapore

Ministry of Health Code of Practice

Estonia

Human Biomedical Research Act (2015)—only applies to research not to clinical practice No

Japan

No

Ethical Guidelines for Human Genome/Gene Analysis Research (2001) AMED-funded projects to develop policy on incidental and secondary findings

Estonian Genome Centre policy Expert board of the Department of Clinical Genetics of Tartu University Hospital

Avoid incidental findings as much as possible but respect autonomy of individual labs and clinics to set own policy Laboratories have some scope to set own policy on returning incidental findings provided pretest genetic counseling is provided Return of findings relating to selected cancers and familial hypercholesterolemia to biobank participants Return of incidental findings as assessed by the expert board, for clinical genomics collaboration between the Estonian Genome Centre and children hospitals of Tallinn and Tartu Research participants have not only right to findings but also a right to opt out. In practice, most findings were not returned New reports in 2017 and 2018 permit return of secondary findings with appropriate consent and counseling

Regional: Quebec Centre for Clinical Genomics

Policy in development

National Precision Medicine Alliance PRISM

National Policy in development PRISM returns project specific list of additional findings Pilot targets specific predefined variants associated with disease risk

2015 National personalized medicine pilot of the Estonian National Biobank

Tohoku Medical Megabank

Return of findings relevant to hypercholesterolemia only

182

CHAPTER 8  Implications of secondary findings for clinical contexts

Medical Genetics, German Society of Human Genetics, or the French Society for Personalized and Precision Medicine. Again, it is not surprising that the majority of these bodies were originally concerned with the clinical provision of genetic services, as the obvious antecedent to clinical genomics. Other entities providing guidance were national bodies concerned with health research—as the originating point of medical genomics—such as the Australian National Health and Medical Research Council, AMED and the Council for Realization of Genomic Medicine (CRGM) in Japan, the German Ethics Council and the Berlin-Brandenburgische Akademie der Wissenschaften, Singapore’s Ministry of Health, or the Estonian Genome Centre. The latter example is also a reminder of the importance of projects, especially state-supported national projects, to build translational infrastructure such as biobanks to support clinical genomics.71 Most of the examples considered earlier have some sort of large-scale nationally supported translational genomics venture. Some such as the 100,000 Genomes Project, Genomics Australia, the Aviesan plan for a network of 12 genome sequencing centers in France, or the Quebec Centre for Clinical Genomics are new endeavors. Others build on existing infrastructure like the Estonian Biobank, the Tohoku Medical Megabank, or the recent announcement that the UK government plans to extend the 100,000 Genomes Project by sequencing a further 500,000 genomes from existing samples contained in the UK National Biobank. In both cases, these translational projects provide a space where clinical implementation of genomics, including returning results, can be tested and evidence of what works or does not work generated and evaluated. There are also sites where future national policies on secondary or additional findings are likely to be formulated.

3 ­Emerging and future scenarios This section looks at a number of areas where additional or secondary findings raise new issues or require a different perspective, in comparison with clinical genomics for adult patients with known diseases. The areas covered are pediatric and prenatal applications of genomics, nutritional and “wellness” genomic services, and the storage and return of sequence data to patients and research participants. Although disparate, each area represents an extension to or, in the case of wellness genomics a spillover from, adult clinical practice. It has been noted that children challenge classifications and standards build around adults because they must be considered both a group in their own right and in terms of the future adults they will become.77 In medicine, children are recognized to require special rules and procedures in domains from giving consent to participating in clinical trials, and returning secondary or additional findings from genomic sequencing in children also raises novel concerns about variants associated with adult- versus childhood-onset conditions. Wellness genomics goes beyond what is usually regarded as “clinical” both not only in the sense that wellness genomic tests provide advice on lifestyle, diet, and other social activities like exercise but also in the way that the online marketing of many w ­ ellness genomic

3 ­ Emerging and future scenarios

services bypasses traditional clinical institutions and the regulations by which they are governed. Finally, return of sequence data is not a novel application of genomics, but rather an additional concern raised by the storage and digital portability of genomic data produced for the clinical applications discussed elsewhere in this chapter. Again, it is a domain where issues of secondary findings inform the broader discussion of what best practice looks like.

3.1 ­Pediatric, neonatal, and prenatal genomics The majority of policies on return of secondary or additional findings described earlier relate primarily to the clinical use of NGS techniques in adults. However, WGS WES and targeted gene panel testing are increasingly being used or considered for use, in pediatric populations.78 The application of genome sequencing in pediatric populations involves additional concerns beyond those discussed for adult populations in terms of how to manage secondary and additional findings. At present the most frequent pediatric use of WES and WGS is in children with rare diseases.79 These tend to be children with highly variable symptoms, such as intellectual disability, developmental delay or congenital malformation, and where there is frequently genetic heterogeneity between cases. Here, sequencing is used primarily as an aid to diagnosis. The meaning of sequence variants is interpreted in light of the patient’s symptoms, as described in more detail in Section 2.2. An example of a possible secondary finding in this scenario, proposed by Klugman and Dolan,80 would be finding a mutation in the BRCA1 gene conveying increased lifetime risk of breast and ovarian cancer, in a 2-year-old child being evaluated for developmental delay using WGS. Returning this result requires consent of the child’s legal guardians. Where these are the genetic parents of the affected child the result also has implications for them and for any related siblings of the affected child. One or both parents may be a carrier of the detected mutation as might an asymptomatic sibling. As with adult patients the implications of genomic sequencing and the potential for findings to affect both patient and direct relatives need to be discussed beforehand so that adequate consent can be given. WGS could also potentially be used to carry out population-wide screening of children for a range of conditions. The most common vision is that this would be integrated into existing public health programs aimed at newborn children. Although not currently practiced the idea of neonatal WGS has been suggested by several prominent sources; it was discussed as far back as a 2003 report on the use of genetic services in the UK’s National Health Service81 and has been raised by the Director of the US National Institutes of Health, Francis Collins, on more than one occasion.82 Screening is not diagnosis. As discussed briefly in relation to the Geisinger Health System MyCode Community Health Initiative (Section 2.1) screening involves systematically testing asymptomatic individuals with the aim of detecting conditions or increased risk of disease in individuals, who can then be followed up with further confirmatory diagnosis and treatment. Current neonatal screening polices vary widely between countries in terms of which conditions are tested for and how many

183

184

CHAPTER 8  Implications of secondary findings for clinical contexts

conditions are included in the screening procedures.79 Moreover, ACMG has explicitly stated that its list of recommended clinically actionable variants does not apply to neonatal screening.81 Any use of genome sequencing for neonatal screening will therefore need a policy on what to look for and which findings to return to parents. Commentaries on potential neonatal use of WGS have stressed that the clinical justification for screening is to identify conditions where preventative action or treatments can be applied during early childhood.79,82 Screening for, or returning secondary or additional findings relating to, adult-onset diseases is contrary to that rationale, although there is a further complication where a mutation indicating an adult-onset condition could also have implications if one or both parents are carriers. Another recent development has been widespread global uptake of genome-based noninvasive prenatal testing (gNIPT) for the detection of fetal aneuploidy, based on cell-free fetal DNA fragments found and analyzable in the circulating blood of a pregnant woman.83 These fragments are reliably available for analysis from 10 weeks of pregnancy so that testing can be done relatively early (although where it is necessary to send samples by post to another country this can take up to 2  weeks).84 Cell-free DNA fragments are assumed to originate from cells of the placenta, so they do pose a problem with confined placental mosaicism. Nonetheless a recent systematic review of gNIPT for fetal chromosomal aneuploidy85 found that genomebased testing methods appear to be sensitive and highly specific for detection of fetal trisomies 21, 28, and 13 in high-risk populations and that there is a paucity of data on the accuracy of gNIPT as a first-tier aneuploidy screening test in a population of unselected pregnant women. The authors concluded that, on the basis of their review, which examined 65 studies of 86,139 pregnant women, gNIPT was not sufficient to replace current invasive diagnostic tests and that invasive fetal karyotyping is still the required diagnostic approach to confirm the presence of chromosomal abnormality prior to decision making. However, Lewis et  al.86 have found NIPT for the common aneuploidy trisomy 21 to be acceptable to a vast majority of women surveyed in England. They concluded that uptake of the test is likely to be high and includes women who currently decline screening and those who will use the test for information only. The discovery of the availability of the whole fetal genome in cell-free form in the circulating blood of a pregnant woman84 also raises questions not only about the ethics of testing a fetus prenatally but also about who should hold the data and who is responsible for “recontacting” the future child and/or parents with genomic information, particularly should it be relevant to the health and/or healthcare of the future child. While some work has been done recently on recontacting in Europe,87 this is clearly an area where consensus about appropriate action is lacking and where resources and desires for “best practice” may not coincide. For both neonatal and prenatal WGS or WES, commentators have also raised the concern that returning data about genomics variants whose significance is uncertain or where they relate to adult-onset disorders could result in stress and unwarranted concern at an already stressful time for many parents and could have potentially deleterious qualities for the developing parent-child bond.79,81

3 ­ Emerging and future scenarios

The meaning of genomic variants depends on correlating sequence variants with phenotypic data from patient populations. This is especially challenging for neonatal and prenatal populations where limited datasets exist and most of the population are asymptomatic. Unlike the diagnostic use of WGS in children with rare diseases, there is no option for physicians to iteratively interpret the neonatal or prenatal sequence data in the context of the patient’s phenotypic profile.81 The meaning and interpretation of variants is likely to change over time as more data from more cases are added to databases,77 and this is especially the case for non-Caucasian groups who are often poorly represented in existing registries and datasets. For all these reasons, working out what, if any, data are cost-effective, clinically useful, and ethically appropriate to return is especially difficult for any attempt to deploy genomics for neonatal or prenatal screening. A further issue is that establishing these genome-phenotype correlations requires making both sequence data and associated clinical information available to researchers and other clinical groups. Where existing data are limited, sharing combined datasets is likely to put the patient at greatest risk of subsequent identification, where they are least likely to receive any direct benefit from the act of sharing.88

3.2 ­Wellness genomics Wellness is a state that encompasses physical, mental, and social well-being and, for some, is a lifelong pursuit. Personal genomic testing, and in particular that based upon nutritional genomics, offers new potential to those hoping to attain an optimal state of wellness. Nutritional genomics refers to the evolving study of gene-­nutrient relationships, including how genetic variations influence the body’s response to nutrients (nutrigenetics) and how nutrients mediate genomic function (nutrigenomics).89 Nutritional genomics has gained increased research attention given its potential to inform an individual’s optimal diet. Tailoring diet to an individual’s genotype is not a new concept; diets low in phenylalanine have been prescribed to people with phenylketonuria, a genetic disorder of amino acid metabolism, since the mid-20th century. The application of nutritional genomics, however, goes beyond the clinical management of monogenic conditions, instead aiming to mitigate polygenic disease risk in otherwise healthy populations. Initially, nutritional genomics was framed to potentially revolutionize the field of human nutrition. Recently the marketing rhetoric of “wellness genomics” has emerged. In many instances, wellness genomic tests expand beyond nutritional genomics, offering consumers not only dietary advice but also insight into fitness regimens, skin care, response to medications (known as pharmacogenomics), and even personality traits. Many of these tests are available online, direct to consumer, although several wellness genomic companies have moved to a direct-to-provider model.90 In this model a healthcare provider facilitates testing for their clients. In these instances the healthcare provider must undergo accreditation with the chosen testing company before offering testing; however, what the training involves and how competency is assessed is unclear. A recent content analysis of online wellness genomic information has

185

186

CHAPTER 8  Implications of secondary findings for clinical contexts

highlighted that in Australia; complementary/alternative medicine providers have enthusiastically adopted this testing.91 Both the wellness genomics testing companies and the associated complementary/alternative medicine providers heavily market their services online. Nutritional genomics is described as the future of healthcare, suggesting that genomic wellness tests are a superior tool for facilitating health and well-being. Websites also claim that genomic wellness tests will reduce the guesswork around diet and lifestyle choices. In particular, websites use language such “optimize” and “transform” to describe how the test results would impact on health. Similar to other direct-to-consumer personal genomic testing advertising, the notion of consumer empowerment was paramount.92 Unfortunately, many of the benefits of nutritional genomics have yet to materialize in most instances. The ACCE framework93 provides a guideline for assessing the quality of genetic tests, especially those used to screen populations, against four criteria. Each letter represents one of the four criteria of the framework. The first, analytical validity, refers to the ability of a test to accurately identify the gene or genetic variation it intends to. The second, clinical validity, considers the reliability of a test to determine disease risk based on a specific genetic variation. The third, clinical utility, refers to the usefulness of test results for informing healthcare decisions. The fourth criterion encompasses ethical, legal, and educational domains. Testing company websites promote the tests as being evidence based and scientifically up to date; some even list the clinical team involved in identifying genetic variants based on the literature.91 However, many gene-nutrient interactions currently lack sufficient clinical validity and yet are still included in these tests.94 Markers of poor health, such as obesity and high cholesterol, which wellness genomics attempts to combat, are mediated by the combination of genetic and nongenetic factors. Wellness genomic tests often do not take into consideration the polygenic nature of noncommunicable diseases, the impact of the environment, or cross-cultural differences.95 Further the nutritional genomics industry currently lacks regulation, meaning testing companies are free to base dietary advice on whatever evidence they choose, resulting in consumers receiving different risk estimates depending on the company. Recognizing this, Grimaldi et al. recently published a set of guidelines for assessing the validity of gene-nutrient interactions, which may produce more standardized tests in the future.94 Regarding clinical utility, research indicates that, even when given genotypebased diets, people are unlikely to make lasting changes to their diet and lifestyle.96 However, interviews with 16 Australian adults have revealed that wellness genomic testing is particularly enticing to the chronically unwell.97 Disheartened by perceived negative interactions with the conventional healthcare system, the majority of these participants turned to complementary/alternative medicine in search of answers. To them, wellness genomic testing represented a new and final hope to get to the “root cause” of their chronic health concerns. Despite describing the test as “empowering” and “validating,” the participants’ self-reported health improvements were small. Most had been prescribed new diets and a variety of supplements but found that the process was more “trial and error” than personalized. While these chronically unwell individuals remained positive about their continuing pursuit of wellness, the

3 ­ Emerging and future scenarios

general, healthy population looking for an “easy fix” to their diet may experience disappointment or “buyer regret.” Given that consumers can access genomic wellness tests online or through healthcare providers who may or may not have sufficient training to interpret the results, issues also arise regarding support given pre- and posttest. Hurlimann et  al.98 recommend all potential consumers be given thorough pretest counseling to ensure informed decisions about testing are made. This is particularly pertinent when considering some genetic variants analyzed in wellness genomics tests also have significant implications for nondiet-related health conditions. Several wellness genomic tests use APOE-ε4 to examine cholesterol regulation, without necessarily communicating the relationship between APOE-ε4 homozygosity and greatly increased risk of Alzheimer’s disease.99 Recently, Janssens et al. published on this issue by describing two nutritional genomic research studies in which APOE-ε4 was included: one where the association between Alzheimer’s risk was described briefly in the participant information sheet and the other where no mention of the association was made at all.99 The authors highlighted the potential psychosocial implications if those research participants homozygous for APOE-ε4 should later learn of their increased risk of Alzheimer’s disease. However, these concerns had already become a reality for one research participant. In 2011 Messner described the case of “Josh,” an otherwise healthy 40 year old, who participated in a study investigating the impact of receiving genetic susceptibility testing on health behaviors.100 After receiving his results, he was shocked to learn that Alzheimer’s disease was one of the health conditions tested for and that he had two copies of the APOE-ε4 allele. Josh described feelings of hopelessness and despair and later criticized the ambiguous consent process involved in the research. APOE-ε4 and Alzheimer’s risk are just one example of secondary or incidental findings that could emerge from genomic wellness testing. As research into genenutrient interactions continues, it is likely clinical validity will improve. Whether or not nutritional genomics will be the answer for those pursuing an optimal state of wellness, however, remains to be demonstrated.

3.3 ­Storage and return of raw sequence data in the clinical and research settings Currently CLIA regulations (section  493.1105) require storage of analytic system records and test reports—including genomics-based tests—for at least 2 years. For more specific suggestions pertaining to storage of NGS technology data, ACMG guidelines recommend that the genomics laboratory consider a minimum of 2-year storage of a digital file type that would allow regeneration of the primary results and reanalysis with improved analytic pipelines. In terms of storage of raw data, it is important to clarify what types of data resulting from WGS could be stored by the laboratories in the clinical and research settings. Data analysis based on WGS generates three file types: (i) FASTQ, which contains raw sequences with corresponding quality scores; (ii) binary alignment/map (BAM), generated by mapping of raw

187

188

CHAPTER 8  Implications of secondary findings for clinical contexts

s­ equences to the human genome reference; and (iii) the variant call format (VCF) file, which contains a list of sequence variants, sorted by genomic position, at which the individual differs from the reference genome. As Evans and colleagues note, “[m]any laboratories produce an annotated VCF with numerous details (such as ­variant type, function, frequency in the population) to aid in the classification and interpretation of each variant. This information, in part, is used to generate the final report for clinicians and patients.”101 Current ACMG clinical laboratory standards for next-generation sequencing states that: “Laboratories should make explicit in their policies which file types and what length of time each type will be retained, and the data retention policy must be in accordance with local, state, and federal requirements.”102 They go on to recommend retention of the VCF and final clinical test report “for as long as possible, given the likelihood of a future request for reinterpretation of variant significance.” Currently, some clinical laboratories do describe their policies relating to storage of raw data in their consent forms, informing the patients about the availability of DNA sequence data for reanalysis and storage of DNA sequence data for various purposes such as test validation or for research. In contrast, some laboratories only indicate the possibility of future policies to incorporate genetic sequence data to permanent medical records.103 With regard to retention of files, the current practices indicate that storing VCF files and, possibly, BAM and FASTQ files by laboratories is necessary. Long-term retention and return of raw genomic data policies may fuel a number of concerns. In terms of the retention, potential unintentional data uses resulting from long-term storage of raw data in patients’ medical records (e.g., by insurance companies or employers) have been underlined as a potential concern. ESHG recommendations, for instance, highlight the potential informational risks that could result from long-term storage of raw genomic data and recommend that the potential implications of access by insurance companies and employers should be addressed.51 Similarly a report by PHG Foundation in the United Kingdom summarizes the main points: “It could be argued that not storing individual genomic data and reanalyzing it in this way would present an enormous missed opportunity to improve both individual and population health. However, storing entire or minimal genome sequences for individual patients would require the use of electronic health records (at least in part), which has major practical and ethical implications. In addition, future technological developments may result in a substantial improvement in the quality of sequencing and genome assembly, and thus make re-sequencing an individual (as required) a better option.”104 In response, storage of raw data privately and outside the patients’ medical records has been suggested as a potential solution. A recent example is MIDATA, a nonprofit cooperative that offers patients private storage for a wide range of health and personal data and allows patients to decide who should have access to the data and for which purposes. In the context of raw data returning policies of WGS research projects, Thorogood and colleagues recently concluded that the current practices of 10 projects show: “Data types and formats may differ depending on the context, sequencing platform, analysis pipelines, and evolution of common file formats. The examples of

4 ­Economic dimensions of returning secondary findings

genomic data formats currently provided to participants include reduced BAM, VCF, and FASTQ.”105 Return of genomic data directly to patents and research subjects also raises the potential for individuals to use third-party websites to seek their own interpretation of the genomic data, which has been perceived by some experts as concerning. Previous studies have shown that individuals may upload their data on online platforms that provide services for interpretation of raw data, such as openSNP, Promethease, GEDMatch, and Genome Mate Pro. The features of such websites vary, ranging from returning health-related results to nonmedical traits and genealogy.106 As with direct-to-consumer genetic testing services such as 23&Me and online wellness genomics platforms, concerns have been raised about returning significant medical information to the individuals without the support and counseling provided by qualified healthcare professionals.107 These questions about how much support individuals should receive from their healthcare providers to interpret the raw data persist.

4 ­Economic dimensions of returning secondary findings from genomics in clinical routine care Although there are many potential opportunities from using NGS technologies, including WES and WGS, there is considerable uncertainty regarding whether they will deliver anticipated improvements in patient health. The health economic case for these technologies requires that their value to healthcare providers can be demonstrated.108 This requires an assessment of their costs and benefits compared with other technologies (standard practice care), the implications for health care systems, and an understanding of patient and other stakeholder views. While the latter is obviously important in its own right, if fewer people undergo sequencing, this could mean that insufficient samples are sequenced in bulk, which would increase costs and reduce the likelihood that sequencing will be cost-effective. There is some evidence that applying NGS in clinical practice might improve the diagnosis and (in some cases) treatment of genetic disease. However, demand is increasing for evidence on the cost-effectiveness of these technologies compared with current practice to ensure that the technologies are not merely an expensive add-on to patient care.109 A recent systematic review identified only 36 papers that reported economic evaluations, cost studies, or outcome studies related to WES and WGS. Most provided little detail on their study methods and generally did not consider the clinical or economic implications for patients after receiving a diagnosis.110 Although these studies looked at sequencing in several genetic conditions, they most commonly examined neurological and neurodevelopmental disorders. Study sample sizes varied from a single child to 2000 patients, with most studies having small sample sizes. There were large ranges in cost estimates for a single test, from $555 to $5169 for WES and from $1906 to $24,810 for WGS. The review concluded there was an urgent need for studies that carefully evaluate the costs, effectiveness, and cost-effectiveness of NGS technologies.

189

190

CHAPTER 8  Implications of secondary findings for clinical contexts

Effectively responding to secondary findings could incur significant initial financial costs but could reduce morbidity, mortality, and overall costs, if this information helps to identify diseases at an earlier stage. While there is limited health economic evidence on the use of NGS generally, there is even less information on specific components of NGS, such as the use of secondary findings, especially with respect to cost-effectiveness and stakeholder preferences. A number of health economic analyses are, however, worth highlighting, including an economic decision model by Bennette et al.111 and a stakeholder survey by Regier et al.112 Bennette and colleagues evaluated the clinical and economic impact of returning the ACMG-recommended secondary finding results.111 They developed a decision model to estimate the likely quality-adjusted life years (QALYs) and lifetime costs associated with returning these findings in three hypothetical cohorts of 10,000 patients. These were patients with hypertrophic or dilated cardiomyopathy (two inherited heart diseases), patients with colorectal cancer or polyposis, and “healthy” individuals undergoing testing because family members have genomic risk factors (or family history indicates a specific disease risk). The authors concluded that returning secondary findings to patients could be cost-effective for certain populations, with QALYs increasing in all three groups. However, screening of generally healthy individuals was not cost-effective based on their calculations, unless genomic sequencing costs are less than $500 per patient. In the current climate, this is still an ambitious price for sequencing the genome of one patient. In fact, sequencing currently costs over $2500 per patient if it is conducted within individual laboratories, as opposed to being centralized and at scale.113 From an economic perspective, personal utility is essentially the “well-being” people experience from choosing a particular healthcare service. A Canadian study by Regier et al. attempted to estimate the personal utility derived from the reporting of secondary findings.112 They used a survey method called a discrete choice experiment to evaluate participants’ personal utility for reporting secondary findings. A discrete choice experiment is used commonly in economics (not just health economics) and market research to gather preferences from stakeholders for different attributes of a good or a service, which gives an indication of how much an individual value that good/service. By breaking down a good/service into its various attributes (characteristics), each having its own corresponding levels, survey participants are able to trade off different attributes against each other, and researchers can then see which attributes are considered the most or least important to individuals and understand whether a particular good/service is preferred.114 Regier et al. used five attributes in their discrete choice experiment investigating preferences for returning secondary findings: disease penetrance, disease treatability, disease severity, carrier status, and cost, which were described in the survey in the context of hypothetical diseases.112 The survey participants were 1200 members of the general Canadian public. Participants indicated that they valued receiving information about high-­ penetrance disorders (larger proportion of individuals with the mutation who have clinical symptoms) but did not value receiving information on low-penetrance disorders (smaller proportion of individuals with the mutation who have clinical

5 ­ Discussion and conclusions

s­ ymptoms). The average willingness to pay to receive secondary findings was $445 in a scenario where clinicians returned information about high-penetrance, medically treatable disorders, but only 66% of participants indicated that they would choose to receive information in that scenario. On average, participants placed importance on having a choice about what type of findings they would receive, including receiving information about high-penetrance, treatable disorders or receipt of information about high-penetrance disorders with or without available treatment. The predicted uptake of that scenario was 76%. Although most of the people completing the survey valued receiving information on secondary findings, personal utility depended on the type of finding, and not all participants wanted to receive this information, irrespective of the potential health implications.112 These survey findings are important because they suggest that a one size fits all approach to reporting secondary findings might not be appropriate. This evidence is interesting given the discussion earlier in the chapter about the use of wellness genomics, given that survey respondents were more likely to value the use of genomics in the context of disease management. The limited number of health economic assessments on secondary findings have generally used health economic decision models or surveys (rather than patient level data) to examine likely costs and effects of returning information on secondary findings to patients. To fully understand the economic value of returning these findings, it will be important in the future to make use of the wealth of clinical and economic information being generated within large sequencing programs such as the UK 100,000 Genomes Project and large centralized biobanks. These initiatives are routinely collecting resource use data, which could be used to assess the costeffectiveness of using secondary findings in routine healthcare.

5 ­Discussion and conclusions This chapter has demonstrated that when it comes to a predefined list of secondary findings to be routinely examined as part of genomics in clinical care, there is considerable variation between countries and between jurisdictions within federated countries such as Australia and Canada. There are different approaches and considerations at play in different settings such as pediatric and adult clinical care and diseasefocused or wellness genomics and even in terms of which technique—WGS, WES, and multiple panel testing—should be employed. The evidence for the economic cost-effectiveness of returning particular sets of secondary or additional findings is equivocal and will require further studies. Much the same can be said of the current study of patient preferences across jurisdictions, conditions, in pediatric versus adult cohorts and in terms of whether attitudes to receiving secondary or additional findings vary according to other sociodemographic characteristics. A number of theories and models have been developed to account for successful or unsuccessful adoption of novel health technologies. These include the field of implementation science,115 normalization process theory,116 and the notion of an “adoption space” where emerging technologies gain an identity as a “cutting

191

192

CHAPTER 8  Implications of secondary findings for clinical contexts

edge,” “lifesaving” or “complicated,” and “expensive” that influences whether and where they are seen as worth adopting by hospital managers and other professional groups.117 Although these concepts are all distinct, they have some features in common. Firstly, they are rarely systematically applied to clinical genomics. Secondly, they all emphasize, in different ways, that any new technology will be operating in an existing environment of organizational frameworks, professional roles and practices, legal requirements and responsibilities, physical and technological infrastructure, social and cultural norms, expectations and values, and political and economic structures, and imperatives. While a machine may “work” according to the way it is designed, making it “workable”—that is making it practically usable in a real world context—also requires embedding the technology in this multilayered existing environment. This may mean adapting both the environment and some features of the technology: “technologies will always need skilled human work, inter-sectoral negotiation and a social infrastructure to ensure that they ‘work’”.118 The case of secondary findings in genomics exemplifies this translational work. As discussed earlier, secondary findings can be understood as a strategy to create a standard of care—an ethically and professionally acceptable way of dealing with the “overspill” of unintended genomic results—that allows clinical implementation of NGS technologies to proceed. Secondary findings affect a separation between genomic variants whose significance is known and agreed to be serious and actionable and those whose import is uncertain and therefore difficult to interpret or act upon. It is intended to enable action, by providing a normative guide on how to use the technology appropriately. The fact that the ACMG guidelines have had an impact on debates far beyond the borders of the United States shows how critical resolving this tension has been in facilitating translation and making clinical genomics “workable.” Equally the fact that several jurisdictions are using or trialing their own lists of secondary findings, different from those specified by the ACMG, demonstrates that, while the idea is extremely useful, clear criteria for demarcating clinically actionable from not (yet) actionable variants are far from universally agreed. Genomics England, for example, employs a significantly shorter list of secondary findings compared with the ACMG list, while the French tradition supports looking for and returning a greater number of variants. The idea of secondary findings is therefore not a simple binary proposition—return them or do not—it is also a question of how additional findings might be returned and how this should best be achieved. Beyond the different criteria for clinical actionability, any attempt to implement return of secondary or additional findings also requires addressing a plethora of related “how” issues: whether to implement “opt-out” mechanisms recognizing patients’ right “not to know” and/or “opt-in” mechanisms operationalizing the corollary “right to know”; whether to provide pretest genetic counseling and/or posttest counseling in all cases or only some; whether to return all findings in one go or separate return of the primary result of genomic testing by leaving return of secondary or additional findings to a later date; what recontact options should be in place, especially for pediatric patients who may wish to consent to receiving additional findings on reaching the legal age of capacity; whether patients can access—or have

5 ­ Discussion and conclusions

a right to access—their “raw” sequence data files and/or transfer them to third-party analysis services; and working out whether genomic testing is being rolled out as an aid to primary diagnosis (as in pediatric rare disease), patient stratification as part of a precision medicine initiative (as in refining cancer diagnoses), or as a screening tool (as with the MyCode Community Health Initiative and the proposals for neonatal and prenatal screening programs)? The feasibility of each of these different options will depend a lot on the existing environment. Are enough genetic counselors available to provide pre- and posttest counseling in all cases, are recontact mechanisms in place or do they have to be created, what do existing legal instruments say about the duty of care or of confidentiality, and what budget is available for different activities? Collecting and supplementing the datasets needed to make reliable correlations between genotype and phenotype also takes clinical genomics into the realms of data transfer between institutions and across borders, and in some cases between public and private sectors. This requires considerations of privacy, data protection, data ownership, and intellectual property.119,120 That these are not simple requirements to negotiate is attested to by the ongoing work of groups like the Global Alliance for Genomics and Health (Ga4GH), which attempts to produce harmonized international standards to enable legal sharing of genomic data.121 The heterogeneity of existing systems and their differing capacity to engage with international data flows and regulations accounts for a further chunk of the variability in the way secondary or additional findings are managed in practice. Translation of genomics is given an additional layer of complexity because it operates in what has been termed a “learning healthcare system.”122 The idea of a learning healthcare system is that new technologies can be translated by using them to provide care in the clinical context while at the same time using the data from their clinical usage to update and improve the way they are used. In the case of genomics, this means integrating genomic data into clinical care while utilizing genomic data collected in the clinical context to update the datasets used to correlate genomic and phenotypic data. It also means collecting evidence for the cost-effectiveness of using genomic data and returning different kinds of findings through the process of rolling these services out into clinical care. In essence, this is a model of ongoing “learning by doing.” A learning healthcare model represents a significant divergence from the clear separation of research and care that characterizes traditional healthcare systems. The challenge of making clinical decisions on an evidence base that is not only subject to change but where the decision to participate actively affects that evidence base raises a number of ethical and organizational challenges, especially with regard to collecting meaningful consent from participants about what returning results is likely to mean (and that this meaning is subject to change and may necessitate future recontact).70,123 This is particularly the case with the statistical risk associated with a particular variant in an asymptomatic individual, which therefore has direct relevance to secondary and additional findings. Ultimately, most jurisdictions described here have adopted some form of exploratory implementation of clinical genomics, dealing with secondary or additional

193

194

CHAPTER 8  Implications of secondary findings for clinical contexts

findings through smaller evaluative studies, or flagship projects that remain separate from mainstream healthcare until sufficient evidence can be collected to inform further development. In this regard, it is not necessarily problematic that different territories approach things in a different manner. Although regulatory harmonization is often eulogized, trying different strategies in different locations affords an opportunity for organizations to learn from a wider range of experience, if—crucially— mechanisms are put in place to report on what works and what is unsuccessful in different locations and to share this information among relevant stakeholders.124 This is a different kind of organizational learning from that envisioned by the learning healthcare system but nonetheless may be equally relevant to clinical genomics. A further complication is the rise of genomic services provided direct to consumer or direct to provider (including private providers), which mix analysis relevant to clinical care with findings relating or pertaining to relate to nutrition, well-being, ancestry, diet, exercise, and other lifestyle factors. As these genomic services can be presented as lifestyle, education, or entertainment products and are often marketed and sold online, they tend to escape many of the challenges of implementing clinical genomics in practice. Although desirable from a strictly economic standpoint, this is less desirable from a health protection perspective as these services may avoid having to make responsible decisions about return of secondary of additional findings and all the implementation measures that accompany them. There is seldom any provision of genetic counseling before or after testing, and policies on pediatric testing, recontact, ownership of genomic data, and other aspects may be lacking or out of step with prevailing ethical and regulatory consensus. The development of coherent and responsible national and international principles for managing secondary and additional findings must therefore also consider those uses of genomics that fall outside the purview of traditional clinical services. There is a strong argument that, as ideas of agreed best practice for testing, evaluating, and implementing return of primary and secondary results of genomic testing results coalesce out of the current exploratory translational research programs, they should also shape the appropriate standards (e.g., provision of genetic counseling) for online and private sector genomic services whether these present themselves as health related, educational, or otherwise.

­References 1. Middleton A, Morley KI, Bragin E, et al. Attitudes of nearly 7000 health professionals, genomic researchers and publics toward the return of incidental results from sequencing research. Eur J Hum Genet. 2016;24:21–29. 2. Tan N, Amendola LM, O’Daniel JM, et al. Is “incidental finding” the best term? A study of patient’s preferences. Genet Med. 2017;19:176–181. 3. Hall M, Bobinski MA, Orentlicher D. Medical Liability and Treatment Relationships. 2nd ed. New York, NY: Aspen Publishers; 2008. 4. Green RC, Berg JS, Grody WW, et al. ACMG recommendations for reporting incidental findings in clinical exome and genome sequencing. Genet Med. 2013;15(7):565–574.

­References

5. Burke W, Antommaria AH, Bennet R, et al. Recommendations for returning genomic incidental findings? We need to talk!. Genet Med. 2013;15:855–859. 6. Wolf SM, Annas GJ, Elias S. Patient autonomy and incidental findings in clinical genomics. Science. 2013;340(6136):1049–1050. 7. Kalia SS, Adelman K, Bale SJ, et al. Recommendations for reporting of secondary findings in clinical exome and genome sequencing, 2016 update (ACMG SF v2. 0): a policy statement of the American College of Medical Genetics and Genomics. Genet Med. 2017;19:249–255. 8. American College of Medical Genetics and Genomics. ACMG Updates Recommendation on “Opt Out” for Genome Sequencing Return of Results, https://www.acmg.net/docs/ Release_ACMGUpdatesRecommendations_final.pdf. 9. Fowler  SA, Saunders  CJ, Hoffman  MA. Variation among consent forms for clinical whole exome sequencing. J Genet Couns. 2018;27(1):104–114. 10. Andrews M. What If Your Doctor Offered Genetic Testing as a Way to Keep You Healthy? Washington Post. May 28, https://www.washingtonpost.com/national/health-science/ what-if-your-doctor-offered-genetic-testing-as-a-way-to-keep-you-healthy/2018/05/ 25/84b69238-5dd5-11e8-a4a4-c070ef53f315_story.html?noredirect=on&utm_term=. ca9dc1d23c07; 2018. (Accessed July 26, 2018). 11. Health  G. MyCode Conditions 2016, https://www.geisinger.org/precision-health/mycode/mycode-conditions; 2018. (Accessed July 26, 2018). 12. Berg JS, Foreman AK, O’Daniel JM, et al. A semi-quantitative metric for evaluating clinical actionability of incidental or secondary findings from genome-scale sequencing. Genet Med. 2016;18:467–475. 13. Collins FS, Varmus H. A new initiative on precision medicine. NEJM. 2015;372:793–795. 14. Genomics Working Group of the All of Us Research Program. Considerations Toward a Comprehensive Genomics Strategy, https://allofus.nih.gov/sites/default/files/gwg_final_report.pdf; 2018. (Accessed August 20, 2018). 15. National Academies of Sciences, Engineering, and Medicine. Returning Individual Research Results to Participants: Guidance for a New Research Paradigm. Washington, DC: The National Academies Press; 2018. 16. Genomics England. The 100,000 Genomes Project, https://www.genomicsengland. co.uk/the-100000-genomes-project/; 2018. (Accessed October 28, 2018). 17. Godfrey  K. Genetics white paper heralds “a revolution in health care”. BMJ. 2003;32(7404):1413. 18. The Royal College of Radiologists. Standards for Interpretation and Reporting of Imaging Investigations. 2nd ed. Clin Radiol. 2018. March, https://www.rcr.ac.uk/system/ files/publication/field_publication_files/bfcr181_standards_for_interpretation_reporting.pdf; (Accessed February 21, 2019). 19. Mitchell  C, Ploem  C, Chico  V, et  al. Exploring the potential duty of care in clinical genomics under UK law. Med Law Int. 2017;17(3):158–182. 20. Hehir-Kwa  JY, Claustres  M, Hastings  RJ, et  al. Towards a European consensus for reporting incidental findings during clinical NGS testing. Eur J Hum Genet. 2015;23(12):1601–1606. 21. Hall A, Hallowell N, Zimmern R. Managing Incidental and Pertinent Findings From WGS in the 100,000 Genome Project. PHG Foundation; 2013, ISBN: 978-1-907198-12-0. 22. Wright CF, Middleton A, Burton H, et al. Policy challenges of clinical genome sequencing. BMJ. 2013;347:f6845, https://doi.org/10.1136/bmj.f6845.

195

196

CHAPTER 8  Implications of secondary findings for clinical contexts

23. Abul-Husn  NS, Manickam  K, Jones  LK, et  al. Genetic identification of familial hypercholesterolemia within a single U.S. health care system. Science. 2016;354(6319), https://doi.org/10.1126/science.aaf7000. 24. Wolf SM, Annas GJ, Elias S. Point-counterpoint. Patient autonomy and incidental findings in clinical genomics. Science. 2013;340(6136):1049–1050. 25. Ormondroyd E, Mackley M, Blair E, et al. “Not pathogenic until proven otherwise”: perspectives of UK clinical genomics professionals towards secondary findings in context of a Genomic Medicine Multidisciplinary Team and the 100,000 Genomes Project. Genet Med. 2018;20:320–328. 26. Davies  S. Annual Report of the Chief Medical Officer 2016—Generation Genome, https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/631043/ CMO_annual_report_generation_genome.pdf; 2018. (Accessed October 28, 2018). 27. NHS Science and Technology Committee. Genomics and Genome Editing in the NHS, https://publications.parliament.uk/pa/cm201719/cmselect/cmsctech/349/34902.htm; 2018. (Accessed October 28, 2018). 28. Hill  S. The Genomic Revolution—Its Future, https://www.england.nhs.uk/blog/genomic-revolution/; 2019. (Accessed February 21, 2019). 29. National Health and Medical Research Council. Principles for the Translation of ‘Omics’-Based Tests From Discovery to Health Care. Canberra: National Health and Medical Research Council; 2015. 30. Human Genetics Society Australasia. HGSA Commentary on ACMG Recommendations, https://www.hgsa.org.au/hgsanews/hgsa-commentary-on-acmg-recommendations; 2018. (Accessed September 27, 2018). 31. Royal College of Pathologists of Australia. Massively Parallel Sequencing Implementation Guidelines, https://www.rcpa.edu.au/Library/Practising-Pathology/ RCPA-Genetic-Testing/MAPSIG; 2018. (Accessed September 27, 2018). 32. Gaff CL, Winship M, Forrest IM, et al. Preparing for genomic medicine: a real world demonstration of health system change. Genom Med. 2017;2(1):16, https://doi. org/10.1038/s41525-017-0017-4. 33. Gaff C. Unpublished research data. 34. Australian Government Department of Health. Health Technology Assessment for Reimbursement, http://www.health.gov.au/internet/hta/publishing.nsf/content/reimbursement-1. (Accessed February 21, 2019). 35. Ackerman  SL, Koenig  BA. Understanding variations in secondary findings reporting practices across U.S. genome sequencing laboratories. AJOB Empir Bioethics. 2017;9(1):48–57, https://doi.org/10.1080/23294515.2017.1405095. 36. Martyn M, Kanga-Parabia A, Lynch E, et al. A novel approach to offering additional findings—testing a two-step approach in the healthcare system. J Genet Couns. 2019, https://doi.org/10.1002/jgc4.1102. [Epub ahead of print]. 37. Bergner AL, Bollinger J, Raraigh KS, et al. Informed consent for exome sequencing research in families with genetic disease: the emerging issue of incidental findings. Am J Med Genet A. 2014;164A:2745–2752. 38. Bernhardt  BA, Roche  MI, Perry  DL, Scollon  SR, Tomlinson  AN, Skinner  D. Experiences with obtaining informed consent for genomic sequencing. Am J Med Genet A. 2015;167A:2635–2646. 39. Downie L, Halliday J, Burt R, et al. A protocol for whole-exome sequencing in newborns with congenital deafness: a prospective population-based cohort. BMJ Paediatr Open. 2017, https://doi.org/10.1136/bmjpo-2017-000119.

­References

40. Vears DF, Sénécal K, Borry P. Reporting practices for unsolicited and secondary findings from next-generation sequencing technologies: perspectives of laboratory personnel. Hum Mutat. 2017;38(8):905–911. 41. Australian Genomics Health Alliance, https://www.australiangenomics.org.au; 2019. (Accessed February 21, 2019). 42. Australian Government Department of Health. National Health Genomics Policy Framework 2018–2021, http://www.health.gov.au/internet/main/publishing.nsf/Content/ national-health-genomics-policy-framework-2018-2021; 2019. (Accessed April 8, 2019). 43. Center for Human Genetics and Laboratory Diagnostics (AHC). Next Generation Sequencing in Molecular Diagnostics, https://www.medizinische-genetik.de/index. php?id=gen-panel-analyse-diagnostik; 2019. (Accessed March 22, 2019). 44. Horak P, Klink B, Heining C, et al. Precision oncology based on omics data: The NCT Heidelberg experience. Int J Cancer. 2017;141(5):877–886. 45. Medical Informatics Initiative Germany, https://www.medizininformatik-initiative.de/ en/start; 2019. (Accessed March 22, 2019). 46. Draft Law on Genetic Testing in Humans (Genetic Diagnostics Act—GenDG 2009) (Germany). Available at: http://dip21.bundestag.de/dip21/btd/16/105/1610532.pdf; 2019. (Accessed March 22, 2019). 47. Berlin-Brandenburgische Akademie der Wissenschaften. Ad-hoc-Stellungnahme— Neue Sequenzierungstechniken: Konsequenzen für die genetische Krankenversorgung. http://www.bbaw.de/publikationen/stellugnahmen-empfehlungen/StellungnahmenGendiagnostik.pdf; 2013. (Accessed March 22, 2019). 48. Deutscher Ethikrat. Die Zukunft der genetischen Diagnostik, https://www.ethikrat.org/ fileadmin/Publikationen/Stellungnahmen/deutsch/stellungnahme-zukunft-der-genetischen-diagnostik.pdf; 2013. (Accessed March 22, 2019). 49. German Society of Human Genetics (GfH). Stellungnahme der Deutschen Gesellschaft für Humangenetik zu genetischen Zusatzbefunden in Diagnostik und Forschung, https:// www.gfhev.de/de/leitlinien/LL_und_Stellungnahmen/2013_05_28_Stellungnahme_ zu_genetischen_Zufallsbefunden.pdf; 2019. (Accessed March 22, 2019). 50. Project EURAT Ethical and Legal Aspects of Whole Human Genome Sequencing. Cornerstones for an ethically and legally informed practice of Whole Genome Sequencing: Code of Conduct and Patient Consent Models. Heidelberg: MarsiliusKolleg der Universität Heidelberg; 2016. 51. Matthijs  G, Souche  E, Alders  M, et  al. Guidelines for diagnostic next-generation sequencing. Eur J Hum Genet. 2016;24(1):2–5. 52. van El  CG, Cornel  MC, Borry  P, et  al. Whole-genome sequencing in health care: recommendations of the European Society of Human Genetics. Eur J Hum Genet. 2013;21(6):580–584. 53. Bertier G, Joly Y. Clinical exome sequencing in France and Quebec: what are the challenges? What does the future hold? Life Sci Soc Policy. 2018;14(1):17, https://doi. org/10.1186/s40504-018-0081-2. 54. Aviesan. Genomic Medicine France 2025, https://solidarites-sante.gouv.fr/IMG/pdf/genomic_medicine_france_2025.pdf; 2019. (Accessed February 27, 2019). 55. Pujol P, Vande Perre P, Faivre L, et al. Guidelines for reporting secondary findings of genome sequencing in cancer genes: the SFMPP recommendations. Eur J Hum Genet. 2018;2:53. 56. Boycott K, Hartley T, Adam S, Bernier F, Chong K, Fernandez BA. The clinical application of genome-wide sequencing for monogenic diseases in Canada: position statement of the Canadian College of Medical Geneticists. J Med Genet. 2015;52(7):431–437.

197

198

CHAPTER 8  Implications of secondary findings for clinical contexts

57. MoHSS. Liste des analyses non couvertes par l’assurance maladie du Québec et non remboursées dans le cadre du Mécanisme d'autorisation et de remboursement des analyses de biologie médicale non disponibles au Québec, Réf.: Circulaire 2011-12. 2015. 58. Génome Québec. Che Sainte-Justine and Génome Québec integrated Centre for Pediatric Clinical Genomics, http://www.genomequebec.com/en/centre-de-genomiqueclinique-pediatrique-integre-genome-quebec.html; 2019. (Accessed March 23, 2019). 59. Human Biomedical Research Act 2015 (Singapore). Available at: https://sso.agc.gov.sg/ Act/HBRA2015; 2019. (Accessed February 27, 2019). 60. Bioethics Advisory Committee Singapore. Ethics Guidelines for Human Biomedical Research, http://www.bioethics-singapore.org/images/uploadfile/reportOnly.pdf; 2015. (Accessed February 27, 2019). 61. Ministry of Health Singapore. Code of Practice on the Standards for the Provision of Genetic/Genomic Testing Servicse and Clinical Genetic/Genomic Testing Services, https://elis.moh.gov.sg/elis/publishInfo.do?task=download&pkId=295; 2018. (Accessed February 27, 2019). 62. Leitsalu L, Haller T, Esko T, et al. Cohort profile: Estonian Biobank of the Estonian Genome Center, University of Tartu. Int J Epidemiol. 2015;44(4):1137–1147. 63. Human Genes Research Act 2000 (Estonia). Available at: https://www.riigiteataja.ee/en/ eli/531102013003/consolide; 2018. (Accessed September 18, 2018). 64. Leitsalu L, Alavere H, Jacquemont S, et al. Reporting incidental findings of genomic disorder-associated copy number variants to unselected biobank participants. Personal Med. 2016;13(4):303–314. 65. Alver M, Palover M, Saar A, et al. Recall by genotype and cascade screening for familial hypercholesterolemia in a population-based biobank from Estonia. Genet Med. 2018, https://doi.org/10.1038/s41436-018-0311-2. 66. Tartu University Hospital (N.D.). NÕUSOLEK KOGU EKSOOMI SEKVENEERIMISE ANALÜÜSIKS, https://www.kliinikum.ee/geneetikakeskus/pildid/saatelehed/Eksoomi_ sekveneerimine_nousoleku_vorm.pdf; 2018. (Accessed September 18, 2018). 67. Ministry of Social Affairs (N.D.). Personalised Medicine, http://www.sm.ee/en/personalised-medicine; 2018. (Accessed September 18, 2018). 68. Ministry of Social Affairs. News Item: Estonia Offers 100,000 Residents Free Genetic Testing. 20th March, http://www.sm.ee/en/news/estonia-offers-100000-residents-freegenetic-testing; 2018. (Accessed September 18, 2018). 69. Tashiro  S. Unintended consequences of ‘soft’ regulations: the social control of human biomedical research in Japan. Int J Jpn Sociol. 2010;19, https://doi. org/10.1111/j.1475-6781.2010.01127.x. 70. Minari J, Chalmers D, Kato K. Return of genetic research results: the Japanese experience and its implication for the international debate. SCRIPTed. 2014;11(2):180–192. 71. Minari J, Brothers K, Morrison M. Tensions in ethics and policy created by National Precision Medicine Programs. Hum Genom. 2018;12(22), https://doi.org/10.1186/ s40246-018-0151-9. 72. Kato  K, Ohashi  N, Yoshida  T. Investigation and Recommendations Regarding the Handling of Incidental and Secondary Findings: A Report by the Sub-Research Group 2 of the AMED Program for the Promotion of Clinical Application of Genome Medicine, http://www.biobank.amed.go.jp/content/pdf/elsi/IF-SF_Nakagama-Kato%20H29.3.24. pdf; Mar 2017. (Accessed March 11, 2019). [in Japanese]. 73. Kuriyama S, Yaegashi N, Nagami F, et al. The Tohoku Medical Megabank Project: design and mission. J Epidemiol. 2016;26(9):493–511.

­References

74. Zerhouni E. The NIH roadmap. Science. 2003;302:63. 75. Woolf SH. The meaning of translational research & why it matters. J Am Med Assoc. 2008;299:211–213. 76. Stokes  E. Nanotechnology and the products of inherited regulation. J Law Soc. 2016;39(1):93–112. 77. Lee  N. The challenge of childhood: Distributions of childhood’s ambiguity in adult institutions. Childhood. 1999;6(4), https://doi.org/10.1177/0907568299006004005. 78. Wade  CH, Tarini  BA, Wilfond  BS. Growing up in the genomic era: implications of whole genome sequencing for children, parents, families and paediatric practice. Annu Rev Genomics Hum Genet. 2013;14:535–555. 79. Howard HC, Knoppers BM, Cornel MC, et al. Whole-genome sequencing in newborn screening? A statement on the continued importance of targeted approaches in newborn screening programmes. Eur J Hum Genet. 2015;23(12):1593–1600. 80. Klugman S, Dolan S. Whole Exome and Whole Genome Sequencing in Prenatal Testing: What’s the Whole Story? GeneWatch, http://www.councilforresponsiblegenetics.org/genewatch/GeneWatchPage.aspx?pageId=551; 2015. (Accessed August 14, 2018). 81. Department of Health (UK). Our Inheritance, Our Future: Realising the Potential of Genetics in the NHS. London: Department of Health; 2003. 82. Johnston J, Lantos JD, Goldenberg A, et al. Sequencing newborns: a call for nuanced use of genomic technologies. Hastings Center Rep. 2018;48(4), https://doi.org/10.1002/hast.874. 83. Minear MA, Lewis C, Pradhan S, Chandrasekharan S. Global perspectives on clinical adoption of NIPT. Prenat Diagn. 2015;35(10):959–967. 84. Lo DYM. Non-invasive prenatal testing using massively parallel sequencing of maternal plasma DNA: from molecular karyotyping to fetal whole-genome sequencing. Reprod BioMed Online. 2013;27:593–598. 85. Badeau M, Lindsay C, Blais J, et al. Genomics-based non-invasive prenatal testing for detection of fetal chromosomal aneuploidy in pregnant women. Cochrane Database Syst Rev. 2017, https://doi.org/10.1002/14651858.CD011767. 86. Lewis C, Hill M, Silcock C, Daley R, Chitty LS. Non-invasive prenatal testing for trisomy 21: a cross-sectional survey of service users’ views and likely uptake. Br J Obstetr Gynecol. 2014;121:582–594. 87. Sirchia  F, Carrieri  D, Deensa  S, et  al. Recontacting or not recontacting? A survey of current practices in clinical genetics centres in Europe. Eur J Hum Genet. 2018;26(7):946–954. 88. Rahimzadeh V, Schickhardt C, Knoppers BM, et al. Key implications of data sharing in pediatric genomics. JAMA Pediatr. 2018;172(5):476–481. 89. Collins  J, Bertrand  B, Hayes  V, et  al. The application of genetics and nutritional genomics in practice: an international survey of knowledge, involvement and confidence among dietitians in the US, Australia and the UK. Genes Nutr. 2013;8(6):523–533. 90. McGowan ML, Fishman JR, Settersten Jr. RA, Lambrix MA, Juengst ET. Gatekeepers or intermediaries? The role of clinicians in commercial genomic testing. PLoS One. 2014;9(9):e108484. 91. Tutty E, Hickerton C, Adamski MM, Metcalfe SA. Personal genomic testing for nutrition and wellness in Australia: a content analysis of online information. Nutr Diet. 2019, https://doi.org/10.1111/1747-0080.12516. 92. Covolo  L, Rubinelli  S, Orizio  G, Gelatti  U. Misuse (and abuse?) of the concept of empowerment: the case of online offer of predictive direct-to-consumer genetic tests. J Publ Health Res. 2012;1(1):3.

199

200

CHAPTER 8  Implications of secondary findings for clinical contexts

93. Sanderson S, Zimmern R, Kroese M, Higgins J, Patch C, Emery J. How can the evaluation of genetic tests be enhanced? Lessons learned from the ACCE framework and evaluating genetic tests in the United Kingdom. Genet Med. 2005;7(7):495–500. 94. Grimaldi KA, van Ommen B, Ordovas JM, et al. Proposed guidelines to evaluate scientific validity and evidence for genotype-based dietary advice. Genes Nutr. 2017;12:35. 95. Guasch-Ferré  M, Dashti  HS, Merino  J. Nutritional genomics and direct-to-consumer genetic testing: an overview. Adv Nutr. 2018;9(2):128–135. 96. Celis-Morales C, Livingstone KM, Marsaux CF, et al. Effect of personalized nutrition on health-related behaviour change: evidence from the Food4me European randomized controlled trial. Int J Epidemiol. 2016;2(1):578–588. 97. Metcalfe SA, Hickerton C, Savard J, et al. Australians’ perspectives on support around use of personal genomic testing: findings from the Genioz study. Eur J Med Genet. 2019, https://doi.org/10.1016/j.ejmg.2018.11.002. 98. Hurlimann T, Robitaille J, Vohl MC, Godard B. Ethical considerations in the implementation of nutrigenetics/nutrigenomics. Pers Med. 2017;14(1):75–83. 99. Janssens  ACJ, Bunnik  EM, Burke  W, Schermer  MH. Uninformed consent in nutrigenomic research. Eur J Hum Genet. 2017;25:789–790. 100. Messner DA. Informed choice in direct-to-consumer genetic testing for Alzheimer and other diseases: lessons from two cases. New Genet Soc. 2011;30(1):59–72. 101. Evans BJ, Dorchner MO, Burke W, Jarvik GP. Regulatory changes raise troubling questions for genomic testing. Genet Med. 2014;16(11):799–803. 102. Rehm HL, Bale SJ, Bayrak-Toydemir P, Berg JS, Brown KK, Deignan JL. ACMG clinical laboratory standards for next-generation sequencing. Genet Med. 2013;15(9):733–747. 103. Shabani M, Vears D, Borry P. Raw genomic data: storage, access and sharing. Trends Genet. 2018;34(1):8–10. 104. Public Health Genomics Foundation Foundation. Next Steps in the Sequence: the Implications of Whole Genome Sequencing for Health in the UK, http://www.phgfoundation.org/documents/283_1323430677.pdf; 2019. (Accessed March 1, 2019). 105. Thorogood A, Bobe J, Prainsack B, et al. APPLaUD: access for patients and participants to individual level uninterpreted genomic data. Hum Genom. 2018;12(1):7, https://doi. org/10.1186/s40246-018-0139-5. 106. Nelson SC, Fullerton SM. “Bridge to the Literature”? Third-party genetic interpretation tools and the views of tool developers. J Genet Couns. 2018;27(4):770–781. 107. Badalato L, Kalokairinou L, Borry P. Third party interpretation of raw genetic data: an ethical exploration. Eur J Hum Genet. 2017;25(11):1189–1194. 108. Gavan SP, Thompson AJ, Payne K. The economic case for precision medicine. Expert Rev Precis Med Drug Dev. 2018;3(1):1–9. 109. Chisholm J, Wordsworth S. The economics of sequencing. In: ‘Generation Genome’ Annual Report of the Chief Medical Officer 2016. London: Department of Health; 2017. [chapter 16]. 110. Schwarze K, Buchanan J, Taylor JC, Wordsworth S. Are whole exome and whole genome sequencing approaches cost-effective? A systematic review of the literature. Genet Med. 2018, https://doi.org/10.1038/gim.2017.247. 111. Bennette CS, Gallego CJ, Burke W, Jarvik GP, Veenstra DL. The cost-effectiveness of returning incidental findings from next-generation genomic sequencing. Genet Med. 2015;17(7):587–595. 112. Regier DA, Peacock SJ, Pataky R, et al. Societal preferences for the return of incidental findings from clinical genomic sequencing: a discrete-choice experiment. CMAJ. 2015;187(6):E190–E197.

­References

113. Schwarze K, Buchanan J, Fermont JM, et al. The complete costs of genome sequencing: a micro-costing study in cancer and rare diseases from a single centre in the United Kingdom. Genet Med. 2019, https://doi.org/10.1038/s41436-019-0618-7. 114. Ryan M. Discrete choice experiments in health care. BMJ. 2004;328:360. 115. May CR, Finch T, Ballini L, et al. Evaluating complex interventions and health technologies using normalization process theory: development of a simplified approach and web enabled toolkit. BMC Health Serv Res. 2011;11:245, http://www.biomedcentral. com/1472-6963/11/245. 116. Bauer MS, Damschroder L, Hagedorn H, Smith J, Kilbourne AM. An introduction to implementation science for the non-specialist. BMC Psychol. 2015;3:32, https://doi. org/10.1186/s40359-015-0089-9. 117. Ulucanlar  S, Faulkner  A, Peirce  S, Elwyn  G. Technology identity: the role of sociotechnical representations in the adoption of medical devices. Soc Sci Med. 2013;98(December):95–105. 118. Greenhalgh T, Procter R, Wherton J, Sugarhood P, Hinder S, Rouncefield M. What is quality in assisted living technology? The ARCHIE framework for effective telehealth and telecare services. BMC Med. 2015;13:91, https://doi.org/10.1186/s12916-015-0279-6. 119. Kosseim P, Dove ES, Baggaley C, et al. Building a data sharing model for global genomic research. Genome Biol. 2014;15(8):430. 120. Kaye J, Terry SF, Juengst E, Coy S, Harris JR, Chalmers D. Including all voices in international data-sharing governance. Hum Genom. 2018;12:13, https://doi.org/10.1186/ s40246-018-0143-9. 121. The Global Alliance for Genomics and Health. A federated ecosystem for sharing genomic, clinical data. Science. 2016;352(6291):1278–1280. 122. Friedman  CP, Wong  AK, Blumenthal  D. Achieving a nationwide learning healthcare system. Sci Trans Med. 2010;2(57):57cm29. 123. Faden RR, Kass NE, Goodman SN, Pronovost P, Tunis S, Beauchamp TL. An ethics framework for a learning health care system: a departure from traditional research ethics and clinical ethics. Hastings Center Rep. 2013;43(S1):S16–S27. 124. Sabel CF, Zeitlin J. Learning from difference: the new architecture of experimentalist governances in the EU. Eur Law J. 2008;14(3):271–327.

201

CHAPTER

Secondary findings: Building a bridge to the future of ELSI

9

Kyle B. Brothersa, Martin Langankeb, Pia Erdmannc,† a

Endowed Chair of Pediatric Clinical and Translational Research, University of Louisville, Louisville, KY, United States b Department of Social Work, Protestant University of Applied Sciences, Bochum, Germany c Faculty of Theology, University of Greifswald, Greifswald, Germany

As we reach the concluding chapter of this book, we recognize that readers will have traveled different paths before reaching this final chapter. Some readers may have read this text from beginning to end. Other readers—and we anticipate the majority of readers—will have engaged with this book by reading a subset of chapters, perhaps even a single chapter, with the goal of digging deeper into specific topics. Regardless of the path taken to this point, we hope learners and scholars who have engaged with this text will continue the journey started here. This text is not intended so much as a destination for those wanting to understand the significant controversy that has developed around secondary findings in recent decades. Rather, it is intended as a point of departure, a first stop, for clinicians, genetics researchers, bioethicists, and others who now have an opportunity to take part in the debates and discourses summarized here. In fact, this book is an appropriate starting point for anyone interested in the broader world of scholarship on the ethical, legal, and social issues (ELSI) raised by genomics technologies. After all the controversies and points of disagreement that were first formulated in debates on secondary findings have since expanded to define a great deal of the discourse that comprises ELSI scholarship. In today’s world, ELSI has grown into a community of scholars and practitioners seeking to make sense of the seismic uncertainties that genomics technologies have created in the worlds of medicine, law enforcement, ancestry, and politics. In seeking to make sense of those repercussions, earlier work on secondary research findings often provide the framework. It is as if this issue of secondary research finding sits at a strategic location in close proximity to the genomics epicenter. In this chapter, with its goal of both wrapping up and looking to the future, it will be useful to briefly consider why the discourse on secondary research findings † 

Deceased.

Secondary Findings in Genomic Research. https://doi.org/10.1016/B978-0-12-816549-2.00009-6 © 2020 Elsevier Inc. All rights reserved.

203

204

CHAPTER 9  Secondary findings: Building a bridge to the future of ELSI

has proven to be such a defining issue for ELSI scholarship. Naturally, this is itself a complex question, but we can venture a few observations. We hope these thoughts will serve both to pull together the threads addressed in this book and to usher readers into the broader world of ELSI issues. If we are to utilize genomics technologies in ways that help most people and hurt as few as possible, we will need diverse, ­creative—not to mention interdisciplinary and multidisciplinary—minds. We hope this book will recruit some such minds into the cause.

1 ­Secondary findings as sentinel debate There was a time, very roughly from 2005 to 2015, when a significant proportion of ELSI discourse centered on the management of secondary findings generated in research contexts. When Francis Collins, director of the National Institutes of Health (NIH) and former lead for the Human Genome Project, testified before the Presidential Commission for the Study of Bioethical Issues in 2011, he observed that the issue was “a hot topic in every conversation about every genetic research protocol that I’m involved in.”1 As discussed in Chapter 1 the term “secondary findings” was uncommon during that time, with most scholars using the term “incidental findings” that had been borrowed from imaging. The term “incidental findings” was using in imaging in both clinical and research settings. In genomics, however, it was first used primarily in research settings. Clinical exome sequencing (ES) and genome sequencing (GS) were uncommon until the latter years of this period, so concern about secondary findings felt more urgent in research contexts. It is also telling that “return of results” or “return of research results” was used as a shorthand to refer to this issue. Implicit in this terminology was an ethical concern not only about secondary findings per se but also about the entire idea of disclosing research findings to participants. Scientific knowledge about the causal relationships between genetic variants and disease was still developing. As a result, many worried that research participants would receive genetic research results that were poorly understood, or even outright incorrect. Views on disclosing research findings to participants was also influenced by the research models that were increasingly being used in genomics during this time. Conventional research studies tended to recruit and consent a new sample of participants for each study. Such an approach was economically infeasible, however, for studies like genome-wide association studies (GWAS) that often required 10,000 or more participants. Genetic biorepositories were thus being built around the world to support these types of large-scale genomics studies. However, this innovation in genomics research brought the challenges of “return of research results” into stark relief. How could researchers be expected to disclose secondary finding to tens of thousands of research participants who were no longer in contact with the research team? It would be a mistake, however, to interpret this history as evidence that the ethical, legal, and societal issues raised by secondary findings were specific to a particular

1 ­ Secondary findings as sentinel debate

time period, a particular type of study design, or a particular stage of scientific development. In truth, nearly all of the issues that were being debated around “incidental findings” in 2005 remain salient today. Genomics technologies are now being used more widely in translational and clinical contexts, designs for large-scale genomics research have evolved, and understanding of genotype-phenotype associations have drastically improved. But we still face many of the same fundamental issues. For those wishing to delve deeper into these ELSI issues, it is more helpful to take a different view of early debates in genomics about “incidental findings” and “return of research results.” According to this understanding, secondary findings became a defining issue for ELSI scholarship at least in part because this issue provided the first concrete opportunity to deal with difficult issues raised by genomics technologies. For example, a single laboratory test—whether it be GS or ES—has the potential to uncover information about 20,000 genes and perhaps thousands of different medical conditions. The capacity to generate numerous results is not unique to genomics technologies, but it is a characteristic that raises real issues: When the potential results a laboratory test can generate are virtually open-ended, how should we approach informed consent? Can current pretest and posttest counseling strategies accommodate numerous results, including both results related to the reason for testing and those unrelated to testing? These were issues that clinicians, genomics researchers, and ELSI scholars were going to be faced with eventually; it was simply that the management of secondary findings in the research context was one of the first settings where these issues started to reach the real world. Something similar could be said for a number of other issues. For example, one dimension of early debates on secondary findings in genomics research focused on clarifying which types of results would be considered for return. This was certainly an ethically relevant issue, as it is difficult to examine whether it might be permissible or even obligatory to disclose secondary findings to participants without first defining the different types of results that might be considered for return. However, this discourse on how results should be classified—and, in fact, how we even define what counts as a “result”—is a fundamental issue for genomics. Genome and exome sequencing technologies are capable of generating millions of data points; translating those data points into something meaningful—a “result”—is an interpretive process. How should laboratories apply informatics tools and human efforts to filter, prioritize, validate, and categorize results? These questions are almost like fractals in mathematics: as one zooms in on each subsequent issue, new layers of complexity come into focus. Early on in the debate on secondary findings, there was an intuition that results should be returned if they are useful. But this simply raised additional issues. What counts as utility? Does only clinical utility count, or are more personal dimensions of utility relevant? Does utility have to be empirically established (such as through comparative effectiveness research), or is it good enough that experts have an intuition that a type of genomic information could lead to “actionability”? Scholars in genomics, medicine, and ELSI research continue to dig through these layers. They first came into focus, however, in the course of debates on the management of secondary findings in genomics research.

205

206

CHAPTER 9  Secondary findings: Building a bridge to the future of ELSI

2 ­Secondary findings as cultural crossroads The discourse on secondary genomic findings in research also provides a window into ELSI scholarship because it reflects broader cultural tensions that remain unresolved. For example, there is a broad, international cultural debate about whether individuals hold something akin to an ownership right over their data. The European Union strongly endorsed this concept in 2016 with the adoption of the General Data Protection Regulation (GDPR), but important questions remain in Europe and elsewhere about how much control individuals should hold over data that were collected from or about them. Critically the debate about data ownership is not only about the data itself but also about findings generated using “my” data. Amazon, for example, utilizes machine learning and other techniques to develop an analysis of each customers’ purchasing history. Companies like Google utilize similar strategies to generate a picture of its users’ individual interests and priorities. For Amazon the goal is to recommend additional items that users might want to purchase. For Google the goal is to deliver both search results and advertisements that anticipate users’ preferences. While both companies hold large amounts of their users’ data, the real “money” lies in these analyses derived from the data. If users have some right of ownership over their data, such as the “right of erasure” proposed under the GDPR, do they also have some rights over the analyses derived from that data? If Google has developed a picture of my interests, do I have a right to see that abstraction of my data? These types of issues have also played an important role in the discourse on secondary findings. One question is whether research participants retain some sort of ownership rights over the biosamples and data they contribute to research. But perhaps an even thornier issue is whether researchers are also obligated to disclose to participants all of the genomic findings they generate. This drives to the heart of debates about the criteria for selecting secondary findings for return. Even if there is a consensus that it is permissible, perhaps even obligatory, to disclose secondary findings that researchers consider “actionable,” it remains unresolved whether participants also have some right to access the other individual findings that researchers generate. When asked which types of results they would like to receive, many research participants state that they want to know “everything.” What is the scope of everything? Does it include results that are considered valid, but would not inform changes in medical care? What about findings that are based on very preliminary science that are not yet well validated? These questions reflect just one way that the discourse on secondary findings reflects broader cultural debates. We can identify a number of other broad cultural controversies that have found their way into the discourse on secondary findings, including how insights from big data analyses should be applied in the real world and how professionals should balance their opportunity to provide benefits to the public with individuals’ right to plot their own course. That the debate over secondary findings has encapsulated so many of these controversies partly explain why this particular debate has proven to be so pivotal to ELSI scholarship. For scholars interested in wrangling

3 ­Secondary findings and the future of ELSI scholarship

with these far-ranging issues in the context of genomics, the discourse on secondary findings provides both a crucial backdrop and a starting point for future work.

3 ­Secondary findings and the future of ELSI scholarship If this book has served to engage readers in dealing with the difficult issues raised by genomics technologies, it may also be helpful to point to some future directions for that work. This book is not intended to serve as a retrospective on a debate now resolved. Quite the opposite, in fact, the discourse over secondary findings in genomics research has only continued to grow and evolve through the years. This book is just one jumping off point for a discourse that will continue for years to come. Although we feel confident there are still many issues that remain unsolved, we are apprehensive about claiming to know with certainty what the future holds. There are many questions that still need to be answered, but we can only speculate about which of those questions will prove the most important and which answers will be the most elusive. To avoid wild speculation, therefore, we will focus on issues that are emerging at the time of this writing. That is to say, we can identify issues that have already been identified as important challenges but will still require substantial scholarly work before substantive answers will have been identified. First, we will examine the challenge of secondary findings in the context of longitudinal reanalyses over time. Second, we will turn to polygenic risk scores as an example of the increasing sophistication in genomics that will continue to stimulate scholars to revisit questions about the management of secondary findings in coming decades.

3.1 ­Reanalysis for clinical purposes As discussed in the preceding text, the discourse on secondary findings in research was, in some ways, a sentinel debate that foreshadowed later challenges raised by secondary findings in clinical settings. At some point, however, the debates on s­ econdary findings in research and clinical contexts became enmeshed with one ­another. Even though there is substantial evidence that research and clinical contexts are different enough from one another to justify somewhat different approaches to secondary findings, it increasingly became common for scholars to elide this distinction. In fact, even today when scholars in genetics or ELSI issues discuss s­ econdary findings, it is not always clear whether they have in mind clinical or research contexts. Another example is the application of guidelines for one context to the other. In 2013 the American College of Genetics and Genomics (ACMG) published guidelines for the management of “incidental findings.”2 The authors of the document were very clear that they intended their recommendations to apply to clinical contexts only, but it did not take long for dimensions of the recommendations to be applied in research contexts as well. In fact, nearly every chapter of this book on secondary findings in research references these 2013 ACMG recommendations intended for secondary findings in clinical contexts.

207

208

CHAPTER 9  Secondary findings: Building a bridge to the future of ELSI

One context where the blurring of this distinction has become important is the question of when sequencing results from ES and GS should be reanalyzed. In clinical contexts, laboratory tests are typically analyzed, and the results are reported to the ordering healthcare provider. The result remains in the patient’s chart, but the biosample it was derived from is typically discarded. At that point the laboratory test is “complete.” Something similar takes place in research contexts. Researchers collect biosamples from participants and then perform their research analyses. Depending on the study design, they might disclose some of these findings to participants or their healthcare providers. At that point the participant’s participation in the study is often considered “complete.” As we have discussed, however, scientific knowledge about the association between genetic variants and medical conditions is evolving rapidly. If a genetic result is disclosed to a patient or research participant today, it is quite possible that conclusions about the meaning of that result would be different if the result were to be disclosed to a different patient or research participant in 1 year. According to one recent analysis, as an example, around 10% of genetic results believed today to be the likely explanation for a medical condition (referred to as a “likely pathogenic” result in this field) will eventually prove not to be the cause of that medical condition.3 The evolving nature of genomic knowledge has created questions in both clinical and research contexts about whether researchers and clinical laboratories should retain raw genomic data and then reanalyze it periodically. The idea is that if the interpretation of a result were to change, or if a result not previously discoverable were to later be discovered, then this information could be provided to the patient or research participant. This practice could certainly prove useful, especially for patients or research participants who had previously received results that would later have been proven inaccurate. It also raises a number of important challenges, including questions about the practicability of this approach and issues of data protection. Although a full discussion of these issues is beyond the scope of this chapter, it will be useful to touch on them briefly to demonstrate how much work still needs to be done. The question of practicability is perhaps the most vexing issued raised by the longitudinal reanalysis of genomic data. In research contexts, funding for research studies typically lasts no longer than 5 years. During this period, it should be possible to reanalyze genomic data and contact participants with updated results, assuming funders like the National Institutes of Health (NIH) make funding available to cover this. It is unclear, however, whether genomics researchers would typically have adequate resources to carry out longitudinal reanalysis after the end of a study. One possible solution for continuing reanalysis after the end of a research study would be to transfer sequencing data to the clinical context so that reanalyses conducted after the end of a research study could be carried out as a part of routine health care. There are institutions that have undertaken this type of transfer from research to clinic. For example, the third iteration of the Electronic Medical Records and Genomics (eMERGE) Network involved the disclosure of research results to clinicians so that these results could be used in clinical contexts.4 Despite some success with this approach in well-resourced academic medical centers, significant work

3 ­Secondary findings and the future of ELSI scholarship

would be needed to make this approach sustainable and scalable. In many countries with robust data protection laws, the transfer of data from a research context to a clinical context is limited by legal standards. In addition, there is very limited infrastructure worldwide to facilitate the routine transfer of research data to clinical contexts. Programs like the German Medical Informatics Initiative have begun work to address these infrastructure barriers,5 but it is clear that a significant amount of work still needs to be done to make this transfer a routine part of genomics research. Even data generated in clinical laboratories for clinical purposes face significant challenges. Typically these laboratories operate on a transactional model: they perform the test and report the results to the ordering provider. They are then paid for that work by the patient’s insurance company. Would it be sustainable for clinical laboratories to perform longitudinal reanalysis when they are only paid for performing the initial test? The funding issue is further complicated by legal and infrastructure issues. In most countries throughout the world, electronic medical records are held primarily within the healthcare system where the information was generated. As patients move throughout their lives, there is no infrastructure in place to allow either genomic data or findings from longitudinal reanalyses to “follow” them. Any solution to this problem would need to address privacy concerns, which includes complying with applicable data protection laws. These practical challenges are critical to an ethical analysis of these issues. If researchers or clinical laboratories have a responsibility to reanalyze genomic data over time to update analyses, that responsibility would need to be weighed against the practical constraints that laboratories face. Recent policy statements have begun to address these types of questions. The American Society of Human Genetics (ASHG) published a document in 2019 to provide guidance for genomics researchers6; in the same year the ACMG published a “points to consider” piece to guide clinicians and clinical laboratories.7 However, neither piece was framed as providing recommendations or best practices, reflecting the lack of consensus on many of these issues. A great deal of work is still needed to understand how the practical constraints of laboratories should be weighed with the potential responsibility they may have to update or correct information that had previously been reported to participants.

3.2 ­Polygenic risk scores Another topic that will require substantial scholarly work in coming years relates to the ethical issues raised by polygenic risk scores. Like the questions surrounding longitudinal reanalysis, this is not an entirely new topic. Genetics researchers have known for decades that when genetic factors contribute to the development of common medical conditions like cancer, type 2 diabetes mellitus, and coronary artery disease, this is not typically due to a single gene. Rather, multiple genes interact with one another, and the environment, to contribute to risk for these conditions. The idea behind polygenic risk scores is simply that predictions about individuals’ risk for developing these conditions would be more accurate if researchers and clinicians could account for the contribution of many genes at the same time.

209

210

CHAPTER 9  Secondary findings: Building a bridge to the future of ELSI

However, this is easier said than done. Although researchers have been successful at identifying numerous genes that contribute to risk for developing many common medical conditions,8 it is typically not known how these risk genes interact with one another. If a patient carries variants on two genes that contribute to risk for the same cancer, are those risks additive with one another? Or do they instead increase risk exponentially when found in combination? Or do they cancel one another out? There are numerous factors that contribute to the complexity of these questions. One particularly challenging issue is that as the number of genes involved increases, it becomes increasingly difficult to find individuals who carry all of the relevant combinations of variants. How can scientists predict the risk conferred by five different genetic variants if they have never previously identified an individual who carries that particular combination of variants? There are, of course, many scientists working to solve these technical challenges, including a great deal of work to utilize machine learning and related technologies to improve the predictive accuracy of polygenic risk scores.9 As these approaches become more successful, it will become increasingly possible to incorporate environmental factors alongside polygenic risk scores to identify individuals who are most at risk for developing medical conditions.10 These examples demonstrate ongoing evolution in the way laboratories utilize genomic information to generate “results” that are potentially relevant to the health of patients and research participants. From an ethical perspective, this is a fascinating challenge. Given the evolving sophistication of these laboratory techniques, ELSI scholars will increasingly need to revisit questions that have been addressed in previous iterations of the discourse on secondary findings. For example, when would we know that a polygenic risk score has reached a sufficient level of validation that it would be acceptable to disclose it to research participants as a “secondary finding?” How should the “actionability” of a polygenic risk score be assessed, and under what circumstances would there be an ethical obligation to disclose a polygenic risk score to a research participant? This technical evolution also creates challenging questions about the way genetic information was used in the past. If well-validated polygenic risk scores turn out to contradict predictions based on single genes, how should we approach patients and participants who were given those kinds of “incorrect” results? For example, if a patient decided to pursue preventive mastectomy based on a single risk gene for breast cancer, what should she be told, if anything, about the availability of polygenic risk scores that might reveal her risk was lower than earlier thought? In the next several years, ELSI scholars will need to help answer these questions raised by polygenic risk scores. But this dynamic will likely continue for many years to come. As technical innovations in genomics and other biomedical sciences evolve, the scientific community and society as a whole will need the help of experts in ethical, legal, and social issues who also understand the science. Perhaps even more importantly, we will need ELSI scholars versed in science working together in collaboration with scientists versed in ELSI issues. The work will never end. We should never forget, however, that in many ways, this ongoing work began with challenging questions about how secondary findings should be managed in research settings.

­References

­References 1. Wolf SM. The past, present, and future of the debate over return of research results and incidental findings. Genet Med. 2012;14(4):355–357. 2. Green RC, Berg JS, Grody WW, et al. ACMG recommendations for reporting of incidental findings in clinical exome and genome sequencing. Genet Med. 2013;15(7):565–574. 3. Harrison SM, Rehm HL. Is ‘likely pathogenic’ really 90% likely? Reclassification data in ClinVar. Genome Med. 2019;11(1):72. 4. Gottesman O, Kuivaniemi H, Tromp G, et al. The Electronic Medical Records and Genomics (eMERGE) Network: past, present, and future. Genet Med. 2013;15(10):761–771. 5. TMF – Technologie- und Methodenplattform für die vernetzte medizinische Forschung e.V. Vernetzen. Forschen. Heilen. Forschung stärken, Versorgung verbessern. Medizininformatik. https://www.medizininformatik-initiative.de/de/start; 2019. (Accessed December 18, 2019). 6. Bombard Y, Brothers KB, Fitzgerald-Butt S, et al. The responsibility to recontact research participants after reinterpretation of genetic and genomic research results. Am J Hum Genet. 2019;104(4):578–595. 7. Deignan  JL, Chung  WK, Kearney  HM, Monaghan  KG, Rehder  CW, Chao  EC. Points to consider in the reevaluation and reanalysis of genomic test results: a statement of the American College of Medical Genetics and Genomics (ACMG). Genet Med. 2019;21(6):1267–1270. 8. GWAS Catalog. The NHGRI-EBI Catalog of Published Genome-Wide Association Studies. https://www.ebi.ac.uk/gwas/; 2019. (Accessed December 19, 2019). 9. Pare G, Mao S, Deng WQ. A machine-learning heuristic to improve gene score prediction of polygenic traits. Sci Rep. 2017;7(1):12665. 10. Clarke  TK, Adams  MJ, Howard  DM, et  al. Genetic and shared couple environmental contributions to smoking and alcohol use in the UK population. In: Mol Psychiatry. 2019:https://doi.org/10.1038/s41380-019-0607-x.

211

Index Note: Page numbers followed by f indicate figures, t indicate tables, and b indicate boxes.

­A Actionability, 49–50, 118–119 Additional findings, xix–xxi, 166 disclosure, 15–17t ethical framework, 19–23 four-layer model, 3–4 in medical research, 1–2 in national policies, 180–181t population-based imaging, 4–19 Adoption space, 191–192 Ad populum, 105 Age-related macular degeneration (ARMD), 44–45 Allele frequencies, 68 All of Us Research Program, 162 Alternative sequencing assay, 70–71 Altruism, 107 Amazon, 206 American College of Medical Genetics and Genomics (ACMG) actionability, 52, 81–82 exome and genome sequencing report, 59–60 56 gene-disease pairs, 56 guidelines for incidental findings management, 207 medically actionable genes, 162 policies development, 107–108 Secondary Findings Maintenance Group of, 161 secondary target genes, 161–162 SFv2.0 genes, 69 variants, 52 American Society of Human Genetics (ASHG), 209 Analytical validity, 33 Annotation, 64–67 APOE-ε4, 187 Association of Molecular Pathology (AMP), 52 Australia, genomics, 165–167

­B Barcodes, 63–64 Batch variant calling, 66 Belmont Report, 20, 79 Binary alignment/map (BAM), 187–188 Biobanks, 37–38, 107, 158, 175 Biobank-based research, 100 Biobank participants, 117 Bioethics, 99–100. See also Empirical bioethics

Breaking bad/sad/serious news algorithms, 14, 15–17t, 142 Broad consent, 93–94 Burrows-Wheeler Aligner (BWA), 65

­C Canada, genomics, 171–173 Canadian College of Medical Geneticists (CCMG), 171–172 Carrier status, 47 Cascade testing, 93 Categorical consent, 93 Cell-free DNA fragments, 184 Centers for Medicare and Medicaid Services electronic health record, 146–147 CLIA regulations, 146, 187–188 Clinical actionability, 55, 159 Clinical exome sequencing, 175 Clinical Genome Resource Actionability Working Group, 50 Clinical genomics, 33, 171–173, 182 Clinical laboratories, large-scale DNA sequencing in, 61–64 Clinical Pharmacogenetics Implementation Consortium (CPIC), 46 Clinical Sequencing Evidence-Generating Research (CSER) Consortium, 90 Clinical Sequencing Exploratory Research (CSER) group, 138 Clinical sequencing, secondary findings in, 59–60 Clinical transformation, 163 Clinical utility, 33–34 Clinical validity, 33, 83 Coherentism, 104 Consent, 32–33, 91–94 Consent and Disclosure Recommendations (CADRe) workgroup, 147, 148f Context-related decisions, 169 Contractualism, 20 Council for Realization of Genomic Medicine (CRGM), 176–177

­D Databases, 37–38, 164 Data protection laws, 31 Decisional capacity, 36

213

214

Index

Decision-making broad consent, 93–94 consent models, 91–94 family consent, 93 modular consent, 92–93 staged consent, 91–92 without formal counseling, 89–91 Democratic fallacies, 105 Diagnostic misconception, 8 Disclosing genomic sequencing adolescent scenario, 145b characteristics of study participants, 136–140 guiding principles, 134t laboratory reports of genomic findings, 146–147 medical implications, 139b modality and content, 140–148 notification to participants of sequence result, 137–140 research context, 134–136 research findings, 134t who will return results, 147–148

­E Electronic Medical Records and Genomics (eMERGE) Network, 90, 138–139, 208–209 e:Med program, 167–168 Empirical bioethics, 101–108 acceptability/practicality, 103–104 coherentism, 104 direct justification, 104–106 elicited preferences, 109–110 gaining insights into current moral stances, 102 guideline development, 106–107 identifying problems, 103 intended use of, 102–106 personal and disease related preferences, 110 policies development, 107–108 solidifying/contextualizing accepted norms/ principles, 103 stakeholders’ preferences, elicitation, 111–112 tools/instruments, 111 validity of, 101 Estonia, genomics, 175–176 Estonian Biobank, 175 Estonian Genome Centre, 175 Estonian Health Insurance Fund in 2014, 175 Estonian Human Genes Research Act, 175 Ethical framework, 19–23 Ethical, legal, and social issues (ELSI), xviii, 203, 207–210 Ethical principles, 30–32 Ethics by opinion poll, 105

Ethics experts, 103, 105 Ethics Guidance for Human Biomedical Research, 173 European Society of Human Genetics (ESHG), 157–158 Excess information, 168 Exome sequencing (ES), xviii, xxiii, 60–64, 62t, 70–71

­F Face-to-face disclosure, 14, 122 Family consent, 93 Fastq files, 64–65, 187–188 Flowcell, 63–64 Formal counseling, 89–91 Four-layer model, 3–4 France, genomics, 171

­G Geisinger Health System MyCode Community Health Initiative screening, 183–184 Genes, 49 Gene-disease pairs, 54 Gene list, 55–56 Gene panels, 34–35 General Data Protection Regulation (GDPR), 31, 37, 206 Generation Genome, 164–165 Genetic biobanking, 106 Genetic counseling, 79, 89 Genetic counselors, 89 Genetic Diagnostics Act (GenDG), 31–33, 35–36 Genetics Expert Network for Enterprises (GENE) consortium, 163 Genome Analysis Toolkit (GATK), 65–66 Genome-based noninvasive prenatal testing (gNIPT), 184 Genome sequencing (GS), xviii, xxiii, 60–64, 62t, 70–71 100,000 Genomes Project, 34–35, 163–165, 167–168 Genome-wide association studies (GWAS), 44, 204 Genomic findings, xxiv carrier status, 47 Mendelian disease, 46–47 multifactorial disease risks, 44–45 pharmacogenomics, 45–46 types of, 44–47 Genomic knowledge, 208 Genomic Medicine Multidisciplinary Team (GM-MDT), 164

 Index

Genomics Australia, 165–167 Canada (Quebec), 171–173 in clinical routine care, 189–191 Estonia, 175–176 France, 171 Germany, 167–170 Japan, 176–178 pediatric, neonatal, and prenatal, 183–185 secondary findings, 13–18 Singapore, 173–174 storage and return of raw sequence data in clinical and research settings, 187–189 United Kingdom, 162–165 United States, 160–162 wellness, 185–187 Genomics England Clinical Interpretation Partnership (GeCIP), 163 Genomic sequencing, 29, 108, 156, 165 Genomic sequencing, disclosing adolescent scenario, 145b characteristics of study participants, 136–140 guiding principles, 134t laboratory reports of genomic findings, 146–147 medical implications, 139b modality and content, 140–148 notification to participants of sequence result, 137–140 research context, 134–136 research findings, 134t who will return results, 147–148 Genomics research, xxiii, 88–89, 100 German Ethics Council (DER), 168 German Gene Diagnostics Act, 168 German Medical Informatics Initiative (MII), 167–168, 208–209 German Ministry of Education and Research (BMBF), 167–168 German NaKo National Cohort study, 11 German Society of Human Genetics (GfH), 169 Germany, genomics, 167–170 Germany’s Genetic Diagnostics Act (GenDG), 32–33, 35–36

­H Headquarters for Healthcare Policy (HHP), 176–177 Health Insurance Portability and Accountability Act (HIPAA), 31, 37 Health-related findings, 156 Health technology assessment (HTA), 166 Heidelberg-based EURAT Group, 169–170

High-level moral theory, 78 Hospital Research Ethics Committees, 178 Human Biomedical Research Act, 2015, 173 Human Genetics Society of Australasia, 165 Human Genome Project, 80, 176, 204

­I Illumina sequencing technology, 60–62, 64 Incidental findings, xx–xxi advantages and disadvantages, 15–17t in clinical and research settings, 204–205 definition, 134t, 156–157 four-layer model, 3–4, 15–17t in medical research, 1–2 Singapore, 173–174 Individual findings actionable, 171 children and adults, 36 handling of, 29–30, 32 return of, 30–33 right of access, 37 Individual genomic findings, xxii consent, 32–33 decisional capacity, 36 how to return, 35–36 international data sharing and return, 37–38 international legal and ethical principles, 30–32 planning, 32 secondary findings vs. right of access, 37 what to return (or not), 33–35 Information process, 121–122 Information-seeking trait, 7–8 Informed consent, xxiii, 32–33 accounting for changing information, 86–87 addressing barriers to understanding, 85–86 broad consent, 93–94 complexity, 85 consent models, 91–94 counseling challenges, 85–89 deciding what to disclose, 81, 83–84 education, 85 family consent, 93 genetic counseling, 89 high-quality practices, 84 misconception, 88–89 modular consent, 92–93 offering options and eliciting preferences, 81–83 positive lists, 86–87 providing information, 81–82 respect for autonomy and, 78–81 therapeutic and diagnostic misconceptions, managing, 87–89

215

216

Index

Institutional review boards (IRBs), xxi, 88, 178 International data sharing and return, 37–38 International legal and ethical principles, 30–32 Internationally as Research Ethics Committees (IRBs), 29–30

­J Japan, genomics, 176–178

­L Lane, 63–64 Large-scale DNA sequencing, 61–64 Learning Healthcare System, 88–89, 193 Legally authorized representatives (LARs), 36 Legal principles, 30–32

­M Magnetic resonance imaging (MRI), xvii–xviii MAPQ score, 65 Medical actionability, 33–34 Medical research, additional findings in, 1–2 MediShield Life, 174 Melbourne Genomics Health Alliance, 165–166 Mendelian disease, 46–47 Moderate contractualism, 20 Modular consent, 92–93 Monogenic disorders. See Mendelian disease Moral authority, 105 Multifactorial disease risks, 44–45 Multiplex genetic technology (MGT), xviii, xx, 43 MyCode Community Health Initiative, 137–138, 142, 142b, 162

­N National Health Genomics Policy Framework, 167 National Health Service (NHS), 162–163 National Institutes of Health (NIH), 204, 208 National Precision Medicine Alliance, 173 Negative lists, 55 Negligence, 30–31 Neonatal genomics, 183–185 New Directions: The Ethics of Synthetic Biology and Emerging Technologies, 21 Next-generation genome sequencing (NGS), 156 Nonmaleficence, 20–21 Normalization process theory, 191–192 North Carolina Clinical Genomic Evaluation by Nextgeneration Sequencing (NCGENES) study, 50 North Carolina Newborn Exome Sequencing for Universal Screening (NCNEXUS) study, 50 NovaSeq, 64

­P Paired-end reads, 64–65 Patient-Centered Outcomes Research Institute (PCORI), 146 Pediatric genomics, 183–185 Personal utility, 118 Pharmacogenomics, 45–46, 69–70 PharmGKB, 46 Polygenic risk scores, 209–210 Population-based imaging actionability and clinical utility, 10–12 diagnostic misconception, 8 disclosing/withholding additional findings, 10 false-positive and false-negative findings, 19 genomic secondary findings, 13–18 information-seeking trait, 7–8 participants, preparation, 18–19 positive list, 11–12 quality of information, 5–7 recommended classification, 12t research, 2 research protocol, designing, 19 to tell/not to tell, 7–10 Population-based studies, 4–5, 29 Preference utilitarianism, 105 Prenatal genomics, 183–185 Prima facie, 105 Primary finding, xx Principles for the translation of omics-based tests from discovery to health care, 165 Principles of Biomedical Ethics (Beauchamp and Childress), 79 Public Health Genomics (PHG) Foundation, 163–164 Public Population Project in Genomics (P3G) Consortium, 50–51

­Q Quality-adjusted life years (QALYs), 190 Quality control (QC) metrics, 65 Quebec Center for Clinical Genomics, 172–173 Quebec, genomics, 171–173 Quebec Ministry of Health and Social Services (MoHSS), 172–173

­R Reportable secondary finding actionability, 49–50 balancing risks and benefits, 53–54 confirmation and reporting, 48 genes, 49

 Index

participant and study factors, 50–52 variants, 52–53 Research context, 134–136 Research ethics committees (RECs), 88 Research findings, 134t Research results, 134t Right to know information, 31–32 Routine genomic screening, 38 Royal College of Pathologists of Australasia, 165

­S Secondary findings, xx–xxi in clinical sequencing, 59–60 as cultural crossroads, 206–207 definition, 156–157 economic dimensions of returning, 189–191 and ethical, legal, and social issues (ELSI) scholarship, 207–210 face-to-face meetings, 116 false-positive/false-negative results, 115–116 in genomics research, 100 identification, 61f impacts and implications of, 115, 120–121 information process, 115–116, 121–122 in national policies, 180–181t participants’ and lay attitudes and preferences, 117–123 polygenic risk scores, 209–210 potential categories of, 69–70 preferences for, 113–114, 118–120 reanalysis for clinical purposes, 207–209 reportable actionability, 49–50 balancing risks and benefits, 53–54 confirmation and reporting, 48 genes, 49 participant and study factors, 50–52 variants, 52–53 return of results, 204 vs. right of access, 37 rights, responsibilities, policies, and practices, 116–117, 122–123 as sentinel debate, 204–205 terminology definition, 134t variant filtration to detect/avoid, 67–69 Sequence read alignment, 64–67 Shared decision-making, 79 Singapore, genomics, 173–174 SingHealth Duke-NUS Institute of Precision Medicine (PRISM), 174 Single-gene disorders. See Mendelian disease Single-labeled nucleotide, 64

Single nucleotide polymorphisms (SNPs), xviii, 44–45 Single-nucleotide variants (SNVs), 65–66 Single-sample variant calling, 66 Soft law, 29 SouthSeq, 90 Special relationship, 31 Stakeholders’ preferences attitudes and perspectives, 112–123 normative ethical questions, 101–108 strategies and tools for eliciting preferences, 109–112 when to elicit, 111–112 Standard of care, 30–31, 160–161

­T Tiered consent, 93 Tohoku Medical Megabank, 177 Translational genomics research, 173 Translational research, 178–179 Tri-Council Policy statement on the Ethical Conduct for Research Involving Humans, 172 Tuskegee study, 88

­U UK Biobank, 158 United Kingdom, genomics, 162–165 The United Nations Convention on the Rights of Children, 109–110 United States, genomics, 160–162 US National Academies of Science, Engineering, and Medicine (NASEM), 162

­V Variant call format (VCF) file, 60–61, 65–66, 187–188 Variant of uncertain/unknown significance (VUS), 52–53, 87, 156 Variants calling, 64–67 filtration, 67–69 secondary finding, 52–53 validation via alternative sequencing assay, 70–71 Victorian Infant Hearing Screening Program, 166–167 Voluntary consent, 32–33

­W Wellness genomics, 185–187 Whole exome and whole-genome sequencing techniques (WES/WGS), 61–62, 77, 80–81, 167, 183 Whole fetal genome, 184

217