An Assessment Guide To Geriatric Neuropsychology 9780805819915, 0805819916, 9781410602916, 1410602915

Table of contents :
Content: Contents: Preface. Issues in Psychogeriatric Assessment. Psychometric Issues. Screening Instruments for Cognitive Impairment. The Wechsler Adult Intelligence Scales. Measures of Intellectual Functioning and Neuropsychological Batteries. Memory Assessment. Attention and Executive Functioning. Language, Visuospatial, Perceptual, and Motor Functioning. Assessing Psychopathology. Age-Associated Conditions Affecting Cognition. The Role of the Caregiver in Neuropsychological Assessment. Ethics in the Assessment of Elderly Persons.

Citation preview

An Assessment Guide to Geriatric Neuropsychology

Holly Tuokko Thomas Hadjistavropoulos

An Assessment Guide to Geriatric Neuropsychology

jy.-si

This page intentionally left blank

An Assessment Guide to Geriatric Neuropsychology

H olly Tuokko University of Victoria and Elderly Outreach Service, Ministry of Health, British Columbia

T hom as Hadjistavropoulos University of Regina

Ip JL

Psychology Press Taylor & Francis Group New York London

First Published by Lawrence Erlbaum Associates, Inc., Publishers 10 Industrial Avenue Mahwah, New Jersey 07430 Transferred to Digital Printing 2009 by Psychology Press 270 Madison Ave, New York NY 10016 27 Church Road, Hove, Fast Sussex, BN3 2FA Copyright © 1998 by Lawrence Erlbaum Associates, Inc. All rights reserved. No part of this book may be reproduced in any form, by photostat, microfilm, retrieval system, or any other means, without the prior written permission of the publisher. Cover design by Kathryn He

Library of Congress Cataloging-in-Publication-Data TuokJko, Holly An assessment guide to geriatric neuropsychology / Holly Tuokko and Thomas Hadjistavropoulos. p. cm. Includes bibliographical references ar.d index. ISBN 0-8058-1991-6 (doth : alk. Paper) 1. Geriatric neuropsychology. 2. Cognition dis­ orders in old age—Diagnosis. 3. Neuropsychologi­ cal tests. I. Hadjistavropoulos, Thomas. II. Title. [DNLM: I . Cognition—in old age. 2. Geriatric Assessment. 3. Neuropsychology—in old age. WT 145 T926a 1997] RC451.4.A5T895 1997 618.97’680475—dc21 DNLM/DLC for Library of Congress 97-34403 CIP

Publisher’s Note The publisher has gone to great lengths to ensure the quality of this reprint but points out that some imperfections in the original may be apparent.

To Don and Aaron — H .T .

To Heather and Nicholas —T. H.

This page intentionally left blank

Contents

Preface

ix

1

Issues in Psychogeriatric Assessment

2

Psychometric Issues

11

3

Screening Instruments for Cognitive Impairment

28

4

The Wechsler Adult Intelligence Scales

44

5

Measures of Intellectual Functioning and Neuropsychological Batteries

75

6

Memory Assessment

90

7

Attention and Executive Functioning

123

8

Language, Visuospatial, Perceptual, and Motor Functioning

144

9

Assessing Psychopathology

161

Age-Associatcd Conditions Affecting Cognition

181

10

1

vii

viii

CONTENTS

11

The Role of the Caregiver in Neuropsychological Assessment

205

12

Ethics in the Assessment of Elderly Persons

222

References

232

Author Index

263

Subject Index

277

Preface

More and more studies of the performance of older persons on various neuropsychological measures have been reported in the literature over the past decade in response to the growing demand for clinical practitioners to become involved in the assessment of the elderly. These studies, and others emerging from the area of cognitive aging research, indicate that issues such as educational background are of particular concern when assessing this cohort, and that some cognitive domains are more susceptible than others to the effects of aging. Despite the appearance of more normative information for older persons, we know of no single source in which the findings to date are summarized. This text is designed to fill the gap. Changes in cognitive functioning occur as a result of aging. Similarly, the prevalence of disorders that adversely affect cognitive functioning increases with age. Neuropsychological measures of cognitive abilities are particularly useful for detecting and monitoring changes in cognitive functioning but only within the context of adequate normative information. This is particularly true when assessing older persons for whom changes in cognitive functioning may or may not be reflective of underlying “pathology.” Neuropsychological assessment of geriatric populations involves not only performance-based measurement of areas of cognitive abilities but heavy reliance on reports of caregivers—both formal and informal—about the functioning of the affected individual. Despite the importance of this information, few standardized and/or validated approaches exist in the field. Those that are available have appeared in scattered reports in the literature. In this text, we bring together relevant information concerning measures designed to obtain information from caregivers. A related issue

X

PREFACE

is the psychological well-being of caregivers and a brief summary is also given of measures useful in the assessment of caregiver well-being. Another important yet neglected area is that of the ethics involved in conducting geriatric neuropsychological assessment. The issues discussed in this text include informed consent, confidentiality, the right to autonomy and self-determination, and issues pertaining to the provision of feedback to clients. Guidelines for the resolution of ethical dilemmas are also provided. ACKNOW LEDGMENTS The preparation of this text reflects the integration of research findings with day-to-day clinical practice issues. Our exposure to these issues within the context of a teaching hospital environment at the Clinic for Alzheimer Disease and Related Disorders at Vancouver Hospital and Health Sciences Centre formed the foundation for our journey into the preparation of this text. We owe our thanks to Dr. Lynn Beattie, Director of the Clinic, for providing a stimulating and fostering environment to staff and students. The support and encouragement provided by Dr. Ian McDowell, Barbara Heliiwell, and Betsy Kristjansson, members of the Canadian Study of Health and Aging Coordinating Office, is much appreciated. Data from the Canadian Study of Health and Aging is reported throughout the text. This project was funded by the Seniors Independence Research Program, administered by the National Health and Research Development Program of Health and Welfare Canada. The study was coordinated through the University of Ottawa and the Canadian government’s Laboratory Centre for Disease Control. We also thank our dedicated, astute, hard-working assistants—Robert Frerichs, Tamara Goranson, Jocelyn Robinson, and Karen Eso—without whom the task would have been less fun and less rewarding. Finally, we acknowledge our families, who were patient and understanding of our need to make this contribution to the field. — Holly Tuokko — Thomas Iladjistavropoulos

Chapter

1

Issues in Psychogeriatric Assessment

T he field o f clinical neuropsychology is broad-ranging, covering diverse problem s and populations. A lthough the field as a whole has experienced trem endous growth over the past two decades, one o f the fastest growing areas w ithin clinical neuropsychology is the assessment o f persons over age 65, or geriatric assessment. T h ere are a num ber o f reasons for this developm ent. First, changes in the dem ographic structure o f the N orth American population have been o ccurring over the last century such that larger and larger p ro p o rtio n s of the population are living past age 65. A variety o f factors, m ost related to im proved health status, are contributing to this change. Im provem ents in public health (e.g., sanitation, sewage treatm ent, nutritional know ledge), the introduction o f antibiotics and im m unizations in the m iddle o f this century, and the developm ent o f effective interventions for the leading causes o f death (e.g., cardiac and cerebrovascular disorders and cancer) are all contributors to this dem ographic shift. M oreover, in the years fol­ lowing W orld War II, a record num ber o f births were recorded in N orth America— the “baby boom .” These “baby boom ers” are now moving toward age 65 and the consequences of the “graying o f the baby b o o m ” has been the source o f m uch econom ic and political debate (e.g., Barer, Evans, & H ertzm an, 1995; D enton & Spencer, 1995), particularly in the context of h ealth and health-related issues. Age-related changes have been docum ented in nervous system structure an d function as well as o th e r physical systems such as vision, hearing, gastrointestinal functions, and the musculoskeletal system (e.g., B irren, Sloane, & Cohen, 1992; Meneilly & Tuokko, 1994). These norm al changes 1

2

CHAPTER 1

in biological, psychological, cognitive, and behavioral systems seem to occur at differing rates in different persons and result in increased variability on m ost measures of functioning with increasing age. Moreover, certain dis­ orders or conditions appear correlated with age or occur commonly in older age groups. Comorbidity of physical illness and m ental disorders (e.g., anxiety is often seen with cardiovascular, gastrointestinal, o r pulm o­ nary disease) is also m ore com m on in an aged population than in younger persons (Birren et al., 1992). Thus, distinguishing the anticipated effects o f aging from the effects of diseases that arc more prevalent with age within the context of com orbid processes is a very im portant and chal­ lenging task. To fulfill their role, practitioners working with elderly persons require a comprehensive knowledge of processes associated with norm al aging (gerontology), as well as a current understanding of the pathological proc­ esses associated with aging (geriatrics). Each of these areas has expanded rapidly, most notably since the early 1980s, and is continuing to grow with the realization of, and speculation as to, the impact of the dem ographic changes within the N orth American society. This book summarizes the areas of knowledge required to address issues specific to elderly persons. Although many of the principles of neuropsychological assessment pre­ sented elsewhere (e.g., I.ezak, 1995) may be com m on to all client popu­ lations, the focus here is on the specific issues that may influence the choice of assessment techniques or interpretation of test findings for per­ sons over age 65. Neuropsychological assessments are typically requested when brain dysfunction is suspected or known to exist. The assessment may focus primarily on concerns over diagnostic issues, docum entation of change over time, or care planning and m anagem ent of behavior. The purpose of this book is to bring together the available normative inform ation for persons over age 65 and to highlight the conceptual and methodological issues pertinent to the use of this inform ation. For exam­ ple, the variety of factors specific to the individual that may influence test perform ance or interpretation are raised in the rem ainder of this chapter. Definitions of what constitutes a norm al sample and procedures for deter­ mining norms are discussed in chapter 2.

FACTORS ASSOCIATED WITH THE AGING PROCESS A num ber of issues specific to the population under investigation must be taken into consideration when selecting and interpreting psychological measures. These include the cognitive, physical, and social processes asso­ ciated with norm al aging.

ISSUES IN PSYCHOGERIATRIC ASSESSMENT

3

Cognitive Processes Cognitive and behavioral changes compatible with norm al aging should not be m isinterpreted as being indicative of abnormal or pathological change in functioning. Although this may seem obvious, it is extremely im portant that practitioners appreciate the complexity of this distinction. Diseases commonly manifested by elderly persons may be superim posed on normal age-related changes. Normal age-related changes influence not only the presentation of the disease, but the expectation of response to treatm ent and the likelihood of potential complications. Cognitive aging research has shed much light on the types of changes in functioning that occur with advancing age. The introductions to chapters 4, 5, 6, and 7 address age-related expec­ tations for various domains of cognitive functioning (i.e., intelligence, memory, attention, language, visuospatial, perceptual, and m otor skills). Briefly, the most striking agc-relatcd changcs observed in cognitive func­ tioning include: 1. Measures dependent on m otor speed (e.g., reaction time, tapping, etc.) arc perform ed less well by older persons. Slowed mental processing appears as the most im portant com ponent in the characteristic behavioral slowing associated with normal aging (Lezak, 1995). Hence, any task that contains decision points, initiation or redirecdon of movement may be slowed and will become pardcularly apparent when timed. 2. Measures of the ability to focus on a simple task and perform it without losing track of the task (e.g., Digit Span Forward) are perform ed well by most older adults (M. S. Albert, 1988). 3. Semantic knowledge changes significantly with age. Age-related de­ clines have been shown on measures of nam ing (M. S. Albert, Heller, & Milberg, 1988; Borod, Goodglass, & Kaplan, 1980; Goodglass, 1980; LaBarge, Edwards, & Knesevich, 1986) and verbal fluency (i.e., generation of words within a time period; M. S. Albert et al., 1988; O bler & M. L. Albert, 1981; Spreen & Benton, 1969). These changes appear primarily after age 70 while other linguistic abilities appear to rem ain relatively intact (M. S. Albert, 1988). 4. Manifestations of changes in memory functioning are d ep en d en t on the type of memory task involved. The ability to retain small am ounts of inform ation over a brief period of time (once called short-term memory) shows hardly any loss with age (Talland, 1965). Secondary, or long-term, memory declines with age. The degree of loss is related to the type of material to be rem em bered and the assessment method. Large age-related declines are found in free recall (Botwinick & Storandt, 1974; Craik, 1977;

4

CHAPTER 1

Gilbert Sc Levee, 1971; Kausler 8c Lair, 1966). Decrements are larger for recall than for recognition of material (Erber, 1974; Harwood 8c N'aylor, 1969; Howell, 1972). 5. Visuospatial ability, as assessed by complex visual reproduction and recognition tasks (e.g., Block Design, Figure Integration), declines as in­ dividuals age (Doppelt 8c Wallace, 1955; Wentworth-Rohr, Mackintosh, 8c Flalkoff, 1974). Similarly, drawing task perform ance (e.g., three-dimensional cube, clock face) also appear to decline with age (Plude, Milberg, 8c Cerella, 1986; Tuokko, I ladjistavropoulos, Miller, H orton, & Beattie, 1995). This seems to be true even when the speed com ponents of the tasks are removed (Botwinick, 1977). 6. Abstraction and conceptualization have been assessed in a variety of ways many of which show age-related declines. O f particular interest is proverb interpretation, which shows substantial deterio ratio n with age (M. S. Albert, 1988; Bromley, 1957). This research emphasizes the need for normative inform ation based on age. Norms can help determ ine what constitutes a change in functioning above and beyond age-related change. M. S. A lbert (1981) n o ted “ageappropriate norms based on a systematic comparison between elderly nor­ mal and pathological populations do not exist for most behavioral tests of brain dam age” (pp. 385-386). This situation has changed with the increas­ ing dem and for this information. From the literature on age-related cognitive changes, issues concerning the appropriateness o f various measures for use with older persons have been raised. The utility of many standard psychological measures for re­ lating to the perform ance of everyday behaviors by older persons has been challenged and a need for m ore “ecologically valid” measures was identi­ fied. Moreover, it has been observed that elderly persons may perceive the measures as “meaningless” within the context of their lives and may be unwilling to take part in an assessment. This controversy has resulted in the recent development of measures incorporating real-life situations into the context o f the assessment process (e.g., shopping list learning as part of a memory battery), thereby expanding the types of assessment tools available for this population. Physical Processes In addition to these age-related cognitive changes, there are a num ber of physical factors that might interact with, influence, or distort the clinical picture. Most notably, changes in sensory processes (i.e., vision and hear­ ing) are com m on with increasing age and may significantly interfere with a person’s perform ance on measures of cognitive functioning. Most persons

ISSUES IN PSYCHOGERIATRIC ASSESSMENT

5

over age 60 experience some form o f visual com prom ise (Fozard, 1990). Declines in hearing parallel those o f vision and approxim ately 70% of persons in the 71- to 80-year-old age range suffer some hearin g loss (Fozard, 1990). It should be noted, however, that even persons who are legally blind can perform well on m any neuropsychological measures containing visual stimuli, because these stimuli are often large and clearly defined. Despite this, the clinician m ust be cognizant o f th e possible effects o f vision and hearing loss on test perform ance and ensure that these influences have been controlled (e.g., ensure individuals wear their glasses o r h earin g aid; use o f pockettalker to assist in com m unicating with h earing im paired) or com pensated for by use o f m easures not d ep e n d en t o n the im paired sen­ sory modality (e.g., use o f a verbal m easure o f m em ory with a visually im paired person). A n o th er potential com plication to test perform ance an d in terp retatio n when w orking with this population is the higher prevalence o f medical problem s than seen in younger individuals. Thus, in addition to norm al changes in physiological functioning associated with age, the prevalence o f a variety o f medical conditions increases with age (e.g., diabetes, h eart disease, arthritis). Many o f these disorders may be related to observed cognitive im pairm ent in older persons and becom e im p o rtan t w ithin the co n tex t o f differential diagnosis. For exam ple, p oor m etabolic control in diabetics may greatly increase the risk o f vascular com plications, including retinopathy, nephropathy, and neuropathy (C. M. Ryan, 1988). Observed cognitive deficits may be a function of these micro- an d m acrovascular changes rath er than suggestive o f an additional disease process (e.g., Alzhei­ m e r’s disease). Too often, the role o f existing medical disease in disrupting cerebral functioning is overlooked an d deficits are attributed to o th er sources o f cognitive im pairm ent. O th er disorders, such as arthritis, may lim it perform ance on tasks requiring g raphom otor skill o r speed o f p er­ form ance, thereby com plicating test interpretation. A related issue is the potential effect o f m edications o n test perform ance. It is n o t uncom m on for elderly persons to be taking a variety o f m edications for com orbid disorders. Moreover, older persons have increased sensitivity to m edications due to altered abilities to m etabolize and excrete m edica­ tions (e.g., Birren et al., 1992). As drug sensitivity and interactions are often specific to individuals, the potential com plicating role o f m edications m ust be kept in m ind. Thus, it is extrem ely im portan t to d eterm in e the m edications a person is taking at the tim e o f the assessment, as well as the dosages. T he most effective way to accom plish this is to have the person o r a family m em ber bring in all m edications for review. It may be necessary to cou n t the tablets to ensure that m edications arc being taken as p re­ scribed. Inappropriate use o f prescription (and nonprescription) m edica­ tions may contribute to the clinical presentation and needs to be addressed.

6

CHAPTER I

N oncom pliance with m edication regimes may result from an inability to m o n ito r the m edications due to m em ory deficits rath er than unwillingness to comply. To ensure optim al functioning, it may be necessary to m onitor the individual’s perform ance as m edications are intro d u ced or withdrawn. It is com m on knowledge that older persons, particularly those with medical problem s, may tire easily. H ence, to obtain an estim ate of m aximal functioning, it may be necessary to lim it the length o f testing sessions o r provide frequent breaks during the testing. This may be do n e by m aking it clear to the person that rest breaks are available as need ed , o r checking with the person at intervals during the assessment as to the n eed for a break. Increasing agitation o r withdrawal from responding may signal that a rest break, o r redirection to casual conversation, is w arranted. For persons who arc very frail, such as those in care facilities, it may be necessary to schedule several sessions to obtain a perso n ’s optim al perform ance. Ex­ am ination o f test results in the context of the o rd er o f adm inistration may be o f assistance, if there is any concern that the person was fatigued by the assessment process. If there appears to be m arked variability between the first and last m easures adm inistered, then it may prove beneficial to rcadm inistcr the latter ones on an o th er occasion to ensure fatigue was n o t the source o f the difficulties. It is also im portant to view the tendency toward fatigue within the context o f the perso n ’s activities o f daily living. An individual who cannot sustain enough focused attention to com plete a 30-minute assessment procedure may have great difficulty perform ing daily activities when living alone in the com munity. Similarly, individuals living in a care facility can n o t be expected to successfully take p art in group recreational activities if they are unable to com plete a sim ple m ental status exam ination w ithout becom ing unduly fatigued. Social Processes W hen assessing older persons, the social context o f the individual m ust be kept in m ind. A lthough this is true when conducting neuropsychological assessments at any age, issues specific to old er persons ce n ter aro u n d social change, loss, and the context o f the psychological assessment. R etirem ent may be a m ajor adjustm ent for som e persons and can result in p ro fo u n d feelings o f loss and isolation. C hange of residence may result in loss of friends and family, social supports, o r a sense o f belonging. It is com m on for older persons to have friends and family m em bers who have recently died or are ill. Certainly grieving an d bereavem ent are processes with pervasive effects on a person’s functioning. Alternatively, the death o f a spouse o r caregiver may bring to light the p oor cognitive functioning o f the rem aining partner, which, in the context o f the struggle to assume new roles, com es to the attention o f others.

ISSUES IN PSYCHOGERIATRIC ASSESSMENT

7

Financial limitations a n d /o r sociocultural issues may also complicate the clinical picture and n eed to be considered in the context o f differential diagnosis and planning care. Many persons of this age co h o rt may lack familiarity with the types o f assessment procedures conducted by n eu ro ­ psychologists, and fear o f loss o f ability o r independ en ce may com plicate the assessment process. Careful explanations as to the in ten t and purpose of the assessment may be required to ensure maximal perform ance.

PURPOSE OF THE ASSESSMENT

Diagnosis A neuropsychological assessment may be requested for a variety of reasons. O ften, diagnostic issues are of primary concern. Identification o f dem entia is perhaps the m ost com m on referral issue in this age group. D em entia may be defined as an overall decline in m enial capacity (one o r m ore cognitive dom ains) that renders the individual unfit to m eet the diverse intellectual dem ands associated with the obligations o f everyday life. W ithin this diagnostic category, there may be a wide variety o f individual patterns o f cognitive disability. D em entia has been subclassified in a variety o f ways: according to most prom inent cognitive features (e.g., am nesic, aphasic, visuoperceptive, global), according to anatom ical location (e.g., cortical, subcortical, axial; Joynt 8c Shoulson, 1979), according to reversibility of the underlying etiologic condition (reversible conditions may include Nor­ mal Pressure H ydrocephalus, d rug toxicity, thyroid dysfunction, n eu ro ­ syphilis, B12 deficiency, liver failure; irreversible conditions may include A lzheim er's disease, vascular disorders, alcohol-related dem entia, H u n t­ in g to n ’s disease, P arkinson’s disease, A m yotrophic Lateral Sclerosis), and severity o f functional deficits (i.e., mild, m oderate, severe). In addition to identifying cognitive im pairm ent (i.e., g reater than n o r­ mal age-related decline), it is necessary to differentiate possible dem entia from o th e r com m on disorders affecting older persons. These include de­ pression, acute confusional state, and cognitive changes associated with a variety o f rem ediable medical conditions. Moreover, it is possible th a t there may be m ore than o n e condition present. For exam ple, it has been n oted th a t depression may be present in various neurological conditions (e.g., poststroke, A lzheim er’s disease, Parkinson’s disease). T h e identification o f coexisting (com orbid) conditions is o f im portance as treatm en t for rem e­ diable disorders may improve the perso n ’s quality of life. W hen addressing diagnostic issues, the factors described previously m ust be taken into con­ sideration.

8

CHAPTER 1

Baseline Test results may also serve as a baseline to m onitor change with intervention or the passage of time. In this context, repeated adm inistrations of the measures would be required. In these situations, issues concerning test-retest effects may need to be addressed and it may be pru d en t to select measures with multiple forms. Care Planning T he assessment results may also play a role in care planning where a careful, detailed delineation of the cognitive strengths and weaknesses of the individual may be translated into m anagem ent strategies and suggestions for care approaches. As yet, there is little research addressing the link between pattern of cognitive abilities and specific approaches/strategies that are most effective. However, there is no question that understanding a person’s capabilities and limitations is essential to a rational care plan. Being able to articulate for individuals and their care providers the nature of the cognitive deficits (e.g., problem retrieving inform ation) is the first step in encouraging creative problem-solving for care providers. For example, a useful premise for assisting care providers in developing problem-solving strategies for persons with dem entia (Tuokko & Purves, 1993) is that persons with dem entia act on incomplete inform ation. T hat is, due to cognitive limitations (e.g., new learning problems or inability to self-evaluate), the individual is only processing part of the inform ation in the environm ent. O nce the nature of the deficit is clarified, care providers can actively seek strategies to help the person with dem entia accomplish tasks. This m ight include restructuring the environm ent, reinterpreting the behavior o f the individual, or modifying expectations o f the individual with dem entia. O ther ways of using assessment inform ation in developing care plans have been discussed by S. H. Zarit and J. M. Zarit (1983). For each of the assessment goals defined earlier, it is necessary to assess a wide range of domains of cognitive functioning, including attention and concentration, m otor functioning and praxis, memory, receptive and ex­ pressive language skills, reasoning and thinking processes, and visuospatial and verbal reasoning abilities. The assessment may be brief (e.g., screening instrum ent; see chap. 3) or more lengthy depending on the nature of the issues under investigation. Generally, the inform ation contained in Table 1.1 would be obtained in one form or another so that issues central to differential diagnosis and care planning could be addressed. The measures selected within each dom ain depend on the context of the evaluation and the preferences of the neuropsychologist. Measures for use in care planning for a 95-year-old frail woman living in a nursing home may be very different from those selected for use with a 65-ycar-old, recently

ISSU ES IN PSY CH O G ERIA TRIC ASSESSM ENT

9

TABLE M Assessment Information A. T e stin g A ro u sal level A tte n tio n L an g u ag e i) ii)

C o m p re h e n s io n E xp ressio n (n a m in g , re p e titio n , fluency, p h ra se le n g th , re sp o n s e sp e e d , w ord fin d in g , p a ra p h a s ic e rro rs [literal: s o u n d s like; se m a n tic : sim ilar m e a n in g ]) M o to r i) M ov em en ts (facial, gait, tre m o r) M e m o ry i) N ew le a rn in g ii) V erbal, visual iii) P erso n al in fo rm a tio n iv) R e m o te m em o ry o r in fo rm a tio n le a rn e d in th e p ast V isuospatial re a so n in g V erb al re a so n in g M ood B. C o llateral in fo rm a tio n fro m family H isto ry i) O n se t ii) D u ra tio n iii) C o u rse iv) D e m o g ra p h ic in fo rm a tio n (e d u c a tio n , o c c u p a tio n ) v) L iving situ a tio n , h o b b ies, a lc o h o l c o n s u m p tio n vi) M edical c o n d itio n (tn ed s, su rg eries, an e sth e tic s, p s y c h ia tric /n e u r o ­ logical status) P re se n t fu n c tio n in g ( p re s e n c e /a b s e n c e a n d d u ra tio n o f sy m ptom s) i) P ersonality ii) Everyday tasks iii) L an g u ag e skills iv) M e m o ry fu n c tio n s v) Self-care fu n c tio n s

retired business executive living alone, who may be exhibiting the early signs o f dem entia. T hroughout the assessm ent process, various hypotheses about the nature o f the deficit and underlying disorder may be generated and exam ined. For exam ple, if a person is unable to follow a simple written instruction, then it may be that vision is im paired, the person never was able to read, or the person can no longer read despite intact vision but can follow the same instruction given orally. In addition to the material gathered through direct assessment, it is necessary to obtain inform ation from a collateral inform ant (usually a family m em ber or spouse) to ensure a com plete picture is obtained (see chap. 10). Particularly in suspected cases o f dem entia, individuals may be unaware o f changes in their behavior that are readily apparent to others (McGlynn Sc Kaszniak, 1991). It is rarely the case that persons with dem entia seek assessm ent for themselves. Typically, it is family or friends w ho make their

10

CHAPTER 1

concerns known to th e prim ary care physician o r others o n b eh a lf o f the affected individual. Material obtained from inform ants may focus on prob­ lems observed in everyday functioning, but would also clarify the nature of the onset, duration, and course o f the presenting problem s. Moreover, infor­ m ation obtained from inform ants may include an appraisal o f the individu­ als’ living situation and th eir capability to function w ithin th e ir personal en vironm ental context. All inform ation gathered is exam ined in relation to the individuals’ p rem orbid level o f functioning (see chap. 4), environ­ m ental context, an d m ust include as m uch relevant m aterial as possible. Assessment over intervals of time is also o f particular im portance as conditions may evolve gradually (e.g., A lzheim er’s disease), may fluctuate (e.g., Lewy Body dem entia), o r may improve (e.g., treated conditions or stabilization poststroke). A lthough there are established sets o f criteria for m any disorders affecting elderly persons (see chap. 9), subtle cognitive deficits may be observed that do not satisfy existing criteria b u t may be associated with em erging conditions. As yet, these mild deficits in cognitive functioning are poorly understood and only repeated assessment can de­ term ine the eventual outcom e. In addition, neurodegenerative disorders progress over time, but the rate o f progression and the individual features th a t em erge may differ substantially between persons. M onitoring change over tim e can be o f particular im portance in the design o f care plans appropriate to the individual’s specific strengths an d weaknesses at the time o f assessment. T he elem ents o f th e care plan may n eed to shift and be altered as the characteristics o f the affected individual change. CONCLUSIONS W hen assessing older persons, clinicians require a com prehensive knowl­ edge o f processes associated with norm al aging (gerontology), as well as a cu rren t understanding o f the pathological processes associated with aging (geriatrics). It is im portant to evaluate the influences th at a variety o f factors may have on test perform ance. These include sensory deficits, physi­ cal lim itations, and sociocultural issues. Many areas o f cognitive functioning m ust be assessed to determ ine w hether o r n o t cognitive im p airm en t is p resent, to differentiate between conditions that may give rise to cognitive im p airm ent in elderly persons and to provide m eaningful contributions to the care plan for individuals. M onitoring change over tim e is often necessary as conditions (i.e., underlying disorders) and the needs o f the individual may evolve. Supplem enting test inform ation with observations o f functioning at hom e o r in social contexts and inform ation ab o u t the n atu re o f the onset, duration, and course o f the behavioral changes o b ­ tained from o th e r sources (e.g., family o r friends) is crucial for diagnostic an d care planning purposes.

Chapter

Psychometric Issues

Psychometric theory underlies test construction and interpretation but practicing psychologists often have inadequate or obsolete training in meas­ u rem e n t theo iy (Retzlaff & G ibertini, 1994). U nfortunately, clinical neuropsychological assessment is far less precise and sophisticated than some other areas of assessment (e.g., educational or personnel testing). In response to the concern about the appropriateness o f measures for clinical use, standards for the developm ent of new measures have been developed (American Psychological Association, 1985). The influence o f these standards is apparent in some of the newer measures discussed later in chapters 4 and 5. However, the same attention to psychometric issues has not been afforded to many of the older, m ore commonly encountered, neuropsychological measures that have been extended to include samples o f older persons. Available psychometric inform ation has been included in the descriptions of specific instruments (see chaps. 3-8, 11) when this inform ation was derived from samples of elderly persons. A discussion of the general psychometric properties of specific measures can be found in Spreen and Strauss (1991) or Lezak (1995). This chapter raises, more generally, issues that need to be considered by the clinician when interpreting individual test results. It is im portant that clinicians be aware of the limits to inferences that can be made from test perform ances as the consequences of m isinterpretations are often extremely powerful for individual persons. Geriatric neuropsychological assessment often centers on the identification of deficits a n d /o r deterio­ ration in cognitive functioning. It also contributes to the process o f deter­ m ining why this deterioration may be occurring. To identify deficits or

11

12

CHAPTER 2

deterioration, som e ideal m ust exist against which the individual's p er­ form ance can be m easured. This com parison standard can be norm ative (derived from perform ance on the m easure by a representative group of persons) o r individual (in relation to the individual’s own history o r p resen t characteristics). This chapter focuses on normative issues an d provides de­ scriptions of the best available norm ative samples. It also discusses funda­ m ental characteristics o f the m easures themselves (e.g., reliability an d va­ lidity). A basic know ledge o f statistical concepts (e.g., m ean, standard deviation) is assumed. Readers are referred to basic statistical textbooks such as Ferguson (1981) o r those on psychometric theory (e.g., Anastasi, 1988; Nunnally, 1970) for a m ore thorough discussion o f these concepts.

COMPARISON STANDARDS AND NORMS Clinical assessment typically involves the adm inistration o f a variety of m easures o f cognitive functioning, as well as gathering inform ation through interview with the client a n d /o r a person familiar with the client. T he exact m easures adm inistered vary depen d in g on the purpose(s) o f the assessm ent (see chap. 1). Direct observation o f the p erso n ’s perform ance d uring test adm inistration provides an additional rich source o f inform a­ tion abo u t the individual’s approach to tasks, tolerance levels, personal style, and coping skills, as well as providing an opportunity for the exam iner to n o te speech and language characteristics an d abnorm alities in move­ m en t that may be clinically significant. Standardized tests are adm inistered to g ather objective, readily replicable data that perm it reliable in te rp re ta­ tion an d m eaningful com parisons (Lezak, 1995). Both forms o f inform ation (i.e., observations and test scores) are essential to the neuropsychological assessment. In isolation, each form o f inform ation is subject to m isinter­ pretation. Test scores may be objective b u t m ust be considered within the specific context o f the individual. Observations lack objective com parability. Notably in geriatric assessment, behaviors associated with norm al aging may be readily identified as deficits by som eone used to w orking with a younger population o r unfam iliar with the behavioral, cognitive, social, a n d /o r physical correlates o f aging. Psychological test scores are most com m only in terp reted in relation to the perform ance o f a standardization sample. A standardization sam ple is a representative group o f individuals who are adm inistered the m easure in a standardized fashion. Standardization refers to uniform ity in adm inister­ ing and scoring the test and is discussed later in the chapter. T o accurately d eterm ine the individual’s perform ance in relation to the standardization sample, the perform ance o f the standardization sample can be converted to a set o f derived scores characterizing the distribution of scores for the

PSYCHOMETRIC ISSUES

13

sam ple (standard scores). This allows individual scores to be exam ined in relation to o th e r persons and to perform ances on o th e r tests (Anastasi, 1988). Thus, an individual’s perform ance is evaluated in relation to norms, o r the perform ance o f the standardization sample. S tandard scores com e in different form s (e.g., z scores, T-scores), b u t are all based on the m ean an d standard deviation of the scores in the stand­ ardization sample (Lezak, 1995). Com parability o f the scores assumes th at the underlying distributions o f scores have essentially the sam e form (i.e., typically a norm al distribution o f scores aro u n d the m ean o r norm al curve). T h e term normalized standard scores is used to identify standard scores that have been statistically transform ed to fit a norm al curve. Scores may also be p resented as stanines, perccntile equivalents, o r merely as m eans and standard deviations. W echsler Intelligence Q uotients (IQs) arc standard scores expressed with a m ean of 100 and standard deviation o f 15, whereas W echsler subtest scores have a m ean of 10 and standard deviation o f 3. Figure 2.1 illustrates relations between these various m ethods o f expressing scores and includes the com m only accepted description term s for classifying ability levels. As noted by Lezak (1995), many m easures o f cognitive function are affected by age an d education (or vocational achievem ent) an d these variables need to be taken into account when generating norms. O ften, the cognitive abilities o f interest in neuropsychological assessment are those likely to be norm ally distributed in the adult population for which stan dard scores can be generated. However, some behaviors o f interest, o r dom ains o f cognitive functioning, are n o t expected to be norm ally distrib­ u ted in the population. Some o f these have been described by Lezak (1995) as species-wide perform ance expectations. T hat is, it is expected th a t all persons o f a certain age (e.g., adults) will m anifest these capabilities. The behaviors, then, are rudim entary com ponents o f cognitive behavior. If an individual cannot perform this task, im pairm ent is assumed. For exam ple, following sim ple instructions o r copying simple behaviors may reflect this type o f task. O th er behaviors can be assessed in term s o f d eterm ining w hether o r n o t the individual has attained o r m aintained the m inim um requirem ents necessary for perform ing a task (e.g., qualifying for a driver’s license). This approach to assessment is a form of criterion-referenced testing (Anastasi, 1988) and is m ore com m only en cou n tered in the fields of speech language pathology an d occupational therapy. C riterion-referenced m easures are designed to assess a clearly defined and delim ited dom ain of skills. In addition, perform ance is evaluated in terms o f achievem ent on the m easures, n o t in relation to o th e r persons (G ronlund, 1973). For exam ple, it may be said that H elen can spell 90% o f the words in the u n it w ord list. A lthough m ost of the discussion focuses on m easures for which the distribu­ tion o f scores approaches normality, these o th e r forms o f m easures may be m ore appropriate for use in certain circumstances.

14

CHAPTER 2

i________ i_______ i_______ i_______ i ______ i_______ i_______i_______ i_______ i_______ i_______ i_______ i ______ i_______ i_______ i ______i

z scores

-4.0

-3.0

-2.0

-1.0

0

*1.0

+2.0

+3.0

+4.0

----- ----1 L_ --- 1 --------------------2.1 -----1 ----- --- 1 -----1 — -i AbilityLevels1 0 -1.---3 -----0.6 +1.3 ■» 0.6 Pota'ded Superior Aveage VerySuperior H ighAvprago LowAverage WechsferScales DonationIQ s

55 1

70 ‘

85 7

100 10

115 1

1^0 16

145 19

FIG. 2.1. T he relations between the norm al curve an d som e com m only used test scores. Data from the Test Service Bulletin No. 48, January 1955 by T he Psychological C orporation. A dapted an d reproduced with perm is­ sion. All rights reserved.

S creening m easures, o r b rie f m easures th a t sam ple a b ro ad ran g e o f behaviors, are often desig n ed to assess rud im en tary behaviors. A lthough som e dispersion o f scores is expected, screen in g m easures te n d to be skewed to th e left such th a t m ost persons achieve p erfec t o r alm ost-perfect scores. C onversion o f these scores to sta n d ard scores is n o t ap p ro p riate. O ften , p ercentiles may be g en e rate d for com parison p u rp o ses fo r a spe­ cifically d efined g ro u p (e.g., by age a n d / o r educatio n ; C rum , A nthony, Bassett, & M. F. Folstein, 1993). C utoff scores are often ap p lied to screen in g m easures, w hich may be d efin ed to m inim ize false positive a n d / o r false negative rates o f classification. A discussion o f cu to ff scores is in clu d ed la ter in this ch ap ter. Because m any screen in g m easures arc highly skewed, they te n d to lack sensitivity to m ild problem s o r to cognitive deficits in

PSYCHOMETRIC ISSUES

15

persons with prem orbid functioning in the superior range (see chap. 3). Thus, screening measures are similar to the species-wide perform ance ex­ pectations in that identification of a deficit lends support to the hypothesis that the individual is showing cognitive impairment. However, if a deficit is n ot detected with a screening measure, im pairm ent may still em erge when measures with a normal distribution of scores (thereby allowing for many gradations of perform ance) are employed. Individual Comparisons As is discussed later, the correspondence between the characteristics of the standardization sample and the individual in question is of im portance when interpreting the m eaning of test results. In addition, comparison to popula­ tion norms is most useful when the individual’s general ability level had been o f average caliber (Lezak, 1995). Thus, it is necessary to determ ine an individual’s prem orbid level o f functioning if interpretations are to be meaningful within the context of the individual. Direct comparisons with test scores (e.g., IQs, school achievement measures) obtained prior to the onset of the cognitive decline are the most desirable, but are rarely available or accessible for this population. Thus, indirect m ethods—which may in­ volve determ ining estimates from dem ographic variables, test perform ances, or some com bination of these (see chap. 4)—are most commonly employed. By definition, the majority o f individuals will fall within the average range with respect to their prem orbid level of functioning. However, com­ paring test perform ance with an estimate of the original level o f cognitive ability, not solely with respect to w hether or not the test perform ance falls within the average range, becomes particularly im portant when assessing individuals of a superior to very superior or borderline to low average endowment. Although the difficulty in identifying cognitive deterioration in persons of superior intellectual prowess has been em phasized (Naugle, Cullum, 8c Bigler, 1990), it is of equal importance that persons of low prem orbid abilities not be overdiagnosed with dem entia. This is of par­ ticular im portance in light of recent findings from epidemiological studies of Alzheimer’s disease, which have shown a very robust association between this age-related disease and low education (D’Arcy, 1994; Mortimer, 1995). W hen com paring test perform ance to estimated prem orbid level of functioning, patterns of deficit may emerge. A single discrepant score is expected and unlikely to be of clinical significance. A pattern o f scores that conforms to that known to be associated with specific neurological or psychiatric conditions lends support to the diagnosis, if this best accounts for a person’s behavioral abnormalities. As noted by Lezak (1995), it is highly desirable to dem onstrate statistically significant discrepancies between ex­ pected and observed perform ance levels when identifying patterns of perform ance. In this context, it must be noted that the patterns o f perform ­

16

CHAPTER 2

ance described in ch ap ter 9 were derived from reviewing the literature describing perform ance deficits associated with specific disorders and from personal clinical experience. Also, it m ust be noted th a t there will be num erous exceptions to these characteristic patterns o f perform ance asso­ ciated with different conditions (Lezak, 1995). N onetheless, p attern analysis is a fundam ental co m ponent o f the assessment, which w hen used in conjunction with inform ation obtained through observation an d interview, may provide im portant insights into diagnostic and care plan n in g issues. Standardization Samples C entral to the appropriate use o f normative standards to identify areas o f deficit o r im pairm ent is the correspondence between the characteristics o f the standardization sample and the individual in question. For exam ple, it is inappropriate to use norm s developed on a sample o f 25- to 34-year-old males as a com parison standard for the perform ance o f an 85-year-old w om an on a m easure o f memory. T he m ore similar the individual is to the standardization sample in term s of im portant dem ographic charac­ teristics (e.g., age, race, education, and occupational background), the m ore likely deficits identified will be true and n o t due to these o th e r factors. Thus, the purpose b ehind standardization (or adm inistration and in terpretation in relation to norm s) is to reduce the am o u n t o f m easure­ m e n t erro r (possibility that the test perform ance is du e to things o th e r than a true deficit). T he m ore closely the test adm inistration and the characteristics o f the individual are to the standardization p ro ced u re and sample, the m ore likely that the result will be valid. It is im p o rtan t for the standardization sam ple to be large enough to yield stable values an d the sam ple to be selected as representative o f the section o f the population for which the test is designed (Anastasi, 1988). As differences exist in the samples used to standardize different m easures, the resulting norm s will n o t be direcdy com parable. As becom es evident in the following chapters, some large-scale n o rm a­ tive projects have been undertaken with elderly persons that have allowed the generation o f age-appropriate norm s for a variety o f measures. Some o f these studies include m ultiple m easures whereas others focus on single instrum ents. Some provide systematic com parisons between elderly norm al an d pathological populations and others do not. W hen evaluating the utility o f these norm s for use in specific situations, it is o f utm ost im portance th at the characteristics o f the normative sample be carefully considered. In the descriptions to follow, it will be ap p aren t th at the definition o f “n o rm al” varies from study to study. T here has been considerable debate in the medical and psychological literature as to the definition o f norm al in the context o f aging. Lezak (1995) noted that with advancing age, every organ

PSYCHOMETRIC ISSUES

17

system, including the brain, undergoes alterations to som e degree and age-related changes have been observed in brain size, brain biochemistry, and cerebral blood flow (Duckett, 1991; Wallin & Gottfries, 1990). These brain changes may be reflected in decline from previous levels o f functioning as a function o f age. However, Rinn (1988) n oted th at cognitive decline is less pron o u n ced in elderly persons who enjoy unusually good physical health an d m ore severe in those who do not. Pathological medical conditions th at occur m ore frequently with advancing age (e.g., diabetes, high blood pressure) are associated with dem entia in their extrem e form s an d may be associated with lesser cognitive decline in their m ilder forms. Many older persons have chronic medical conditions, and m any m ore may have a m edical disorder, o f which they may n o t even be aware. Rinn (1988) argued that these disorders result in the convergence o f m ultiple small pathological brain changes reflected as norm al age-related decline. W hen com parisons are m ade to a random ly selected representative sample o f the population o f old er persons, the individual is being com pared to a group o f people with a high incidence o f m edical disorders that may com prom ise m ental function­ ing. Thus, the person may be declared average in relation to persons who are markedly d eteriorated o r even close to death. Conversely, a com parison g roup com prised only o f older individuals who were highly active and who exhibited extrem ely good health would not be representative o f the norm al elderly population. Rinn (1988) recom m ended to developers o f norm s that the status of elderly participants be checked about 3 years after data collection and the norm s adjusted to elim inate those who were close to death at the time o f exam ination. At this point, R inn’s approach has n o t been ad o pted for any o f the norm s described in this book. Thus, in selecting appropriate norm s for use in a specific context, it is im p o rtan t for the clinician to establish the “best” com parison group: super healthy active volunteers, random ly selected persons screened for m ajor m edical illness, o r persons deem ed “cognitively n orm al” through screening a n d /o r clinical evaluation. A lthough the effect o f health status may be a m ore p ronounced potential source o f erro r for elderly samples, Lezak (1995) n oted that m any neuropsychological m easures do n o t have exten­ sive norm s and may be standardized on an undefined m ixed o r no n ran d o m sample. Com parison o f the individual’s test perform ance with several sets o f norm s may reveal no significant difference in findings. If im pairm ent is identified using som e sets of norm s but not others, th en the clinician m ust rely on rational judgm ent. As a num ber o f large-scale projects are referred to thro u g h o u t the rem aining chapters, the norm ative samples are described here. R. R. Bomstein (1985). A n um ber o f m easures were adm inistered to 97 C anadian volunteers age 60 and older. No inform ation ab o u t health status was provided.

18

CHAPTER 2

Framingham Heart Study. Farm er et al. (1987) presented norm ative data for various neuropsychological measures collected on 2,123 participants in the Fram ingham H eart Study. W hen specific subpopulations were ex­ cluded (e.g., those with h earing im pairm ents, the blind, those with docu­ m ented strokes, those residing in care facilities, and those for whom English was n o t their first language), 1,195 participants rem ained. They presented the m eans and standard deviations for age groups (in 5-year intervals) by education levels (8-11 years, high school, m ore than high school), and by gender. Mitrushina and Satz. M itrushina and Satz (1989) assessed 156 healthy Caucasian volunteers between age 57 an d 85 who com pleted a self-report medical questionnaire. No persons had a history o f neurological o r psy­ chiatric disorders an d all scored above 24 on the Mini-Mental State Ex­ am ination (see chap. 3 for a description o f this m easure). T hese individuals were well-educated persons functioning independently and living autono­ mously in the community. Johns Hopkins Teaching Nursitig Home Study o f Normal Aging (e.g., Bolla, L indgren, Bonaccorsy, & Bleecker, 1990; Bolla-Wilson & Blceckcr, 1986). Participants in this study were Caucasian volunteers ranging from age 39 to 89 who were free from neurological and psychiatric disorders an d who had never had uncontrolled hypertension, pulm onary edem a, liver failure, renal failure, congestive h ea rt disease, uncontrolled thyroid dysfuncdon, electroconvulsive therapy, sleep disorders, com a, o r a lco h o l/d ru g abuse. Ryan, Paolo, and Associates. T hese researchers developed norm s and exam ined norm ative issues prim arily for the W echsler Adult Intelligence Scale-Revised (Paolo & J .J . Ryan, 1993c; J. J. Ryan & Paolo, 1992a, 1992c; J. J. Ryan, Paolo, & Brungardt, 1990). T heir sam ple o f 130 persons was chosen to m atch the 1983 U.S. Census Bureau population figures for p ersons age 75 an d older. O th e r studies from this g ro u p (e.g., Paolo & J. J. Ryan, 1993a; J. J. Ryan, Paolo, & Brungardt, 1992; J. J. Ryan, Paolo, P ehlert, & Coker, 1991) vary in sam ple sizes. However, in all o f these studies, participants lived in the midwestern U nited States (i.e., Kansas, Missouri, an d Iowa), primarily in urban settings (i.e., 90% ), and were recruited from senior citizen organizations, retirem ent com m unities, news­ p ap er advertisem ents, and by word o f m outh. All participants were screened as healthy an d w ithout evidence o f cu rren t o r past psychiatric, neurological, o r m ajor systemic illness on the basis o f a b rief health questionnaire and the G eriatric D epression Scale (Brink et al., 1982). All participants pos­ sessed adequate hearing and vision. M inor age-related conditions (e.g., senile diabetes, essential hypertension) were not excluded.

PSYCHOMETRIC ISSUES

19

T he norm ative approach used paralleled that o f the W echsler Adult Intelligence Scale-Revised (J. J. Ryan et al., 1990). Raw scores were con­ verted to scaled scores on the basis o f the 20- to 34-year-old reference group. T he m eans and standard deviations o f the scaled scores on the Verbal, Perform ance, an d Full Scales for the 75 to 79 an d 80+ age groups were exam ined in relation to the WAIS-R standardization sample. T he general trend o f m ild decreases in the sums o f scaled scores was m aintained with these additional age groups, suggesting they are an app ro p riate ex­ tension to the national sample. T o generate IQ scores, M atarazzo’s (1972) equation was used to convert each sum o f scaled scores to a new score with a m ean o f 100 and a standard deviation o f 15. T he distributions were norm alized and sm oothed (Angoff, 1971) to elim inate m inor irregularities. Tables for age-corrected subtest scores were constructed, converting the m eans and standard deviations o f the raw scores for each subtest to scores with a m ean of 10 an d standard deviation o f 3 using M atarazzo’s (1972) form ula. T he distributions were adjusted to lit a norm al curve as closely as possible. Iimik and Associates at the Mayo Clinic. T hese researchers provide agespecific norm s for a variety o f m easures (e.g., W echsler M em ory S caleRevised: Ivnik et al., 1992b; Auditory Verbal Learning Test: Ivnik et al., 1990; Visual Spatial L earning T e s t Malec e t al., 1992; W echsler Adult Intelligence Scale-Revised: Ivnik et al., 1992a). Sample sizes vary across the studies from 394 to 512 participants. T hese persons ranged from age 56 to 97 and were recruited from a large sam ple o f volunteers for Mayo’s O ld er A m ericans Norm ative Studies (MOANS). They lived an d functioned ind ependently in the com munity. They had a variety o f medical conditions com m on to elderly persons but were considered norm al by their prim ary care physicians. Ivnik et al. (1990) noted th a t care was taken to ensure the representativeness o f their data by m inim izing selection bias in their sam pling, thoroughly docum enting health status and specifying a definition o f “norm al.” Ivnik et al. used overlapping, m idpoint age ranges (Pauker, 1988) to provide the m axim um am ount o f normative inform ation in th eir studies. Raw scores earned on each subtest score were converted to percentile ranks and age-specific scaled scores for m idpoint ages occurring at 3-year intervals from 61 through 88 years. T he age range around each m idpoint was 5 years. For each m idpoint range, this approach provides the broadest possible norm ative base. T h at is, an 80-year-old person would be com pared to the sample with a m idpoint age of 79 years for which scores are derived from the perform ance o f all persons 74 to 84 years o f age. A 74-year-old person would be com pared to the sample with a m idpoint age o f 73 years for which scores are derived from the perform ance o f all persons 68 to 78

20

CHAPTER 2

years old. In this way, the size o f com parison groups is maximized, thereby ensuring a relatively large n in each subcategory cell. Heaton and Associates. T hese researchers developed norm s for an ex­ p an ded H alstead Reitan Battery (H eaton, G rant, & Matthews, 1991) based o n 486 participants age 20 to 80 with 0 to 20 years o f education. The battery included the WAIS, Wisconsin Card Sorting Test, and the Boston N am ing Test, as well as a variety o f o th e r measures. Participants were derived from several sources (e.g., a n u m b er o f U.S. states such as Cali­ fornia, W ashington, Colorado, Texas, O klahom a, Wisconsin, Illinois, Michi­ gan, New York, Virginia, and the Canadian province o f M anitoba) and included both urban and rural areas. All participants were interviewed and d en ied the presence of psychiatric and neurological disorders. All were rated by their exam iners as having put forth sufficient effort to yield valid test results. A regression approach was em ployed to develop the norm s, using age, g ender, an d years o f formal education as predictor scores (H eaton et al., 1991). This approach requires that raw scores be transform ed to scaled scores with a m ean o f 10 and a standard deviation o f 3. A stepwise multiple regression was perform ed for each test m easure using age, education, an d sex to p redict the scaled scores. Predicted scores, based on age, gender, an d education, were generated using this regression form ula. For each individual, the predicted score is then subtracted from the o btained scaled score. This difference is th en divided by the standard deviation o f the residuals for that m easure. T he resulting value is then m ultiplied by 10 and added to 50 to yield a 7-score. Scaled score equivalents o f raw scores and 7:score equivalents o f scaled scores for persons up to age 80 by 5-year intervals an d levels o f education (6-8, 9-11, 12, 13-15, 16-17, 18+ years) by g en d er are included in their com prehensive norms. H eaton (1992) also used this approach to develop norm s for the WAISR. In this project the WAIS-R standardization sample (excluding persons age 16 an d 17 years, A'= 1,680) was used. In the WAIS-R standardization sample, the sam e stratified sam pling basis was used for the old er groups as for the younger age groups. T he normative data for persons age 55 and over, however, continued to be based on sm aller samples than o th e r age groups and persons o f advanced age continued to be excluded (i.e., over age 75). Thus, norm s were only generated up to age 74. H eaton, C helune, Talley, Kay, and Curtiss (1993) exam ined the Wis­ consin Card Sorting Test perform ance using a sample that included 384 persons age 20 and older. O ne h u n d red and fifty o f these persons were participants in H eaton et al.’s (1991) norm ative project. Two additional samples were included for this project. O ne h u n d red an d twenty-four com­ mercial airline pilots between age 24 and 65 were adm inistered the Wis­

PSYCHOMETRIC ISSUES

21

consin Card Sorting Task (WCST). All but 5 o f these persons were from Colorado. Education levels ranged from 14 to 20 years. T he o th e r sample con tained 73 of the 80 persons exam ined by A xelrod an d H enry (1992). T hese were healthy adults recruited for participation in a health prom otion project from in d e p en d e n t living retirem ent residences in the D etroit area. Based on self-report, none o f these individuals had a history o f psychiatric treatm ent, neurological disease o r injury, substance abuse, o r significant m edical illness requiring treatm ent (e.g., diabetes, C hronic obstructive pulm onary disease). They ranged from age 50 to 89. Education levels ran ged from 6 to 20 years. T h e m e th o d o f co n tin u o u s n o rm in g (A ngoff & R o b ertso n , 1987; Zachery & G orsuch, 1985) was used to derive normative data for a separate census age-m atched sam ple and for the entire sample g rouped by age and age by education. This m ethod is recom m ended to correct for irregularities in the distributions o f scores, o r the trends in the m eans an d standard deviations across groupings when sample sizes are 200 o r less. T his ap­ pro ach requires that the line (or curve) o f best fit for the data be deter­ m ined by polynomial regression. T he m ean, standard deviation, skewness, an d kurtosis o f the distribution o f scores arc estim ated an d percentiles and standard scores are calculated on the basis o f this estimate. Norms were generated u p to age 90. The Canadian Study of Health and Aging (CSHA). As a nationw ide study o f the prevalence o f dem entia in Canada, the CSHA yielded norm s for a variety o f measures (e.g., A uditory Verbal L earning Test, individual subtests from the W echsler A dult Intelligence Scale-Revised, Buschke’s C ued Recall paradigm for m em oiy assessment; T uokko & W oodward, 1996). This demographically corrected norm ative inform ation was based on the perform ­ ance o f 215 individuals age 65 and older living in the com m unity (i.e., no t institutionalized), who com pleted all o f the neuropsychological meas­ ures used in the CSHA. All o f these individuals scored 78 or above on the M odified Mini-Mental State Exam ination (3MS; T eng & Chui, 1987) in the screening exam ination an d were diagnosed with no cognitive loss after com plete m ultidisciplinary clinical exam inations (see Tuokko, Kristjansson, & Miller, 1995). T he perform ance o f 187 persons with dem entia, as d e­ term ined by m edical and neuropsychological exam ination, was exam ined using the norm s to determ ine the sensitivity and specificity o f each m easure (Tuokko & W oodward, 1996). T he participants identified with no cognitive loss an d d em en tia in the CSHA differed in im portant ways from norm al volunteers and clinic-based d em entia samples m ore com m only en countered in o th e r neuropsychologi­ cal studies o f dem entia. All participants were initially random ly selected from all persons over age 65 living in each region of the country. This

22

CHAPTER 2

being the case, the CSHA sam ple may be m ore representative o f the general p opulation than o th e r studies. T h at is, the CSHA sample is less biased in term s o f self-selection than m ost norm al samples. A regression approach to the developm ent o f the norm s, using age, gender, and years of form al education as predictor scores was used. An approach, essentially identical to that o f H eaton et al. (1991) was followed in the developm ent o f these norms. Kaufman Batteries. Kaufman developed a series o f batteries that m easure aspccts o f cognitive functioning and contain norm s for persons age 85 and older (A. S. Kaufman & N. L. Kaufman, 1990, 1993, 1994). T h e stand­ ardization sample for cach of these m easures was described as 2,000 persons from age 11 to 94 selected using a stratified multistage random sam pling p rocedure, to get a cross-sectional representation of the U.S. population as reflected in the 1988 census data. T he sample was stratified within each age group by gender, geographic region, socioeconom ic status (SES de­ fined by education level), and race o r ethnic group. O ne h u n d red persons were selected at each 5-year age group between age 65 and 74 (55% and 58% female, respectively) and 100 subjects over age 75 (65% fem ale) were included.

PSYCHOMETRIC ISSUES Reliability and Validity T h e concepts o f reliability and validity are fundam ental in test construction. T h ere are various forms o f reliability and validity that provide inform ation ab o u t the m easure for interpretation purposes. T he read er is directed to o th e r sources (e.g., Anastasi, 1988; Cronbach, 1990; Nunnally, 1970) for a full discussion o f these concepts. Only a brief discussion is included h ere to highlight a few issues of particular im portance with respect to geriatric neuropsychological assessment. A lthough reliability and validity are im portant to test in terp retatio n , n o t all neuropsychological m easures easily satisfy the usual requirem ents for adequate reliability and validity (Kaszniak, 1990). This is n o t to say that these m easures should not be used, but that lim itations to the m easures m ust be kept in m ind. Moreover, a m easure may be valid for on e purpose (e.g., identifying persons with dem entia) but not an o th er (e.g., predicting rehabilitative success o f a person following a stroke). H ence, the adequacy o f reliability and validity arc relative, and interpreters o f test perform ance arc obligated to thoroughly understand these constructs when interp retin g test scores. In practical terms, m isinterpretation of test data can be highly

PSYCHOMETRIC ISSUES

23

costly. O verinterpretation o f a po o r perform ance on a single, less than adequately reliable o r valid m easure can potentially place older, highly vulnerable persons at risk of losing their autonom y an d self-determ ination if, on the basis o f this perform ance, it is successfully argued th a t these individuals are incom petent to m ake decisions on their own behalf. In general, reliability refers to the consistency with which the underlying behavior is m easured. For exam ple, it reflects the degree o f similarity in the scores obtained by the same person on the same m easure adm inistered at two points in time, o r the similarity o f scores obtained by two raters scoring the same test protocol. Various forms o f reliability have been iden­ tified. M ethods to exam ine the consistency within a m easure o r test (i.e., internal consistency) may be addressed by developing alternate forms, dividing the test items in half and exam ining the correlations between the halves (i.e., split-half), o r by exam ining the intercorrelations o f all items with each o th e r (i.e., C ronbach’s alpha). A nother form o f reliability evaluates the degree to which two judges (or raters) arrive at the same conclusion regarding an individual’s per­ form ance (i.e., intcijudgc reliability o r inteijudge concordance). This type o f reliability is of particular im portance when the scoring o f a test protocol requires som e degree of subjective evaluation on the p art o f the rater. Most neuropsychological tests boast highly standardized scoring criteria th a t minimize inteijudge differences, but this is n o t true for all measures. Variations in the adm inistration and scoring guidelines for the original Mini-Mental State Exam ination (M. L. Folstein, S. E. Folstein, & McHugh, 1975) have given rise to m any concerns regarding the replicability o f this m easure from one practitioner to an o th er (Tom baugh & McIntyre, 1992). A third form of reliability exam ines the stability o f scores over time (i.e., te st-re test). This form is o f particular interest in geriatric assessment, be­ cause it is often the goal o f the assessment to identify true and significant change in perform ance across time (e.g., deterioration of significant mag­ nitu de to suggest an underlying neurodegenerative process). Several co­ efficients, analogous to the reliability coefficient, have been developed that identify an in stru m en t’s ability to detect change within individuals o r groups o f individuals (S treiner & N orm an, 1995). As n o ted by Retzlaff and G ibertini (1994), differences across time arc related to both the reliability o f a test and true and natural changes in people’s scores across time. N atural changes in test perform ance over time may be particularly evident with increasing age an d it is o f great im portance that norm al changes n o t be identified as indicative o f pathological change. It is far too com m on to see practitioners becom e co n cern ed with a m odest (e.g., 3-point) decline on a screening measure such as the MiniM ental Stale Exam ination (MMSE) over a period w ithout consideration for the norm al age-related expected change in scores. Formulas for d eter­

24

CHAPTER 2

m ining significant changes in test perform ance over time at various levels o f significance are available to assist the practitioner in n o t overinterpreting m odest changes in test perform ance (e.g., Cahan, 1989; Payne & Jones, 1959; Silverstein, 1981). In fact, when assessing the test-retest reliability o f a m easure in a population with dem entia, the time interval is kep t sh o rt (e.g., a few days) as changes in test scores over longer periods o f time could be reflective o f cognitive decline as opposed to a lim itation o f the test. Validity refers to the degree to which the test m easures th at for which it was intended. For exam ple, how effective is the m easure in identifying a particular diagnostic group, o r how well does this test m easure depres­ sion? T hree forms of validity are typically described: content, construct, and predictive validity. C ontent validity describes w hether or n o t the co n ten t of a particular test is m easuring a particular dom ain. T h e prim ary purpose o f som e m easures is to capture o r represent a specific universe o f content. Classroom exam inations, designed to reflect the course co n ten t, are ex­ am ples o f this type of measure. Procedures for evaluating c o n ten t validity are described by Anastasi (1988), who stressed that content validation is particularly appropriate for criterion-referenced tests where perform ance is in terp reted in term s o f co n ten t m eaning. C ontent an d face validity are n o t to be confused. Face validity refers to what the test appears superficially to m easure, n o t necessarily what the test actually m easures. Face validity plays a role in determ ining appropriate measures for use with the elderly in that test com pliance may be jeopardized if the m easure appears irrele­ vant, silly, inappropriate, o r childish. As n oted in chap ter 4, the face validity o f the W echsler A dult Intelligence Scale for use with old er persons was questioned by K endrick (1982b). Construct validity refers to the exten t to which a test may be said to m easure a particular theoretical construct o r trait. In this case, there is no obvious co n ten t corresponding to the behavior of interest (e.g., depres­ sion). T he construct u n d e r investigation is hypothetical an d abstract rath e r than, concrete (as with co n ten t validity) and m ust be shown to relate to o th e r m easures o f the same dom ain. C onstruct validation is established w hen test scores relate to dom ains in ways that make theoretical sense. F or exam ple, a m easure o f depression should be related to g en d er because depression is known to be m ore com m on in women than m en. C onstruct validation is determ ined by the gradual accum ulation o f inform ation from various sources to support this assertion. Some o f the procedures providing such supporting evidence include correlations with o th e r tests hypothesized to m easure the same construct, factor analysis, an d interventions to improve certain abilities (Anastasi, 1988). Thus, a new m easure o f m em ory should correlate with existing m easures o f m em ory o r should have high loadings o n the same factor as two o r three o th e r well-known m em ory scales. Retzlaff

PSYCHOMETRIC ISSUES

25

and G ibertini (1994) noted that construct validity is particularly im p o rtan t within neuropsychology w here idiosyncratic test construction techniques are often used. Criterion-related validity indicates the effectiveness o f a test in determ ining an individual’s perform ance on a criterion measure. Two types o f criterion validity are generally discussed: concurrent and predictive. In both cases, perform ance on the test is checked against a criterion to d eterm in e a statistical relation between them . T he difference between co n cu rren t and predictive validity is the tim ing of the adm inistration o f the test and the criterion m easure. T he most com m on types of criteria used in geriatric neuropsychology are diagnostic groupings o r o th e r biological variables. Because the criterion is available, it is im portant to determ ine the function served by the test. For exam ple, does the test provide a sim pler, quicker, o r less expensive substitute for the criterion data? Two issues concerning criterion validity that have received relatively little attention in neuropsychological assessment are criterion contam ina­ tion and operating characteristics o f measures. Both o f these topics arc o f im portance when determ ining the criterion validity o f a m easure. W hen d eterm ining the criterion status o f individuals, it is o f im portance that the test scores do not themselves influence the identified criterion status. T hat is, if a test score is used to identify groups o f cognitively im paired (presum ed dem ented) and cognitively intact persons, o n e cann o t then look for dif­ ferences between the groups on that test score an d infer criterion validity (i.e., the m easure is capable of discrim inating between the groups). Simi­ larly, when exam ining the predictive validity of a test score (e.g., w hether o r n o t a m easure o f m em ory adm inistered at Tim e 1 can p redict diagnosis of dem entia at Tim e 2), the exam iner m ust n o t be aware o f the Tim e 1 test score w hen m aking the diagnosis at Tim e 2. Recently, determ ination o f criterion validity in psychology has been ap p ro ach ed through exam ination o f the operating characteristics o f a m easure. W ithin this framework, the constructs o f sensitivity an d specificity are central. Sensitivity and specificity assess how well test scores discrim inate between two groups. In a clinical context, die usual goal is to identify those who are disease-positive (e.g., im paired) from those who are disease-nega­ tive (e.g., norm al), according to som e known external criterion (i.e., gold standard). T he gold standard is the definitive diagnosis o f the disease u n d e r consideration (e.g., CT-identified lesion, clinically diagnosed A lzheim er’s disease, etc.). T he sensitivity o f a test at any given cutoff score is the p ro p o rtio n of disease-positive (D+) persons with scores in the test-positive (T+) range. T hus, the sensitivity is the probability of a positive test result given a dis­ ease-positive person. T he specificity o f a test score is the p ro p o rtio n of disease-negative (D -) persons with scores falling in the test-negative range

26

CHAPTER 2

(T -). Thus, the specificity is the probability of a negative test result given a disease-negative person. In Table 2.1, N is the sample size and A, B, C, and D are the cell frequencies. Thus, A represents the num ber of diseasepositive persons with a test score falling in the test-positive range. Then, the sensitivity and specificity are calculated as follows: sensitivity = A / (A + C) and specificity = D / (B + D). From Table 2.1, the probability of finding a positive test result in a disease-negative person (false positive; i.e., test identifies a normal person as im paired) can be calculatcd as 1-specificity. The probability of finding a negative test result in a disease-positive person (false negative; i.e., test identifies the impaired person as normal) can be calculated as 1-sensitivity. A test that discriminates perfecdy between the two groups would have both sensitivity and specificity equal to 1.0 at a certain cutoff score (i.e., nonover­ lapping distributions of scores) and this would be the opdm al cutoff score. However, some overlap in the distribution of psychological scores for im paired and nonim paired persons often exists. In theory, a good clinical test is both highly sensitive and highly specific at a certain optimal cutoff score. In pracdce, there may need to be a trade-off between high sensitivity and high specificity (Zweig & Campbell, 1993). The choice, then, of an optimal cutoff score depends on many factors, such as die ultimate cost of false positive and false negative results in the situation under consideradon. Usually, high sensitivity is most desirable for screening tools (Essex-Sorlie, 1995), because the point is to make sure no cases of the disease are missed and follow-up testing will then identify the false positive cases (norm als). However, in some cases, measures that are both highly sensitive and specific are most desirable. For example, in a disorder such as dem entia where no treatm ent or cure is available, it is im portant that clinicians can confidendy rule out the presence of disease as well as detect it (Essex-Sorlie, 1995). Needless to say, the approach of identifying a cutoff score to determ ine who requires further assessment is of practical value and is often a key elem ent of screening tests. Unfortunately, scores on screening tools are often interpreted at face value without consideration of either the need for further evaluation or the context of the assessment. By definition, false TABLE 2.1 M ethod for D eterm ining the Sensitivity and Specificity o f a Measure in Relation to Diagnostic Criteria Gold Standard T a t Raulls T+ 7-

/)+

D-

A C A + C

B D B + D

A + B C + D N

PSYCHOMETRIC ISSUES

27

positive cases and false negative cases are expected to som e (p red eter­ m ined) degree. Moreover, cutoff scores often need to be adjusted by age, education, prem orbid intelligence, and race-ethnicity (Adams, Boake, & Crain, 1982). A related issue raised by Meehl and Rosen (1955) is the base rate at which the condition occurs in the sample from which the cutting score was developed. T he usefulness o f a cutoff score o r how m uch the test score contributes to the diagnostic process is jointly d eterm in ed by its association with the disease and the base rate o f the condition. For exam ple, the use o f any valid test will improve predictive o r diagnostic accuracy as the base rate approaches 50%. With extrem e base rates, the im provem ent gained by using the test may be negligible (W edding & Faust, 1989).

CONCLUSIONS This chapter has addressed som e o f the psychometric issues o f im portance to test interpretation. A lthough many o f these issues are com m on to psy­ chological assessment (e.g., Anastasi, 1988; C ronbach, 1990; Nunnally, 1970) o r neuropsychological m easurem ent (e.g., Lezak, 1995; Retzlaff & G ibertini, 1994), in general, particular issues com m only co n fronted in geriatric neuropsychological assessment were identified. T he samples and m ethods em ployed in some o f the large-scale norm ative studies for persons over age 65 were described. Because it is often difficult to obtain sufficiently large samples o f very old individuals (e.g., over age 90), approaches that provide m axim um inform ation from these samples are particularly useful (e.g., regression, overlapping cells). T he complexities in h e re n t in assessing samples known to show change in function over relatively sh o rt periods o f tim e an d for whom many factors are interacting to create the observable behaviors attests to the need for a solid understand in g o f psychometric theory. Knowledge o f m ore general life-span developm ental issues is also im p o rtan t in geriatric neuropsychological assessment.

Chapter

Screening Instruments for Cognitive Im pairm ent

Screening instrum ents for cognitive im pairm ent have been used extensively by ncuropsychologists, neurologists, and o th e r health professionals working with elderly populations. Most of these are brief, easy to adm inister and score, and are used widely in both clinical an d research settings. T h eir overall diagnostic accuracy can be im proved by com bining data from sev­ eral instrum ents (Eisdorfer & C ohen, 1980). W hen using screening instrum ents, cutoff scores are frequently em ­ ployed. If perform ance falls below the m inim um level expected, dem entia is usually suspected and follow-up testing, as well as additional clinical inform ation, is necessary to establish an appropriate diagnosis. It is n o t appropriate to derive a diagnosis o f dem entia solely based on screening in strum ent perform ance. No screening instrum ent is perfectly valid and, consequently, som e elders with no clinically significant deterioration will be identified as im paired (false positives) and some individuals with de­ m entia will n o t be identified (false negatives). P erhaps the m ost widely known screening instrum en t is the Mini-Mental State Examination (MMSE; M. L. Folstein et al., 1975). T h e MMSE is one o f the tests recom m ended by the N ational Institute o f N eurologic and Com m unicative Disorders and Stroke and the A lzheim er’s Disease and Related Disorders Association (NINCDS-ADRDA) to facilitate the diagno­ sis of A lzheim er’s disease (McKhann ct al., 1984). M. L. Folstein e t al. (1975) included the instrum ent in its entirety in an appendix. T he MMSE is a quick screening tool for the assessment o f cognitive im p airm en t in elderly persons. It is divided into two sections, the first o f which requires only oral responses and covers orientation, memory, and attention with a 28

SCREENING INSTRUMENTS

29

m axim um score o f 21. T he second p art tests ability to nam e, follow verbal an d written requests, write a sentence spontaneously, an d copy interlocking pentagons. T he m axim um score for the second section is 9, m aking the highest possible score 30. T he items are differentially w eighted for the purposes o f scoring. T he MMSE takes less than 10 m inutes to adm inister an d has satisfactory reliability an d validity (Tom baugh & McIntyre, 1992). In terms o f internal consistency, com m unity samples yield coefficients ranging from .68 to .77, w hereas higher internal consistency is evident when m ixed samples o f m edical patients are being exam ined (Tom baugh & McIntyre, 1992). T estretest reliability, with intervals sm aller than 2 m onths, range between .80 an d .95 for both cognitively intact an d im paired elderly (Tom baugh & McIntyre, 1992), although a study o f delirium patients resulted in m uch lower reliability (Anthony, Le Resche, Niaz, Von Korff, & M. L. Folstein, 1982). Traditionally, scores of 23 o r lower are considered to fall within the im paired range (Tom baugh & McIntyre, 1992). A lim itation o f this in strum ent is that sensitivity and specificity decrease as the subject’s age increases and level o f education decreases (La Rue, 1992). For this reason, in this chapter, norm ative inform ation is broken down by age and educa­ tional level (see T able 3.1; Tom baugh, McDowell, Kristjansson, & Hubley, 1996). T om baugh et al. (1996) presented extensive inform ation on the sensitivity and specificity o f the MMSE. In a recent systematic exam ination o f race on the MMSE, it was concluded that there were n o significant race effects on total MMSE scores between W hites and Blacks o f sim ilar ed u ­ cational background (M urden, McRae, Kaner, & Buchm an, 1991). T h e effects o f hearing loss on MMSE perform ance have also been investigated (U hlm ann, Teri, Rees, Mozlowski, & Larson, 1989). It was concluded that hearing loss was related to significantly reduced cognitive p erform ance in clients with probable A lzheim er’s disease regardless o f w hether the screen­ ing m easure was adm inistered in written o r oral form . T hey also poin ted o u t that the research participants’ visual acuity was at least adequate. T heir findings suggested that dim inished MMSE perform ance associated with m ild to m oderate hearing loss is n o t necessarily an artifact of the m eth o d o f adm inistration. In a related investigation, Fillenbaum, George, and Blazer (1988) assessed w hether it is preferable to score nonresponses on the MMSE as correct o r as errors. Based on a com parison of resp o n d ers’ an d n o n resp o n d ers’ abilities to perform activities of daily living, as well as on exam ination o f the relative difficulty level, they concluded th at scoring nonresponses as errors is m ore likely to be correct. Several variations in MMSE adm inistration exist (see T om baugh & McIn­ tyre, 1992). Specifically, in the original version, participants are asked to nam e the hospital an d the floor they are on. Alternative questions have b een used outside hospital settings. A lthough the words apple, penny, an d

30

CHAPTER 3 TABLE 3.1 MMSE N o rm s (P e rce n tile S cores) S tratified fo r Age a n d Years o f E d u catio n for P artic ip a n ts D iagnosed as N o C ognitive Im p a irm e n t Age 65-79

MMSE Score 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 12 Years Education)

Similarities C om prehension Block Design

14.67 14.73 14.33

14.10 14.35 13.63

Unoer Education Grtnip (MT. Scaled scores with a m ean of 10 and standard deviation o f 3, percentile ranks and descriptive categories are presented for the oth er three subtcsts and the Recall/Closure Composite score. Standard scores with a mean of 100 and standard deviation of 15, perccntile ranks and descriptive categories arc presented for the K-SNAP Composite score. An Im pairm ent Index score of 3 or above suggests the presence o f neu­ rological im pairm ent and m ore com prehensive evaluation is recom ­

MEASURES OF INTELLECTUAL FUNCTIONING

83

m ended. K-SNAP norm ative subtest scores were based on 14 age groups for persons age 11 through 85+. K-SNAP norm tables present confidence intervals at the 90% and 95% confidence level. T he K-SNAP Com posite score for persons over age 65 shows high in­ ternal consistency. Split-half reliability coefficients and test-retest reliability coefficients are available in the m anual (A. S. Kaufman & N. L. Kaufman, 1994). An average gain o f six IQ points was observed for the K-SNAP Com posite score. T he M ental Status data from KAIT standardization also pertain to its use as part o f the K-SNAP. Exam ination o f age changes on the subtesLs revealed th at the m ean scores on all subtesLs showed declines after age 50 with som e declines beginning in early adulthood. This is consistent with previous literature in that all three subtests reflect abilities vulnerable to the aging process. T he K-SNAP was factor analyzed with the KAIT, WAIS-R and K-FAST. No K-SNAP subtests loaded on the first V erbal/Crystallized factor. O ne subtest was associated with the Perceptual O rganization factor, on e subtcst was associated with the Fluid factor, and one subtcst was associated with the m em ory/sequencing dim ension that included WAIS-R Digit Span. C orrelations o f the K-SNAP with the KAIT were higher than K-SNAP correlations with the WAIS-R. T h e K-SNAP p erfo rm an ce o f a h etero g en e o u s clinical sam ple o f neurologically im paired persons (« = 49), including those with A lzheim er’s disease (n = 10), was com pared with m atched control subjects. It is n ote­ worthy that the clinical group (n = 49) scored significantly lower than the control group on the K-SNAP Com posite score but the o th e r differences fell sh o rt o f significance. T he Im pairm ent Index was exam ined to determ ine its usefulness in discrim inating between samples with known cognitive/neurological im pair­ m ent and those w ithout such im pairm ent. W hen A lzheim er’s patients were com bined with m entally handicapped persons and com pared to their m atched controls, an Im pairm ent Index o f 3 o r m ore points occurred over 90% o f the time in relation to 11.9% o f the controls. W hen this group was com bined with o th e r neurologically im paired persons (n = 91), less than 60% earn ed scores o f 3 o r m ore in relation to 8.8% o f the controls. Certainly, m ore o f the persons obtaining high Im pairm ent Index scores (i.e., 4 thro u g h 8), were neurologically im paired than were norm al controls. T he Kaufman Brief Intelligence Test (K-BIT; A. S. Kaufman & N. L. Kauf­ m an, 1990) was designed as a brief (15-30 m inutes), individually adm in­ istered m easure o f verbal and nonverbal intelligence com posed o f two subtests: Vocabulary and Matrices. This m easure is not viewed as a substitute for a com prehensive m easure o f intelligence. T he K-BIT was designed for use w hen a brief m easure will suffice and trained professionals may be unavailable (e.g., large-scale screening, estim ate of IQ is o f lesser im por­

84

CHAPTER 5

tance in the assessment, time constraints, research purposes). Because the m easure is easy to adm inister, Kaufman and Kaufman suggested th at tech­ nicians o r paraprofessionals may be trained by qualified personnel to ad­ m inister this m easure. Vocabulary is viewed as a m easure of Crystallized thinking and contains two parts: Part A, Expressive Vocabulary, consists o f 45 items an d requires the person to provide the nam e o f pictured objects; Part B, D efinitions, consists o f 37 items and requires the person to provide the word th at best fits: a phrase description (e.g., a dark color) and a partial spelling o f the target word (e.g., BR_W_). T he Definitions subtest is an alternate form of the identical task on the KAIT Crystallized L earning Stale but the actual item s on the two tests do n o t overlap. T he Matrices subtest is a 48-item m easure o f nonverbal skills and the ability to solve new problem s. It is viewed as a m easure o f fluid thinking. T he m easure is com posed o f several types o f visual stimuli (e.g., m eaningful and abstract) an d all items require an understanding o f the relations am ong the stimuli. All items are m ultiple choice an d require the person to point to the correct response o r say the letter associated with it. T he majority o f the items involve abstract stimuli and require th e solution of 2 x 2 or 3 x 3 matrices o r to com plete a d o t p attern. Each item dem ands nonverbal reasoning an d flexibility in applying a problem-solving strategy. As with the o th e r Kaufman batteries, age-based standard scores having a m ean o f 100 and a standard deviation o f 15 are provided for each subtest an d the overall score o r K-BIT IQ Composite. Age-based standard scores were developed by dividing the sample into 17 separate age groups (i.e., 4, 5, 6 ,. . . 13-14, 15-16, 17-19, 20-24, 25-34, 35-44, 45-59, 60-90). Later a linear interpolation was applied to expand the linearly equ ated raw scores to tables for 53 age groups (including 65-69, 70-74, and 75-90). T h e two subtest standard scores were sum m ed to develop the com posite standard scores and sm oothed within the 17 original age groupings. Split-half reli­ ability coefficients an d test-rctcst reliability coefficients, with a m ean tcstretest interval o f 15 days, are described in the m anual (A. S. Kaufman & N. L. Kaufman, 1990) for 50 persons age 55 to 89. A lthough 20 validity studies arc reported in the m anual, none contain persons over age 50. T he Quick Test yields M ental Age and IQ scores but is actually a m easure th at exam ines vocabulary in situational contexts. A card containing four pictures is shown to individuals and they are instructed to p o in t to the picture that best fits the word read by the exam iner. T h ere arc three forms o f this m easure, each containing 50 words. T h e original norm ative sample (R. B. A m m ons & C. H. Ammons, 1962) contained persons age 2 to 43. A dult norm s are applicable only up to age 45. Sinnctt, H olcn, and Davie (1988) adm inistered the Q T to 100 volun­ teers, 20 (10 males and 10 females) from each o f five age categories (60-64,

MEASURES OF INTELLECTUAL FUNCTIONING

85

65-69, 70-74, 75-79, an d 80+ years). W ithin each gender, five persons of relatively higher socioeconom ic backgrounds and five with relatively lower socioeconom ic backgrounds were chosen. T en percen t (one m an an d one woman in each age category) were m em bers o f identifiable eth n ic/racial minorities. T he Q T IQs were derived from the sum o f scores for Forms 1, 2, and 3, given in that order. T he overall IQ values (i.e., all age categories) had a m ean o f 109.71 and a SD o f 15.7. T he m eans for the age categories did n o t differ significantly from one an o th er except for the oldest group (age 80 to 96), which was signilkandy below all o th e r age categories. Sinnett et al. suggested a correction o f approxim ately 10 points m ight be ad o pted as a guideline to equate the scores o f the oldest age group with the o th e r age categories. Levine (1971) adm inistered the QT to 50 volunteers (25 males an d 25 females) age 60 to 100 from old age hom es, senior citizen clubs, residents o f a retire m en t village, or referred to the exam iner in relation to the WAIS. A m ean level o f education o f 10.6 years was attained for this sample. All three forms o f the QT were adm inistered to all subjects (as was the WAIS). T h e QT was significantly correlated with all WAIS IQs (e.g., QT with FSIQ r = .88) but m ore highly with the VIQ (r = .89) than the P IQ (r = .78). Levine concluded that the QT taps essentially the same construct as the WAIS IQ and may be a useful, brief instrum ent for assessing the intellectual level o f older persons.

NEUROPSYCHOLOGICAL TEST BATTERIES T h e Halstead-Reitan Neuropsychological Test Battery (HRB; H alstead, 1947; Reitan & Davidson, 1974; Reitan & Wolfson, 1993) began as seven measures chosen to discrim inate between brain-dam aged persons an d norm al sub­ jects: Category Test, Tactual Perform ance Test, Rhythm Test, Speech Sounds Perception Test, Finger Oscillation Test (i.e., Finger T apping), Critical Flicker Fusion Test, an d the Tim e Sense Test. T he latter two m easures have b een dro p p ed from m ore recent m odifications o f the battery because they have n o t been shown to identify those with brain dam age with sufficient accuracy. In addition to the individual test scores, an Im p airm en t Index is calculated by applying cutting scores to determ ine the n u m b er o f m easures falling within the im paired range. T he present battery yields seven scores an d abnorm al functioning is identified if 60% o r m ore of the m easures fall within the im paired range. O th er m easures often included in this battery arc the Trail-Making Test an d the Aphasia Screening Test. Reitan and Wolfson (1986) reviewed am ple evidence that the m easures within the H alstead-R eitan Battery arc sensitive to age even u n d e r age 65. Over age 65, these dccrcm cnts arc even m ore m arked. Meyerink (1982)

86

CHAPTER 5

exam ined 25 persons with no past o r present evidence o f cerebral disease or dam age in each o f five age decades beginning at age 20. Com parisons between the age groups revealed that com plex tasks and those m ost heavily d e p e n d e n t on abstract reasoning were observed to decline m ost significandy with age. Based on diese findings, Reitan an d Wolfson (1986) recom m ended the inclusion of the Category Test, T actual Perform ance Test-Localization C om ponent, and Part B o f the Trail Making Test as m easures o f general integrity of the cerebral hem ispheres when assessing for age effects. T he original norm s for the HRB (Halstead, 1947) were n o t well founded (Lezak, 1995) and consisted o f only 28 persons (age 14 to 50). This has raised serious concerns about the appropriateness o f the cutting scores for older populations as most o f the m easures decline with age (e.g., Bak & G reene, 1980). M ore recently, norm s for the HRB have been developed (H eaton ct al., 1991) based on 486 participants age 20 to 80 using a regression approach (see chap. 2 for description o f sample an d m ethods). It has been noted (Feurst, 1993) that transform ing raw scores into dem ographically corrected scores using the H eaton et al. (1991) m anual can be laborious an d a com puter program has been developed to autom ate this procedure. However, Feurst (1993) noted some problem s with the co m puter program , as well, which may d eter all bu t the m ost avid HRB enthusiasts. It is to be expected that these problem s will be corrected in future versions o f the software. Despite im provem ents to procedures for using the norm s o f H eaton et al. (1991) for old er age groups, Lezak (1995) noted th at the HRB may n o t be chosen for use with elderly persons for o th e r reasons. Most notably, older persons may be unwilling o r unable to tolerate the length of the testing session required to adm inister the HRB. Moreover, the difficulty level is such th a t older persons may withdraw from tesdng or be unable to perform sufficiendy well on m easures to w arrant the cost in term s o f dm e an d effort. Certainly older persons suspected o f neurological disorders would be fatigued by the HRB easily with little useful inform ation concerning strengths and weaknesses yielded for diagnostic o r rehabilitative purposes. Faibish, A uerbach, and T hornby (1986) addressed these issues an d pro­ posed a form o f the HRB m odified for older adults. T h e m easures chosen for alteration included: the Category Test, the Tactual Perform ance Test, the Seashore Rhythm Test, the Speech Sounds Test, an d the Trail Making Test. T he Category Test for Younger C hildren was selected an d m odified for use with elderly persons. T he Tactual Perform ance Test was changed radically with the presentation o f two 5-form boards for each han d (i.e., four trials). Faibish et al. (1986) reported that over 95% o f the trials were com pleted in less than 5 m inutes with a m oderate age effect being seen (r = .53). Modifications to the Speech Sounds Test focused on reduced length an d response intervals were increased on the Seashore Rhythm

MEASURES OF INTELLECTUAL FUNCTIONING

87

Test. T he Trail Making Test for O lder C hildren was adopted in favor o f the standard adult form. W hether o r n o t these changes are sufficient to create a battery o f suitable length th a t is sensitive to forms o f d eterio ratio n o f concern in an aging population (e.g., dem entia) is yet to be determ ined. The Luria Nebraska Neuropsychological Battery (LNNB) is m ade up o f ex­ am ination techniques described by Luria, the preem in en t Russian n eu ro ­ psychologist, and organized by Christensen. These techniques were co n ­ verted into test items in this battery, which as noted by Speirs (1981) is n o t to say th a t “the test is an operationalization o f L uria’s m ethod" (p. 339). G olden, Moses, G raber, an d Berg (1981) selected items from Chris­ te n sen ’s m anual, which discrim inated between norm al subjects an d an unspecified group o f neurologically im paired persons. The 269 items were organized into 11 clinical scales: C l = m otor; C2 = rhythm ; C3 = tactile; C4 = visual; C5 = receptive language; C6 = expressive speech; C7 = writing; C8 = reading; C9 = arithm etic; CIO = memory; C l 1 = intellectual processes; and C l 2 = interm ediate memory. Form II o f this m easure largely parallels Form I b u t contains an extra subtest, Interm ediate Memory (10 item s), which assessed delayed recall o f some o f the previously adm inistered short-term m em ory items. Perform ­ ance on each item is evaluated on a scale from 0 (norm al) to 2 (severely im paired). T he 11 clinical scales are derived from the sums o f the items within each scale. Five summary' scales (Pathognom ic, Right H em isphere, Left H em isphere, Profile Evaluation, and Im pairm ent scales) are also de­ rived by sum m ing particular items. For example, the Right H em isphere scale is com posed o f the sums o f all tactile an d m otor function items for the left side o f the body. O th er scales, derived from sum m ing particular items, have been generated since the publication o f this m easure (see Lezak, 1995), including a 66-item list o f qualitative aspects o f test perform ­ ance to allow for the exam ination o f the nature o f a failure. Initial norm ative data were lim ited (i.e., 26 female and 24 m ale hospi­ talized medical patients). T he critical level is defined by calculating the p erso n ’s age x .214 for every age between 25 and 70 years. T he n u m b er o f years of education (0 through 20) is m ultiplied by 1.47, which is then subtracted from the critical level. T hese corrections assume sim ple linear relations between age, education, an d perform ance on every skill o r func­ tion that does not correspond to studies on cognitive changes with aging. O th er lim itations o f this m easure include the heterogeneity w ithin scales (e.g., verbal, visual, visuospatial, auditory all within one scale) and the lim ited b readth within each dom ain o f each scale (e.g., some aspects of m em ory n o t addressed) (Lczak, 1995). Spitzform (1982) first exam ined the perform ance o f persons age 65 to 83 on the LNNB (n = 14). M aclnncs, Gillen, G olden, an d G raber (1983) ex tended this look at the perform ance of norm al elderly persons to 78

88

CHAPTER 5

volunteers with a m ean age o f 72.2 years. T h e m ean perform ance o f this group fell within the average range for 14 clinical scales, localization scales, an d factor scales. W hen this group was divided into y oung-old (60-74) an d o ld -o ld (75+) groups, the m ean perform ances o f both groups fell within the average range an d the only statistically significant differences between groups on the clinical scales were in favor of the o ld -o ld group (Expressive Speech and W riting). W hen males ( n = 26, m ean age = 73.4) and females (n = 52, m ean age = 71.6) were exam ined, the males perform ed significantly b etter than the females on the Visual scale an d the females perform ed significantly better than the males on the Expressive Speech and Pathognom ic scales. W hen a group o f 100 neurologically im paired o ld er persons was com pared with this norm al sample using the objective rules rep o rted by G olden et al. (1981), 72 of the 78 norm al persons were correctly classified and 86 o f the 100 neurologically im paired persons were correctly classified. G olden et al. noted that the lack o f age-related cognitive decline on the LNNB may be the result of sam pling (i.e., 75+ were selected to be as healthy as 60- to 74-year-old group); sim pler items on the LNNB (o th er neuropsychological batteries contain m ore com plex tasks th at are affected by norm al aging); o r m ore crystallized items on the LNNB (agerelated effects are m ost p ronounced for fluid tasks). They concluded that the LNNB may be a useful m easure o f neuropsychological functioning with elderly persons. A norm ative study for Form II of the LNNB was co nducted by W ong, Schefft, and Moses (1990) with 100 norm al individuals age 17 thro u g h 70 (M = 38.5, SD= 14.7). T hey presented the m eans and standard deviations for each scale and n oted that the average m ean predicted perform ance level score o r baseline (critical level) = .171 (age) - 1.16 (education) + 59.7. They noted that m ore research is needed before the norm s could be used in clinical situations on a regular basis. T he effects o f age, sex, race, education, and health status on test perform ance were n o t exam ined. Moses, Schefft, Wong, and Berg (1992) extended these norm s to include a total of 392 persons age 12 to 80. T hey recom m ended the use o f uniform T-scores over standard T-scores as this m ethod makes percentile equivalents o f the scaled score the same for the clinical scales an d m aintains the sensitivity o f the test to outliers. T he uniform T-score value for 316 persons in the norm ative sam ple was then predicted using a m ultiple regression p ro cedure with age and education en tered as the in d e p en d e n t variables. T h e prediction equation that estimates the baseline was: 57.482 - (1.629 X educational level) + (0.078 x age). T he application o f this form ula was validated in a sample o f 55 control subjects (age 18 to 71) and 55 neurologically im paired subjects (age 18 to 73). The formula (baseline value plus 10 T-scorc points = critical level) correctly classified 85.5 % (47/55) of the normal participants when four or

MEASURES OF INTELLECTUAL FUNCTIONING

89

fewer scales were elevated above the critical level (including C12, In ter­ m ediate Memory and excluding C7 and C9, W riting and A rithm etic, re­ spectively, from the sum ). Using this criteria, 87.3% o f the neurologically im paired subjects were correctly classified. Moses et al. n oted the differ­ ences in m ean values between this sample and that rep o rted earlier (W ong et al., 1990) indicate that the earlier norm s were too strin g en t to account for the range o f norm ality likely to be encountered in clinical practice. They also recom m ended the use o f this uniform T-score approach rath er than the previously used norm ative prediction m ethod recom m ended by G olden and colleagues for the LNNB II. A short form containing approxim ately half (141) o f the original items (269) has been proposed for use with elderly persons (McCue, Shelly, & G oldstein, 1985). T his short form was designed to specifically address re­ ferral issues o f dem entia, differential diagnosis o f type o f d em en tia and distinguishing between dem entia and depressive pseudodem entia. Thus, the Memory, Intellectual Processes, and Pathognom ic Scales o f the LNNB were adm inistered in their entirety and the Rhythm Scale was deleted. T he rem aining scales were shortened an d modified. W hen a heterogeneous sample o f persons age 55 an d old er (n = 247) was exam ined, correlations between the 7-scores generated from the abbrevi­ ated scales and those derived from the full battery ranged from .8 to .9. McCue, G oldstein, an d Shelly (1989) adm inistered this sh o rt form o f the LNNB to 79 elderly persons, 34 diagnosed with probable A lzheim er’s disease an d 45 m eeting D SM -III criteria for depression. D iscrim inant function analysis yielded a correct classification rate of 86.1 %. Exam ination o f specific scales (e.g., M emory scale, Overall M ean T-score, and Pathognom ic scale) yielded slightly p o o rer classification (i.e., 79.7%, 81%, 79.8%, respectively). O f note, the m ean profile fo r the depressed group was com pletely norm al w hen critical level age and education adjustm ents were applied. T he LNNB has been roundly criticized for various flaws, including p o o r norm s and diagnostic insensitivity (Lezak, 1995). T he advantages o f this m easure for use with older adults is the simple, generally n o n th reaten in g n atu re o f the items an d its brevity. CONCLUSIONS This chapter highlighted efforts to extend knowledge ab o u t m easures o f global functioning to persons age 65 and older. As seen with the WAIS-R, floor effects (e.g., HRB) may lim it the use o f some m easures with old er persons. New m easures o f intelligence have em erged (i.e., K aufm an’s m easures) based on different theoretical constructs than the WAIS-R. W hether or n o t these batteries will be of particular use in the assessment o f older persons is yet to be determ ined.

Chapter

6

Memory Assessment

Assessment o f m emory is the central issue in m ost referrals o f elderly persons for neuropsychological evaluation. Many com prehensive sources are available that detail proposed theories, classifications, an d subdivisions o f m em ory functioning (e.g., Squire, 1987). Reeves and W edding (1994) n o ted that m em ory processes involve registration, encoding, storage, con­ solidation, and retrieval, at least. Moreover, m emory tasks may be classified within a tem poral framework as involving im m ediate, recent, an d rem ote memory. Because the structure of m em ory assessment tools may reflect these concepts, some o f the m ore com m on constructs are discussed briefly. It is generally assum ed that inform ation m ust be registered, o r encoded, if it is to be rem em bered. Specific encoding operations are perform ed on what is perceived an d inform ation is processed fu rth er into short- an d long-term storage. Short-term m em ory is a lim ited capacity store with ex­ trem ely rapid decay (i.e., seconds to a m inute). Working memory, a type of short-term memory, is m ost related to active processing, and is discussed with m easures o f attention and concentration in chap ter 7. Inform ation th at has been encoded into short-term storage m ust then be m aintained o r elaborated if it is to be consolidated into long-term storage (held for m inutes, hours, o r longer). Only inform ation that has been adequately en co d ed and stored is available to then be retrieved from storage. Retrieval may be facilitated by sem antic prom pting, cued recall, an d recognition procedures. This facilitation o f m em ories implies that some o f the to-berem em bered m aterial m ust have been adequately enco d ed an d stored for it to be revealed u n d er these conditions. Differences in the du ratio n o f m aintenance o f the inform ation in long-term m emory (e.g., 5 to 5 m inutes, 90

MEMORY ASSESSMENT

91

15 m inutes, o r longer), speed of memory decay or retention oner time may also be o f significance for characterizing m em ory functioning. A b undant evi­ d ence to support a relation between anatom ical substrates an d various com ponents o f m em ory processing has been described an d detailed else­ w here (e.g., Lynch, McGaugh, & W einberger, 1984) b u t as yet is n o t fully understood. T he majority o f memory' assessment instrum ents for which norm ative inform ation has been collected were n o t derived from any particular theo­ retical m odel. However, some of the newer clinical instrum ents include constructs that extend beyond the simple presentation o f m aterial for retrieval at a later time. In addition, m easures o f subjective appraisals (e.g., Crook & Larrabee, 1992) o f m em ory functioning have been developed because com plaints o f m em ory change may or may n o t be related to m em ory perform ance on objective measures. It is im portant to be able to distinguish between norm al m em ory changes and declines in m em ory functioning resulting from a variety o f medical conditions, including A lzheim er’s disease. Such conditions may affect m em ­ ory functioning o f older persons and im pact on their abilities to m eet the com plex dem ands of everyday life (see chap. 10).

MEMORY BATTERIES T he first battery o f m easures designed specifically to assess m em ory was the Wechsler Memory Scale (WMS; D. Wechsler, 1945; see T able 6.1). T he seven subscales are com bined into a M emory Q uotien t with m ean o f 100 and SD o f 15, m aking the scores com parable to W echsler’s IQ measures. T he original norm ative sam ple did n o t include people over age 60 (D. W echsler, 1945). Klonoff and K ennedy (1966) collected norm ative data on 115 Canadian m ale veterans between age 80 and 92 living in the com ­ munity. T he same year, Hulicka (1966) exam ined age differences on the WMS and included persons age 60 to 89 who were hospitalized veterans, residents of hom es for the aged, o r m em bers o f G olden Age Clubs. C authen (1977) was the first to collcct norm s for a 60-minute delayed recall for the Logical M emory subscale in older persons. Since then, norm s have been collected from a n um ber o f sources addressing a variety o f procedural concerns. Differences between the norm s in terms o f sam ple com position an d scoring systems illum inate the need for care in applying these norm s in clinical settings (see Table 6.2). Age effects have been identified for most subscales o f the WMS across studies. Hulicka (1966) noted age differences between scores for Logical Memory (LM), Digit Span (DS), and Associate L earning (AL). Perform ­ ance on the easy pairs o f the AL subtest shows far less change over age

92

CHAPTER 6 TABLE 6.1 C o n te n t o f th e WMS a n d W M S-R

Subscale

WMS

Personal an d C urrent Inform ation O rientation

Age, date o f birth, curren t and recent public officials T im e and place

M ental Control

Automatisms and simple conceptual tracking Im m ediate recall of verbal ideas from two paragraphs

Logical Memory (I-M)

Digit Span (DS)

Repetition o f forward and backward digit series

Visual Reproduction (VR)

Im m ediate visual memory drawing

Associate L earning (AL)

Verbal retention of easy and hard word pairs

Visual Paired Associates (Visual PA)

Figural Memory

Visual Memory Span (VMS)

W MS-R Extended with additional questions Extended with additional questions Speed credits elim inated from WMS scoring system Stories are m ore nearly equivalent in difficulty. Scoring system improved. Delayed recall trial added. Easier items added to both forward and backward series. Scoring procedures changed. Item content m odified and subtest lengthened. Scoring rules improved. Delayed recall trial added. R enam ed to Verbal Paired Associates. Two easy paired associates deleted. Scoring system simplified. Material learned to a constant criterion. Delayed recall trial added. New subtest corresponding to Verbal Paired Associates (Verbal PA). Contains delayed recall trial. New subtest m easuring recognition o f abstract visual patterns New visual-spatial subtest corresponding to the verbal Digit span subtest

Note. WMS = W echsler Memory Scale (D. Wechsler, 1945) an d WMS-R = W echsler M em­ ory Scale-Revised (D. W echsler, 1987). W echsler Memory Scale-Revised. Copyright © 1945, 1974, 1987 by T he Psychological Corporation. Adapted with permission. All rights reserved. “Wechsler Memory Scale” and "WMS” are trademarks of The Psychological Corporation.

th a n rcca ll o f h a rd pairs (d e sR o sic r s & Iv iso n , 1 9 8 6 ). P e r so n s a g e 6 0 to 6 9 r e c e iv e d a n a v era g e sc o r e o n th e ea sy pairs (1 5 .0 5 + / — 2 .6 2 o f a total o f 18 p o in t s ) , m o r e th a n fo u r tim e s g r e a te r th a n th e ir av era g e p e r fo r m a n c e o n th e h a r d ite m s (3 .3 2 + / - 2 .6 2 o f a to tal o f 1 2 ). A g e e ffe c ts w er e e v id e n t fo r V isu a l R e p r o d u c tio n (V R ) a n d D e la y e d R eca ll fo r b o th LM p a ra g ra p h s in C a u t h e n ’s (1 9 7 7 ) sa m p le . H a a la n d , L in n , H u n t, a n d G o o d w in (1 9 8 3 )

93

MEMORY ASSESSMENT

TABLE 6.2 WMS Normative Studies Author

N

Trahan et al. (1988)*

255

18-91 years; 50-69 (« = 51), 70-91 ( n = 26)

Abikoff et al. (1987)b

339

18-83 years

Haaland ct al. (1983)

175

65-69 (n = 49), 70-74 (n = 74), 75-79 ( n = 40), 80+ (n = 13)

Ivnik el al. (1991)

99

65-97 years

Mitrushina & Satz (1989)

156

57-85 years

Klonoff & Kennedy (1966)

115

80-93 years

Cauthen (1977)

64

Sample

Age

60+ years

Healthy, nonhospitalized, neurologically normal with no evidence of major psychiatric illness o r mental deficiency. Age and education norms for gist, verbatim, immediate, delayed, and 24hour recall; interrater reliability. Volunteers in a longitudinal study on immunology and aging; no prescription medication; living independently; self-report and medical examination to assess health status including history, physical, and appropriate laboratory tests. Normal volunteers participating in an Alzheimer’s disease patient registry; white. Very intelligent (high average to superior IQ ), optimally functioning residents of a retirem ent community in Southern California. Canadian male veterans residing in an urban community (Vancouver, BC). Most subjects had good global health ratings (from internist) and were moderately or very active; majority had unskilled/semiskilled occupational backgrounds. Canadian volunteers, living in institutional settings.

Note. WMS = Wechsler Memory Scale (D. Wechsler, 1945). “Using only the Visual Reproduction subtest of the WMS. bUsing only the Logical Memory subtest of the WMS.

re p o r te d that e d u c a tio n a n d in c o m e d iffer en ce s d id n o t a c c o u n t fo r the e ffe c ts o f a g e o n R u ssell’s variant o f th e LM an d VR subtests (i.e., im m ed ia te a n d 3 0 -m in u te d ela y ). A b ik o ff e t al. (1 9 8 7 ) n o ted effects o f a g e o n the LM su b sca le for im m e d ia te , 20- to 3 0 -m in u te delay and 24-h ou r d elayed gist a n d verbatim recall— e x c e p t for im m ed ia te recall Form II. Ivnik ct al. (1 9 9 1 ) fo u n d a lim ited co rrela tio n o f age w ith th e W MS m easu res a c c o u n t­ in g fo r o n ly 10.7% o f the variance in th e total W M S raw sco re. H e n c e , age stratification was n o t w arranted fo r th eir sam p le. T a b les 6 .3 a n d 6.4 su m m a rize so m e o f th e available norm ative data for p erso n s over age 60 o n th e various subtests o f th e WMS.

94

CHAPTER 6

TABLE 6.3 Logical Memory and Visual R eproduction Scores (SD) for Persons Age 65 and O lder From H aaland et al. (1983) Age. Group

Variable

65-69 (n = 49)

70-74 (n = 74)

75-79 (n = 40)

80+ (n = 13)

5.9 (2.5) 3.6 (2.5) 56.3 (23.9)

6.1 (1.7) 4.0 (1.4) 65.7 (13.2)

4.9 (2.0) 4.2 (1.9) 86.8 (31.5)

3.3 (2.3) 2.8 (1.9) 92.6 (42.8)

Logical Memory Im m ediate Delayed Percentage Retained*

7.4 (2.5) 4.8 (2.4) 63.1 (21.3)

6.7 (2.6) 4.2 (2.4) 60.2 (23.0)

Visual Reproduction Im m ediate Delayed Percentage Retained*

6.0 (2.1) 5.4 (2.5) 89.3 (27.6)

5.1 (2.0) 4.3 (2.3) 83.4 (37.2)

Note. Im m ediate an d delayed scores for Logical M em ory are based on the average o f the details recalled from two stories, whereas sim ilar scores for Visual R eproduction are based on the total n u m b er of details recalled, as proposed by D. W echsler (1945). From “A Normative Study of Russell's Variant of the W echsler Mem ory Scale in a Healthy Elderly P opulation,” by K. Y. H aaland, R. T. Linn, W. C. H unt, an d j . S. Goodwin, 1983, Journal of Consulting and Clinical Psychology, 51, p. 879. Copyright © 1983 by the Am erican Psychological Association. A dapted with perm ission. •‘Percentage retained = Total n u m b er of details recalled on delay / Total n u m b er o f details recalled im m ediately (x 100).

B oth C au th e n (1977) an d Ivnik e ta l. (1991) n o te d stro n g er co rrelatio n s betw een IQ a n d WMS perfo rm an ce th a n age. T hey suggested th a t IQ ra th e r th an age deserves consideration in adjustm en t for n o n m em o ry fac­ tors w hen using th e WMS with o ld e r persons. Ivnik e t al. (1991) also fo u n d co rrelations with edu catio n , w hich they suspected arose from th e co rrela­ tion o f b o th WMS an d ed u catio n with IQ an d p rese n t norm ative d ata by ed u catio n a n d IQ (sec T able 6.5). T h e LM subtest has also received additional atte n tio n as a result o f p ro blem s in in te rp re tatio n a n d in terstudy com parisons arising from im ­ precise scoring instructions. T h e no rm s g en e rate d by C au th e n (1977), H aalan d e t al. (1983), an d H ulicka (1966) are based o n W echsler’s im ­ precise scoring instructions. A bikoff et al. (1987) ap p lied o p eratio n alized sco ring systems for verbatim a n d gist recall for im m ed iate, in te rm ed ia te (25 m in u tes), a n d long-term (24-hour) recall intervals. Both age an d edu-

95

MEMORY ASSESSMENT TABLE 6.4

Summary of Hulicka (1966) Data for Persons Age 60 and Older on the Wechsler Memory Scale WMS Subsaila

Age

60-69

n

M SD

Logical Memory

5.47 1.16

6.24 2.29

7.34 2.90

5.24 1.03

5.63 2.46

5.60 0.24

6.92 2.02

Digit Span

Visual Reproduction

Associate Learning

9.91 1.58

6.03 3.72

11.94 4.53

7.35 3.83

9.91 2.51

4.95 3.42

10.98 4.78

6.80 3.19

10.00 2.35

4.00 2.38

9.98 3.28

46

M SI)

80-89

M ental Control

70

M SI)

70-79

Information

25

Note. From "Age Differences in Wechsler Memory Scale Scores.” Journal o f Genetic Psyclwfogy, 1 0 9 , 135-145,1966. Adapted with permission of the Helen Dwight Reid Educational

Foundation. Published by H eidref Publications, 1319 Eighteenth St., N. W., W ashington, D.C. 20036-1802. Copyright © 1966.

cation effects were found and Abikoff et al. (1987) provided prediction equations, taking age and education into account (see Table 6.6). T he p rediction equations can also be used to generate z scores, thereby enabling com parison of individuals at different ages and educational levels. This standard score is obtained by subtracting the predicted score from the actual score an d dividing by the SD (given in Table 6.6). Ivnik e t al. (1991) expressed delayed free recall on the LM an d VR subtests as “p ercent rete n tio n ” scores based on the am o u n t o f inform ation originally learned an d found no relation with age, education, or IQ. T h ere­ fore, norm s by age were n o t considered necessary. H aaland et al. (1983) rep o rted a similar finding. This feature o f m em ory perform ance may prove to be useful clinically, if this lack o f relationship to age, education, and IQ can be replicated. D ’Elia, Satz, and Schrctlcn (1989) critically reviewed the existing studies th at provide norm s for the WMS. They noted differences between studies in the am o u n t o f detail provided concerning the sam ple and procedural variables such as scoring m ethods. In addition to these lim itations, they com m ented that there are no reliable delayed recall data for the AL subtest for any age group. Moreover, there are no data available regarding recog­ nition testing on delay for any o f the original o r m odified WMS subtests. They go on to indicate that w ithout assessment of delayed recognition for m aterial not recalled, it is not possible to ascertain w hether the m anifest deficit is one o f encoding or retrieval.

96

CHAPTER 6

TABLE 6.5 Means and Standard Deviations on WMS by WAIS-R Full Scale IQ Group ______________________ IQ._____________________

Inform ation Orientation Mental Control Logical Memory Digit Span Visual Reproduction Associate Learning Raw Total Delayed Logical Memory Delayed Visual Reproduction

110 (n = 15) 5.7 (.7) 6.0 (0 ) 7.9 (1.4) 10.5 (2.9) 10.9 (2.0) 9.3 (2.6) 14.4 (3.4) 63.6 (6.4) 8.1 (3.1) 8.1 (3.5)

Note. WMS = Wechsler Memory Scale (D. Wechsler, 1945) and WAIS-R = W echsler Adult Intelligence Scale-Revised (D. Wechsler, 1981). From “Wechsler Memory Scale: IQ-dcpcndent Norms for Persons Ages 65 to 97 Years,” by R. J. Ivnik, G. E. Smith, E. G. Tangalos, R. C. Petersen, E. Kokmen, and L. T. Kurland, 1991, Psychological Assessment, 3, p. 160. Copyright © 1991 by the American Psychological Association. Reprinted with permission.

To address some o f the shortcom ings o f the WMS, an extensive revision was u n dertaken in 1987 with m odifications to the original WMS subscales (as indicated in Table 6.1). In addition to im proving scoring procedures for some subtests, delayed recall com ponents were added for LM, Verbal P aired Associates (VerbalPA), and VR. Two o f the three new subtcsts (Visual Memory Span: VMS; and Visual Paired Associates: VisualPA) were designed as visual counterparts to the DS and VerbalPA subtests. T he subscalcs from the Wechsler Memory Scale-Revised (WMS-R; D. Wech­ sler, 1987) are com bined into four Index scores: A tten tio n /C o n cen tratio n , Verbal Memory, Visual Memory, Delayed Recall. A G eneral M emory index score is obtained using a sum of the weighted raw scores for the Verbal M emory and Visual M emory com ponents. Each of the five index scorcs has a m ean o f 100 and a standard deviation o f 15. In addition, the m anual

97

MEMORY ASSESSMENT

TABLE 6.6 Equations to Predict WMS Subtest Scores From Age and Education Recall Measure Gist Immediate Delayed 24-hour Verbatim Immediate, Form I Immediate, Form II Delayed, Form I Delayed, Form II 24-hour, Form I 24-hour, Form II

Equation

SD

6.72 + .20 (age) + .93 (education) - .003 (age2) 3.40 + .26 (age) + .90 (education) - .003 (age2) 3.00 + .19 (age) + .98 (education) - .003 (age2)

6.01 6.56 6.41

4.39 5.20 2.37 2.92 1.31 2.37

4.83 4.32 4.27 4.09 3.77 3.91

—.03 - .01 - .05 - .02 —.05 - .05

(age) (age) (age) (age) (age) (age)

+ + + + + +

.62 .57 .55 .56 .58 .65

(education) (education) (education) (education) (education) (education)

Note. WMS = Wechsler Memory Scale (D. Wechsler, 1945). From “Logical Memory Subtest of the Wechsler Memory Scale: Age and Education Norms and Altcmate-form Reliability of Two Scoring Systems,” by H. Abikoff, J. Alvir, G. Hong, R. SukofT, J. Orazio, S. Solomon, and S. Saravay, 1987, Journal o f Clinical and Experimental Neuropsychology, 9, p. 446. Copyright © 1987 by Swets & Zeitlinger Publishers. Used with permission.

contains tables of the following data: cum ulative frequencies for the n u m ­ b er o f Inform ation and O rientation questions answered correctly by each age group (i.e., 55-64,65-69, 70-74), the p ercent passing each Inform ation an d O rientation question by age, percentile equivalents o f raw scores o f DS an d VMS (Forward an d Backward) by age, percentile equivalents o f raw scores for LM I by age group, percentile equivalents o f raw scores for LM II by age group, percentile equivalents o f raw scores for VR I by age group, and percentile equivalents o f raw scores for VR II by age group. T he norm ative sample did n o t include people over age 74. Moreover, data are only available for som e o f the individual (i.e., n o t com bination) subscale scores. Test users, then, have lim ited flexibility in selecting specific subtests from the WM S-R for use. In many circumstances, it is n o t necessary o r possible to adm inister the entire battery, although specific subtests would be o f interest to clinicians. A dditional norm ative inform ation from the WMS-R has been collected o n older age groups. Cullum, Butters, T roster, and Salmon (1990) exam ­ ined forgetting rates for verbal and nonverbal m aterial for 47 persons age 50 to 70 an d 32 persons age 75 to 95. Despite equivalent scores on measures o f global cognitive status an d atten tio n /c o n cen tra tio n , the old er group d em onstrated significantly m ore rapid forgetting rates on the VR, Ver­ balPA, and VisualPA subtcsts. T he severity and pattern o f losses ap p eared useful in differentiating “abnorm al” forgetting from th at exhibited by n o r­ mal elderly persons. L ichtenberg an d Christensen (1992) ex ten d ed the norm ative data on the LM subtests o f the WMS-R by exam ining 66 cog­

98

CHAPTER 6

nitively in tact geriatric medical patients (age 70 to 99) from an urban hospital. All participants scored 129 o r greater on the D em entia Rating Scale (DRS; see chap. 3) and showed no evidence o f medical conditions o f sufficient severity to affect their cognitive functioning on medical history review. Analyses indicated that LM scores for Forms I and II were u n co r­ related with education, race, sex, o r age (Form I, M = 17.8, SD = 5.8; Form II, M = 13.7, SD = 6.6). Ivnik e t al. (1992b) provided age-specific norm s for the WM S-R o n a sam ple o f 441 cognitively norm al persons age 56 to 94 (see chap. 2 for a description o f this sample and m ethods). Some differences in adm inistra­ tion and scoring of the MAYO adaptation o f the WMS-R exist. T he only difference in adm inistration procedure from the WMS-R is th at only th ree learning trials were allowed during both the VisualPA an d VerbalPA I subtests. This adm inistration difference, then, precludes com putation of the standard WMS-R Delayed Recall index for this sam ple and these norms. However, delayed recall (30 m inutes) data for this adm inistration o f the subtest were obtained and com bined with the LM and VR Delayed Recall scores to define the MAYO Delayed Recall Index. T he MAYO system contains several m ajor com putational differences from the WMS-R but the overall results are similar, though n o t identical, using the two systems. First, all subtest raw scores are converted to agecorrected and norm alized scaled scores (MOANS Scaled Scores). These age-corrected scaled scores have a m ean o f 10 and a standard deviation o f 3 thereby allowing intersubtest com parisons. Second, the MOANS Scaled Score for each subtest is w eighted using different weights from the tradi­ tional WMS-R. T hird, the weighted MOANS Scaled Scores are grouped and sum m ed into the same subtest groupings as the traditional WMS-R. Fourth, the subtotal for the Visual subtests (Figural M emory + VR I + VisualPA I) is divided by 3 before specifying the w eighted MOANS Scaled Score sum to be converted to the MAYO Visual M emory Index. Fifth, on e new index, the MAYO P ercent R etention Index, is com puted. This m easure evaluates delayed recall as a function o f the am ou n t o f d ata originally learned as distinct from die absolute am ount o f data rem em bered (already captured by MAYO Delayed Recall Index). To obtain this score, the LM II and VR II scores are expressed as a percentage o f the LM I (LM II/LM I x 100) and VR I (VRII/VRI x 100) scores. T hese values are th en converted to age-corrected and norm alized MOANS Scaled Scores an d weighted. Finally, the six weighted MOANS Scaled Score sums are converted to their respective indices. Ivnik et al. (1992b) provided tables to convert raw scores to MOANS scaled scores and the w eighted MOANS Scaled Scorc Sums to MAYO Indices. As indicated by Spreen and Strauss (1991), the WMS-R takes longer to adm inister than the WMS. T here is no parallel form and the new “nonverbal”

MEMORY ASSESSMENT

99

tests are n o t pure m easures of visual learning/m em ory. T h e co n tin u ed absence o f recognition tasks limits the capabilities o f the scale as n o ted for the WMS and, although delayed recall tasks have been added, the WMS-R co n ten t does n o t reflect developm ents in m emory theory since the publica­ tion o f the WMS (D. W echsler, 1945). As published, the WMS-R lacks norm s for persons over age 74 an d the lack o f norm ative data for individual subscales (e.g., Verbal PA) limit the user’s ability to select specific com po­ nents o f interest from this m easure. However, the availability o f the MOANS norm ative data for persons over age 60 overcomes some o f these lim itations and has m ade the WMS-R the choice o f m any clinical practitioners working with older persons. Being able to generate norm alized scaled scores for the individual subtests increases the flexibility o f the m easure and the inclusion o f the P ercent R etention score makes the WMS-R m ore contem porary in relation to developm ents in m emory theory. However, the differences in adm inistration and scoring between the traditional WMS-R an d the Mayo version m ust not be overlooked if the obtained scores arc to be m eaningful. Care m ust be taken to ensure that only three learning trials are adm inistered d u ring the VisualPA and VerbalPA I subtests if the MOANS Scaled Scores are to be calculated for these measures. T he Guild Memory Test was introduced in 1968 (Gilbert, Levee, & Catalano, 1968) to assess im m ediate and delayed recall o f paragraphs, im m ediate and delayed recall of paired associates, digit span (forward an d reversed), and recall o f geom etric designs. T he m easure was originally standardized on 834 persons between age 20 an d 69. Crook, Gilbert, and Ferris (1980) extended the norm ative data collection to include 228 persons age 60 to 80 residing in the com munity. Persons with psychiatric disorders a n d /o r m edical problem s that may have interfered with the assessment were excluded, as were individuals who scored less than a WAIS Vocabulary scaled score of seven. Cutoff scores (i.e., next possible subtest score on e standard deviation below the m ean score attained) were determ ined for each subtest. They n o ted that “we do n o t suggest that the Guild criterion is suitable for any application o th e r than clinical research” (p. 1318). More recently, batteries o f m easures have been developed that address constructs arising from m em ory theory such as encoding strategies, learning rates, consistency o f recall, degree of vulnerability to interference, retention of inform ation over tim e, e rro r types, and im provem ent with cued recall and recognition testing. T he Memory Assessment Scales (MAS; Williams, 1991b) rep resen t a battery o f tasks derived from the m em ory assessment literature in clinical psychology, cognitive psychology, and neuropsychology. A lthough no one m em ory battery could reasonably assess all o f the tasks identified in the literature, the MAS includes m easures o f verbal and nonverbal attention, co ncentration, and short-term memory; verbal and nonverbal learning and im m ediate memory; m em ory for verbal an d nonverbal m aterial following

100

CHAPTER 6

delay; recognition, intrusions during verbal learning recall, and retrieval strategies. T he 12 subtest scores, based on seven m em ory tasks, are sum m a­ rized in T able 6.7. In addition to 12 subtest scores, three Summary Scale scores an d a Global Memory Score are calculated: the Short-term Memory score (based on the Verbal Span an d Visual Span subtests); the Verbal M emory Summary Scale (based on the List Recall an d Im m ediate Prose Recall); an d the Visual Memory Summary Scale score (based on the Visual R eproduction and Im m ediate Visual Recognition scores). T he Global M emory Scale score is based on the Verbal and Visual M emory Summary Scale scores. Verbal Process scores, which probe strategies underlying perform ance on the list learning subtcst, can also be exam ined. For exam ple, freq u en t Intrusions indicates problem s discrim inating relevant from irrelevant responses. The co n ten t o f intrusion (e.g., within sem antic category, irrelevant) may provide clues as to type of learning strategy being used. Clustering indicates w hether o r n o t sem antic categories arc being used to aid recall, with a high score indicating use of efficient strategy to assist recall. If Cued Recall is intact in the context o f p oor free recall, a retrieval problem is evident. However, if Cued Recall is also poor, an encoding deficit may be im plicated. W hen List Recognition is intact in the context o f p oor free recall, retrieval problem s are evident. If both free recall and List Recognition arc poor, th en a deficit in encoding may be apparent. T he norm ative sam ple (Williams, 1991b) consisted o f 843 adults age 18 to 90. In the age groups 60-69 years and 70+ years, th ere were 190 an d 156 subjects in the norm ative sample, respectively. T h ree sets of norm ative tables are provided in the m anual; 467 subjects reflecting the distribution of the U.S. population, 843 subjects grouped by age decade up to 70+ years, an d 843 subjects grouped by age an d three levels o f education (less th an o r equal to 11 years, 12 years, an d g reater than o r equal to 13 years). Raw scores are converted to scaled scores (M = 10, SD = 3) for each subtest. Scaled scores are sum m ed to derive the four sum m ary Standard scores (M = 100, SD = 15 for Short-term Memory, Verbal Memory, Visual Memory, an d Global M emory). T he m ethod o f continuous n o rm ing (see chap. 2 for a description o f this m ethod) was used to derive the separate norm ative data for age decade and age by education classifications for all subtests, Summary Scales, and Global Memory Scale. Verbal Process scores were too highly skewed to w arrant this process. Thus, norm ative data for the Verbal Process scores were determ ined by calculating raw score ranges above and below the m ean m inus one standard deviation. This battery is new an d little research has, as yet, been generated. In its favor, the norm s for older persons arc substantial an d it allows for a variety o f m em ory processes to be exam ined. Lezak (1995) n o ted that som e inform ation seems to be lost by n o t scoring the story recall or the

TABLE 6.7 M em ory A ssessm ent Scales (W illiam s, 1991) S ubtests Subtcst

List Learning

Prose Memory

Description Auditory verbal learning task of 12 com m on words; 3 of each from 4 sem antic categories (countries, colors, birds, cities); 6 recall trials o r until recalls all 12 Auditory verbal prose recall task—recall a sh o rt story; then asked 9 questions about details o f story

List Recall

Recall list learning; recall within cued sem antic categories; recognition from prim ed list of 24

Verbal Span

Short-term auditory memory; repeat increasing longer series o f num bers; forward and backward like Digit Span Points to series of stars in specified sequence Recognition m em ory for geom etric (nonverbal) designs. 5 = sam e/d ifferen t; 5 = recognition from an array Two trials, reproduce nonverbal geom etric design, distractor between presentation and reproduction. Scored for presence o r absence of details Associate verbal an d nonverbal m aterial; learn the nam es of people in photographs, after learning trials must recognize nam e from brief list of alternatives Recall words from List L earning, th en cued sem antic recall Asked to recall story, then asked nine questions 20 printed geom etric designs, 10 from Visual Recognition Asked to identify nam es from photos

Visual Span Visual Recognition Visual R eproduction

Nam es-Faces

Delaved List Recall Delayed Prose Recall Delayed Visual Recognition Delayed Narnes-Faccs Recall

Score Total = List acquisition score; additional scores = intrusions, successive clustering Im m ediate Prose Recall score = n u m b er o f correct responses to questions; serves as interferen ce task for next List Recall List Recall score = num ber words recalled. A dditional scores = intrusions, successive clustering, list recognition Verbal Span score is com bination of F a n d B

Visual Span = longest sequence Im m ediate Visual Recognition score = score o f all 109 trials Visual R eproduction score = scores for the two drawings

Im m ediate N am es-Faces = sum of two trials

Delayed List Recall = n u m b er o f words recalled, additional scores = intrusions, successive clustering N um ber correct to nine questions = score N um ber recognized = score N um ber correctly identified

102

CHAPTER 6

delayed design recall. Moreover, it was suggested that the designs seem to be fairly verbalizable despite being referenced as nonverbal tasks in the MAS m anual. Finally, Lezak (1995) questioned the conceptual pooling o f subtest scores to generate sum m ary scores.

VERBAL MEMORY MEASURES In addition to the aforem entioned m em ory batteries, th ere are a n u m b er o f o th e r traditional m em ory m easures for which norm ative data have been collected for older samples. List learning tasks are perh ap s the m ost com ­ m only used m easure o f m em ory functioning in older populations. D uring inform al m ental status screening, the exam inee is often given a list o f three o r four words to repeat, with instructions to rem em b er these words for later recall. Repetition is required to ensure that the words have been registered appropriately. O nce the words have been adequately registered, individuals engage in o th e r tasks for approxim ately 5 m inutes after which they are asked to recall the words. A lthough thorough norm ative data do n o t exist for this m easure, errors arc not uncom m on in large studies o f norm al elderly persons an d such sim ple tasks most likely lack sensitivity an d specificity. O th er list learning tasks are also available for use with old er persons across a range of difficulty levels. T he Hopkins Verbal Learning Test (HVLT) was designed as a very brief test o f verbal memory. Six equivalent forms have been developed and prelim inary standardization an d validation data were presented by B randt (1991). Each o f the six forms o f the HVLT consists o f a 12-item word list com posed o f four words from each o f three sem antic categories. T he exam inee is instructed to listen carefully to the words an d try to m em orize them . T he instructor reads the words at a rate o f o n e w ord every 2 seconds. T he exam inee's free recall o f the list is recorded. This procedure is repeated for a total o f th ree trials. After the th ird recall trial, the exam inee is read a list o f 24 words an d asked to indicate, after each word, w hether o r not that word was contained in the recall list. H alf of the 12 distracter words were from the same sem antic categories as the targets words (related distracters) and the rem ain d er were from o th e r categories (unrelated distracters). F.quivalcncy of the six forms an d the sensitivity o f the HVLT to condi­ tions that may affect m em ory (i.e., A lzheim er’s disease, global am nesia) have been dem onstrated by B randt (1991). T he availability o f alternate form s makes this m easure attractive. However, relatively little inform ation is as yet available addressing possible age differences o n this m easure and it may be that, with only 12 items, it lacks sensitivity to m ild m em ory im pairm ents, particularly in younger people. O n the o th e r han d , its brevity

MEMORY ASSESSMENT

103

makes it an attractive task for use with frail individuals or in settings where lim ited assessment tim e is available. T he Auditory Verbal Learning Test (AVLT; Rey, 1964) was originally de­ signed in French for children but has been translated into English and used with samples o f all ages. A 15-word list is read at a rate o f o ne word p e r second imm ediately after which free recall o f the words in any o rd er is requested. This procedure is followed for five learning trials. A second list o f 15 words is then read and recall of th a t list is requested (List B). Following recall o f this second list, recall of the original list is requested without re-presentation o f the first list (List 7). A 30-item recognition list may th en be adm inistered o r a 30-minute delayed recall o f the first list may be requested. In 1983, Q uery and M egran observed age-related declines in AVLT scores for recall, recognition, and learning. Subjects were 677 male inpa­ tients ranging in age from 19 to 81 years. O ne h u n d re d an d six o f these subjects were age 60 o r older. They provided m eans an d standard deviations for ages 60-64 (n = 57), 65-69 (« = 26), and 70+ (n = 23). In contrast, Bolla-Wilson and Bleecker (1986) found age was only related to perform ­ ance on the first two trials in a sam ple o f 114 healthy volunteers between age 40 and 80. H igher verbal intelligence was associated with better p er­ form ance on the AVLT and women consistently perform ed b etter than m en (see Table 6.8). G. Geffen, Moar, O ’H anlon, Clark, and L. B. Geffen (1990) exam ined the AVLT perform ance o f 153 adults g rouped into age categories with approxim ately equal num ber of males an d females per g roup and m atched for intelligence, education, and occupation. Forty-four persons age 60 an d over were included, and classified into two age groups (60-69 an d 70+ years o f age). Overall perform ance declined with age and females perform ed b etter than males. A num ber o f verbal learning indices were derived from the AVLT scores, including recall for each trial, the total recall sum m ed over the five learning trials, total n u m b er o f repeated words over five trials, the n u m b er o f extra-list intrusions over the five learning trials, List B recall, List 7 recall, and a 20-minute delayed recall. Following the 20-minute delayed recall trial, the subject was asked to iden­ tify the words from the original list from a printed list o f 50 words con­ taining the 15 target items, the 15 words from List B and an additional 20 phonem ically o r semantically similar distracter words. G. Geffen et al. (1990) presented m eans and standard deviations for age by g en d er group­ ings. Data on the AVLT were collected as p art of the C anadian Study of H ealth and Aging (sec chap. 2 for a description o f this sam ple an d m eth­ ods). T he T rue Positive and T rue Negative scores from the Recognition co m ponent were not exam ined in the same fashion as the rest of the data due to nonnorm ality o f regression residuals. Instead, the frequency distri­ bution for each of these m easures was exam ined with scores o f 12 or less

104

CHAPTER 6

TABLE 6.8 Perform ance Means and Standard Deviations on the AVLT by Sex and WAIS-R Vocabulary Subtest Raw Scores Men on Voiabulary

Tri/d

1 2 3 4 5 Recognition

Women on VoHibultiry

50 (n = 39)

50 (n = 43)

5.6 (1.6) 7.7 (1.9) 9.6 (2.0) 9.7 (2.7) 10.6 (2.2) 13.6 (1.8)

6.6 (1.5) 9.5 (2.2) 10.6 (2.7) 11.4 (2.5) 12.6 (2.2) 14.3 (1.1)

5.9 (1.4) 9.3 (1.9) 11.2 (1.6) 12.0 (1.8) 12.4 (2.0) 14.5 (1.0)

7.1 (1.4) 9.9 (1.8) 11.4 (2.3) 12.4 (2.1) 13.2 (2.1) 14.3 (1.0)

Note. AVLT = Auditoiy Verbal Learning Test and WAIS-R = Wechsler Adult Intelligence Scale-Revised (D. Wechsler, 1981). Vocabulary groups derived from the raw scores on the WAIS-R Vocabulary subtest. From "Influence of Verbal Intelligence, Sex, Age, and Education on the Rey Auditory Verbal Learning Test,” by K. Bolla-Wilson and M. L. Bleecker, 1986, Deuelofimmlal Neuropsychology, 2, p. 207. Copyright © 1986 by Lawrence Erlbaum Associates, Inc. Reprinted with permission.

(T rue Positives) and 13 o r less (True Negatives) falling one standard de­ viation below the m ean. T uokko and W oodward (1996) provided exact p redicted score values for age and education by g en d er and T-scores for which provided the optim al balance between sensitivity and specificity for each derived score. T hree sets o f norm ative data for the A V IT have been generated from the Mayo Clinic group. Ivnik et al. (1990) presented age-specific norm s based o n a sam ple of 394 persons age 55 years and over (see chap. 2 for a description o f this sam ple an d m ethods). A lthough Ivnik et al. (1990) p resented descriptive statistics on a trial-by-trial basis, they in troduced four sum m ary scores for the A V IT (see Table 6.9) and included norm ative inform ation for each one. Since the Total L earning (TL) scorc docs n o t capturc inform ation about learning occurring after Trial 1, a L earning Over Trials (LOT) score was calculated as an estim ate o f actual im provem ent over trials using inform ation from all five learning trials. T he two m easures of p erc en t retention reflect short-term retention (data recalled at Trial 7) and the o th e r reflecting long-term retention (i.e., 30- m inute delayed free recall). R. C. Petersen, G. Smith, Kokmen, Ivnik, an d Tangalos (1992) exam ined 161 com m unity dwelling, cognitively norm al individuals age 62 to 100

105

MEM ORY ASSESSM ENT

TABLE 6.9

Description of Ivnik et al.’s (1990) AVLT Summary Scores Summ ary Score Title

Content

T otal L earning (TL) L earning Over Trials (LOT)

Short-term Percent R etention (STPR)

L ong-term P ercent R etention (LTPR)

Sum of words recalled across Trials 1 through 5. TL - (5 x Trial 1 recall); that is, T L corrected for im m ediate word span o r nu m b er o f words recalled on Trial 1. 100 x (List A T rial 6 R ecall/Trial 5 Recall); n u m b er o f words recalled on List A Trial 6 expressed as a proportion o f the n u m b er of words recalled on Trial 5. 100 x (30-m inute Delayed Recall S co re/T rial 5 Recall); nu m b er of words recalled on 30m inute Delayed Recall expressed as a proportion o f the num ber of words recalled on Trial 5.

recruited as part of the Mayo Clinic Alzheimer’s Disease Patient Registry. Total Learning (TL), or acquisition, scores were found to decline with age hut were not related to education. Delayed recall, expressed as a change from the fifth learning trial to the 30-minute delayed recall trial (or for­ getting) remained stable over age and education level. Means and standard deviations for delayed recall from combined groups were presented. Ivnik et al. (1992c) updated their AVLT norms to be consistent with the methodology applied to the WAIS-R (see chap. 4) and the WMS-R. That is, AVLT scores (see Table 6.10) were converted to age-corrected and TABLE 6.10 Ivnik et al.’s (1992) AVLT Summary Scores Summ/iry Index

Content

MAYO Auditory-Verbal L earning Efficiency Index (MAVLEI)

MOANS Trial 1 Scaled Score + MOANS L O T Scaled Score

MAYO Auditory-Verbal Delayed Recall Index (MAVDRI)

MOANS List A Trial 6 Scaled Score + MOANS 30-m inute Delayed Recall Scaled Score

MAYO Auditory-Verbal P ercent R etention Index (MAVPRI)

MOANS STPR (see Table 6.9) Scaled Score + MOANS LTPR (sec T able 6.9) Scaled Score List B Recognition

106

CHAPTER 6

normalized MOANS Scaled Scores (M = 10, SD = 3), thus making them directly com parable to the MOANS WAIS-R and WMS-R data. The Scaled Scores were then grouped and sum m ed to yield three MAYO AuditoryVerbal Indices: a learning efficiency index, a delayed recall index, and a percent retention index. Tables converting the Scaled Score sums to MAYO Index scores are also presented. The inclusion of summary scores and the elim ination of the Total Learning (TL) score are departures from Ivnik et al.’s previous AVLT work, but were done to enable the direct comparison of AVLT scores with WAIS-R and WMS-R scores. They argued (Ivnik et al., 1992c) that TL and LOT arc redundant and, within the updated system, Trial 1 and LOT are not redundant and together provide the basis for the Learning Efficiency Index that is conceptually similar to the now aban­ doned TL. Perhaps the greatest drawbacks of the AVLT for use with elderly persons arc its length and its reliance on auditory input. H earing problem s are com m on in the elderly and may affect perform ance on the AVLT. Although most cognitively intact elderly persons are able to com plete the AVLT, tolcrancc of its length and repetitive nature decreases as fatigue and cog­ nitive im pairm ent increase. Thus, the AVLT may be best used with high functioning persons or those for whom the presence of memory impair­ m ent continues to be suspect (i.e., very mild deficits). This is n o t to say that the AVLT cannot be used with persons across the course of a dem entia but that compliance problems may increase with disease severity thereby limiting its utility as a monitoring tool. The basic form at o f the California Verbal Learning Test (CVLT; Delis, Kramer, Kaplan, & O ber, 1987) was modeled after the AVLT (i.e., list learning). The CVLT builds on the knowledge derived from cognitive psychology’s understanding of memory functions and incorporates these principles into a test with a scoring system capable of quantifying both the spared and impaired com ponent processes of memory functioning. The CVLT quantifies num erous cognitive com ponents of verbal memory within a single test and provides normative inform ation on how a task is solved, including different strategies, processes, and errors. In the CVLT, a “shopping” list (Monday list) consisting of 16 words, four words from each of four semantic categories, is presented and followed immediately by a free recall trial. This is repeated for a total of five free recall trials. A second “shopping” list (Tuesday list) is then presented followed immediately by a free recall trial. A “short-delay free recall” of the Monday list is then requested. Next, the participant is provided with a cue for each of the four semantic categories in succession and asked to rep o rt all of the items from the Monday list in that category. Nonverbal testing (e.g., block construction or finger tapping) is then adm inistered for 20 minutes, after which a “long-delay” free recall of the Monday list is

MEMORY ASSESSMENT

107

requested and followed by a “long delay” cued recall o f this list. Finally, a list o f 44 words is read to participants an d they are asked to rep o rt w hether o r n o t the item was on the Monday list. T he level o f th e participant’s perform ance on a n u m b er o f CVLT pa­ ram eters can then be generated using standard score tables (e.g., short an d long delay free recall, clustering, primacy and recency effects, consis­ tency of item recall, perseverations, intrusions, response bias). Delis et al. (1987) described how differences in perform ance between the various trials may shed light on the underlying type o f m em ory problem manifest. For exam ple, a large negative difference in standard scores between Trial 1 o f List A and List B may reflect an unusually high degree o f proactive interference. T he norm ative sam ple for the CVLT consists of 273 neurologically intact individuals (104 males, 169 females; m ean age = 58.93, SD — 15.35; m ean educational level = 13.83, SD = 2.70). Because this reference group is a com bination of several independently collected samples, sm oothed age curves were fitted to raw data using m ultiple regression. T he curve provides an estim ate o f the m ean score at each age for sex, because age an d sex accounted for a substantial portion of the variance on all recall measures an d for sm aller (but statistically significant) proportions on 11 o f the 14 o th e r variables. A similar curve-fitting technique was used to estim ate the standard deviation at each age and sex. T he Standard Score shows the n u m b er o f standard deviations any raw score deviates from the expected m ean for that age an d sex. A dditional data available in the m anual include inform ation on internal consistency, test-retest reliability, validity, factor structure, criterion-related validity, and score patterns o f selected n eu ro ­ logical groups including A lzheim er’s disease and Parkinson’s disease. Norm ative inform ation is available to age 80 by gender. T he scoring systems allow for at least 19 different aspects o f m em ory functioning to be exam ined. T he shopping list nature o f th e task lends some ecological validity to this m easure and may be viewed as assessing som ething m ore relevant to daily functioning than o th e r learning lists. However, like the AVLT, the CVLT may be too long for some patient populations, particularly the very old and those with cognitive im pairm ent. According the Lezak (1995), Kaplan developed a 9-item version (3 words from 3 categories) th a t may be m ore suitable for these populations, although no norm s are as yet available. A nother variant o f the list learning task differs in p rocedure from the afo rem entioned m easures in that selective rem inding takes place after the initial recall trial. T he earliest and m ost frequently cited version o f this type o f m easure is the Selective Reminding Test (SRT; Buschke, 1973) using a 12-word recall list. Following each trial, only those words n o t recalled by the subject are presented again by the exam iner. Typically, this procedure

108

CHAPTER 6

is repeated until a criterion is m et (e.g., 2 consecutive trials com pletely recalled u n d er the Free Recall condition) o r until 12 learning trials have taken place. However, a variety o f adm inistration procedures has been em ployed. In addition, a variety of scores has been derived from the SRT. O th er variants (Pictorial cues) have also been developed th at use a selective rem inding procedure. Banks, Dickson, and Plasay (1987) adm inistered the SRT to 60 volun­ teers from seniors’ centers, retirem ent clubs, and churches in Mississippi age 65 to 75. T he m ean age was 69.2 years (SD = 3.3) an d the m ean n u m b er o f years o f education for this sample was 15.15 (range 6-18.8, SD = 3.45). A 12-word list devised by Levin, Benton, an d G rossm an (1982) was read at a rate o f one word every 2 seconds. A m axim um of 12 learning trials or 2 consecutive trials com pletely recalled u n d er free recall was em ­ ployed. T he following salient d ep e n d en t m easures were selected for use in this study: Sum Recall (num ber o f words recalled across all trials), Long-Term Storage (num ber of words recalled on at least two consecutive trials), Consistent Long-Term Retrieval (num ber o f words recalled on all subsequent trials w ithout rem inding), Short-Term Recall (n u m b er o f words never recalled on two o r m ore consecutive trials), Long-Term Retrieval (n u m ber o f words not in Consistent Long-Term Retrieval). No age effects were ap p aren t across the age categories 65-69 (n = 32), 70-75 ( n = 28). Females (n = 37) perform ed b etter than males (n = 23) on all sum m ary m easures already noted except Long-Term Storage and a significant de­ crease in recall was app aren t from the last recall trial to a 15-minute delay recall trial. Means and standard deviations by g ender were presented. Ruff, Light, and Q uayhagen’s (1989) norm ative study o f the SRT con­ tained 33 individuals from age 55 to 70. All subjects were screen ed with an interview questionnaire and selection criteria included no history of neuropathology, no hospitalization fo r psychopathology, no significant his­ tory o f drug o r alcohol abuse, and English as the prim ary language. T h e exam iner read a list o f 12 discrete words that were individually rep eated by the subject to assure initial com prehension o f each word. Recall o f the list was requested and rem inders were subsequently given only for the words not recalled on the free recall trial. This procedure was co n tin u ed until all 12 words were recalled on two successive trials or when 12 recall trials had been exhausted. Following a 1-hour delay, recall o f the list was again requested. No prior w arning that a 1-hour delay recall would be requested was given. No significant age o r education effects were found within this sample, although a gender effect, with w om en’s perform ances su p erior to those o f m en, was noted. Larrabee, T rah an , Curtiss, and Levin (1988) provided norm s for 7 age groups from age 18 to 91, which include all 11 o f the scores defined in Spreen an d Strauss (1991). Participants were 271 healthy, nonhospitalized,

MEMORY ASSESSMENT

109

n o npaid volunteers with no reported history of neurological o r psychiatric disorders. T he Verbal SRT, Form 1, was adm inistered following the pro­ cedures o f Buschke (1973) (i.e., words read at 2-second intervals, selective rem inding after first recall trial). Learning criterion was recall o f all 12 words on 3 consecutive trials o r com pletion o f 12 trials. After the final recall trial, a cued recall trial was adm inistered in which cards containing the first two o r three letters of the stimulus word were p resented an d the exam inee was asked to identify which word on the stimulus list began with these letters. This was followed by a m ultiple-choice recognition task. T he exam inee was asked to discrim inate the list word from three foil words (a phonem ic foil, a sem antic foil, and an unrelated foil) that were p rin ted on 5-by-8 index cards. Thirty m inutes later, free recall, w ithout cueing or rem inding was assessed. Both age and sex appeared as salient subject variables with education being relatively unim portant. G en d er corrections were applied to the data. T he norm ative data for persons age 60 an d older arc presented in Table 6.11. Note that m ean values do n o t always necessarily sum to the exact totals (e.g., LTR and STR). T hese small discrepancies ap p ear related to the g en d er corrections for the different scores (Spreen & Strauss, 1991). M asur ct al. (1989) adm inistered one o f four different 12-word versions o f the selective rem inding task to 134 participants from a longitudinal d em entia risk factor study. All persons screened as cognitively norm al for 2 consecutive years. T he average age o f the sam ple was 80 years with over two thirds being female and less than o n e third with less than 9 years of education. T he words were presented visually (on index card) and were sim ultaneously read to the persons at 5-second intervals. T he person was asked to repeat each word. A fter the presentation o f all 12 words, the person was asked to recall as m any words as possible in any order. After 60 seconds, the exam iner verbally rem inded the individual o f words that were n o t recalled u n d e r the free recall condition at 2-second intervals. This procedure was repeated for 6 trials. Im m ediately after the last trial, the individual was asked to perform 60-second trials of spontaneously gen­ erating words within the categories o f fruits, animals, flowers, vegetables, an d nam es. Following com pletion o f the 5 category retrieval tasks, subjects were asked to recall as many words from the original list as possible. For words that were n o t retrieved, during this Delayed Recall (DR) trial, the person was asked to recognize each word from an array o f four alternative choices presented verbally and consisting o f the word, a semantically re­ lated distracter, a phonem ically related distractcr, and an u n related distractcr. A m axim um correct scorc for Delayed Recall + Delayed Recognition was 12. N o significant effects o f sex, age, education, o r test form were found. Table 6.12 contains the collapsed m eans and standard deviations for this sample.

110

CHAPTER 6

TABLE 6.11 Normative Data on the SRT for Persons Age 60 and Older Age

VSRT Measure LTR STR LTS CLTR RLTR Rem inders Intrusions C ued Recall M ultiple Choice Delayed Recall

6(1-69 (n = 50)

70-79 (n = 59)

80-91 (n = 27)

101.52 (24.68) 13.52 (9.52) 107.00 (21.79) 88.92 (35.85) 14.66 (11.83) 28.12 (15.16) 3.90 (7.29) 9.58a (1.93) 11.96 (.20) 9.58 (2.46)

89.95 (29.23) 17.47 (10.41) 95.54 (24.86) 69.68 (35.96) 20.71 (14.37) 36.95 (15.17) 4.22 (5.76) 8.95b (2.12) 11.85 (.58) 9.05 (2.62)

77.22 (26.26) 20.74 (9.62) 87.48 (25.26) 54.96 (29.04) 1*2.19 (10.70) 43.96 (15.77) 3.30 (5.09) 8.16c (2.22) 11.93 (.27) 8.37 (2.45)

Note. C orrection values for raw scores of males (calculate before e n te rin g norm ative tables). T otal = +5; LTR = +9; STR = -4 ; LTS = +7; CLRT = +13; RLTR = -5 ; Rem inders = -5 ; Intrusions = 0; Cued Recall = 0; M ultiple Choice = 0; Delayed Recall = +1. Caution: Do not correct LTS o r CLTR if raw score is 0. From “Normative Data for the Verbal Selective R em inding Test,” by G. J. Larrabee, D. E. T rahan, G. Curtiss, an d H. S. Levin, 1988, Neuropsychology, 2, p. 179. Copyright © 1988 has been transferred to the A m erican Psychological Association. A dapted with permission. an = 31. bn = 38. cn = 19.

Deptula et al. (1990) administered five forms of a 16-word SRT to normal young adults and 45 normal elderly persons (age 60 to 79). Of note, these groups had relatively high levels of education (elderly mean = 14.2, SD = 2.9) and age-adjusted performance on the Vocabulary subtest of the WAISR (elderly mean = 14.1, SD = 2.4). Words were read at a rate of 3 seconds per word. Seven recall trials of the same list were administered. Each subject received three of the five forms at approximately 1-week intervals (as baseline measures in a drug trial). Scores were corrected for practice effects and the equivalency of the five forms was examined. The five forms showed adequate reliability for the Total Recall measure. Forms 1 and 4

111

MEMORY ASSESSMENT TABLE 6.12 P erfo rm an ce o n C o m p o n en ts o f the SRT for M asur et al.’s (1989) N orm al Elderly Sam ple ( N = 134) Measure Sum Recall LTR STR Storage Estimate CLTS CR CLTR Intrusions SDR SDR + DMCR

M

SD

45.53 34.97 10.56 9.05 39.63 24.09 21.34 0.82 6.79 11.70

7.94 11.51 4.77 5.01 11.84 9.43 11.82 1.16 2.68 0.54

Nut*. From “Distinguishing Normal and Dem ented Elderly With the Selective Reminding Test," by D. M. Masur, P. A. Fukl, A. D. Blau, L. J. Thai, H. S. Levin, and M. K. Aronson, 1989, Journal of Cliniml and Experimental Neurttfisychology, 11, p. 621. Copyright © 1989 by Swets & Zeitlinger Publishers. Used with permission.

showed the lowest correlation. Strongest correlations were between Forms 1, 2, an d 3. T he subcom ponents were less consistently reliable th an the total score. LTR an d CLTR ten d ed to be the most reliable m easures. Forms 1 and 4 were again the weakest. T rahan an d L arrabee (1993) exam ined rate of forgetting using various indices derived from the Verbal SRT in 287 adults age 18 to 91. O ne h u n d re d and fifty-nine o f these individuals were age 50 o r older. A for­ getting score was defined as an A cquisition Score m inus the n u m b er of words freely recalled during the 30-minute delayed recall task. T h ree ac­ quisition scores were used and exam ined by age: the n u m b er o f words recalled on the final learning trial, the n um ber o f words in long-term storage in the final learning trial, and the num ber o f words in consistent long-term retrieval on the final learning trial. Significant decrem ents were observed across age groups (18-29, 30-49, 50-69, 70+). O lder persons exhibited a higher rate o f forgetting than younger persons using the second forgetting index, although the am ount forgotten was very m odest. W hen the third forgetting index was used, older persons actually exhibited no forgetting, obtaining higher scores on the delayed task than d u ring acqui­ sition. Percentile ranks for each forgetting score are presented for persons age 50 and over. T rahan and L arrabee (1993) suggested that the second forgetting index (i.e., Trial 12 LTS) appears to be the m ost useful because it provides the best fit with m ost traditional models o f long-term memory. T he strength and weakness o f the SRT lies in the variety of adm inistra­ tion and scoring procedures that have been used with this recall paradigm . N one of the set o f norm s presented here are derived from the same

112

CHAPTER 6

adm inistration and scoring procedures. Some studies use 6 learning trials, others 7, and still others 12. Some studies use 12 words, others 16. Some delayed recall intervals are 60 minutes, others 5 minutes. Not all papers clearly identify their departures in administration and scoring procedures. It has been noted that SRT scores are often redundant, perhaps measuring similar com ponents of memory (Larrabee et al., 1988). Moreover, Spreen and Strauss (1991) noted that the operational definitions for some of the subscales may not, in fact, be tapping the hypothesized underlying con­ structs adequately. T he selective rem inding paradigm has been used in a variety of other formats. Schmidt, Tom baugh, and Faulkner (1992) developed measures o f memory for verbal passages, list, and word-pair learning using a selective rem inding paradigm for each. The Passage measure consisted of an athematic paragraph containing 31 bits of verbal inform ation. After presenta­ tions of the passage, a free recall condition was employed, followed by a cued recall condition (i.e., specific questions asked about the content o f the passage). Two learning trials were administered. After a 20-minute delay, free recall of the passage was requested. This was followed by cued recall and a recognition task (4-item multiple-choice format) for items missed under both free and cued recall conditions. The Word List measure contained 15 unrelated words presented five times. Following each free recall trial, semantic cues were provided for items missed un d er free recall. The delayed recall trial included free recall, cued recall, and recognition (i.e., 4-item multiple-choice) com ponents. The Word Pairs measure con­ tained 3 easy and 11 hard (8 concrete and 3 abstract) word pairs. Four learning trials were administered. Prior to the first recall trial, the entire list was presented, but only pairs that were missed were presented on the subsequent trial. Following each presentation, the first word of the pair was presented and the person was required to recall the second word. Again, a delayed recall trial was adm inistered consisting of free recall, and recognition (4-item multiple-choice format) com ponents. Age effects were apparent for all three measures. O lder persons benefited m ore from cueing and recognition than younger subjects. Retention scores also yielded significant age effects that may have been due to the lower initial learning levels of older subjects. W hen retention scores were exam ined in terms of percent savings (retention trial score/last learning trial scorc x 100), age effects typically continued to be evident (except for passage free recall). Tables summarizing the normative data for subjects age 60 and older for these three verbal memory tasks were presented. These measures are com ponents from a m ore extensive battery of memory tests, the Learning and Memory Battery (LAMB; Tombaugh & Schmidt, 1992). The Cued Recall paradigm for memory assessment, another measure based on a selective rem inding procedure, was introduced by Buschke (1984)

MEMORY ASSESSMENT

113

specifically for use with elderly populations. In his original study, Buschke displayed line drawings o f 12 com m on objects in ran d o m o rd er to 10 persons over age 60 an d asked them to locate and nam e each picture in response to a stimulus cue (e.g., W hich one is a piece o f clothing?). Fol­ lowing this search condition, the pictures were rem oved an d the person was asked to co u n t backward from 100 for 60 seconds as a distraction task. A fter 60 seconds had passed, the person was asked to recall as many o f the pictured items as possible in any order. W hen the subject was n o longer able to provide additional responses (> 20 seconds), the sem antic stimulus cues were given for all items missed on free recall. If any items were n o t recalled u n d e r free o r cued recall conditions, then the pictures were once again displayed an d the person was asked to search for an d nam e only those item s (i.e., missed u n d e r both free and cued recall). T he pictures were once again rem oved from view. T he same recall procedures were followed for six learning trials. In developing this m easure for use with the elderly, the search procedure was em ployed to ensure or enhan ce encoding, because encoding has been implicated as contributing to m em ­ ory deficits o f older persons. T hree different channels were em ployed to assist in encoding (i.e., visual pictures, category cues an d nam ing). Thus, unlike many oth er m em ory measures, this procedure was designed to fa­ cilitate and control initial encoding o f the to-be-rem em bered item s so th at deficits in retrieval and acquisition could be m ore clearly exam ined. A retrieval score was generated by sum m ing the items across trials u n d er the free recall condition. An acquisition score was derived by sum m ing items across trials for both free and cued recall conditions. This basic pro ced u re was later m odified by T uokko et al. (Tuokko & Crockett, 1989; Tuokko, Vernon-W ilkinson, Weir, & Beattie, 1991) to standardize the layout o f the stimulus figures, reduce the n um ber of learning trials to three, and include a delayed learning trial (5-m inute delay later m odified to 15-minute delay). Four sum m ary scores were identified by Tuokko et al.: FR o n Trial 1, Retrieval (sum o f free recall over th ree trials), Acquisition (sum o f FR + CR over three learning trials), and R etention (sum o f FR + CR for delayed recall trial). Data on Tuokko e t al.’s (1989, 1991) m odification o f the Cued Recall paradigm were collected as p art o f the C anadian Study o f H ealth and Aging (see chap. 2 for a description o f this sample an d m eth o d ). T he Acquisition and R etention scores were n o t exam ined in the same fashion as the rest o f the data due to nonnorm ality o f regression residuals. Instead, the frequency distribution for each o f these measures was exam ined with a R etention score o f 0-11 falling below the 16th percentile. For persons age 65 to 84, an Acquisition score o f 0-35 fell m ore th an o ne standard deviation below the m ean, whereas for persons 85+, scores o f 0-34 fell in this range. Tuokko and W oodward (1996) provided exact predicted score

114

CHAPTER 6

values for age and education by gender. T he 'Ascore for Retrieval th at provided the optim al balance between sensitivity (.83) an d specificity (.94) was 34. G rober, Buschke, Crystal, Bang, and D resner (1988) m odified the basic cued recall paradigm to include 16 items, presenting 4 item s at a time. After each set o f four items was searched, identified, an d nam ed, im m ediate verbal cued recall o f ju s t those four items was tested by presenting each cue. O nce all four items were recalled correctly in response to the cues, the next set o f four items was presented for the search procedure. After all 16 items were searched and recalled through im m ediate cued recall, three recall trials, each preceded by 20 seconds of interference (counting backward) were conducted. Each recall trial consisted o f two parts: 2 m in­ utes for free recall and cued recall o f those item s n o t recalled u n d e r the free recall condition. If an item was not retrieved within 10 seconds o f category cue presentations, a rem inder o f the item was given. Im m ediately following the last recall trial, the word and an equal n u m b er o f related and unrelated foils were presented to assess recognition memory. O ne w ord at a tim e was presented and the participant was asked to decide w h ether o r not the w ord was in the m em ory list. This pro ced u re has been effective in correcdy classifying 7 1/73 dem entia cases an d 4 7 /4 7 cognitively in tact individuals (G rober & Buschke, 1987; G rober et al., 1988). R. C. Petersen et al. (1992) collected data for 161 persons age 60 and old er with this basic procedure. Six recall trials were em ployed an d a delayed recall trial was adm inistered after 30 m inutes. No recognition task was em ployed. Correlations with age were noted for a variety o f scores derived from this m easure. Mean and standard deviations for a variety o f scores by age were presented. A nother le a rn in g /re te n tio n m easure developed specifically for use with old er persons is the Aronson Shopping List (ASL; A ronson, 1985). Tsang, A ronson, and Hayslip (1991) noted that m ost existing m easures o f m em ory have little face validity with respect to personal relevance for an elderly individual and m otivational/em otional factors may confound th e ir per­ form ance on these measures. T he ASL was designed specifically to help differentiate learning/m em ory dysfunction o r im pairm ent from states o f benign forgetfulness am ong older individuals. ASL training requires the person to learn a shopping list o f 10 noun-adjective pairs with nouns used as the stimulus for recall. Pairs were chosen to balance g en d er interests an d to form unusual com binations (e.g., wide magazine). T h e list is pre­ sented at a rate o f 3 seconds between pairs and 1 second between n o u n ­ adjectives. All pairs are presented in a p redeterm ined o rd er in the first thro ugh third trials. T hereafter, pairs are elim inated once they are recalled accurately. Im m ediate feedback is given after each recall attem pt. O ne p o in t is given for each e rro r in the first three trials an d two points are

MEMORY ASSESSMENT

115

given for each error thereafter. The test is term inated after an errorless trial or after 7 training trials. Delayed recall, or ASL retention, is assessed after approximately 30 minutes of additional testing, which does not in­ clude verbal memory tasks. Delayed recall responses are given 10 points for each correct adjective, 5 points for each synonym, and 2 points for any adjective that is recalled but mispaired. Unrelated responses or guesses are given 0 points. Tsang et al. (1991) provided normative inform ation for a group o f 81 persons between the age of 61 and 87 recruited through newspaper ads and visits to organizations such as American Association of Retired Persons. They were paid a small hourly wage for participation. Eighty-three percent had m ore than a high school education and 93% reported their health status as “good” to “excellent.” Participants under age 75 differed from those over age 75, when other dem ographic variables were controlled, on ASL training erro r score and num ber of adjectives recalled after delay. Females made fewer errors than males. Training error and retention scores, adjusted statistically to control for the effects of other dem ographic variables, were presented. Tsang et al. (1991) suggested that training error scores greater than 10 be used to identify im pairm ent for females under age 75 who graduated from high school. Two additional points could be allowed for those whose education was below eighth grade and one point for those who com pleted m ore than grade eight but did not com plete high school. They also suggested that two points be added to scores for males, and one point for each 5 years of age beyond 75. Tsang et al. (1991) asserted that the presence of a problem perform ing either of these tasks is suggestive of processes other than norm al aging but, as yet, it is unclear whether or not the measure will be sensitive to mild memory impairments. The Recognition Memory Test (RMT) is actually two tests, parallel in form but providing verbal (words) and relatively nonverbalizable (faces) stimuli for assessing material specific memory deficits. Each test contains 50 stimu­ lus items and 50 distracters. Diesfeldt and Vink (1989) extended the normative data on the RMT for faces to include 89 subjects between age 69 and 93. The standard recognition for faces was used (i.e., all faces are male and for the recognition trial each item is paired with a photo of man o f similar age and degree of hairiness with random ized positions). As in the original version, the stimuli were presented at 3-second intervals. The subject was to judge the pleasantness of the stimulus. Recognition memory was assessed immediately after presentation by asking the subject which item of a pair (item and distracter) had been seen earlier. In W arrington’s (1984) original version, raw scores are converted to percentile scores for three age groups (18-39, 40-54, 55-70, or to stand­

116

CHAPTER 6

ardized scaled scores with a range of 3 to 18), thereby limiting their use with older individuals. Diesfeldt and Vink (1989) found that the RMT for faces was acceptable for use with norm al elderly persons in that it was not overly time consuming and that the distribution of scores revealed no ceiling effects. Distribution of scores on the RMT for faces shows a decline across age groups. Diesfeldt and Vink (1989) noted that the RMT for faces is not independent of intellectual ability and this needs to be considered when interpreting perform ance on this measure.

VISUAL LEARNING MEASURES A num ber of visual learning measures have normative data for elderly persons. The most frequently used visual recall test, the Benton Visual Re­ tention Test (BVRT), has three forms and various adm inistration procedures. The most com mon of these is a 10-second exposure of each stimulus card followed by immediate recall by drawing (Administration A). Normative data are provided in the test manual for age 8 to 64 (Benton, 1974) with a progressive decline in recall after age 40. Robcrtson-Tehabo and Arenberg (1989) provided means and standard deviations for persons age 60 to 69 ( n= 28), 70 to 79 (n= 47), and 80 to 89 (n= 15) from the Baltimore Longitudinal Study of Aging. A marked decline in perform ance, particu­ larly after age 70, was noted. Youngjohn, Larrabee, and Crook (1993) presented data for persons age 18 to 70+ at three education levels. The BVRT Administration A (10-second exposure, immediate recall) was ad­ m inistered to 1,128 volunteers who reported no evidence o r history of physical, psychiatric, or neurological conditions that could affect memory on a health history questionnaire and who scored within the nondepressed range on the Affective Rating Scale (Yesavage, 1986). Table 6.13 summ a­ rizes these data for persons over age 60. Youngjohn et al. (1993) generated the following predicted score equations for specific individuals: Predicted BVRT # correct ( + / - 1.57) = 7.87 - .045 (age) + .098 (years of education); Predicted BVRT # errors ( + / - 2.88) = 1.73 + .088 (age) - .126 (years of education). Administration M (10-second exposure, multiple-choice format) has been used with elderly persons to overcome limitations imposed by the drawing response. A num ber of physical conditions (e.g., arthritis, tremor) may prohibit use of drawing tasks in older persons. Norms in the manual only extend to age 55. K. M. Montgomery (1982) used the multiple-choice version of the BVRT with a 5-second exposure to assess healthy elderly subjects age 65 to 89. Data arc summarized in Table 6.14. The BVRT multiple-choice form al was adm inistered as part of the Canadian Study of Health and Aging (see chap. 2 for a description of this sample and m ethod).

TABLE 6.13 N orm ative Data on the BVRT for Persons Age 60 an d O lder By E ducation BVRT Performance Education

n

Mean Correct

Mean Errors

Age 60-69 12-14 years

129

15-17 years

159

18+ years

134

6.18 (1.67) 6.70 (1.47) 6.80 (1.55)

5.55 (2.74) 4.99 (2.78) 4.93 (2.87)

5.62 (1.73) 6.06 (1.84) 6.22 (1.57)

7.28 (3.55) 6.74 (4.34) 6.33 (3.63)

Age 7(H12-14 years

53

15-17 years

54

18+ years

49

Note. From “New Adult Age* and Education-Correction Norms for the Benton Visual Retention Test,” by J. R. Youngjohn, G. J. Larrabee, and T. H. Crook, 1993, Clinical Neuropsychologist, 7, p. 158. Copyright © 1993 by Swets 8c Zeitlinger Publishers. Used with permission.

TABLE 6.14 Perform ance o f Elderly Persons on the BVRT M ultiple C hoice With 5-second Exposure Score 6 7 8 9 10 11 12 13 14

Percentile 2.4 4.7 12.9 24.7 42.4 55.3 82.4 95.3 100.0

Note. From A Normative Study of Neuropsychological Test Performance of a Normal Elderly Sample (p. 79) by K. M. Montgomery, 1982, University of Victoria. Unpublished master’s thesis. Adapted with permission.

117

118

CHAPTER 6

Raw score to scaled score conversions an d exact prediction scaled score for the C orrect BVRT-MC scores are provided by age, gender, an d education (Tuokko & W oodward, 1996). A nother m easure o f visual m em ory that requires m inim al m o to r dem ands is the Visual Spatial Learning Test (VSLT; Malec, Ivnik, & H inkeldey, 1991). T h e VSLT requires a person to learn to recognize seven designs th at are difficult to verbally encode an d to recall the correct placem ent o f these designs on a 6 x 4 matrix. A fter seeing the design placed on the 6 x 4 grid, persons are given an em pty grid and 15 designs from which to select and place the 7 targets as they were when seen on the grid. Five learning trials are adm inistered and a 30-minute delayed recall trial. Scores determ in ed for each trial include: the n um ber o f correct designs chosen (D), the nu m b er o f correct positions chosen (P), the num ber o f correct designs placed in the correct position (PI)), and the num ber o f incorrect designs (i.e., intrusion errors) (E). Scores were sum m ed over the five learning trials to yield sum m ary scores identified as DSUM, PSUM, and PDSUM. ESUM is the sum o f intrusion errors across all trials (i.e., 5 learning + delayed recall trial). Malec et al. noted that because PSUM and PDSUM were highly correlated, only PDSUM was exam ined. A norm ative sam ple o f 455 persons age 56 to 97 from the Mayo O lder Americans Norm ative Studies (see chap. 2 for a description o f this sample and m ethod) was adm inistered this task (Malec et al., 1991). Conversion o f raw ESUM, DSUM, and PDSUM scores to Scaled Scores with a m ean o f 10 and SD o f 3 is presented. ESUM did n o t correlate with age and was standardized on the entire sample. T he D an d PD scores for the delayed recall trials were handled differently because o f the high correlation between these scores an d their respective Trial 5 scores. F or each possible score on Trial 5, a delayed score was given an im p airm en t score (IS) o f - 2 (falling below 15th percentile), -1 (falling between 16th and 35th p ercentile), or 0 (scoring above 35th percentile). This pro ced u re was developed for the entire sample because it was argued th a t rem oving the redundancy with Trial 5 would also remove m ost o f the variance associated with age. This procedure was adopted instead o f expressing delayed recall scores as a percentage o f original learning as percentage scores are often correlated with both the num erato r and den o m in ato r an d thereby only partially reduce the redundancy with Trial 5. This m easure is quite new so little inform ation is yet available. G. E. Sm ith et al. (1992), in factor analyzing a battery o f m easures including the VSLT, questioned the validity o f the construct o f nonverbal m em ory as distinct from verbal memory, especially in an older norm al adult p o p u ­ lation. However, this does not mitigate the im portance o f the developm ent o f this n onm otor d ep e n d en t m easure o f visual memory. F u rth er research and clinical practice will determ ine the relative usefulness o f this m easure in relation to others currently available.

MEMORY ASSESSMENT

119

Visual m em ory abilities have also been assessed using the Rey Osterrieth Complex Figure (ROCF) and the Taylor Complex Figure (TCF) across age ranges. Typically, persons are asked to copy the figure an d are n o t told th at they will be asked to reproduce the figure from m em ory at a later time. Delayed recall is typically required and some investigators also use an im m ediate recall condition. If an im m ediate recall condition is used, this serves as a facilitator for future recall trials (i.e., Delayed Recall), consequently it is imperative to use appropriate norm s (i.e., same proce­ d ure) for these tasks. Spleen and Strauss (1991) provided data for persons over age 60 on the copy with delayed recall version o f the ROCF using the scoring criteria provided in their book. In addition to providing data for the 3-m inute recall for persons age 45 to 83, Boone, Lesser, Hill-Cuticrrcz, Berm an, and D'F.lia (1993) com puted a percen t reten tio n scorc. A sample o f English-speaking volunteers were screened for history o f psy­ chotic o r m ajor affective disorder, cu rren t o r past history o f alcohol or o th e r substance abuse, and docum ented neurological illness and significant medical illness that could affect the central nervous system. In addition, persons who were found to have abnorm al neurological exam ination, m e­ tabolic abnorm alities detected with blood tests, or abnorm al findings on m agnetic resonance testing were elim inated from the sample. T h e final sam ple consisted of 91 individuals (34 males, 57 females) with a m ean educational level o f 14.5 (SD = 2.5). Berry, Allen, an d Schm itt (1991) adm inistered both the ROCF and the TCF to 107 persons age 50 to 79. Interrater, alternate forms, test-retest, an d internal consistency reliabilities were all adequate to good for the recall trials o f the ROCF. T he ROCF was significantly correlated with age an d education but n o t gender. No norm ative data were p resen ted but the observed correlations with age an d education suggest th at these factors be considered when developing adequate norms. A subsequent study by Berry a n d C arpenter (1992) found no differences when the length o f the delay interval was varied (15, 30, 45, o r 60 m inutes) between age an d g en d er equivalent groups o f persons age 60 and older. T om baugh, Schmidt, and Faulkner (1992) adopted a different adm inis­ tration procedure for the TCF to control for the duration of stimulus exposure an d provide inform ation on rate o f learning an d level o f reten­ tion. New scoring m ethods were also developed. Participants were told th at they would be shown a design they would have to draw from mem oiy, they would be given four tries at this, and they would be required to recall the design later. O n each o f the four trials, the TCF was exposed for 30 seconds. T h e figure was rem oved and the subject was allowed 2 m inutes to reproduce the figure from m emoiy. Approximately 15 m inutes after the last acquisition trial, during which o th e r measures were adm inistered, delayed recall was requested. Following this, the subject was given 4 m inutes

120

CHAPTER 6

to copy the TCF with the stimulus figure present. T h e average am o u n t o f time to adm inister the test, excluding the delay interval, was approxim ately 13 m inutes. An item ized scoring system (Tom baugh, 1989) yielded a maxi­ m um o f 69 points. Age effects were ap p a ren t (see T able 6.15). Data were also expressed as a p erc en t o f the copy score. This may be useful for d eterm ining w hether or n o t a person with a mild constructional im pair­ m e n t also has a m em ory deficit over and above the constructional one. As noted by Lezak (1995), the revised procedure for T o m b au g h ’s version to the TCF allows for the generation o f a learning curve, thereby adding im p o rtan t inform ation not obtained on o th e r com plex figure tasks. T om ­ baugh et al. (1992) asserted that, if com parison between visual an d verbal m em ory m easures is im portant to neuropsychologists, then this new p ro ­ cedure m ore closely parallels the verbal learning m easures typically adm in­ istered and thereby should yield m ore com parable data than the traditional ROCF or the TCF. Moreover, the inclusion o f norm ative data assisting in TABLE 6.15 S um m ary o f Data for Persons Age 60 a n d O ld e r o n th e T C F Age Measure Trial 1 M SD Trial 2 M SD Trial 3 M SD Trial 4 M SD Total Recall M SD Delayed Recall M SD Copy M SD

60-69

70-79

28.5 9.9

20.7 10.8

44.3 13.2

31.0 13.3

51.5 12.0

39.1 14.5

55.0 11.6

43.6 14.1

179.3 43.7

134.9 49.6

54.6 11.0

43.7 14.8

65.3 3.5

63.0 4.4

Note. TCF = Taylor Complex Figure. From “A New Procedure for Adm inistering the Taylor Complex Figure: Nonnative Data Over a 60-year Age Span,” byT . N. Tombaugh, J. P. Schmidt, and P. Faulkner, 1992, Clinical Neuroj/.sychologist, 6, p. 69. Copyright © 1992 by Swcts & Zcttlinger Publishers. Used with permission.

MEMORY ASSESSMENT

121

the differentiation o f copy versus m em ory deficits may prove very useful to clinicians. T he detailed scoring system may prove to be a lim iting factor for those with little scoring time available.

SELF-REPORT MEASURES A n other type o f m em ory m easure that may be o f use in the assessm ent o f o ld er adults is the rep o rt o f self-perceived m emory functioning. Com plaints o f changes in m em ory functioning may accom pany the norm al aging proc­ ess and may be o f use when com pared to objective m easures o f m em ory functioning in the diagnostic assessment process (see chap. 10). Gilcwski and Zelinski (1986) critically reviewed 10 questionnaires developed to as­ sess m em ory com plaints in older adults. O f these, the psychom etric p ro p ­ erties o f 6 had been investigated. A lthough they generally reco m m en d ed the Memory Functioning Questionnaire (MFQ; Gilcwski, Zelinski, Schaie, & T ho m pson, 1983) and the Metamemory in Adulthood instrument (MIA; Dixon & H ultsch, 1984), both of these m easures were primarily designed as re­ search instrum ents an d are probably o f excessive length (64 an d 120 items, respectively) for m ost clinical situations. A sh o rter m easure has been developed that although less conceptually sophisticated than the MFQ an d the MIA, may be m ore practical in m ost assessment situations. C rook and L arrabee (1992) rep o rted norm ative data for a m easure containing 21 items reflecting a perso n ’s ability to rem em ber specific types o f inform ation and 24 items reflecting how often specific m em ory problem s occur (MAC-S). Both types o f items were recorded on a 5-point Likert scale with choices ranging from “very p o o r” to “very good” o n the Ability Scale, and from “very o ften ” to “very rarely” on the Frequency scale. V olunteers age 18 to 92 were recruited to take p art in the norm ative study. All persons identified through interview as having neurological dis­ o rd ers that could produce cognitive changes, m ajor psychiatric distur­ bance, o r a history o f substance abuse were excluded from the sample. T h e MAC-S was adm inistered to 1,106 persons. T he data were grouped into five Ability factors (Rem ote Personal Mem­ ory, N um erical Recall, Everyday Task-O riented Memory, W ord R ecall/Se­ m antic Memory, Spatial and T opographic Memory) an d five Frequency factors (W ord and Fact Recall o r Semantic Memory, A ttention an d Con­ centration, Everyday-Task O riented Memory, G eneral Forgetfulness, and Facial Recognition) (Crook & Larrabec, 1992). Total scores for the Ability (max = 21 x 5) and Frequency scales (max = 24 x 5) were also calculated, as was the sum for four item s rating global m em ory functioning (e.g., In general, com pared to the average individual, how would you describe your memory?; How would you describe your memory, on the whole, as com ­

122

CHAPTER 6

p ared to the best it has ever been?). Norm ative data for persons age 60 to 69 (n = 346) and 70+ years (n = 157) were presented. A parallel version for family ratings o f patient m em ory functioning (M AC-F; Feher, M ahurin, Inbody, Crook, & Pirozzolo, 1991) has also been used to exam ine differ­ ences between patient and family perceptions o f m em ory functioning in persons with A lzheim er’s disease. This m easure is quite new and so little inform ation is, as yet, available. T h ere is reason to suspect that dissociations between objective an d sub­ jective m easures of m em ory functioning may be o f use in a diagnostic context. For exam ple, persons suffering from depression may com plain o f m em ory problem s but yet perform within the average range on objective measures. Conversely, som e persons with A lzheim er’s disease seem u n ­ aware o f their markedly poor perform ances on m easures o f m em ory func­ tioning. W hether o r n o t self-report measures o f m em ory functioning will yield enough m eaningful inform ation to w arrant their inclusion in clinical practice is yet to be determ ined.

CONCLUSIONS It is apparen t that the assessment o f m em ory has bro ad en ed substantially from the original form ulation o f the WMS. Many m easures, reflecting a variety o f theoretical perspectives, are available to the clinician. T h e issues o f age-related changes in m em ory perform ance and sensitivity to age-related m em ory disorders have been incorporated into the construction of m any o f these new er measures. Selecting the m ost effective an d efficient m easure (s) for use in a particular context rem ains the challenge for the clinician.

Chapter

Attention and Executive Functioning

Attention refers to the ability to focus on a simple task an d perform it w ithout losing track o f the task. If attention is im paired, the individual will also have difficulty perform ing m any o th e r tasks. A lthough the concept o f attention has been the focus o f m uch debate by cognitive psychologists (see P arasuram an & Davies, 1984, and H asher & Zacks, 1979, for reviews), norm ative data are available for relatively few com m only em ployed meas­ ures o f attention. It m ust be n oted that these m easures tend to assess sustained attention rath er than o th e r form s o f attention (e.g., selective or divided) and, hence, are not com prehensive measures o f attention or attentional capacity. T hese m easures have also been considered as m ethods for assessing working m em ory (Lezak, 1995) o r im m ediate m em ory (Tra­ han, G oethe, & Larrabee, 1989). Executive functions refer to those capacities that enable a person to engage in indep en d en t, purposive, an d self-serving behavior (Lezak, 1995). Execu­ tive disorders can affect cognitive functioning directly by com prom ising the developm ent o f strategies to approach, plan, and carry o u t activities, or thro ugh defective m onitoring o f the perform ance (Lczak, 1995). From a psychosocial viewpoint, persons with executive disorders may lack initiative o r be unable to plan o r carry o u t the activity sequences th at m ake up goal-directed behaviors. Persons with these disorders often perform well in structured assessment sessions because m ost m easures o f cognition only require responses to focused questions and the person is directed (or follows the “plan” provided) by the exam iner. Executive functions are often ascribed to the frontal lobes b u t arc also sensitive to dam age in o th e r areas of the brain, m ost notably subcortical 123

124

CHAPTER 7

areas (Lezak, 1995). T he tasks included here are similar in th at they require p lanning and abstraction o f underlying principles for successful com ple­ tion. However, not all persons with executive disorders will have difficulty perform ing these tasks. As n oted by Lezak (1995), executive disorders may­ be m anifest in m any ways. For exam ple, individuals may be able to perform tasks w hen asked but may be unable to perform these tasks unless asked (e.g., getting dressed). It is also im portant to keep in m ind th at p o o r perform ance on these tasks is n o t synonymous with frontal lobe dam age. Deficits in planning may arise for a n um ber o f reasons (e.g., o th e r areas o f im paired cognition). Persons with dense m em ory im pairm ent will have difficulty with com plex "planning” tasks like Wisconsin Card Sorting (dis­ cussed later) because they may n o t rem em ber the instructions to the task o r the category they were using to search by. Similarly, the person with p o o r m em ory may forget to cat an d thereby appear unable to perform a self-serving behavior. O bservation or report o f behavior is often the way in which deficits in executive function com e to light. M arked discrepancies between the ob­ served test perform ances and the rep o rt o f everyday functions may suggest underlying disorders o f executive functions.

MEASURES OF A TTENTION Reverse spelling is one exam ple o f the ability to reverse a sequential o rd er th at is difficult for persons with aphasia (language im pairm ent) o r o th er organic m ental disorders (Bender, 1975). Reverse spelling has been incor­ p o rated into som e screening m easures for cognitive im pairm ent (e.g., the Mini-Mental State Exam ination an d the Modified Mini-Mental State Ex­ am ination, see chap. 3) and exam ined, in its own right, in relation to age. T he prevalence an d degree o f disability in perform ing reverse spelling of 5-, 4-, 3-, and 2-letter words was exam ined in three groups o f persons over age 60 by B ender (1979). G roup 1 consisted o f 156 persons betw een age 60 an d 88. Thirty-two o f these persons were age 75 and older. G roup 2 consisted o f 100 persons between age 65 and 90. T he 5-letter reverse spelling was, by far, the m ost difficult o f the tasks and the incidence in errors increased with age. Similarly, Jenkyn ct al. (1985) reported that the incidence o f failure on reverse spelling o f WORLD increased from 6% at age 50 to 54 years to 21% over age 80 in a sample o f 2,029 persons age 50 to 93. These individuals were volunteers from a population o f cu rren t and retired employees o f F,. I. Du P ont de N em ours and Com pany, Inc., in W ilm ington, Delaware. N one o f these persons had a known history of any central nervous system disease, psychiatric disorder, o r medical illness with which neurologic a n d /o r psychiatric dysfunction is frequently associated.

ATTENTION AND EXECUTIVE FUNCTIONING

125

B ender (1979) noted th a t a deficit in reverse spelling may be d ue to lack of familiarity with language, defective education, and unrecognized congenital dyslexia; im pairm ent o f memory; acquired defect in language o r aphasia; disability in learning a new o rd e r o f reverse sequences; perseverative process; specific disability to reverse a series o f m eaningful o r n o n ­ m eaningful symbols; o r a locatable lesion for defect in reversibility. He suggested that the higher incidence o f reverse spelling errors for old er persons m ight relate to rigidity in perform ance o r m ild new learning defi­ cits seen as a function o f norm al aging. Serial sevens, a popular m easure o f the ability to atten d and concentrate on a task, requires the individual to co u n t backward from 100 by 7s (i.e., subtract 7 from 100 and continue subtracting 7 from the rem a in d er). Deficits have been observed in perform ance on this task by psychiatric an d n eu ro ­ logical patients (Hayman, 1942; Luria, 1966), although no com parisons with the incidence and types of errors seen in norm al adults were made. As an alternative for reverse spelling, serial seven perform ance is present o n the Mini-Mental State Exam ination (see chap. 3) b u t was rem oved from the Modified Mini-Mental State Exam ination. At least p art o f the reason for its removal relates to the lack o f com parability between serial sevens an d reverse spelling (Ganguli et al., 1990). Moreover, as age increases, perform ance decreases—as does its potential utility for differentiating be­ tween age-related decline an d pathological changes in functioning, such as dem entia. For exam ple, A. Smith (1967) exam ined serial seven per­ form ance in 132 adults, including 9 persons between age 56 and 63. O f the total sample, 42.5% (i.e., 56/132) perform ed the task w ithout error. O nly 1 o f 9 persons over age 55 perform ed w ithout error. A nother study exam ined the serial seven perform ance o f 506 users o f care services in the Capital Regional District (CRD) o f British Columbia. Persons were screened for cognitive im pairm ent using the A dult Lifestyles an d Function Interview (ALFI) version o f the MMSE (Roccaforte, Burke, Bayer, & Wengel, 1992) in which only the nonw ritten com ponents are adm inistered. Using data collected on separate samples o f clinically defined norm al and d em ented participants in the Canadian Study o f H ealth and Aging, a cutoff score that yielded sensitivity an d specificity o f .70 and .75, respectively, was determ ined for the ALFI items, w ithout including tile serial sevens, and applied to the CRD sample. T he distribution o f errors by age group is shown in T able 7.1 and the distributions o f erro rs for the norm al and "cognitively im paired” groups are shown in Table 7.2. T he use o f serial sevens as a screening instrum ent for d em entia should be ap p roached with caution as errors are n o t uncom m on in norm al elderly persons for this task. A task that reportedly shows little change with age is the repetition of digit series o f increasing length. T he Digit Span subtests o f the W echsler

126

CHAPTER 7 TABLE 7.1 F re q u e n c y o f E rro rs o n th e First Five R esponses o n S erial Sevens fo r Cognitively In ta c t Individuals (N = 425) by Age G ro u p Age

Number of Errors

65-69 (n = 28)

70-74 (n = 50)

75-79 (n = 81)

80-85 (n = 111)

85+ (n = 155)

0 1 2 3 4 5 Refused the task

57.1 10.7 0.0 3.5 7.3 0.0 10.7

64.0 18.0 6.0 2.0 2.0 0.0 8.0

45.7 25.9 9.9 2.5 7.5 0.0 7.5

44.1 19.8 8.2 5.4 10.8 0.0 11.7

42.6 14.8 15.5 7.7 12.9 0.7 5.8

Note. From T. G oranson, H. T uokko, L. R oscnblood, an d R. Frerichs (O ctober 1997). Srrial Sevens: Myths and Realities. Poster session presented at the m eeting o f the C anadian Association oil Gerontology, Calgary, Albcria, C anada. Used with perm ission.

A d ult Intelligence Scale Revised (WAIS-R, see chap. 4) an d th e W echsler M em ory Scale-Revised (WMS-R, see chap. 6) assess digits rep etitio n in forw ard an d reverse o rd er. G enerally, little effect o f age is seen on rep e­ tition o f digits (forw ard) (Craik, 1990; Jarvik, 1988). A lthough th e digits backw ard task is som ew hat m o re sensitive to th e effects o f aging, it is less sensitive th an o th e r m easures o f m em ory (G ranick Sc F ried m an , 1967). K lonoff a n d K ennedy (1966) re p o rte d th e p erfo rm an ce o f a g ro u p o f 115 com m unity dw elling C anadian m ale veterans random ly selected from a la rg e r g ro u p (n = 172) re p o rte d in K lonoff an d K ennedy (1965;see chap. 6) on the D igit Span subtest o f th e W echsler M em ory Scale. T his g ro u p ran g e d in age from 80 to 92 with a m ean o f 83.61 years (SD = 2.48). A m ean p erfo rm an ce o f 10.20 (SD = 1.97) was o b ta in ed on th e Digit Span subtest. TABLE 7.2 P e rc e n t o f N orm al ( N = 425) a n d Im p a ire d ( N = 81) Cases R efusing th e T ask o r M aking 0 T h ro u g h 5 E rro rs o n th e First Five R esponses to Serial Sevens Errors 0 2 3 4 5 Refused th e task

Normal

Impaired

47.1 18.4 11.1 5.2 9.6 0.2 8.4

16.0 12.3 8.6 6.3 17.3 0.0 39.5

ATTENTION AND EXECUTIVE FUNCTIONING

127

Mitrushina and Satz (1989) investigated the perform ance of 129 persons over age 65 on the num ber of digits repeated in forward order from the WAIS-R. Although there was no correlation between perform ance on this task and age, they reported the means and standard deviation by age group: age 66-70 (M = 7.9, SD= 2.7); age 71-75 (M = 7.6, SD = 2.9); and age 76-85 (Af = 6.9, SD = 3.9). The WMS-R provides norms separately for the forward and reverse sequences to age 74. Approximately 50% of their sample age 65 to 69 (n = 55) was able to repeat 7 digits forward (52%) and 6 digits backward (60%). Less than 5% repeated 4 digits forward and 2 digits backward. Approximately 50% of their sample age 70 to 74 (n = 50) was able to repeat 7 digits forward (46%) and 5 digits backward (48%). Less than 5% repeated 3 digits forward and 2 digits backward. Farmer etal. (1987) presented normative data for the Digit Forward and Digits Backward com ponents o f the WAIS collected as part of the Framing­ ham H eart Study (see chap. 2 for description of this sample). They included data (mean num ber of digits recalled) by age group (in 5-year intervals), education levels (8-11 years, high school, m ore than high school) and gender (see Table 7.3). Scores were very similar for males and females. These data reflect poorer perform ance than that reported by Mitrushina and Satz (1989) and that seen on the WMS-R subtests at the same ages. Given the lack of age effects in the traditional digit recall tasks already discussed and their insensitivity to various neurological disorders, Trahan et al. (1989) adm inistered a memory span measure containing m ore than 8 items (i.e., verbal supraspan) to 301 persons from age 18 to 91. All of these persons were screened through clinical interview and persons with neurologic or psychiatric disorders were excluded from the sample. Free recall of a list o f 12 unrelated words taken from the Verbal Selective Reminding Test (VSRT; Levin et al., 1982; see chap. 6) was identified as a measure of verbal supraspan. Significant age differences were found with older persons perform ing below the levels of younger persons. Table 7.4 shows the cumulative percentile ranks for persons over age 65. The Paced Auditory Serial Addition Task (PASAT) is considered a measure o f information-processing speed, efficiency, and concentration skills. Ex­ tended normative data have been collected on the Galveston version of this measure (Roman, Edwall, Buchanan, & Patton, 1991). In this version, the individual is requested to listen to a series of single-digit numbers, presented at a fixed speed, and add the last digit presented to the digit immediately preceding it. Four separate series o f num bers are presented, each separated by a brief pause, with each series presented at a faster pace. The 41 older adults (age 60-75) perform ed m ore poorly than the younger age groups (i.e., 18-27, 33-50). The study included the m ean num ber correct, percent correct for each trial, and the total for all trials.

128

CHAPTER 7 TABLE 7.3 M eans a n d S ta n d a rd D eviations by Age G ro u p an d E d u catio n Ixvel for W om en o n WAIS Digits F orw ard a n d B ackw ard (F a rm e r e t al., 1987) Digits f t ’Hoard

Age Group 65-69 Af SD n 70-74 M SD n 75-79 M SD n 80-84 M SD n 85-89 M SD n

Digits Backward

8-11 Years Education

12 Yean Erluaitian

13+ Years Education

8-11 Years Education

12 Yean Education

13+ Years Education

5.4 1.3 37

6.1 1.0 51

6.1 1.3 67

3.8 1.2 38

4.6 1.1 52

4.5 1.2 70

5.5 1.3 35

5.9 .9 36

6.4 1.2 50

3.9 1.1 36

4.1 1.0 36

4.8 1.2 50

5.4 1.0 28

6.1 1.4 36

6.1 1.1 20

4.0 1.2 29

4.4 1.3 36

4.4 1.4 22

5.5 1.3 16

6.1 1.2 15

5.7 1.1 19

3.8 1.2 16

4.6 1.1 15

4.4 1.0 19

5.1 .7 7

5.0 1.0 3

5.8 1.0 4

3.3 .8 7

4.3 1.5 3

3.8 .5 3

Note. R eproduced with perm ission o f authors an d publisher from Farm er, M. E., W hite, L. R., Kittner, S. J., Kaplan, E., Moes, F.., M cNam ara, P., Wolz, M. M., Wolf, P. A., & Feinleib, M. N europsychological test perform ance in Fram ingham : A descriptive study. Psychological Ref/orts, 60, 1987, pp. 1023-1040. Copyright © Psychological Reports, 1987.

B rittain, La M arche, R eeder, R oth, a n d Boll (1991) also develo p ed n o rm s fo r th e Galveston version o f th e PASAT a n d in c lu d ed a g ro u p o f p ersons (n = 122) over age 54. N orm ative data fo r th e PASAT by IQ (Shipley Institute o f Living Scale) are presen ted . D etailed in stru ctio n s for th e ad m inistration o f th e PASAT are in clu d ed as an ap p en d ix .

MEASURES OF EXECUTIVE FUNCTIONING T h e Stroop Test (Jensen & Rohw er, 1966; S troop, 1935) contrasts p erfo rm ­ ance o n a basic task, like co lo r nam ing, with p erfo rm an ce o f th e task in th e p resen ce o f conflicting o r in c o n g ru e n t stimuli. T h e S troop T est com es in d iffe ren t variations b u t m ost involve color w ords (e.g., red , blue, green)

129

ATTENTION AND EXECUTIVE FUNCTIONING TABLE 7.4 Cum ulative P ercentile Ranks for P ersons O ver Age 65 on Recall o f th e 12 U n related W ords From th e V erbal Selective R em in d in g T est Cumulative Percentile Ranks by Age Group

Seme

9 8 7 6 5 4 3 2 1 M SD

6 6 -7 7 years (n = 83)

100.0 97.6 92.8 88.0 60.2 39.8 18.1 7.2 2.4 4.94 1.67

78+ yean (n = 40)

100.0 92.5 77.5 52.5 30.0 5.0 —

4.43 1.36

Note. From "An Examination of Verba) Supraspan in Normal Adults and Patients With Head Traum a or Unilateral Cerebrovascular Accident," by D. E. T rahan, K. F.. Goethe, and G. J. Larrabee, 1989, Neuropsychology, 3, p. 85. Copyright © 1989 has been transferred to the American Psychological Association. Adapted with permission.

p rinted in congruent or in congruent colors (see Lezak, 1995, for a full discussion o f Stroop C olor-W ord form ats). T he task is eith er to read the words or nam e th e colors. Poor perform ance on this task has been in ter­ preted as reflecting im paired selective attention, a failure o f response in­ hibition, o r concentration effectiveness (Lezak, 1995). Moreover, it is often viewed as a general m easure of cognitive flexibility (see Spreen & Strauss, 1991). Graf, Uttl, an d Tuokko (1995) presented a picture-w ord version o f the S troop task containing congruent and incon g ru en t picture-w ord pairs to determ ine w hether o r n o t two versions of the Stroop task (C olorW ord an d Picturc-W ord) in fact reflect the same general, as opposed to test-specific, index o f cognitive flexibility in a group o f old er adults. In addition, they developed normative tables for their measures. Regardless o f the variant o f the Stroop Test u n d er investigation, slowing with advancing age has been dem onstrated. Sprcen an d Strauss (1991) rep o rted that slowing o f color nam ing and an increase in the Stroop interference effect (difference between congruent and in co n g ru en t con­ ditions) appear as the m ajor changes with increasing age. T hey rep o rted the m ean an d standard deviation for 67 persons over age 60 using the Modified Stroop Test (Spreen & Strauss, 1991), which uses a 24-item /card form at with a prelim inary colored d o t nam ing trial. G raf et al. (1995) presented norm s (see Tables 7.5 an d 7.6) for their color-w ord and picture-w ord versions of the Stroop task arran g ed by mid-

TABLE 7.5 M eans a n d S ta n d a rd D eviations (in p a re n th e se s) fo r th e C o lo r-W o rd S tro o p T est P erfo rm an ce (in m sec p e r item ) o f V arious M id p o in t O v erlap p in g A ge G roups Age. Group Midpoint and Range

Score. W ord C ongruent Color Incongruent Color-W ord IncongruentColor C olor/W ord Incongrucnt/ Color

68 65-71 (n = 50)

71 68-74 (n = 53)

74 71-77 (n = 58)

77 74-80 (n = 44)

80 77-83 (n = 32)

83 80-86 (n = 19)

86 83-95 (n - 11)

421 (81) 416 (73) 592 (109) 1087 (238) 170 (96) 495 (230) 1.42 (.24) 1.85 (.40)

430 (76) 433 (71) 633 (129) 1157 (304) 202 (120) 526 (262) 1.49 (.30) 1.84 (.40)

433 (68) 437 (70) 659 (132) 1211 (322) 224 (119) 554 (256) 1.53 (.30) 1.85 (.37)

439 (68) 436 (63) 655 (127) 1266 (365) 215 (108) 610 (292) 1.50 (.26) 1.95 (.41)

466 (70) 463 (63) 730 (156) 1379 (332) 254 (145) 646 (267) 1.58 (.33) 1.92 (.40)

488 (76) 468 (69) 803 (186) 1502 (387) 315 (153) 699 (305) 1.65 (.32) 1.89 (.39)

490 (88) 468 (77) 854 (250) 1564 (547) 364 (204) 710 (366) 1.74 (.36) 1.83 (.37)

Note. From “Color- and Picture-word Stroop Tests: Perform ance Changes in O ld Age,” by P. Graf, B. U ttl, and H. Tuokko, 1995, Journal of Clinical and Experimental Neuropsychology, 17, p. 399. C opyright © 1995 by Swets 8c Zeitlinger Publishers. Used with permission.

TABLE 7.6 M eans a n d S ta n d a rd D eviations (in p a re n th e se s) fo r th e P ic tu re -W o rd S tro o p T est P erfo rm an ce (in m sces p e r item ) fo r V arious M id p o in t O v erlap p in g Age G roups Age C/roup Midpoint and Range

Score Picture C ongruent Incongruent PictureC ongruent IncongruentPicture P ictu re/ C ongruent In c o n g ru e n t/ Picture

68 65-71 (n = 50)

71 68-74 (n = 53)

74 71-77 (n = 58)

77 74-80 (n = 44)

80 77-83 (n = 32)

83 80-86 (n = 19)

86 83-95 (n = 11)

784 (209) 484 (111) 1008 (197) 302 (189) 232 (256) 1.66 (-43) 1.34 (.30)

812 (213) 484 (99) 1052 (208) 330 (201) 240 (219) 1.72 (.46) 1.34 (.26)

881 (236) 491 (101) 1135 (241) 391 (217) 259 (281) 1.82 (.47) 1.34 (.30)

906 (248) 504 (107) 1146 (268) 395 (201) 246 (312) 1.79 (.41) 1.32 (.33)

993 (273) 526 (114) 1224 (279) 457 (225) 240 (318) 1.87 (.45) 1.29 (.32)

982 (247) 563 (121) 1282 (290) 418 (190) 300 (263) 1.76 (.32) 1.34 (.31)

1000 (270) 591 (132) 1280 (334) 409 (211) 280 (225) 1.70 (.34) 1.30 (.36)

Note. From ‘‘Color- and Picture-word Stroop Tests: Perform ance Changes in O ld Age,” by P. Graf', B. Uttl, and H. Tuokko, 1995, Journal of Clinical and Experimental Neuropsychology, 17, p. 400. C opyright © 1995 by Swets & Zeidinger Publishers. Used with permission.

132

CHAPTER 7

p o in t overlapping age groups (see chap. 2 for a description of this m e th o d ). T hese tasks were adm inistered to 129 com m unity dwelling adults between age 65 and 95 who participated in the V ancouver C enter co m p o n en t o f the C anadian Study o f H ealth and Aging (CSHA) an d were screened as cognitively intact (see chap. 2 for inform ation about the overall sam ple). In the color-w ord task, four cards were used, each containing 27 items th a t were arranged in 3 colum ns and 9 rows. T he first card (W ord) con­ tained color words (e.g., red, blue) printed in black ink. T he second card (Color) presented 27 colored (i.e., red, blue o r green) series o f Xs in the length o f the word corresponding to the color o f the ink (e.g., XXXX p rin ted in blue, X X Xprinted in red). T he third card (C ongruent) showed the color words printed in their corresponding colors an d the fo u rth card (incongrucnt) showed the color words (i.e., red, green, blue) p rin ted in in co n g ru en t ink colors (e.g., the word “red" printed in green ink). For the picture-w ord task, three cards containing four colum ns and five rows were developed. In each cell was a line drawing o f a com m on object. T he first card (Picture) contained only pictures. T h e second card (C ongruent) contained pictures with the nam e corresponding to the object p rin ted over it. T he third card (incongruent) contained the same pictures as the Picture and C ongruent cards and nam es as on the C ongruent card b u t the objects and nam es were deliberately m ism atched (e.g., the word BALLOON p rin ted over a picture o f an um brella). As noted by G raf e t al. (1995), the C olor-W ord and P icture-W ord tasks ap p eared to m easure different constructs in th eir elderly sam ple. T he C olor-W ord interference score, b u t not the Picture-W ord scores, ap p eared related to o th e r presum ed m easures o f cognitive flexibility (e.g., Digit Symbol and Similarities subtests o f the WAIS-R). They also n o ted that analyses o f the raw scores and derived scores yielded different age effects. T hey cautioned against the use of a com m only used index o f Stroop Test perform ance (MacLeod, 1991), the interference difference score (i.e., Incongruent-C olor) that was insensitive to age-related slowing. They favored the use of a ratio index o f interference (i.e., In co n g ru en t/C o lo r) to capture age-related changes. Lezak (1995) noted that the Stroop task is unpleasant for participants and recom m ended a m axim um time o f 5 m inutes for the interference trial. O f particular im portance to perform ing this task is visual com petence. Because color-blindness is age related in m en and visual difficulties in general increase with age, it is imperative that these factors be evaluated (e.g., as in the Spreen & Strauss, 1991, version) w hen d eterm in in g the appropriateness o f this m easure for a specific elderly individual. T he Trail Making Test. (TMT) contains two sections: Trails A an d Trails B. Trails A is a sim ple join-the-dots task where the participant m ust draw a line connecting circles containing num bers, placed ab o u t a page, in

133

ATTENTION AND EXECUTIVE FUNCTIONING

sequence. Trails B requires the individual to draw a line connecting circles containing both num bers and letters by alternating between the two se­ quences (i.e., 1, A, 2, B, and so forth). Visual scanning an d tracking play im p o rtan t roles in perform ing these tasks. Perform ance on this task may relate to the person’s ability to follow a sequence mentally, deal with m ore than one stimulus at a time (Eson, Yen, & Bourke, 1978), o r to be flexible in shifting the course o f an ongoing activity (Pontius & Yudowitz, 1980). S preen and Strauss (1991) provided detailed adm inistration procedures. Perform ance times increase significantly with age and are affected by edu­ cation in younger samples. Seconds to com pletion and num ber o f errors have been exam ined for a variety of samples o f older adults (see Table 7.7). Stanton, Jenkins, Savageau, Zyzanski, and Aucoin (1984) adm inistered the TMT to 102 p er­ sons age 60 to 69 in a prospective study o f recovery after elective coronary bypass-graft o r valve-replacem ent procedures. R. R. Bornstein (1985) adTABLE 7.7 M eans a n d S tan d ard Deviations (in p arentheses) for the T rail M aking T est D rsc T ip lw n

Trails A

Trails 11

Time, in .1ectnuls for 60- to 69-year-old medical surgical patients ( n = 102) Tima in seconds for 60- to 69-year-old

49.3 sydiobgy, 4, p. 32. Copyright © 1988 by l.awrence Erlbaum Associates, Inc. Adapted with permission.

th eir base sam ple and 107 persons (m ean age = 42.5, SD = 17.1) from th eir validation sample. T he total nu m b er o f seconds req u ired to place 25 pegs was the score used in their norm ative system (see chap. 2 for a description o f their approach). Age appears to exert som e effect, though n o t dram atic, in term s o f T-score values (e.g., a raw score o f 65 is equivalent to a scaled score o f 10; for m en with a grade 12 education, T-scores o f 62, 67, an d 70 are obtained by persons aged 60, 70, and 80, respectively). A nother com m on m easure o f m anual dexterity is the Finger Oscillation, or Finger Tapping Test (FTT). For this task, the resp o n d en t is asked to tap a key attached to a device for recording the n u m b er of taps as quickly as possible over a 10-second interval. Typically, five trials are perform ed for each h an d an d the score is the average o f the five trials. Some variations in adm inistra­ tion and scoring have been noted (see Lezak, 1995, for a discussion of these). As with the PP, an age-related decline would be expected for the FTT. R. R. Bornstein (1985) adm inistered the FTT to 57 persons age 60 and older. Scores were based on the average o f 5 trials within 5 points of each o th e r to a m axim um o f 10 trials (see Table 8.7). O f note, the m en perform b etter on this m easure of m anual dexterity than the women. CONCLUSIONS T he two groups o f tasks included in this chapter, those assessing lan­ guage/com m unication and those assessing v isuopercep tu al/m o to r func­ tions, each show changes with increasing age. T he role o f m em ory has

160

CHAPTER 8 TABLE 8.7

Means and Standard Deviations on the Finger Tapping Test for Persons Between Age 60 and 69 Mals,

Hand Preferred Hand Nonpreferred Hand

Female

Less Than High School Education (n = 16)

High School Education or Better (n = 23)

Less Than High School Education (n = 22)

High School Education or Better (n = 34)

39.1 (5.7) 35.2 (5.2)

43.0 (4.7) 39.3 (6.2)

29.7 (6.2) 29.8 (5.6)

32.2 (6.0) 32.0 (4.9)

Note. From “Normative Data on Selected Neuropsychological Measures From a Nonclinical Sample," by R. A. Bornstein, 1985, Journal o f Clinical Psychology, 41, p. 656. Copyright © 1985 by John Wiley & Sons, Inc. Adapted by permission o f Jo h n Wiley & Sons, Inc.

b e e n n o te d w ith r e sp ect to th e la n g u a g e m easu res a n d th e role o f m o to r s p e e d has b e e n n o te d as a p o ssib le c o n fo u n d in th e p erfo r m a n c e o f so m e v isu o p erccp tu a l tasks. G e n d e r an d e d u c a tio n effects are also o b serv ed for so m e o f th e se m easu res. F or th ese m easu res as for th o se d iscu ssed in th e o th e r ch a p ters, th e use o f age-ap p rop riate n o rm s is im perative, particularly after a g e 75.

Chapter

Assessing Psychopathology

The geriatric neuropsychologist often wishes to assess for the presence of psychopathology, because such conditions can often mimic or exacerbate symptoms of dem entia (Lezak, 1980). it is estimated that a m inim um of 10% of people over age 65 have emotional and cognitive problem s of clinical severity (Birren & Sloane, 1977). Although there is litde research to support the assumption that mental disorders manifest themselves in similar ways in younger and older adults (La Rue, Dessonville, & Jarvik, 1985), the DSM-IV (American Psychiatric Association, 1994) criteria generally do not differentiate between these two groups. In addition, risk factors for psychiatric conditions may not be the same for young and old persons, as the latter may be m ore likely to suffer from sensory deficits, physical illness, bereavement, and social isolation ( l a Rue et al., 1985). Traditionally, research work with elderly psychiatric patients has not been very popular, although in recent years there has been a tendency for such work to em erge (Gatz, 1995; I .a Rue et al., 1985). The relative lack of inform ation in this area may complicate the assessment of psycho­ pathology in elderly persons. For instance, somatic complaints are generally more com m on in later life but may also be evident in depression (Davison & Neale, 1994). Differentiating somatic symptoms of depression from symp­ toms due to physical decline may be difficult. Despite em erging interest in this field (e.g., Birren et al., 1992), attem pts to develop psychometric instrum ents and tailor existing ones for use with elderly persons are rela­ tively rare. Moreover, available inform ation lacks detail and is relatively primitive in relation to work pursued with younger persons. Rather than including an extended discussion of psychopathology in elderly persons, a brief overview of major psychopathology syndromes is 161

162

CHAPTER 9

p resented and their assessment with reference to available psychometrically developed instrum ents is discussed. Also included is a selective review and norm s pertaining to specific instrum ents that have b een used with elders. Generally, it is recom m ended that the geriatric neuropsychologist refer to the D S M -IV in attem pting to diagnose m ental disorders. A lthough the system is not specifically geared toward the elderly patient, it provides explicit definitions an d behavioral criteria and is probably the diagnostic system that m eets with the highest consensus in N orth America. T he D SM IV section m ost relevant to the geriatric neuropsychologist is “Delirium, D em entia and O ther Cognitive D isorders.” Psychological disorders often considered in neuropsychological assessment include depression, delu­ sional (paranoid) disorders, schizophrenia and psychotic disorders, sleep disorders, hypochondriasis, and substance related disorders. A brief over­ view o f these categories is provided. A variety o f instrum ents designed to assess adult psychopathology are available to the clinician. G reat caution should be exercised when using such instruments, however, because most lack representative norms for elderly persons and this may limit their validity. Even where norm s for elders exist, the best way to in te rp re t findings is n o t always clear. Specifically, w hether o r n o t separate norm s for elders m ust be used depends b o d i o n the m agnitude and the m eaning o f age differences (Butcher et al., 1991). In some instances, it may be that scores obtained by elderly persons may n o t be statistically different from scores obtained by younger age groups. Such findings would argue against the n eed for separate norm s for the elderly. U nfortunately, some o f the studies reporting norm s did n o t rep o rt statis­ tical com parisons am ong different age groups. It is, therefore, difficult to evaluate the utility o f the repo rted norm s. In addition, as norm ative studies on psychopathology scales are typically cross sectional, age effects can n o t be disentangled from cohort effects, and many o f the differences may be due to historical effects an d differences in physical health rath e r than personality and psychopathology p e r se (B utcher et al., 1991). Finally, norm s derived from cognitively intact groups o f subjects may be o f lim ited usefulness in the assessment o f people with cognitive im pairm ents. This chapter reviews instrum ents th a t have been developed for adm ini­ stration to the individual. First it discusses som e m ultidim ensional instru­ m ents (i.e., MMPI-2, Brief Symptom Inventory and projective tests). Such instrum ents may be useful for the assessment o f a variety o f conditions. Following the discussion o f the m ultidim ensional instrum ents, an exam ina­ tion o f specific disorders and psychometric instrum ents th at can be used to assess these is included. (C hapter 11 focuses on the role o f collaborative inform ants in the assessment process and discusses proxy-adm inistered m easures that may assist in the assessment o f psychopathology in elderly persons).

ASSESSING PSYCHOPATHOLOGY

163

MULTIDIMENSIONAL PSYCHOPATHOLOGY INSTRUMENTS The Minnesota Multiphasic Personality Inventory-2 (Butcher, Dahlstrom, Gra­ ham, Tellegen, 8c Kaemer, 1989) is a veiy widely used, empirically developed measure of psychopathology. Although it consists of 10 clinical scales, a configural interpretation of the instrum ent is recom m ended (Graham, 1993). Reliability varies from scale to scale but test-retest reliability for scales ranges from .58 to .92 (Butcher ct al., 1989). The data also support the validity of the instrum ent (Graham, 1993). In a review of the literature, G raham (1993) concluded that elderly persons obtain somewhat higher MMPI scores on Scales 1 (hypochondriasis), 2 (depression), 3 (hysteria), and 0 (social introversion) and lower scores on Scales 4 (psychopathic deviate) and 9 (mania). Graham suggested that such elevations in the elderly should not be taken to reflect increased psychopathology but concern about health and decreases in activity levels. Graham also suggested that the decision as to whether age-specific norms should be used on the MMPI or the MMPI-2 depends on the context of the assessment. Does the assessor wish to determ ine w hether the elderly person has somatic concerns relative to the population at large or relative to other persons the same age? Although the normative sample of the MMPI-2 included older adults (Butcher et al., 1991), they were underrepresented. The com mittee did not include separate norm s for elderly persons in the manual, but conducted appropriate analyses and concluded that age-specific norms are n o t needed (at least for older m en). Specifically, Butcher et al. (1991) concluded that of the 567 MMPI-2 items, only 14 differed by m ore than 20% in endorsem ent and such differences probably reflected cohort effects. Based on current knowledge, Graham (1993) and Butcher et al. (1991) recom m ended that the use of age-specific norms for older adults does not seem appropriate at this time. Table 9.1 presents some norm s for older m en based on Butcher et al. (1991). Readers can refer to these for comparison purposes but it may be prem ature to use these norm s at the present time as they are not the basis for the MMPI validation research. Butcher etal. (1991) presented additional statistical inform ation pertinent to these norms. Priest and M eunier (1993) studied a relatively small sample (valid N= 26) o f well-educated elderly women and presented MMPI-2 norms based on their findings. Although some differences in the perform ance of this sample and younger adults were found, these were not large. Quite rightfully, they cautioned that their results are not large enough to warrant, in and of themselves, the disregarding of traditional normative inform ation. Clearly, the question as to w hether separate MMPI-2 norms are needed for elderly women, can be addressed using a larger and m ore diverse sample of adults. A lthough cautions for the use o f the previous version o f the MMPI (L. W. Smith, Patterson, Grant, & Clopton, 1989) for the elderly have

164

CHAPTER 9 TABLE 9 .1 Age D ifferences in M M PI-2 Clinical Scales for the B oston N orm ative Aging Sam ple

Scale. Lie (L) Fake Bad (F) Subtle Defensiveness (K) Hypochondriasis (Hs) Depression (D) Hysteria (Ily) Psychopathic Deviate (Pd) Masculinity-Fcmininity (Mf) Paranoia (Pa) Psychasthcnia (Pt) Schizophrenia (Sc) Hypomania (Ma) Social Introversion (Si)

60-69 Years of Age (n = 591)

70-91 Years of Age (n = 241)

M

SD

M

.SD

4.94 3.67 17.21 5.24 19.08 21.50 14.78 22.69 8.82 8.38 7.95 14.52 25.61

2.36 2.62 4.65 4.10 5.07 4.93 4.19 4.25 2.87 6.61 6.04 3.71 8.61

5.10 3.69 16.74 5.83 20.03 21.20 13.77 23.35 8.46 8.91 8.26 14.17 27.32

2.58 2.69 4.50 4.06 5.24 4.99 3.75 4.15 2.66 6.43 5.88 3.88 8.94

Note. MMPI-2 = revised Minnesota Multiphasic Personality Inventory. From “Personality an d Aging: A Study of the MMPI-2 Among O lder M en,” b y j. N. Butcher, C. M. Aldwin, M. R. Levenson, Y. S. Ben-Porath, A. Spiro III, and R. Bosse, 1991, Psychology and Aging, 6, p. 366. Copyright© 1991 by the American Psychological Association. A dapted with permission.

been raised in the literature, relevant norm s on a sh o rt version o f the original MMPI were presented by L. W. Sm ith et al. (1989). T h e clinician may wish to plot the MMPI profile based on norm s for elderly persons (see, e.g., W. G. Dahlstrom , Welsh, & L. E. Dahlstrom , 1972, pp. 384—385) in addition to the profile based on the norm ative sample. However, R. L. G reen (1980), in his interpretative guide for the MMPI-2, reco m m en d ed th at the actual interpretation should be based on the general norm ative sample profile as the in terpretation and validation inform ation is based on this sample. Thus, it is n o t clear what would be the validity o f in ter­ p retations based on norm s for the aged. Overall, m ore research is need ed to d eterm ine the utility of the MMPI-2 with the elderly. However, the scale has clearly proven useful in the assessment o f younger adults. T he Brief Symptom Inventory (BSI) is an abbreviated form o f the Symptom C hecklist-90 (SC I-90; Derogatis, 1977). It consists o f 53 items and m eas­ ures psychopathology on the dim ensions listed in T able 9.2. T hese d im en­ sions correspond and are highly correlated to the dim ensions o f the SCL90 (Derogatis, 1977). H ale, C ochran, and H edgepeth (1984) p roduced norm s for persons over age 59 (see Table 9.2). They poin ted o u t th at the SCL-90 includes somatic items but these do not confo u n d the depression scale as they appear in a separate som atization scale. Because the scores

165

ASSESSING PSYCHOPATHOLOGY TABLE 9.2 N orm ative Info rm atio n on th e B rief Symptom Inventory Female (n = 364)

Age Somatization Obsessive-Compulsive Interpersonal Sensitivity Depression Anxiety Hostility Phobic Anxiety Paranoid Ideation Psychoticism

Mat* (n = 201)

M

SD

M

SD

73.54 .50 .83 .40 .53 .48 .29 .25 .37 .26

7.04 .53 .67 .56 .59 .54 .41 .42 .46 .37

73.92 .45 .73 .32 .43 .30 .34 .17 .44 .25

7.06 .48 .59 .47 .50 .39 .42 .36 .51 .36

Note. Ninety-five percent of the subjects were residing independently in the community. T he rem aining were living in nursing homes. From "Norms for the Elderly on the Brief Symptom Inventory,” by W. D. Hale, C. D. Cochran, and B. E. H edgepeth, 1984. Journal of Consulting /mil Clinical Psychology, 52, p. 321. Copyright © 1984 by the American Psychological Association. Adapted with permission.

for males and females differ along som e scales, H ale et al. p ro d u ced sepa­ rate norm s for each gender. In addition, their overall conclusion was th at the use o f special norm s for the elderly would be appropriate as higher scores are obtained for older adults than young adults on many scales. In term s o f projective approaches, it is probably fair to say th at they are n o t used com m only in geriatric neuropsychological assessments. They may prove useful, however, in som e cases (e.g., in situations w here an elderly p erson lacks testwise skills and is uncom fortable in answering direct ques­ tions). A lthough som e norm ative work in the use o f projective tests with the elderly has been done, the samples are generally small an d validity evidence still rem ains to be accum ulated (Hayslip & Lowman, 1986). Fur­ th erm ore, the ability o f such tests to assist in the diagnosis o f psychopa­ thology in the aged rem ains to be d eterm ined (Hayslip & Lowman, 1986). UNIDIMENSIONAL PSYCHOPATHOLOGY INSTRUMENTS Anxiety Scales Estimates o f prevalence for anxiety disorders in the elderly range from 3.5% to 10.2% (Flint, 1994) an d tranquillizer m edications are used by the elderly in disproportionately higher rates than by younger individuals (H ersen &

166

CHAPTER 9

Van Hasselt, 1992). Nonetheless, Flint (1994) concluded that most studies show that anxiety disorders are somewhat less com mon in the elderly than they are in young adults. H e also concluded that generalized anxiety and phobias account for most late-life anxiety. Agoraphobia (and possibly ob­ sessive compulsive disorder for women) may occur for the first time late in life. Panic disorder, obsessive-compulsive disorder in males, and phobias can persist from younger years, or they may be attributable to another medical or psychiatric disorder. Finally, Flint (1994) concluded that co­ morbidity of anxiety with medical illness and alcoholism is lower in late life than it is in younger people. Professionals must differentiate anxiety from physical disorders and con­ ditions associated with anxietylike symptoms (e.g., hyperthyroidism, small stroke or ischemic attack, use o f substances such as caffeine and alcohol). Several psychometric instruments arc available that may be used to measure anxiety and fear in older adults within the context o f the clinical assessment. These include the BSI, the MMPI-2 (discussed earlier), the State-Trait Anxiety Inventory (Spielberger, Gorsuch, & Lushcne, 1970; Spielberger, Gorsuch, Lushene, Vagg, & Jacobs, 1983), the Stale Trait Anxiety Inventory for Children (Spielberger, Gorsuch, Lushene, M ontouri, & PlaLsek, 1973), and the Self-Rating Anxiety Scale (W. K. Zung, 1971). However, most developed assessment techniques assume that the elderly patient docs n o t present with cognitive im pairm ent and the norms that exist have been based primarily on community dwelling elders. Thus, they may n o t be appropriate for institutionalized persons. The progression of dem entia can limit the opportunity for the direct assessment of the elderly person and m uch of the assessment inform ation can be better derived from interviews with a proxy or by observation. Given the limited norms that are applicable to elders, objective behavioral assessment procedures such as behavioral avoidance tests, functional analysis, self-monitoring, and other related pro­ cedures (see Bellack & Hersen, 1988) may be particularly useful for this group. Finally, as depression is often com orbid with anxiety, the clinician may also wish to use a depression scale that has been norm ed for elders. The State-Trait Anxiety Inventory (STAI; Spielberger et al., 1970) consists of two main scales: a State and a T rait scale. When filling out the State scale, individuals are asked to respond to the items in reference to how they feel at the m om ent the questionnaire is being filled out. In the Trait part of the instruments, the person is asked to indicate how they generally feel. T he items in the two scales are not identical. Responses are provided along 4-point scales. The m anual of this instrum ent reports data supporting its reliability and validity (Spiclberger et al., 1970). Limited norms for the STAI (Spielberger et al., 1970) are available and are listed in Table 9.3 (Knight, Waal-Manning, & Spears, 1983). However, the normative samples arc rather small and the STAI New Zealand norms

167

ASSESSING PSYCHOPATHOLOGY TABLE 9.3 N orm ative Inform ation for O ld er A dults on the S tate T rait Anxiety Inventory Measure Age Croup

Sl'A I-State

ST A I-T rait

Male

60-69 (n = 59) 70-79 (n = 29)

30.39 (7.18) 32.44 (8.49)

32.94 (7.26) 32.15 (7.49)

Female

60-69 (n = 80) 70-79 (n = 37)

32.59 (9.49) 33.49 (9.05)b

35.15 (8.40)a 34.47 (8.45)

Note. STAI-State = State Trait Anxiety Inventory-State; STAI-Trait = State Trait Anxiety Inventory-Trait. The data are based on a sample recruited from a small New Zealand community. From "Some Norms and Reliability Data for the Statc-Trait Anxiety Inventory an d the Zung Self-Rating Depression Scale,” by R. G. Knight, H. J. W aal-Manning, and G. F. Spears, British Journal o f Clinical Psychology, 22, p. 247. Copyright © 1983 by the British Psychological Society. Adapted with permission. an = 78. hn = 36.

o f Knight (1984) may be o f lim ited generalizability to N orth A m erican samples. In addition, representative norm s for elderly persons are need ed using the m ore rec en t version o f the STAI (Spielberger et al., 1983). T he K night (1984) normative inform ation suggests, however, th a t (at least in th eir sam ple) there were no substantial differences in scores over the life span. A lthough research has supported the factorial validity o f the STAI in a psychogeriatric sample (Nesselroade, Mitteness, 8c T hom pson, 1984), it has been argued that the form may be too com plicated for some elderly persons to com plete (Patterson, O ’Sullivan, 8c Spielberger, 1980; Rankin, Gfeller, 8c Gilner, 1993). T h e sim plified three-choice form at for the chil­ d re n ’s version of the STAI (STAIC; Spielberger et al., 1973) may be m ore appropriate for use with some elders. Patterson e t al. (1980) presented evidence in support o f the convergent an d discrim inant validity o f the STAIC when used with elderly patients. Normative data (Rankin c t al., 1993) for elders are presented in Table 9.4. T he norm s p resen ted by Rankin et al. (1993) on the STAIC are based on a larger sample than K night’s (1984) STAI norm s, but include only people recruited in seniors’ centers. Rankin et al. concluded in their study o f six age cohorts th at anxiety scores on the STAIC do not increase with age. T he Self-Rating Anxiety Scale (W. K. Zung, 1971) has also been n o rm cd for elderly persons. It contains 20 items and was designed to assess anxiety symptoms. Respondents arc asked to indicate how m uch each item applies

168

CHAPTER 9 TABLE 9.4 N orm ative Info rm atio n for O ld er A dults on the State T rait Anxiety Inventory fo r C h ild ren -S tate

Age Group 60-69 70-74 75-79 80-84

years years years years

(n = 21) ( n = 32) ( n = 31) (n =27)

M

SD

28.2 30.2 29.7 29.4

4.3 3.8 4.5 5.1

Note. The data are based on a sample of 38 male and 89 female subjects recruited from senior’s centers and similar organizations. The subjects were not receiving active treatm ent for anxiety or depression nor had they any significant clinical illnesses. Reprinted from the Jo u rn al of Psychiatric Research, 27, by E. J. Rankin, J. D. Gfcllcr, and F. H. Gilncr, Measuring anxiety states in the elderly using the Statc-Trait Anxiety Inventory for Children, p. 113, 1993, with kind permission from Elsevier Science Ltd, The Boulevard, Langford Lane, Kidlington 0X5 IGB, UK.

to them (none or little o f the time, some o f the time, good p art o f the time, m ost o r all o f the tim e). Zung presented validity evidence for the scale. W. K. Zung and R. L. G reen (1973) reported that persons age 65 an d o ld er (n = 47), ju d g e d to be norm al on the basis o f their ability to carry o u t age-appropriate activities, obtained a m ean score o f 40 (SD = 7.0). M ultidim ensional psychopathology instrum ents such as the BSI have norm ative inform ation based on larger an d m ore representative samples th an the unidim ensional inventories reviewed. Overall, the lack of adequate norm ative inform ation for the m easurem ent of anxiety symptomatology in the elderly is striking. It appears that the scales reviewed h ere can be used with the elderly b u t clinicians should exercise great caution b o th in in terp retin g individual scores relative to the norm ative data p resen ted h ere and in draw ing diagnostic and prognostic conclusions from the instru­ m ents. Given the lack o f adequate psychometric inform ation, objective behavioral inform ation such as behavioral avoidance tests, functional analy­ sis, self-m onitoring, and o th e r related procedures (see Bellack & H ersen, 1988) may often prove to be m ore useful indicators o f anxiety than psy­ chom etric scales. F uture research exam ining the psychom etric charac­ teristics o f such scales in elders may improve th eir clinical utility. Depression Scales A ccording to the Epidem iologic C atchm cnt Area Study, the prevalence o f clinically diagnosablc m ood disorders in persons over age 65 ranges be­ tween .5% and 2.2% for m en and 3.1% and 5% for w om en (Myers et al., 1984). W ith bipolar conditions being extrem ely rare in com m unity dwelling elderly persons and believed to alm ost always have an onset before age 65

ASSESSING PSYCHOPATHOLOGY

169

(e.g., Dawson & Neale, 1994;Jam ison, 1979), these estimates reflect largely u n ipolar conditions. If bipolar-like symptoms appear in late life, the pos­ sibility of organic im pairm ent should be assessed. N onetheless, prevalence figures for m ood disorders vary from study to study with some investigators rep orting higher rates than others (Blazer, H ughes, & George, 1987). A lthough prevalence estimates for affective disorders in elders are som e­ w hat lower than those repo rted for younger people, it has been suggested th at elderly persons may experience symptoms o f depression (as opposed to clinical depression) m ore frequently (Norris, G allagher, A. Wilson, & W inograd, 1987). A lthough depression accounts for many acute psychiatric adm issions o f elders (G urland & Cross, 1982), it seems to be less com m on in old er than it is in younger adults (Regier et al., 1988). F urtherm ore, there may be differences in the way symptoms o f depression m anifest themselves in the young an d the old (Davison & Neale, 1994). Blazer (1982) tabulated a com parison o f symptoms based on age. T he inform ation was derived from the preexisting literature an d suggested that elderly persons seem less likely to re p o rt suicidal ideation an d guilt feelings and m ore likely to rep o rt somatic symptoms. Blazer (1982) based his conclusions on only 19 patients over age 60. N onetheless, Musetti et al. (1989) studied a large sam ple and rep o rted findings indicating less suicidal ideation, greater m o to r retard a­ tion, m ore weight loss, less hostility, and m ore general decline in elderly patients with depression. Despite such reports o f reduced suicidal ideation in elders, the num bers o f actual attem pts and com pleted suicides are high in this population (Davison & Neale, 1994). This should be taken into account when assessing suicide risk. O btaining inform ation from caregivers can be helpful in the assessment o f p atien t symptomatology. T eri an d W agner (1992) concluded, however, th a t caregiver and p atient reports o f symptoms o f depression do n o t always correspond and generally patients with A lzheim er’s disease ten d to rep o rt fewer symptoms than their caregivers. Thus, when depression in such pa­ tients is assessed, it is im portant to consider both p atien t an d caregiver reports. In addition to obtaining inform ation about depressive sym ptom a­ tology, eliciting m aterial o n cu rren t an d previous health status as well as m edication use could assist in the interpretation o f the interview responses (G allagher, 1986). It is im portant to note some problem s associated with studies o f depres­ sion in o ld e r adults. Specifically, T hom pson, Ileller, an d Rody (1994) concluded that studies o f depression in older adults often exclude 25% to 35% o f potential participants. They suggested that depressed o ld er adults may be especially likely to refuse participation. This could lim it the generalizability o f the conclusions drawn by many researchers. It also implies th at norm ative studies o f depression in older adults would also be affected and

170

CHAPTER 9

clinicians may wish to keep that in m ind w hen com paring a p atien t’s scores with the norm ative group. Regardless o f this lim itation, T eri an d W agner (1992) argued that alm ost no norm ative inform ation on established depres­ sion m easures is available for people with dem entia and th at difficulties with language abilities may interfere with the validity o f these instrum ents. In assessing for depression, the clinician may supplem ent interview in­ form ation with a variety o f psychometric instrum ents. As suicide rates am ong the elderly are quite high, th e clinician m ust assess carefully for the presence o f suicide risk. This can be done both thro u g h careful in ter­ viewing o f the person and caregiver but may also be evaluated through the exam ination o f critical item responses in psychomctric instrum ents such as the Beck D epression Inventory (Beck, Ward, M endclson, Mock, & Erbaugh, 1961) and the M innesota M ultiphasic Personality Inventory-2 (B utchcr et al., 1989). W hen using any psychometric instrum ent, on e m ust consider the type o f population the patient represents. As an exam ple, H arper, Kotik-Harper, and Kirby (1990) concluded th at psychomctric in ­ strum ents may underestim ate depression in specific subgroups o f elders. Specifically, they found that psychometric instrum ents ten d ed to und eres­ tim ate depression in elders who were thought by physicians an d relatives to show significant and unexplained deterioration in their functioning. Thus, special care should be taken to reduce false negative rates when working with such patients. T he Schedule for Affective Disorders and Schizophrenia (SADS; Spitzer & Endicott, 1978) is an instrum ent that has good reliability an d validity. G allagher (1986) m ade extensive use o f the scale with reliable an d valid findings in older adults. This schedule explores aspects o f affective distress b u t is tim e consum ing an d requires considerable training o n the p art o f the exam iner. G allagher (1986) suggested that clinicians w orking in set­ tings in which tim e and financial considerations may preclude the use o f the SADS, could constnict a com prehensive b u t sh o rter interview based o n SADS questions. SADS questions th a t would n o t be applicable to indi­ vidual clients could be om itted. W ith respect to self-report instrum ents, it is generally reco m m en d ed to avoid those with high somatic co n ten t because this may be co nfounded with the increasing num ber o f physical problem s th at often accom pany old age (e.g., Yesavage, 1986). T he Geriatric Depression Scale (GDS) is a brief self-report depression scalc that has been specifically designed to assess depression in older people (Brink et al., 1982; Yesavage et al., 1983). It consists of 30 items, has a yes-no form at, and assesses affective an d behav­ ioral symptoms of depression. Vegetative symptoms are n o t assessed. O ne o f its advantages over o th e r screening instrum ents is th at it focuses on psychological aspects o f depression by n o t em phasizing somatic items. T he scale can also be read to the subject as its yes-no form at makes it am enable

ASSESSING PSYCHOPATHOLOGY

171

to such adm inistration when the patient has problems that could interfere with reading. The measure has established reliability and validity am ong ind ependent community residents, those receiving medical or psychiatric treatm ents in outpatient and inpatient settings, as well as institutionalized elders (Koenig, Meador, Cohen, & Blazer, 1988; Norris et al., 1987; Parmallee, Lawton, & Katz, 1989; Rapp, Parisi, Walsh, & Wallace, 1988; Yesav­ age et al., 1983). In fact, Parmalee et al. (1989) concluded that there were no differences in the reliability and validity of the instrum ent for cognitively im paired and intact patient groups and the correlation o f the CDS with the Blessed Dementia Rating Scalc (Blessed ct al., 1968; sec chap. 3) was o f negligible magnitude. Overall, the CDS appeared to be quite robust when used with mildly to moderately cognitively impaired as well as cog­ nitively intact aged (Parmallce ct al., 1989). Yesavage (1983) reported sensitivity and specificity inform ation corre­ sponding to different cutoffs. The criterion for depression was based on Spitzer, Endicott, and Robins’ (1978) Research Diagnostic Criteria. W hen a cutoff of 11 is used, the sensitivity and specificity values are 84% and 95%, respectively. A cutoff score of 13 results in 80% sensitivity and 100% specificity. Normative data for groups of elders appear on Table 9.5. Tabic 9.6 contains original and m ore extensive GDS normative data for persons with and without dem entia who were referred to a dem entia assessment clinic. The Beck Depression Inventory (BDI; A. T. Beck et al., 1961) has also been used as a screening instrum ent for the identification of depression in the elderly (Gallagher, Breckenridge, Steinmetz, & Thom pson, 1983). The instrum ent has satisfactory reliability and validity (see Spreen & Strauss, 1991) and it is brief and easy to administer. Allen-Burge, Storandt, Kinscherf, and Rubin (1994) studied a sample of 191 geriatric inpatients and concluded that both the GDS and the BDI were less effective in identifying depression in male than female geriatric patients. The BDI has been criti­ cized because its somatic content may be inappropriate for older adults (Hyer & Blount, 1984; Rapp et al., 1988), but the effects of these items in the num ber of false positives is not entirely clear (e.g., Olin, Schneider, Eaton, Zemansky, & Pollock, 1992). Nonetheless, there is some evidence th at the GDS is easier for older adults to com plete (Olin et al., 1992) and for that reason may be preferred. A. T. Beck and R. W. Beck (1972) introduced a short, 13-item version o f the Beck Depression Inventory for use in screening medical patients for depressive features. The short form correlates with the full version at .96 (A. T. Beck 8c R. W. Bcck, 1972) and has satisfactory reliability and validity (Reynolds Sc Gould, 1981). Knight (1984) presented norm s col­ lected in a New Zealand community on the Short Form of the Beck De­ pression Inventory arguing that the A. T. Beck and R. W. Bcck (1972)

TABLE 9.5 N orm ative In fo rm a tio n on D epression Scales fo r E lderly Subjects Depression Scale

Source.

Geriatric Depression Scale

Yesavage et al. (1983) Hadjistavropoulos et al. (1994)

Hamilton Rating Scale for Depression C entre for Epidem iological Studies Depression Scale

Yesavage et al. (1983) Data derived from proxies associated with case controls in the C anadian Study on H ealth and Aging

Zung Self-Rating Scale for Depression

Yesavage et al. (1983) W. K. Zung & R. L. Green (1973) Knight et al. (1983)

Short Form o f Beck Depression Inventory

Knight (1984)

Beck Depression Inventory Memorial University of N ewfoundland Scale of H appiness

G allagher et al. (1983)

M

(SD)

( /i= 40)

5.75 8.01

(4.34) (5.61)

( » = 40)

5.43 5.95 6.27 5.85 4.86 6.71 34.31 48 33.46 37.18 33.86 34.83 2.29 2.89 3.00 2.83 5.54 16.62

(4.98) (8.37) (7.16) (6.22) (5.87) (6.80) (6.66) (10) (6.49) (7.51) (6.93) (7.43) (2.22) (2.90) (1.75) (2.20) (4.67) (6.59)

Sample Elderly persons recruited from seniors’ centers Elderly persons referred for assessment to an outpatient dem entia clinic ( n = 136) Elderly persons recruited from seniors’ centers 65-69 years (n = 42) 70-74 years (n = 63) 75-79 years (n = 58) 80-84 years (n = 42) 85+ years (n = 14) Elderly persons recruited from seniors’ centers Persons over age 65 (n = 169) Males 60-69 (n = 58) Females 60-69 (n = 78) Males 70-79 (n = 28) Females 70-79 ( n = 34) Males 60-69 (n = 52) Females 60-69 (n = 64) Males 70-79 (» = 27) Females 70-79 (n = 47) Elderly persons recruited from seniors’ centers Elderly persons over age 66 (w = 100)

(« = 40)

( n = 82)

Nate. T he d ata associated with the citation K night et al. (1983) are adapted from Table 2 o f “Some Norms and Reliability Data for the Statc-Trait Anxiety Inventory and th e Z ung Self-Rating Depression Scale," by R. G. Knight, H. J. W aal-Manning, and G. Spears, 1983, British Journal of Clinical Psychology, 22, p. 248. Copyright © 1983 by the British Psychological Society. A dapted with permission. T he data associated with the citation K night (1984) are based on an adaptation of a portion of Table 1 from “Some G eneral Population Norms for the Short Form o f the Beck Depression Inventory,” by R. G. Knight, 1984, Journal of Clinical Psychology, 40, p. 752. Copyright © 1984 by Jo h n Wiley & Sons, Inc. A dapted by perm ission o f Jo h n Wiley & Sons, Inc.

173

ASSESSING PSYCHOPATHOLOGY

TABLE 9.6 Original Geriatric Depression Scale Norms Based on Outpatients Referred to a Diagnostic Dementia Clinic Age Group

M

SD

Individuals Diagnosed With Dementia

65-69 (n 70-74 (n 75-79 (n 80-84 (n 85+ (n =

= 75) = 130) = 108) = 55) 12)

65-69 70-74 75-79 80-84

= = = =

8.50 8.10 6.30 7.60 6.00

5.76 5.61 4.28 5.07 4.00

Individuals Judged as "Not H aving Dementia"

(n (n (n (n

37) 35) 30) 13)

8.00 9.10 8.30 8.00

6.58 6.13 5.86 4.76

Note. The GDS was adm inistered as part of the assessment at the Clinic for Alzheimer Disease and Related Disorders, Vancouver Hospital and Health Sciences Centre, Vancouver, British Columbia, Canada. Diagnostic conclusions (e.g., dem ented vs. not dem ented) were drawn after the adm inistration of the GDS.

sam ple was smaller an d atypical in that it was collected at a hospital setting. Table 9.5 presents data for persons overage 60 derived from K night (1984). K night’s data suggest th a t average scores o f norm al persons do n o t change substantially across the life span but inform ation from clinical samples varying in age would be useful in determ ining w hether a depression by age interaction plays a role in th e scores. T h e generalizability o f these New Z ealand norm s to N orth American samples may also be limited. Norris e t al. (1987) com pared the sensitivity and specificity o f the BDI an d the GDS in a sample of elderly outpatients with medical problem s. T hese estimates are different from sensitivity and specificity estimates re­ p o rted using o th e r samples (e.g., Brink et al., 1982). W ith a cutoff score o f 10 for eith er scale and D SM -III diagnosis o f depression as criterion, sensitivity and specificity were 89% and 82%, respectively, for the BDI and 89% and 73%, respectively, for the GDS. W hen using a cutoff o f 17 for the BDI, sensitivity and specificity were 50% and 92%, respectively. A cutoff o f 14 on the GDS leads to 78% sensitivity and 86% specificity. Thus, lower cutoff scores yield higher sensitivity and lower specificity estimates than high cutoff scores. Clinicians often prefer to achieve high sensitivity because o f the potential costs o f false positive misclassification from a screening in strum ent are potentially lower than that o f false negative misclassification. Researchers, however, may often desire higher specificity an d may select h ig h er cutoff scores (Norris et al., 1987).

174

CHAPTER 9

The Hamilton Depression Rating Scale (Hamilton, 1960) has been used in many studies involving elders, but some reservations about its use have been expressed (Gallagher, 1986) on the basis that validity inform ation for samples of elders is lacking. The scale is used within the context o f an interview with the patient. Based on the interview, the clinician makes ratings on 23 dimensions, such as depressed m ood (ranging from 0 = absent to 4 = patient reports virtually only these feeling states in his spontaneous verbal and nonverbal com munication) and work and activities (ranging from 0 = no difficulty to 4 = stopped working bccausc of present illness). Yesavagc et al. (1983) reported that with a cutoff score of 11, sensitivity and specificity on this instrum ent is 86% and 80%, respectively, when using the Research Diagnostic Criteria for depression (Spitzer & Endicott, 1978). Normative inform ation is presented in Table 9.5. A related scalc that has been evaluated with a sample of cognitively intact elderly medical patients (Rapp, Smith, & Britt, 1990) is the Extracted Hamilton Depression Rating Scale (XHDRS; Endi­ cott, Cohen, Nee, Flciss, & Sarantakos, 1981). This 17-item instrum ent matches items from the SADS interview to the original Hamilton Depression Scale by content. T hat is, the 7-point rating scale of the SADS was converted to the original 4-point rating scale of the Hamilton. Rapp et al. (1990) presented evidence of good reliability and validity. O ne hundred and nine geriatric/surgery patients who did not m eet the Research Diagnostic Criteria (Spitzer et al., 1978) for a psychiatric diagnosis obtained a mean total score o f 4.94 (pooled SD= 6.01). These norms, however, may n o t be applicable to cognitively impaired and community dwelling elderly persons. Rapp et al. (1990) concluded that the XHDRS had improved overall perform ance com pared to the BDI and GDS for their sample of medical patients largely because of its better specificity. For example, when sensitivity equals 83%, the Beck and the GDS both had specificity of 65%, whereas the XHDRS had specificity of 92%. O ther screening instrum ents include the Carroll Rating Scale for Depression (Carroll, Feinberg, Smouse, Rawson, & G reden, 1981), which is a transformation of the Hamilton Rating Scalc for Depression (Hamilton, 1960) into a self-rating format. Like the Hamilton, the Carroll contains somatic items. The Center for Epidemiologic Studies-Depression Scale (CES-D; Raddloff, 1977) requires estimation of symptom frequency that may make it cum­ bersome for some elders to complete. Nonetheless, respondents are given a card with four possible alternatives ranging from “Rarely o r none o f the time (less than once a day)" to “Most or all of the time (5-7 times per day).” The four response alternatives are applied to 20 statements (e.g., “During the past week I did not feel like eating; my appetite was p o o r”). The psychomctric properties of this measure have not been investigated adequately with elderly persons.

ASSESSING PSYCHOPATHOLOGY

175

T he CES-D was adm inistered to elderly proxies associated with the case controls o f th e C anadian Study o f H ealth an d Aging (C anadian Study of H ealth and Aging W ork G roup, 1994). Cognitively norm al case controls were drawn at random from the CSHA com m unity sam ple who screened 78 or above on the 3MS. Original norm s on the proxies associated with these individuals are presented on T able 9.5. Norm s also exist for the Zung Self-Rating Depression Scale (W. K. Zung 8c R. L. G reen, 1973). This 20-item scale has a 4-point form at and has been rep o rted to have satisfactory internal consistency and validity (Yesavage ct al., 1983) w hen used with older adults. T hese investigators rep o rted that using the Research Diagnostic Criteria (Spitzer 8c Endicott, 1978) for de­ pression, a score o f 46 achieves 80% sensitivity an d 85% specificity in elderly samples. Some relevant norm s on the Zung are p resen ted on Table 9.5. From this table, it is a p p a re n t th a t th e sam ple o f W. K. Z ung and R. L. G reen (1973) obtained higher scores than the o th e r samples reported. U nfortunately, W. K. Zung an d R. L. G reen ’s (1973) sample was n o t de­ scribed explicitly. H ence, it is difficult to in terpret the discrepancy in the scores between the samples. T he Memorial University of Netvfoundland Scale of Happiness (MUNSH; Kozma 8c Stones, 1980) is a m easure that can be used to assess level o f well-being in elders. T he m easure has 24 items and is in ten d ed to assess affect w ithin the range from happiness to subclinical depression. Responses are obtained using a y es-no-I d o n ’t know form at. Several studies have shown the internal consistency o f this m easure to be high (Kozma, Stones, & McNeil, 1991). In addition, the scale correlates highly with psychopa­ thology scales and particularly depression (Kozma, Stones & Kazarian, 1985) as well as with o th e r life satisfaction an d m orale scales, an d avowed happiness (Kozma 8c Stones, 1983, 1987, 1988). T he effects o f social de­ sirability and acquiescence biases are m inim al (Kozma 8c Stones, 1987, 1988; Kozma et al., 1991). Norm s o n this instrum ent are p resented in T able 9.5. T he norm s are based on a volunteer sample o f com m unity residents o f an eastern C anadian urban center (Kozma, di Fazio, Stones, 8c H annah, 1992). Generally, scores on this instrum ent ap p ear to stay quite stable across the life span. T he G eriatric D epression Scale is recom m ended for the psychometric assessment o f depression in the elderly not only becausc o f the availability o f norm ative inform ation but also because m ore extensive research has b een conducted on this scale than any o th e r for samples o f elders. This scale could supplem ent interview questions directed toward both the pa­ tient and a proxy. W hen tim e allows, a m easure o f well-being, such as the MUNSH, may be used to supplem ent the inform ation obtained from uni­ dim ensional depression scales.

176

CHAPTER 9

Substance Use Measures Illicit drug use am ong the elderly is very rare. The m ajor substance-related disorders in the elderly typically involve the use of alcohol and prescription tranquillizers. As noted in chapter 11, various patterns of cognitive func­ tioning may be associated with alcohol use. Recognition o f alcohol use disorders in elderly persons is im portant both from a differential diagnostic viewpoint and for determ ining the validity of the assessment results (e.g., if the person is intoxicated during the assessment). To assess for the pres­ ence o f alcoholism, Zimberg (1987) recom m ended asking specifically w hether use of alcohol in the elderly has been associated with marked changes in behavior/personality, memory loss and confusion, social isola­ tion, argumentativeness, neglect of hygiene and self-care, missing medical appointm ents, neglect of medical regimen, interference with income m an­ agem ent, problem s with the law, and problems with neighbors. Problems with drug overdose involving prescription medications may be seen in elders who are estimated to comprise 2% to 6% of overdose problems seen in acute care medical settings (Heller 8c Wynne, 1975; La Rue et al., 1985; D. M. Petersen 8c Thomas, 1975). As cognitive im pairm ent and other factors (e.g., sensory deficits) may interfere with com pliance to the medical prescription, it is necessary for the geriatric ncuropsychologist to obtain detailed inform ation about any and all drugs that a patient may be using. Such drugs may have effects on cognitive function. In addition, misuse could potentially be prevented if it is discovered early that the patient is not following the physician’s recommendations. In addition to interview inform ation that can be obtained by the clini­ cian, a variety of assessment procedures for substance abuse are available. Sobell, Toneatto, and Sobell (1994) described structured interviews and self-report methods that have been used with success in adults. Nonetheless, the utility of such m ethods with samples of elders has not been sufficiendy addressed (C. J. King, Van-Hasselt, Segal, & Hersen, 1994). In addition, organic im pairm ent can interfere with the patient’s ability to provide ac­ curate self-report through interviews and questionnaires. The utility of a small num ber of screening tools for alcoholism has been investigated with samples o f elderly persons. The CAGE (Kwing, 1984) is a brief alcohol screening measure that consists of four questions (e.g., “Have you ever had a drink first thing in the m orning to steady your nerves o r get rid of a hangover?”) and has been shown to have excellent sensitivity and specificity for young persons (M. King, 1986). Buchsbaum, Buchanan, Welsh, Centor, and Schnoll (1992) studied 323 general medical patients age 60 or older, 33% o f whom were found to have problems with alcohol use as assessed using DSM-III criteria. With a cutoff score of 1, the CAGK had sensitivity and specificity of 86% and 78%. With a cutoff score of 2, specificity improved to 90% but sensitivity dropped to 70%. Thus, it appears

ASSESSING PSYCHOPATHOLOGY

177

that a cutoff score o f 1 is appropriate when screening samples of elderly medical patients with a high incidence of alcohol problems, whereas a cutoff score of 2 may be preferable when screening community dwelling elders (King et al., 1994). Data are also available on the Michigan Alcoholism Screening Test (MAST; Selzer, 1971). The 25 items of this true-false inventory can be differentially weighted to yield a highest possible score of 50. Ten- (B-MAST) and 13-item (S-MAST) short versions o f the instrum ent have also been de­ veloped (B. J. Zung, 1979). The use of these scales in elders was studied by W illenbring, Christensen, Spring, and Rasmussen (1987). These inves­ tigators adm inistered the long and short versions of the MAST to 52 elders (m ean age 64 + / - 2.8 years) hospitalized for substance abuse treatm ent and a similarly aged control group o f 33 elders who were hospitalized for problem s unrelated to alcohol. Using a cutoff score of 6 or greater, the MAST had a sensitivity of 100% and a specificity of 90%. W hen weighted scoring was employed, the sensitivity and specificity were 96% and 86% when a cutoff scorc of 3 was used. The briefer B-MAST yielded sensitivity and specificity o f 91% and 83% when a cutoff score of 4 was employed. A cutoff of six positive responses resulted in sensitivity of 82% and specificity of 100%. A cutoff score o f 2 in the S-MAST resulted in sensitivity of 98% and specificity of 72%. Given the overall better sensitivity of the full version of the MAST, it is recom m ended th a tit be preferred over the briefer versions (C.J. K ingetaL , 1994). In summary, it appears that the initial data on the use o f screening tools for alcohol-related problems in samples of elders are very promising. Proxy report and biochemical procedures (breath alcohol tests, urine tests, saliva tests) can also be of value in detecting and m onitoring substance consum ption (Sobel et al., 1994), particularly in cases where there are doubts about the accuracy of self-reports.

ASSESSMENT FOR OTHER FORMS OF PSYCHOPATHOLOGY Personality Disorders Vaillant and Perry (1980) suggested that with aging, a true decrease in the incidence of personality disorders occurs. La Rue et al. (1985) suggested, however, that such estimates of age-related reduction in personality disor­ ders may be due to patterns of service utilization rather than prevalence changes per sc. Although early community surveys (reviewed by La Rue et al., 1985) estimate the prevalence of personality disorders in older adults to be 2.8% to 11% research using recent DSM criteria is needed.

178

CHAPTER 9

Christinson and Blazer (1988) discussed the evolution of personality disorders in the elderly. They pointed out that interpersonally focused coping mechanisms that a person uses may evolve into personality disorders u n d er the stress of physical o r psychiatric problems. Persons with paranoid features may experience an exacerbation of characteristics, such as suspi­ ciousness and hypersensitivity, because of problems developing with various sensory modalities. They also suggested that persons with histrionic features may find that the effectiveness of seductive and dramatic behaviors they previously relied on may be reduced. As the stresses of life may make it difficult for persons with narcissistic features to maintain their fragile self-esteem, such individuals may be expected to experience dysthymia. Individuals with obsessive-compulsive traits may become increasingly rigid and perfectionistic with difficulties associated in decision making. Finally, Christinson and Blazer concludcd that illnesses associated with aging may become legitimate means of gratifying the dependency needs of people with depended personalities. Such views, however, are based largely on clinical experience and intuition. Longitudinal empirical investigations could shed m ore light on the evolution of personality disorders across the life span. Nonetheless, in their review o f the literature, Christinson and Blazer also concluded that core personality traits show relative stability over a lifetime. There is research in support of this view (Costa & McCrae, 1980). Assessment for personality disorders is approached much the same way in older adults as for younger persons (e.g., MMPI-2). The added caution is, once again, that the presence of cognitive im pairm ent may limit the validity of some established assessment methods. The clinician must be cautious, however, as personality disorders may manifest themselves differ­ ently in the old than they do in the young. As research is lacking, it is difficult to determ ine what type of age differences in personality disorder manifestations exist. Hypochondriasis and Health Anxiety The frequency of hypochondriasis am ong the aged is sometimes considered high (e.g., Straker, 1982). Data suggest, however, that this condition is no m ore frequent am ong the aged than it is am ong the young (Siegler & Costa, 1985). Although actual health problems increase with age, there is no corresponding increase in health concern. Regier et al. (1988) found that only . 1% of persons over age 65 can be diagnosed with somatization disorder, a rate com parable to rates reported with younger age groups. Costa and McCrae (1985) concluded that age influences the num ber of health complaints only insofar as it increases actual disease. H ealth con­ cerns of the elderly are no m ore exaggerated than are health concerns of people in other age groups. Regardless of w hether hypochondriasis is

ASSESSING PSYCHOPATHOLOGY

179

suspected, reports o f physical symptoms and health-related concerns should be followed up with appropriate medical examinations. Only if a medical examination does not support the validity of the elderly person’s health concerns, a diagnosis of hypochondriasis may be made. Insomnia It is well established that the prevalence o f insomnia increases with age (Mellinger, Balter, & U hlenhuth, 1985), with 45% of persons between age 65 and 79 having some difficulty with sleeping. Furtherm ore, use of sleep­ ing medications am ong the elderly is com mon (Mellinger et al., 1985). Bootzin and Engle-Friedman (1987) concluded that frequent awakenings are very com mon am ong elders, as are other related problems. The medi­ cations that elderly persons take to com bat difficulties with sleep typically result in quick tolerance (Mellinger et al., 1985). In addition, some tran­ quillizers (such as the benzodiazepines) often result in impairments in learning new inform ation and thinking clearly the next day (Ghoneim & Mewaldt, 1990). It is im portant for the geriatric neuropsychologist to take note of any such medications elderly persons might be receiving and the effects these may have on cognitive function. Cognitive behavioral treat­ m ent programs have been shown to be effective alternatives in treating insom nia in older adults (Edinger, Hoelscher, Marsh, Ipper, 8c IonescuPioggia, 1992), but may be less effective in persons with cognitive impair­ ments. Behavioral assessment procedures for insomnia have been developed (e.g., sleep diaries, polysomnography, mechanical devices, and ancillary m easures). Refer to Lacks and Morin (1992) for a review. Such assessments may be tailored for the elderly. For the geriatric psychologist, it is im portant to rem em ber that insom nia can have a variety of causes, including stress and situational factors, psy­ chiatric disorders, substance withdrawal, sleep-induced respiratory impair­ m ent (i.e., apnea), myoclonus, and resdess leg syndrome (Spielman, 1986). It is im portant to assess for the presence of sleep apnea and neurological conditions that could interfere with sleep. In many instances, polysomnog­ raphy and a neurological referral will be indicated in order to com prehend the factors that cause and maintain the condition and to assist in a treat­ m ent regiment.

CONCLUSIONS Clinical research on the psychometric assessment of psychopathology in elders with cognitive impairments is very limited. Limited age-appropriate norms exist for some standardized instruments, but the clinical utility of

180

CHAPTER 9

such norm s has n o t been adequately established. Thus, clinicians should exercise great caution when they assess psychopathology in the elderly in general and when deciding on appropriate norm ative com parison groups in particular.

Chapter

Age-Associated Conditions Affecting Cognition

O ne of the primary purposes of conducting a neuropsychological evaluation in persons over age 65 is to assist in diagnostic decision making. In this context, it is necessary to determ ine if cognitive im pairm ent (i.e., changc in cognitive functioning greater than that expected as part of the norm al aging process) is present. W hen cognitive im pairm ent is identified, exam ination of the pattern of perform ance, within the context of historical inform ation concerning the em ergence of the difficulties and dem ographic charac­ teristics of the individual, may suggest or support a particular diagnosis. T he Diagnostic and Statistical Manual for Mental Disorders (DSM-IV, Ameri­ can Psychiatric Association, 1994) criteria for specific forms of cognitive disorders are probably the most commonly applied, but other sets of criteria exist for specific disorders and, where appropriate, are highlighted in this chapter. Dementia refers to an overall decline in mental capacity (one or m ore cognitive domains) that renders the individual unfit to m eet the diverse intellectual dem ands associated with the obligations of everyday life. Central to this diagnosis, as defined by the DSM-IV, is im pairm ent of memory. At least one other area of functioning—including disturbances in language (aphasia), m otor planning (apraxia), visual perception, and recognition (agnosia)—a n d /o r executive functions—must also be im­ paired. It becomes evident that there are instances where memory impair­ m ent does not appear central to a neurodegencrative condition and where a single primary area of cognitive im pairm ent may be apparent long before o th er areas of deficit emerge. The DSM-IV provides categories for such disorders (e.g., cognitive disorders not otherwise specified, amnestic dis­ orders) and this chapter refers to these as circumscribed (focal), or mild, cognitive/behavioral deficits. 181

182

CHAPTER 10

Disorders that em erge in late life are the focus o f this chapter. These have been organized as circumscribed (focal), or mild, cognitive/behavioral defi­ cits, and pervasive disturbances of cognition and behavior. These syndrom es o f cognitive and behavioral functioning may o r may n o t be linked to specific etiologies. W here possible, each syndrom e is described briefly within the framework o f elem ents o f the neuropsychological assessment identified in ch apter 1 (see Table 10.1). In this context, it is im po rtan t to note th at the cognitive and behavioral presentations may change o r evolve across the course o f the disorder. Moreover, w hat is known ab o u t the disorders af­ fecting cognition in later life has grown trem endously an d is likely to co n tinue to evolve. A lthough these disorders are presented in “pure form ” as discrete en­ tities, it m ust be noted that it is possible that o n e o r m ore conditions may be present. For exam ple, it has been noted that depression may accom pany various neurological conditions (e.g., poststroke, A lzheim er’s disease, Par­ kinson’s disease). T he identification o f coexisting (com orbid) conditions is o f im portance, because treatm ent for rem ediable disorders may improve the quality o f life o f the individuals a n d /o r those aroun d them . In addition, the factors described in chapter 1 (e.g., vision, hearing, m edications, iso­ lation) may have an effect on how a specific disorder is m anifested. TABLE 10.1 Key Features for Differential Diagnosis Testing Arousal level Attention Language

Motor Memory

Visuospatial reasoning Verbal reasoning Mood History

i) ii)

Com prehension Expression (naming, repetition, fluency, phrase length, response speed, word finding, paraphasic errors [literal: sounds like; semantic: similar m eaning]) i) Movements (facial, gait, tremor) i) New learning ii) Verbal, visual iii) Personal information iv) Remote memory or information learned in the past

i) ii) iii) iv) v) vi)

O nset Duration (bourse Demographic inform ation (education, occupation) l.iving situation, hobbies, alcohol consumption Medical condition (meds, surgeries, anesthetics, psychiatric/neurological status)

AGE-ASSOCIATED CONDITIONS AFFECTING COGNITION

183

MILD AND FOCAL COGNITIVE AND BEHAVIORAL DEFICITS Mild Cognitive Decline T he significance o f cognitive im pairm ent that does n o t m eet criteria for d em entia in elderly persons is, as yet, unknow n. It may be th at within this g roup som e persons are exhibiting the very early signs o f a degenerative process (e.g., Dcrcnzi, 1986; H eath, Kennedy, & Kapur, 1983), whereas others have stable o r reversible conditions. A lthough a n u m b er o f diag­ nostic term s incorporating the concept o f mild im pairm ent o f cognition have em erged in the literature (Dawe, Procter, & Philpot, 1992), th ere is, as yet, no one accepted criteria (Berg, 1990; Morris & Rubin, 1991). Little is known abo u t features that m ight differentiate am ong reversible conditions, those with a benign course, and those that may show progressive deterioration over time. M inor cognitive deficits may result from many factors, including fatigue, depression, visual, or auditory im pairm ent, or nonneurological physical disability— none o f which imply dem entia (Berg, 1990). Zaudig (1992) described m ethods for diagnosing m ild cognitive im pair­ m e n t based on the D SM -III-R and International Classification o f Diseases10th edition (ICD-10; W orld H ealth O rganization, 1992) criteria for de­ m entia. Memory im pairm ent is central to both the D SM -III-R an d ICD-10 criteria and, hence, is reflected w hen using these criteria for diagnosing Mild Cognitive Im pairm ent (MCI). T he proposed criteria for MCI identify subtypes with m em oiy im pairm ent only, m em ory an d intellectual im pair­ m ent, o r m em ory and intellectual im pairm ent coexisting with deterioration o f em otional control, social behavior, o r motivation. A lthough m easures o f m em ory functioning have proved m ost useful in differentiating between groups o f individuals who d eteriorate an d m eet criteria for dem entia over the study tim e and those who did n o t (e.g., Katzman et al., 1989; Linn et al., 1995; Rubin, Morris, G rant, & V endegna, 1989; Tuokko et al., 1991), m em ory im pairm ent is n o t always the primary feature in disorders affecting cognitive functioning. Thus, Z audig’s criteria may only capture this specific subgroup o f persons with mild cognitive decline. Moreover, with the introduction o f changes to the D SM -IV criteria for dem entias, the utility o f Zaudig’s criteria may be b ro u g h t into question. R efinem ent o f the criteria em ployed to diagnose and subclassify MCI is clearly im portant with respect to the early identification of degenerative processes. In addition, it has been observed that this diagnostic group is slightly m ore com m on than dem entia (C anadian Study o f H ealth and Aging Work G roup, 1994). O ne goal o f CSHA-2, the longitudinal follow-up

184

CHAPTER 10

study to CSHA-1, is to exam ine the utility of a variety of m ethods for exam ining the relations between MCI and early dem entia. Age-associated Memory Impairment Crook et al. (1986) proposed criteria to identify the loss of m em ory func­ tion that occurs as a consequence o f norm al aging. Age-Associated M emory Im p airm en t (AAMI) describes otherwise healthy persons over age 50 whose m em ory is p o o rer than that o f healthy young adults. It is expected that AAMI would affect m ost o f the population over age 50 years (McF.ntree & Crook, 1990; G. Smith ct al., 1991). Despite the age-consistent nature o f this “disorder,” in practice many clinicians inadvertently use the term A A M I to refer to circum scribcd m emory im pairm ent. T h at is, m em ory functions im paired in relation to o th e r persons o f the same age. AAMI was n o t inten d ed to identify that specific subset o f individuals. A case has been m ade for using this diagnostic category in the context o f the developm ent o f pharm aceutical agents as treatm ents for norm al aging-related m em ory decline that would “benefit millions o f elderly people who are n o t d em ented o r disabled, but feel at a disadvantage in the co n d u ct o f their lives” (M cEntrcc & Crook, 1990, p. 528). O f note, the AAMI criteria are n o t useful as an indicator for the developm ent o f d em entia over time. In an 8-year longitudinal study o f a random sam ple o f 146 persons over age 65, Snowdon and Lane (1994) found the mortality rate an d develop­ m e n t o f dem entia am ong those fulfilling criteria for AAMI to be similar to o th e r n o n d em en ted groups. Blackford and La Rue (1989) suggested revisions to the criteria for AAMI. In addition to restricting the age range and reco m m ending the use o f a standardized self-report m em ory questionnaire to d eterm in e the ex ten t o f subjective com plaints, they proposed guidelines for identifying m ore select subgroups with age-consistent m em ory im pairm ent (ACMI) an d latelife forgetfulness (LLF). W hether or not these distinctions will be o f benefit is yet to be determ ined. Amnestic Disorders Amnestic disorders, as defined by the DSM-IV, are conditions in which only m em ory functions arc im paired. T hese disorders may ap p ear in re­ lation to general medical condiuons, substance abuse, o r exposure, o r for unknow n reasons. T h ere is am ple evidence to suggest th at exposure to specific toxins (H artm an, 1995) may result in circum scribed m em ory p ro b ­ lems and, for this reason, it is o f the utm ost im portance to obtain accurate historical inform ation concerning exposures (e.g., carbon m onoxide poi­ soning, toluene exposure in house painters).

AGE-ASSOCIATED CONDITIONS AFFECTING COGNITION

185

An often underestim ated potential source o f cognitive deficits, an d spe­ cifically m em ory problem s, in older adults is alcohol abuse. It has been estim ated that alcohol abuse affects som e 10% o f old er m en and 2% of o ld er w omen (Larkin & Seltzer, 1994). However, it has also been noted th at there is insufficient inform ation for defining and delineating the preva­ lence an d patterns o f alcohol use disorders in geriatric populations (Rains & Ditzler, 1993). Elderly persons with alcohol use disorders may p resen t with falls, confusion, self-neglect, unusual behaviors, injuries, m alnutrition, incontinence, and o th e r m anifestations characteristic of a dem en tia (Rains 8c Ditzler, 1993). A lthough th ere is potential for significant recovery o f cognitive functioning with abstinence (G rant, 1987; Salm on, Butters, & H eindel, 1993; W illenbring, 1988), many older persons with alcohol use disorders also have m ultiple medical problem s, polypharm acy, an d o th e r confounding events (e.g., a history o f traum a and falls). Despite this, the role o f alcohol in the presentation o f cognitively im paired persons is an im p o rtan t diagnostic consideration that needs to be addressed even when regular consum ption is as low as 1 drink p er day (Rains & Ditzler, 1993). Wernicke-Korsakoff (WK) syndrome is perhaps the m ost striking disorder associated with chronic alcohol consum ption. T he initial stage o f the dis­ o rd er (W ernicke’s encephalopathy) is characterized by abnorm al eye move­ m ents, gait ataxia, an d a state o f global confusion. Typically, the person em erges from the acute effects o f the W 'emicke’s encephalopathy with K orsakoff s syndrom e. This is characterized by a severe and p erm a n en t am nestic condition with m arked apathy and passivity (Salm on et al., 1993; see Table 10.2). T he cardinal features o f the syndrom e are severe inabilities to learn new inform ation and to recall events occurring within a 20-year p eriod prio r to the onset of the disorder (retrograde am nesia). A tendency to confabulate is often p resent (though n o t necessary) d u rin g the acute phases o f the disorder. However, general intellectual abilities rem ain rela­ tively well preserved. T he p rofound am nesia seen in Korsakoff syndrom e has been attributed to deficits in episodic m em ory (m em ory for specific events an d episodes d e p e n d e n t on tem poral o r spatial cues for their retrieval). Sem antic m em ­ ory o r knowledge o f general principles, associations, and rules th at are in d e p e n d e n t o f the context in which they were learned (e.g., arithm etical procedures, vocabulary knowledge) typically rem ains intact. Persons with Korsakoff’s syndrom e may also have difficulty perform ing o th e r tasks d ep e n d en t on new learning o r requiring the types o f concep­ tualization and problem solving often associated with integrity o f the frontal lobes. For exam ple, im pairm ents have been dem onstrated o n visuospatial processing tasks such as digit symbol substitution (Glosser, Butters, 8c Kaplan, 1977; K apur & Butters, 1977) and on tasks requiring the learning and shifting o f problem-solving strategies (Salmon ct al., 1993).

186

CHAPTER 10

TABLE 10.2 Presentation o f W ernicke-Korsakoff E ncephalopathy and Chronic Alcoholism Domain A ttention an d Arousal L anguage M otor Memory

P roblem Solving

Everyday a n d Social E m otional, Personality, T h o u g h t C o n ten t

History Features O nset Course Medical condition

Wemickfi-Korsakoff Syndrom*

Chronic Alcoholism

Low m otivation, per­ severation

R ecent very im paired; rem ote m emory discontinuous; islands of preservation

Im paired if requires new learning or visuoperceptual processing Inability to plan, initiate, o r consistenUy apply an optim al problem-solving strategy Lack initiative Unaware o f or undisturbed by problem s, placid, cooperative, apathy, som etim es confabulation S udden o r gradual Stable N utritional deficit; chronic alcoholism

Relatively mild; may be lim ited to visual modality; little evidence of retrograde am nesia; not susceptible to proactive interference Im paired if requires new learning o r visuopcrccptual processing Inability to plan, initiate, or consistently apply an optim al problem solving strategy

Gradual ; 5

Chronic alcohol consumption th a t does n o t result in W ernicke-K orsakoff syndrom e may also m anifest with m em ory im p a irm en t (see T able 10.2). H ow ever, m em ory im p a irm en t seen with ch ro n ic alcohol co n su m p tio n is relatively m ild, m ay be lim ited to th e visual m odality, an d does n o t in clu d e re tro g ra d e am nesia. As with WK, these persons typically d o n o t m anifest sem antic m em ory deficits. T h e o th e r specific cognitive deficits (i.e., visuoperc ep tu a l a n d plan n in g ) n o te d fo r WK may also be seen with ch ro n ic alcoholism . T h e n o n m em o ry deficits seen in W'K a n d ch ro n ic alcoholics have b een ascribed to th e d irect toxic effects o f alcohol to th e association co rtex (Salm on e t al., 1993). T h e truly am nesic state seen in WK has b een attrib ­ u te d to th e develo p m en t o f small hem orrh ag ic lesions in th e m idline d ie n ce p h alo n region following thiam ine deprivation (Victor, A dam s, &

AGE-ASSOCIATED CONDITIONS AFFECTING COCNITION

187

Collins, 1989). T he specific structures most often affected are the dorsomedial nucleus of the thalamus and the mamilliary bodies of the hypo­ thalamus. Damage to the basal forebrain has also been implicated in the profound memory deficits seen in WK and alcoholic dem entia (e.g., Lishman, 1986, 1990). It has been suggested (Salmon et al., 1993) that persons particularly vulnerable to the toxic effects of alcohol may develop alcoholic dem entia rather than WK due to m ore widespread cortical damage and perhaps greater involvement of the basal forebrain structures. Disorders Primarily Affecting Language Since Mesulam (1982) first brought the syndrome of primary progressive aphasia to clinical attention, there have been a num ber of reports o f cases presenting with isolated progressive language dissolution (e.g., Chawluk et al., 1986; H eath etal., 1983; K irshner,Tanridag,T hurm an, &Whetsell, 1987; Mesulam, 1987; Sapin, Anderson, & Pulaski, 1989). The language disorder has been characterized in a variety of ways: as an anomic aphasia with reduced verbal output (Mesulam, 1982), as a Broca-type aphasia (Craenhals, Raison-Van Ruymbeke, Rectem, Seron, & Laterre, 1990), and as a hypoki­ netic dysarthria with anomia (K em pleret al., 1989). None of these fits neatly into the Benson-Geshwind classification scheme for aphasia (Mesulam, 1987). Thus, it appears that any specific disorder primarily affecting speech a n d /o r language is captured by this syndrome designation. The early isolated deficits seen with progressive aphasia may evolve into a m ore global dem entia over time (J. Green, Morris, Sandson, McKeel, & Miller, 1990; Kirshner, Webb, Kelly, & Wells, 1984; Poeck & Luzzatti, 1988; Wechsler, 1977). Kempler et al. (1990), in summarizing findings about the syndrome, note that a uniform symptom complex is not apparent, evolution to a full-blown dem entia complex is not necessarily seen, and the rate of progression varies gready from case to case. It has been suggested that “quantifiable neuropsychological assessment” may reveal subtle disturb­ ances in other areas of cognitive functioning and the lack of such evidence may be a source of discrepancy between studies (Foster & Chase, 1983). EEG and PET scan findings have tended to reflect focal abnormalities in the perisylvian region of the left hem isphere (Chawluk et al., 1986; Craenhals et al., 1990; Mesulam, 1982, 1987). Biopsy o r autopsy findings have confirm ed the presence of Pick’s disease (Cracnhals et al., 1990; A. F. W echsler, Verity, R osenchein, Fried, & Scheibel, 1982), spongiform changes and astrocytosis of cortical layer 2 (K irshner et al., 1987), Creutzfeldt-Jacob disease (Mandell, Alexander, & Carpenter, 1989), Alz­ heim er’s disease (Kempler et al., 1990) with disproportionate involvement o f the left inferior parietal cortex (J. Green et al., 1990), o r nonspecific changes with an absence o f specific histopathological markers (J. Green et al., 1990; Kirshner et al., 1987).

188

CHAPTER 10

Disorders Primarily Affecting Behavior A lthough prim ary disorders affecting frontotem poral fu n cd o n in g have been categorized as forms o f dem entia, the typical m em ory disturbances central to D SM -IV definitions o f dem entia are n o t present. Instead, frontolemporal dementia (FT D ), is ch a racterize d by progressive personality change, and breakdow n in social conduct. Investigators in M anchester, UK, and L und, Sweden, have produced a set o f criteria to aid in identifying this disorder (The L und and M anchester Groups, 1994). O ften the earliest and m ost striking features o f FTD (see T able 10.3) are m arked changes in social behavior, which may include the em ergence o f disinhibited behaviors (e.g., violent behavior, unrestrained sexuality), m isdem eanors (such as shoplifting), changes in oral behaviors (e.g., ex­ cessive eating o r drinking), o r lack o f initiative (e.g., withdrawal from social activities, deterioration in hygiene and groom ing). Behaviors may be stereo­ typed o r perseverative (e.g., wandering, ritualistic preoccupations such as hoarding, toileting, dressing, and unrestrained exploration o f objects in the environm ent). T he person lacks concern about actions and may ap p ear em otionally indifferent o r withdrawn. Speech ou tp u t is often reduced thro ugh lack o f spontaneity. Because apathy, inertia, excessive sentim entality, and suicidal o r fixed ideation may also be present, the role of depression may be questioned early in the course. Clients with FTD may be m isdiagnosed as eccentric o r as suffering from psychodc disorders. A pathetic and withdrawn individuals may be described as having “Diogenes syndrom e,” a condition characterized by extrem e self-neglect, social retreat, and hoarding o f worthless objects. D isinhibited consum ption o f alcohol may be m isidentified as alcoholism. TABLE 10.3 Features Associated With Frontotem poral Dementia (FTD) Domain Attention and Arousal Language Motor Memory Problem Solving Everyday and Social Emotional, Personality, T hought Content O nset Course History

FID Distractible Often first sign; anomia, circumlocution, paraphasias, repetitive use of stereotyped responses, ccholalia, com prehension Relatively well preserved until late in the course Visuospatial relatively well preserved until late in the course Everyday behavior affected; social behavior changed Marked personality change; blunting, apathy, irritability Cognitive deficits minimal (motivation to perform?) Insidious Progressive More comm on in females; onset age 40-60; prevalence after age 65 unknown

AGE-ASSOCIATED CONDITIONS AFFECTING COGNITION

189

Clients with FTD are an enigm a for care providers. A lthough they may be forgetful, they do not exhibit m arked m em ory im pairm ents o r disorienta­ tion. T he forgetfulness appears m ore a function o f p o o r abilities to plan and initiate behavior rath er than “tru e” im pairm ents o f new learning. Moreover, visuospatial functioning an d praxis are fairly well preserved, although thinking may be concrete. This, in fact, is the opposite picture to th at o f classical A lzheim er’s Disease (AD), which exhibits early m em ory im pairm ent and little o r no disturbance in social skills until late in the disease’s course. Care providers may feel that because the individual is capablc o f learning and rem em bering, requests for assistance and rem inders are reflections o f “bloody-mindcdncss" o r “m anipulative” behaviors. Similarly, the often m arked apathy may be viewed as an unwillingness to engage o r take p art in activities rather than an inability to self-initiate behaviors. T he lack o f insight into everyday problem s may result in uncooperative behavior, skep­ ticism, or withdrawal from the "interference” o f concerned parties. O nset before age 65, a positive family history of a sim ilar d isorder in a first-degree relative, an d evidence o f m otor neuron disease (e.g., bulbar palsy, m uscular weakness an d wasting, fasciculations) arc supportive fea­ tures for this diagnosis. EEC studies arc norm al, despite clinically evident behavior disturbance, an d brain im aging studies show pred o m in an t frontal a n d /o r an terio r tem ­ poral abnorm ality. According to the L und and M anchester G roups (1994), two types o f histological change underlie the frontotem poral cerebral at­ rophy seen in this disorder. T he first an d m ost com m on, designated as frontal lobe degeneration, is characterized by nerve cell loss an d spongi­ form changes (microvasculation) with m ild to m oderately severe astrocytic gliosis in the o u te r cortical layers. T he second is the typical Pick-like his­ tology characterized by intense astrocytic gliosis with in tran eu ro n al inclu­ sion bodies o r inflated neurons in all cortical layers. FTD results from bilateral pathology' in the frontotem poral regions. Asymmetrical deg en era­ tion may result in prim arily linguistic rath er than behavioral symptoms if the language-dom inant area is affected. H ere then is the link to the clinical entity o f primary progressive aphasia with both disorders form ing p a rt o f the clinical spectrum o f lobar atrophy an d sharing the same spectrum o f his­ tology (Neary, Snowden, & M ann, 1993). T he underlying pathology dis­ tinguishes the clinical syndrom e o f FTD from o th e r disorders th a t may also affect frontotem poral structures (e.g., AD, Creutzfeldt-Jacob disease, subcortical vascular disease). Disorders Primarily Affecting Movement M ovem ent disorder may be present in som eone suspected o f cognitive decline o r im pairm ent. W hen this is the case, it is im p o rtan t th at disorders affecting subcortical structures be considered. Bradykinesia (i.e., slowed

190

CHAPTER 10

m ovem ents), rigidity, p o o r coordination, a variety o f hyperkinesias (such as trem or, chorea, myoclonus, and tics), and dysarthria are characteristic m anifestations of the involvement o f subcortical brain structures (e.g., striatum , thalam us, substantia nigra, subthalam ic nuclei, and deep white m atter tracts). T he term subcortical dementia has been used to refer to the cognitive deficits seen in relation to these disorders and has attracted m uch discus­ sion (e.g., M. L. Albert, 1978; Cum m ings & Benson, 1984; W hitehouse, 1986). T he typical clinical features associated with the subcordcal d em en ­ tias (see Table 10.4) include im pairm ent of recall, slowness o f th o u g h t process (bradyphrenia), p oor abstraction and strategy form ation, apathy, and disturbances of m ood (e.g., depression). O f note, deficits in language, praxis, an d recognition (aphasia, apraxia, and agnosia) are rarely present (M. L. Albert, 1978). A variety o f specific disorders may affect subcortical regions o f th e brain including: (a) degenerative disorders such as Parkinson’s disease, H u n ­ tington's disease, idiopathic calcification of the basal ganglia, progressive su p ranuclear palsy; (b) vascular disorders; (c) m etabolic disorders such as W ilson’s disease o r hypoparathyroidism ; or, (d) dem yelinating disorders such as m ultiple sclerosis (see M. L. A lbert & Knoefel, 1994; Cum mings, 1990). Two disorders that carry with them particularly im p o rtan t specific treatm en t im plications are briefly described here: Leury Body disease, a rela­ tively new diagnostic entity th a t often presents with the extrapyram idal features o f P arkinson’s disease and, hydrocephaly. TABLE 10.4 Features Associated With Subcortical Dementias Domnin Attention and Arousal Language M otor

Memory Problem Solving Everyday and Social Emotional, Personality, T hou g h t C ontent History Features Onset Course

Subcfntic/il Dementia May be poor No aphasia but may be nonspecific word finding when condition is severe Slowness of movement, stooped posture, small-stepped gait, difficulty initiating movement (bradykinesia), trem or, tics, dystonias present, speech may be dysarthric Recall may be affected but recognition generally intact Calculation preserved until late; visuospatial im paired (with or without m otor com ponent) Apathy, depression Typically mild deficits Insidious Slow and progressive

Note. Includes inform ation from Cummings (1990).

AGE-ASSOCIATED CONDITIONS AFFECTING COGNITION

191

A variety o f term s (e.g., diffuse cortical Lewy Body disease, diffuse Lewy Body disease, senile dem entia o f the Lewy Body type) have been used to refer to this disorder, which is characterized by Lewy Body form ation in the brain stem and cerebral cortex. It has been estim ated th at dem entia associated with Lewy Bodies is the second m ost com m on cause o f cognitive im pairm ent in elderly persons (H ansen & Galsko, 1992) an d exceeds the prevalence o f vascular dem entia. Given the frequency with which Lewy Bodies are im plicated as contributing to cognitive im pairm ent, operational criteria for senile dem entia o f the Lewy Body type (SDLT) have been proposed by McKeith, Perry, Fairbairn, Jabeen, and Perry (1992; see Table 10.5). In com parison to persons with AD, those with confirm ed Lewy Bodies m ore often show fluctuating cognitive im pairm ent, psychotic features (in­ cluding visual and auditory hallucinations and paranoid delusions), de­ pressive symptoms, frequent falling, and unexplained losses of conscious­ ness. M oreover, persons with SDLT frequently show often irreversible adverse reactions indicative o f neuroleptic sensitivity syndrom e w hen ad­ m inistered neuroleptic m edications. T he survival time o f persons treated with neuroleptic m edication was reduced by 50%. Hydrocephaly also carries specific treatm ent implications. It has been estim ated that between 6% an d 12% o f all cases presenting with dem entia may have a hydrocephalic etiology (Strub 8c Black, 1988). H ydrocephaly TABLE 10.5 Proposed Operational Criteria for Senile Dementia of Lewy Body Type (SDLT) A. Fluctuating cognitive im pairm ent affecting both memory and higher cortical actions (such as language, visuospatial ability, praxis, or reasoning skills). The fluctuation is marked by the occurrence of both episodic confusion and lucid intervals, as in delirium , and is evident either on repeated tests of cognitive function or by variable perform ance in daily living skills. B. At least one o f the following: (1) visual a n d /o r auditory hallucinations, which are usually accom panied by secondary paranoid delusions; (2) mild spontaneous extrapyramidal features or neuroleptic sensitivity syndrome, i.e., exaggerated adverse responses to standard doses of neuroleptic medication; (3) repeated unexplained falls a n d /o r transient clouding or loss of consciousness. C. Despite the fluctuating pattern the clinical features persist over a long period o f time (weeks or months) unlike delirium, which rarely persists as long. D. Exclusion o f any underlying physical illness adequate to account for the fluctuating cognitive state, by appropriate exam ination and investigation. E. Exclusion of past history of confirmed stroke a n d /o r evidence of cerebral ischaemic dam age on structural brain imaging. Note. From "Operational Criteria for Senile Dementia of Lewy Body Type (SDLT)," by I. G. McKeith, R. H. Perry, A. F. Fairbaim, S. Jabeen, and E. K Perry, 1992, Psychological Medicine, 22, p. 920. Copyright © 1992 by Cambridge University Press. Reprinted with permission.

192

CHAPTER 10 TABLE 10.6 F eatures Associated W ith C om m unicating (N orm a! P ressure) H ydrocephalus

Dnmiiin Attention and Arousal Language Motor Memory Problem Solving Everyday and Social Emotional, Personality, T hought C ontent History Features O nset Course History

Nttrmtd Pressure Hy/lrucejih/ilus Low motivation, perseveration; generally impaired early in illness Generally preserved Slowness, gait disorder (spastic, apraxic), incontinence, shuffling, retropulsive psychomotor retardation Deficits are mild to m oderate in severity Visuospatial problem-solving deficits frequendy but not consistendy Apathy, loss of initiative, and spontaneity the cardinal feature

Sudden or gradual Progressive Idiopathic or secondary to traum a (hem orrhage, tum or, infection, etc.)

refers to enlargem ent o f the ventricles resulting from im paired absorption o f the cerebral spina! fluid (CSF). O bstructive hydrocephaly results from a blockage o f m ovem ent o f the CSF, w hereas nonobstructive (i.e., com m uni­ cating o r norm al pressure hydrocephalus) results from an im balance of the p ro d uction an d absorption o f CSF. This latter condition is traditionally recognized clinically by a specific triad o f disturbances: gait and balance disturbance, urinary incontinence, an d relatively m ild cognitive im pairm ent characterized by “frontal” o r subcortical deficits (including slowness in m entation, loss of initiative and spontaneity; see Table 10.6). T he prim ary m ethod o f treatm en t is surgical insertion o f a sh u n t to divert the CSF, allowing it to be absorbed by the body m ore effectively. T h e results o f surgical intervention show substantial variability (see Stam brook, Gill, Cardoso, & M oore, 1993). Stam brook et al. (1988) n oted th at despite postsurgical im provem ent in gait, balance, incontinence, and m any m ental functions, cognitive functioning did n o t necessarily retu rn to norm al. A lthough there are a n um ber o f factors that n eed to be taken into consideration regarding the treatm ent for this condition, accurate early identification is an im portant first step in the treatm en t process. Disorders Primarily Affecting Attention D isturbance o f attention may be viewed as the hallm ark o f acute confusional state (ACS') o r delirium. T he p erso n ’s ability to focus, sustain, an d shift attention, selectively an d voluntarily, is im paired (Lipowski, 1994). Thus,

AGE-ASSOCIATED CONDITIONS AFFECTING COGNITION

193

the person is distractible an d awareness o f self and environm ent is blurred. T h e condition tends to fluctuate over the course o f the day. Closely asso­ ciated with this disturbance o f attention is disturbance o f the sleep-w ake continuum . T he person with an ACS may be drowsy d uring the day an d insom nic at night. A bnorm al psychom otor activity is also an essential fea­ ture o f ACS (Lipowski, 1994). T hree clinical subtypes o f ACS have been described that incorporate the dom ains o f alertness (or readiness to re­ spond to stimuli) an d psychom otor activity level: hyperalert-hyperactive, hypoalert-hypoactive, and a m ixed type (Lipowski, 1994). ACS related to withdrawal from alcohol o r sedative-hypnotic drugs may be hyperactive, w hereas hypoactivity (sluggishness) is the m ost com m on ACS reaction seen in older persons. T he o nset is typically sudden (i.e., hours to days) an d may be p receded by a p eriod characterized by difficulty concentrating, irritability, restless­ ness, o r fleeting illusions o r hallucinations. ACS is a serious indicator of illness in an elderly person and is to be viewed as requiring im m ediate mcdical attention. In many cases, persons recover when the underlying disorder is treated appropriately. However, ACS in an old er person may also signal the exacerbation o f a chronic, perhaps life-threatening, disease. M oreover, survivors do n o t necessarily recover fully (Lipowski, 1994). O f particular note is that older persons may take far longer to recover from ACS than younger persons an d m onitoring o f this im provem ent is im por­ ta n t to ensure am ple recovery tim e before far-reaching decisions are m ade on the p erso n ’s behalf. For exam ple, postoperative ACS occurs in 10% to 15% o f older persons undergoing surgery and may last for weeks (Mesulam & Geschwind, 1976). ACS may be superim posed on an underlying dem entia and the signal o f an em ergent m edical condition (e.g., urinary tract infection o r p n eu ­ m onia). Thus, it is o f utm ost im portance that the symptoms o f ACS be identified. It is an attentional disorder that affects global cognitive func­ tioning, has an acute onset, fluctuates in severity over the course o f the day, is m ost severe at night, often presents with hallucinations, and often includes restlessness o r sluggishness. T h e m ost com m on precipitating events are physical illness (e.g., infections, metabolic disorders, cardiovas­ cular disorders, cerebrovascular disorders), injury, surgery, anticholinergic, o r recen t change in m edication use. Disorders Primary Affecting Mood or Thought Processes Many disorders affecting m ood and thought processes com m on to younger adults may also be seen in the older adult and have been described in detail elsewhere (e.g., Birren et al., 1992). Discussion h ere is lim ited to two conditions that show little o r no evidence o f cognitive deficits b u t m arked behavior disturbances: depression and paraphrenia.

194

CHAPTER 10

Depressive syndromes are very com m only seen in later life even though the prevalence o f depressions with m arked symptomatology is m uch lower in o ld er persons than in younger adults (Birren et al., 1992). T he D SM -IV provides criteria for diagnosing a n u m b e r o f forms o f m ood disturbance (e.g., m ajor depressive episode, dysthymic disorders, adjustm ent disorders, bereavem ent). T he term pseudodementia has b een used to describe reversible cognitive im pairm ent associated with depression (Wells, 1979). However, longitudinal evaluations o f persons with these apparendy reversible condi­ tions (Alexopoulos, 1989) and the acceptance o f the sim ultaneous presen­ tation o f depression and dem entia (Reifler, 1982) have h elped to clarify the significance o f cognitive deficits coexisting with depressive symptomatology. C ertain clinical features may help to distinguish between persons with depression and those with dem entia o f the A lzheim er’s type (see Table 10.7). However, it m ust be noted that interm ittent depressive symptoms have been repo rted in up to 50% o f persons with dem entia (P. Ernst, Badash, Beran, Kosovky, & Kleinhauz, 1977) and 20% to 30% o f persons with AD (Rcifler, Larson, Teri, & Poulsen, 1986). W hether depression TABLE 10.7 Features Associated With Alzheimer’s-type Dementia and Depression Domnin

Alzheimer's-type Dementia

A ttention an d Arousal

Distractible

Language

Decreases over time

Motor

Dressing apraxia; other apraxias over time Recent and rem ote memory im pairments Abstract reasoning; visuospatial over time

Memory Problem Solving Everyday and Social Emotional, Personality, T hought Content O th er Features

History Features Onset Course

Defrression (pseudodementia) Low motivation; decreased attention to detail Limited spontaneous elaboration Psychomotor retardation Secondary to attention Mental processing slowed

Apathy; personality change may be paranoid o r sexually inappropriate Rarely self-referred; female > male; no identifiable cause

Preoccupation with affective state

Insidious

Relatively sudden preceded by dysphonic symptoms Variable

Progressive

“D on’t know” responses; transposition errors; cautious errors; typically higher functioning than complaints indicate

AGE-ASSOCIATED CONDITIONS AFFECTING COGNITION

195

presents with or w ithout cognitive impairment, treatm ent is warranted as dem ented depressed persons may show a reversal of affective symptoms even though their cognitive functioning remains im paired (Reifler et al., 1989). Thus, the excess disability created by the depression can be alleviated and a better quality of life can be achieved. Psychotic phenomena, including delusions and hallucinations, have been reported in late life in elderly persons with early-onset schizophrenia (e.g., Cohen, 1990); persons with a variety of dem entia syndromes (e.g., Absher & Cummings, 1993); and elderly persons for whom there is a clear absence o f an affective syndrome, progressive cognitive impairment, o r other ob­ vious organic cause (Pcarlson & Petty, 1994). These latter syndromes have been referred to as late-life schizophrenia, or late-onset. paraphrenia. There has been m uch debate conccrning the use of separate criteria for late-onsct paraphrenia as distinct from late-life schizophrenia (e.g., Almeida, Howard, Forstl, & Levy, 1992; Almeida, Howard, Levy, & David, 1995a; Munro, 1991). Both the DSM-IV and the ICD-10 support the classification o f per­ sons presenting in this m anner into the general diagnostic category of schizophrenia or a variety of other diagnostic categories d ependent on the im pact of the symptomatology and duration of the condition (e.g., Delu­ sional Disorder, in which daily functioning is not markedly impaired; Schi­ zophreniform , if symptoms present for less than 6 months; Brief Psychotic Episode, if symptoms present for less than 1 m onth in duration). No distinctions are m ade by age of onset. Almeida et al. (1995a) argued that including late paraphrenia in the diagnosis of schizophrenia or delusional disorder has poor empirical and theoretical bases and late paraphrenia (see Table 10.8) may still be the best option for the classification of these late-onset psychotic states. Munro (1991) also argued that this group of persons, who differ from cases of delusional disorder and schizophrenia, are potentially being “lum ped” into the indeterm inate category of “Psychotic disorder, not otherwise specified.” He proposed a similar set of criteria to that of Almeida et al. (1995a) for paraphrenia. Although, by definition, this group of individuals does n o t exhibit cognitive deficits, it is im portant for neuropsychologists to be aware of the diagnostic issues surrounding this symptom presentation. It is not uncom ­ mon for persons presenting in this way to be referred for evaluation of their cognitive status. It is of interest that there appear to be m ore women than men presenting with this condition, whereas m ore men than women present with earlier onset schizophrenia. Moreover, sensory and social isolation are often present to a greater extent than seen in the general age-related population, as are neurological soft signs, which may suggest an organic basis for this condition (Almeida, Howard, Levy & David, 1995b).

196

CHAPTER 10 TABLE 10.8 In clu sio n a n d E xclusion C riteria fo r D iagnosis o f L ate P a ra p h re n ia

Inclusion (a) O nset of symptoms at o r over age 55 (b) Symptoms m ust have been present for at least 6 m onths (c) At least o n e o f the following: • Delusions o f any kind that are in d ep en d en t o f affective symptoms (delusion must rem ain un ch an g ed in absence o f symptoms) • T h o u g h t echo, insertion, broadcasting, o r withdrawal • P ersistent hallucinations in any modality, when accom panied by c ith er fleeting or half-formed delusions without clear affective content, or by persistent overvalued ideas, or when occurring every day for weeks o r m onths on end Exclusion (a) Lack of adequate corroboration o f history by medical notes (b) D em entia an d o th er diagnosablc organic disorders (ICD-10) • Evidence o f neurological, metabolic, or sim ilar disorder that could cause psychotic symptoms . MMSE < 24 • Presence o f m ajor brain lesions on MRI scan such as stroke o r tum ors (c) Presence of past or cu rre n t m odcratc/scvcrc depressive episode (ICD-10) (d) H arm ful substance use or d ependence (ICD-10) (e) Past history as psychiatric in-patient o r treatm ent with neuroleptics, antidepressants, or lithium at any time before age 55 (f) Illiteracy (g) Severe visual im pairm ent (unable to read 24-point print) (h) N o inform ed consent A’obi. From “Psychotic States Arising in Late Life (Late P araphrenia): Psychopathology an d Nosology," by O. P. Almeida, R. J. Howard, R. Levy, an d A S. David, 1995a, British Journal o f Psychiatry, 166, p. 206. Copyright © 1995 by Royal Society Medicine Press. R eprinted with perm ission.

PERVASIVE DISTURBANCES OF COGNITION AND BEHAVIOR As n o te d earlier, the term dementia, as d efin ed in th e DSM -IV, refers to deficits in m o re th a n o n e are a o f cognitive function in g . As m o re areas o f fu n ctio n in g becom e im paired, it becom es m o re difficult to differen tiate am o n g co n d itio n s on the basis o f cognitive p erform an ce. T h e history an d course characteristics are o f prim ary im p o rtan ce in th e diagnostic process. A lth ough m any conditions p ro d u ce d em en tia, cognitive profiles typical of, o r specific to, these disorders may n o t be readily ap p a ren t. In ad d itio n , th e evolution o f cognitive sym ptom atology over tim e may m ask o r obscure d ifferen tiatin g features am o n g these conditions. O n ce d eterio ratio n in global cognitive functio n in g is evident, diagnostic issues te n d to be o f som ew hat less co n cern an d m a n ag em en t issues typically b eco m e th e focus fo r th e affcctcd individuals an d th eir care p rovider (s). H ow ever, observa­

AGE-ASSOCIATED CONDITIONS AFFECTING COGNITION

197

tion (through neuropsychological evaluation) o f the em ergence o f new features may help clarify the diagnosis in persons initially presenting with m ore circum scribed o r focal disorders (as described previously). Two con­ diuons that typify how the patterns o f characteristics (e.g., onset, course, history) can be useful in m aking etiological distinctions w hen pervasive disturbances o f cognition an d behavior are p resent are A lzheim er’s disease (AD) and Vascular dem entias. Alzheimer’s Disease T he m ost com m on source o f generalized cognitive im pairm ent (or d em en ­ tia) o f insidious onset and progressive course affecting persons over age 65 is A lzheim er’s disease. A m ong the m ost com m only applied criteria for identifying AD are those described by the N ational Institute o f Neurological and Com m unicative D isorders and Stroke-A lzheim er’s Disease and Related D isorders Association (NINCDS-ADRDA; McKhann et al., 1984), which include the presence o f the dem entia syndrom e as established by clinical exam ination and confirm ed by neuropsychological testing, deficits in two or m ore areas o f cognition, progressive worsening o f m em ory an d o th er cognitive functions, no disturbance o f consciousness, onset between age 40 an d 90 and m ost often after age 65, and absence o f systemic disorders or o th e r brain diseases that could account for the progressive deficits in m em ory and cognition. O f note, im pairm ent in a dom ain o f functioning is operationally defined as a score falling below the fifth percentile com pared to appropriate norm ative data controlling for age, gender, and education. T he pathological hallmark o f this neurodegenerative disorder is a prolif­ eration o f neuritic plaques an d neurofibrillary tangles distributed through cortical and subcortical regions o f the brain. A lthough variations in the p attern o f distribution o f these plaques an d tangles may occur, typically the parietotem poral and limbic regions are predom inantly affected (T. L. Kem per, 1994). However, it has been noted that m ultiple n eu ro tran sm itter an d neuro m o d u lato r abnorm alities exist in the AD brains. T he prim ary pathological process is unknow n. Behaviorally, AD typically presents with im pairm ent o f new learning. Initially deficits may only affect the retrieval o f new inform ation. As the disease progresses, im pairm ents in the acquisition an d reten tio n o f new m aterial also becom e apparent. In addition, the individual may also begin to have difficulty recalling inform ation learned in the past a n d /o r fam iliar persons and places. As a consequence o f this m em oiy im pairm ent, indi­ viduals with AD may have difficulty concentrating o r may loose track of w hat they are doing and appear distractible. It is not uncom m on for m arked disturbances in visuospatial processing also to be ap p a ren t relatively early in the course of the disorder. Thus, copying visual m aterial may prove difficult (e.g., drawings o r block designs) and the com prehension o f spatial

198

CHAPTER 10

relations may be poor. Disturbances in language may also em erge charac­ terized by em pty speech, circum locutions, and inappropriate w ord selec­ tion as the disease progresses. Eventually both expression and co m p reh en ­ sion are affected. M otor functions, p er se, rem ain intact well into the course o f the disorder, but problem s com pleting m o to r sequences (e.g., dressing apraxia) may be apparen t in m oderate to late stages. T he m ost com m on change in character o r personality associated with AD is apathy (Bozzola, Gorelick, & Freels, 1992). Persons loose interest in hobbies and activities and may sit unoccupied, o r perform ing a repetitive task, for long periods of time. T he “stages” of AD have been captured in unidim ensional (e.g., Reisberg, Ferris, De Leon, & Crook, 1982) and m ultidim ensional scales (e.g., Clinical Rating Scalc: Berg, 1988; Functional Rating Scale: Tuokko, 1993). T he advantage of m ultidim ensional scales over unidim ensional scales is th a t individual differences in presentation and course can be reflected. For exam ple, Marlin (1987) noted that persons with AD may p resen t with different patterns o f perform ance. Persons presenting with relatively m ore im paired word-finding problem s than visuospatial deficits showed signifi­ cantly g reater hypom elabolism in the left tem poral lobe than o th e r cortical regions. Conversely, persons with greater constructional deficits than wordfinding problem s displayed significantly greater hypom etabolism in the rig h t tem poral and parietal areas. It has been noted th a t measures, useful for the early detection o f AD, may n o t necessarily be those m ost sensitive to change across the course o f the disease. Floor an d ceiling effects for specific instrum ents will em erge at different stages o f the dem entia (e.g., C hristensen, Hadzi-Pavlovic, & Jacom b, 1991). Various batteries o f tests have been recom m ended for use when evaluating persons with dem entia (see Zee, 1993, for a review). In addition, specific tools have been designed for assessing severe AD (e.g., M. A lbert & Cohen, 1992; Volicer, Hurley, Lathi, & Kowall, 1994) o r frail elderly persons (Coval e t al., 1985), which are m ost useful for defining areas of cognitive strengths and weaknesses for care planning. Thus, the p urpose of the assessment is o f utm ost im portance w hen choosing m easures to assess cognition in persons with suspected AD (e.g., early detection, m onitoring behavior change, providing inform ation for care plans o f se­ verely im paired persons). D ifferent m easures are m ost appropriate for different purposes. Vascular Disorders T he relations between vascular disorders and cognitive functioning have u n d ergone m uch scrutiny and continue to be a difficult area to define clearly. F undam ental to understanding this issue is the realization th at

AGE-ASSOCIATED CONDITIONS AFFECTING COGNITION

199

there are a n um ber o f conditions affecting the circulatory system, which supplies the brain an d spinal cord, that may be related to cognitive im­ p airm en t (e.g., throm bo-em bolic disorders, anoxic-ischemic disorders, hem orrhagic disorders). M oreover, the p redom inan t location o f the re­ sulting brain dam age is related to the type o f cognitive im pairm ent dem ­ onstrated. Cum m ings and Benson (1992) described two principal vascular dem entia syndromes: disorders resulting from the occlusion o f arteriolesize vessels irrigating the d ee p gray an d white m atter structures (e.g., lacu­ n ar states and Binswanger’s disease), which primarily affect subcortical structures; and disorders involving extracranial (carotid) o r intracranial vessels o r the small vessels supplying cortical regions. Subcortical states may affect the basal ganglia, thalam us, internal capsule, o r subcortical white m atter and com m on features in the fully developed syndrom e include rigidity, spasticity, pseudobulbar palsy, lim b weakness, exaggerated muscle stretch reflexes, and extensor plantar responses. T he characteristics o f the cognitive im pairm ent are variable and not well defined (Cum m ings & Benson, 1992), but many persons show features typical o f o th e r subcortical dem entias, including apathy and loss o f tact. Cortical insults may result in well-defined focal syndromes reflective o f the location o f dam age. W hen a n u m b er o f infarctions have occurred (i.e., m ultiple and often bilateral), the clinical picture may be that o f a m ore generalized dem entia syndrom e. T he two m ethods m ost widely used in the past to define vascular disor­ ders were the multi-infarct criteria in the D SM -III-R and H achinski’s is­ chem ic score (Hachinski, Lassen, & Marshall, 1974). Each o f these meas­ ures has been found to be lacking (e.g., D ening & Berrios, 1992; M etter & R. S. Wilson, 1993; Rosen, Terry, Fuld, Katzman, & Peck, 1980) and new sets o f criteria have em erged that m ore accurately reflect the diversity o f vascular conditions that may affect cognition. In Chui et al.’s (1992) system for probable ischem ic vascular dem entia, radiologic evidence o f at least one supratentorial lesion and evidence o f a stroke tem porally linked to the dem entia o r two o r m ore o th e r strokes is required. T he N ational Institute o f N eurological Disorders and Stroke-Association Internationale p o u r Recherche et L ’E nseignem ent en N eurosciences (NINDS-AIREN) system (Roman et al., 1993) for probable vascular dem entia requires clini­ cal (i.e., history, exam ination) or radiological (neuroim aging) evidence of cerebrovascular disease, and a plausible tem poral relation between the cognitive dysfunction and cerebrovascular disease. It would be expected, given the diversity o f brain regions potentially affected by vascular disorders, that there may be many different types of clinical presentation associated with vascular disorders. Despite the in h er­ e n t differences between the forms o f vascular disorders th at may affect cognition, th ere are characteristics seen com m only in vascular disorders in general that may aid in their identification (see Table 10.9) an d differ-

200

CHAPTER 10

TABLE 10.9 Features Associated With Alzheimer's-type D em entia a n d Vascular Dem entia Domain

Alzheimer’s-type Dementia

A ttention an d Arousal L anguage M otor

Distractible Decreases over time Dressing apraxia; o th er apraxias over time

Memory

R ecent an d rem ote m emory im pairm ents Abstract reasoning; visuospatial over time

Problem Solving Everyday an d Social Em otional, Personality, T ho u g h t C ontent

Apathy; personality change may be paranoid or sexually inappropriate

O th e r Features

History Features O nset Course History

Vascular Dementia

O ften sensory-m otor signs secondary to infarction in the basal ganglia, internal capsule, thalam us; dysarthria, dysphagia Relatively well preserved WAIS perform ance scores p o o rer than in AD Pseudobulbar affect

Subtle changes; hypertension; focal neurological signs (num bness, tingling, weak­ ness, slurred speech) Insidious Progressive Rarely self-referred; fem ale > male; no identifiable cause

S udden o r insidious Stepwise o r sm ooth progression Steep increase in occurrence after age 65; m ale > female; strokes; TIAs; Mis

en tiatio n from AD. T hese include historical features such as a b ru p t onset, stepwise d eterio ratio n , fluctuating course, previous h y p erten sio n an d a history o f neurological sym ptom s o f tran sien t ischem ic attacks (e.g., Erkinju n tti e t al., 1988; H achinski e t al., 1975). However, as n o te d by C um m ings an d B enson (1992), a gradual d eterio ratio n , as seen in n eu ro d eg en erativ e disorders, a n d th e absence o f typical historical features does n o t p reclu d e th e diagnosis. Typical factors associated with vascular origins o f cognitive declin e in c lu d e hypertension, h e a rt disease, cigarette sm oking, diabetes m ellitis, alcohol con sum ption, an d hyperlipidem ia (M eyer, M cClintic, Rog­ ers, Sims, & M ortcl, 1988). F or the m o st part, neuropsychological research has n o t d ifferen tiated am o n g th e vascular syndrom es and, alth o u g h som e research exists o n specific w ell-defined subtypes (B ernard et al., 1990, 1992), it is n o t possible to m ake specific com m ents a b o u t th e neuropsychological differen tiatio n

201

AGE-ASSOCIATED CON DITIO NS AFFECTING COG NITIO N

o f these conditions from others. T h e m ost com m o n ch aracteristic n o te d is a “patchy” distribution o f intellectual deficits th at may d iffer am o n g individuals (A m erican Psychiatric Association, 1987). SUMMARY OF KEY FEATURES A lthough the p a tte rn o f cognitive a n d behavioral fu n ctio n in g has been th e focus o f the previous discussion, key features may be o f assistance w hen fo rm u latin g diagnostic hypotheses. T hese include th e type o f o n se t (i.e., su d d e n o r g rad u al), a n d course (i.e., rapid decline, insidious decline, step ­ wise decline, fluctuating), in add itio n to the p attern o f stren g th s a n d weak­ nesses observed d u rin g the assessm ent process. For exam ple, th e p resen ce o f a relatively su d d en o n se t o f sym ptom atology would be m ost suggestive o f acute confusional states, depression, trau m a (e.g., m in o r blow to the h e a d o r fall), o r signify a cerebrovascular acciden t (CVA). For m o st d e­ generative disorders, such as A lzheim er’s disease, th e o n se t is slow an d insidious an d inform ants may have g reat difficulty identifying w hen the first changes in fu n ctio n in g w ere noted. T able 10.10 describes som e o f the conditions associated with various course characteristics. A cute confusional states (ACS o r d eliriu m ) may be associated with a variety o f underlying conditions an d may p re se n t as a rap id d e te rio ra tio n in fu n ctio n in g o r a p e rso n ’s level o f fu n ctio n in g may fluctuate. In geriatric p opulations, acute confusional states are life th re a te n in g and req u ire p ro m p t a n d im m ediate m edical intervention . P ractitio n ers w orking in acute care o r long-term care settings may com e in co n tact with p e rso n ’s p re se n tin g with ACS m o re th an practitio n ers in o th e r settings. TABLE 10.10 Course Characteristics Associated With Specific Conditions Course Rapid decline Stepwise d eterioration Stable (following sudden decline)

Progressive decline Im provem ent with time or intervention

Fluctuating

Disorder Acute Confusional State (ACS) Creutzfeld-Jacob Vascular dem entia W em icke-K orsakoff encephalopathy Anoxia Large vessel cerebrovascular accident N eurodegenerative disorders (e.g., A lzheim er’s disease, P arkinson’s disease) Vascular conditions ACS Depression ACS Frontal-type D em entia Lewy Body D em entia

202

CHAPTER 10

A stepwise course is often heralded as a key identifying feature o f vascular dem entia (Hachinski et al., 1974). In som e forms o f this condition, the individual repeatedly suffers small CVAs, which may not p resen t with the classic features o f m otor weakness seen in larger CVAs. Episodes o f flu-like symptoms may be reported following which the individual no longer is able to perform at th eir previous level o f functioning. This p attern of decline differs from th a t m ore typically associated with CVAs w here there is a large vascular event with m arked disturbance of functioning, followed by a p eriod o f “recovery” o f some function an d a relatively stable course thereafter. O ften depression an d ACS will improve with intervention an d so too will the functioning o f the individual. Particularly in cases w here depression coexists with an o th er condition (e.g., poststroke o r A lzheim er’s disease), it is im p o rtan t to distinguish between im proved functioning du e to alle­ viation o f depression o r ACS symptomatology and im provem ent (or rever­ sal) o f the coexisting neurological condition. M isinterpretation o f the na­ tu re and extent o f im provem ent in functioning may place persons at risk (e.g., sent hom e from hospital w ithout supports even though underlying deficits still exist). M arked fluctuations in functioning may occur with ACS b u t may also be a feature o f o th e r degenerative disorders such as Lewy Body disease an d frontotem poral dem entias. As noted earlier, ACS requires im m ediate medical intervention and m ust be ruled o u t as the source o f fluctuations in behavior. T he fluctuations seen in LBD and FTD ten d to occur across days (or testing sessions), w'hereas ACS fluctuations may be m ore m om entto-m om ent in nature. Table 10.11 describes som e o f the diagnoses suspected when particular cognitive strengths and weaknesses are observed d u rin g the assessment process. As a general rule, m ovem ent (including speech) disorders suggest dam age to subcortical structures of the brain though some m ovem ent disorders may be benign o r idiopathic (e.g., trem or and facial dyskinesias). It should also be noted th a t m ovem ent disorders may arise for o th e r reasons than disorders o f the central nervous system (e.g., arthritis) and it is im p o rtan t to distinguish between these sources when m aking diagnostic decisions. M arked fluctuations in attention and concentration d u ring the course o f the assessment may suggest the presence o f an ACS, particularly if consciousness is clouded. Distractibility during the assessment a n d /o r re­ sp onding to irrelevant stimuli (e.g., distracted by noise outside testing room o r objects within the room ) may be seen with FTDs. W hen language deficits are the primary presenting feature, the syn­ drom e of Progressive Aphasia may be considered, although the underlying neurological disorder characterized by this behavior presentation may be

AGE-ASSOCIATED CONDITIONS AFFECTING COGNITION

203

TABLE 10.11 Behavioral Presentations Associated W ith Specific C onditions Observed Impairment Arousal and Attention Clouded consciousness or fluctuations in concentration High distractibility Language Motor Gait

Trem or Tics, facial dyskinesias Slowness in response Initiation problems Speech

Memory Subjective complaint exceeds objective deficits New learning only

Relatively spared M ood/E m otion/T hought Content Preoccupadon with affective state U nconcern or lack of awareness Pseudobulbar affect Marked personality alteration Delusional ideation Depressive symptomatology

R/iises the Question of:

Acute confusional state (ACS) Disorders affecting the frontal lobes; acute confusional state Primary progressive aphasia Communicating (Normal Pressure) hydrocephalus (NPH) subcortical disturbances including Parkinson’s disease, progressive supranuclear palsy (PSP), multisystem degeneration Benign; Parkinson’s disease; Lewy Body dementia Neuroleptic medications; idiopathic Basal Ganglia calcification Psychomotor retardation of depression Disorders affecting the subcortical a n d /o r frontal regions Disorders affecting subcortical structures such as amyotrophic lateral sclerosis or vascular disorders Depression Amnestic disorder (e.g., Wernicke-Korsakoff encephalopathy, WK); early Alzheimer’s disease (AD) Vascular disorders; disorders affecting the frontal lobes Depression WK; AD; disorders affecting the frontal lobes Vascular conditions Disorders affecting the frontal lobes Paraphrenia; various dem entia syndromes (e.g., AD, vascular); ACS May be present in a variety of dem entia syndromes as well as stand alone

u n clca r. A lz h e im e r ’s d isease m ay p r e se n t in this fa sh io n b u t so m ay o th e r n o n sp e c ific n e u r o d e g e n e r a tiv e disorders. T ypically, th e card in al fea tu re o f A lz h e im e r ’s d isease (A D ) is m em ory im p a irm en t. Early in th e cou rse this m ay b e relatively restricted to d eficits in n ew lea rn in g th at evolve th r o u g h o u t th e cou rse to in c lu d e m o st aspects o f m em o ry fu n c tio n in g . H ow ever, o th e r d isorders m ay also p re se n t w ith

204

CFIAPTER 10

circum scrihcd new learning deficits, including various subcortical d em en ­ tias and W ernicke-K orsakoff syndrom e. W ithin the DSM-IV, there is a category o f am nestic disorder that may be related to specific medical con­ ditions o r may be idiopathic (not otherwise specified). W hen m em oiy functioning is relatively spared in the context o f o th e r areas o f cognitive im pairm ent, this may suggest the presence o f a vascular origin to the deficits, o r FTD. M oreover, when subjective com plaints of m em ory im pair­ m en t (on the p art o f the person being assessed) exceed the objective evidence of little if any m em ory deficits, then depression may be suspected. A lthough typically assessed through observation rath er than form al test­ ing, disorders o f em otion, th o u g h t content, and personality may also be suggestive o f specific disorders. For exam ple, preoccupation with affective status is seen m ost com m only in persons with a m ajor depression. Persons who show pseudobulbar affect (i.e., strong affect triggered by m ild stimu­ lation) may have sustained dam age to the bulbar region o f the brain thro ugh traum a o r vascular events. M arked apathy a n d /o r lack o f concern ab o ut their situation may be seen with AD, FTD, o r W ernicke-K orsakoff syndrom e. M arked personality change early in the course o f the disorder is th o u g h t to be primarily associated with FTD and may bring the person into contact with police for the first time late in life. Delusional ideation may accom pany a variety o f organic conditions (e.g., AD, FTD) b u t may also be present in ACS. In the context of intact cognition, delusions may be present in latc-o n set schizophrenia, o r paraphrenia. D ifferentiation of these conditions is o f im portance for identifying all rem ediable sources of cognitive change. M oreover, clarification o f the diagnosis may also assist in care planning through the establishm ent o f app ropriate expectations concerning the present capacities o f the indi­ viduals and their future needs.

Chapter

The Role of the Caregiver in Neuropsychological Assessment

Caregivers and proxies o f palienls can play an im portant role in the as­ sessm ent process. In the case o f elderly persons with cognitive im pairm ents, these individuals may serve as legal guardians and provide consent for assessment a n d /o r treatm ent on behalf o f the affected individual. In all cases, they can provide valuable background inform ation ab o u t the p er­ so n ’s condition and, if applicable, rate o f cognitive decline. This ch ap ter focuses on nonprofessional caregivers, typically family m em bers (i.e., usu­ ally spouses o r children). Two m ain areas o f caregiver assessment are covered here. T h e first area relates to the role o f the caregiver as a provider o f inform ation ab o u t the cognitively im paired individuals and their functioning. T h e second area focuses on the assessment o f caregiver stress that can interfere with effective provision o f care.

THE CAREGIVER AS AN INFORMATION PROVIDER Regardless of the degree o f cognitive im pairm ent, a caregiver has the potential o f being a rich source o f inform ation about an individual’s func­ tioning. In som e m ultidisciplinary settings, family and genetic histories are often collected by som ebody o th e r than the neuropsychologist. Even in such circum stances it is im portant for the neuropsychologist to interview the caregiver to obtain m ore specific inform ation ab o u t the p erso n ’s cog­ nitive functioning. Q uestions concerning the onset and duration o f symp­ toms noticed by the caregiver are im portant, as arc those concerning 205

206

CHAPTER 11

disruptions o f specific behaviors (e.g., W hen were the first changes noticed? W hat types of changes have you noticed in language functions? W hat types of changes have you noticed in m em ory functions? W hat types o f changes cam e first? Was the deterioration gradual o r were there relatively sudden changes noticed? How is the perso n ’s everyday functioning? How did it change recently?). In addition to a personal interview, standardized, valid, and reliable instrum ents may be useful for collecting inform ation from caregivers. It is im portant to note that the correspondence between the responses o f proxies an d aged persons in the com m unity varies d ep en d in g on the n ature of the inform ation sought, characteristics o f the person, charac­ teristics o f the proxies, and the proxies’ ability to observe the care recipient (Zim m erm an & M agaziner, 1994). Specifically, Zim m erm an an d M agaziner (1994) reviewed the literature on some o f the factors th at affect proxy responses and found the following: (a) T he m ore objective and concrete the question, the higher the correspondence between care recipient and proxy reports; (b) proxy-person agreem ent on ratings o f cognitive ability is relatively good (but not as good as it is for m easures of physical abilities), whereas proxy-subject agreem ent on ratings o f affective status varies across studies; (c) if subjects arc in good health, then agreem ent is b etter on m easures o f satisfaction and instrum ental activities o f daily living (e.g., ability to use the telephone) than for physical activities o f daily living (e.g., eating) an d agreem ent is m aximized when the subjects are im paired; an d (d) although generally proxies indicate m ore im pairm ent and disability th an respondents do, there are some exceptions (Basset, M agaziner, & H ebe, 1990). Overall, the m ost concrete finding seems to be that the highest agreem ent between respondent an d proxy reports is o btained for objective items that ask about specific and observable aspects o f functioning an d that proxies tend to rep o rt m ore disability than the care recipient (Zim m erm an & M agaziner, 1994). Norm ative inform ation on some im portant instrum ents that are used to assist in diagnosis by obtaining inform ation from a caregiver is presented here. T he goal of using such instrum ents is typically to d eterm ine w hether o r n o t the functioning o f the elderly person differs from that o f norm al elderly persons. Thus, norm ative data collected from inform ants respond­ ing abo u t norm al elders would constitute an ap pro p riate com parison group. Some inform ation reported by proxies of people with dem entia were also included because they may prove useful in some cases. Measures o f Cognitive and General Functioning T he Presen t Functioning Questionnaire (PFQ; Crockett, Tuokko, Koch, & Parks, 1989) is an interview-administered m easure that consists of 60 items devel­ op ed to form five scales: personality (assesses for psychopathology symp­

207

THE ROLE OF THE CAREGIVER

tom s), everyday functioning (e.g., “problem s handling m oney” an d “p ro b ­ lems shopping"), language skills, memory, an d self-care (e.g., “Must be b ath ed by som eone else”). A copy o f the questionnaire appears in Tuokko and C rockett (1991). Scores reflect the num ber o f problem s en d o rsed by the caregiver. T he questionnaire was developed for adm inistration to collaborative inform ants (typically caregivers of people with dem entia) and represents the caregivers’ im pressions o f the subjects’ deficits. This m easure has satisfactory reliability and validity (Crockett e t al., 1989; H adjistavropou­ los, Taylor, Tuokko, & Beattie, 1994). Such an instrum ent could be adm in­ istered in conjunction with a caregiver interview and contains very specific questions centered around symptomatology that could be followed u p by the interviewer. Normative data for this instrum ent (based on a com m unity sam ple) are presented on Table 11.1. A lthough the neuropsychologist can obtain inform ation through the PFQ about the activities o f daily living (ADL) and instrum ental activities o f daily living (IADL), several additional instru­ m ents exist for that purpose and are presented later. T he Revised Memory and Behaviour Problems Checklist (RMBPC; Teri et al., 1992) is an o th er instrum ent designed for caregivers rep o rtin g on the symptoms exhibited by care recipients. Each behavior is rated o n two scales: (a) Frequency, representing the rate o f occurrence o f specific behaviors (0 = never occurs; 1 = occurs infrequently and not in the last week; 2 = occurred 1-2 times in the last week; 3 = occurred 3-6 times in the last week; an d 4 = occurs daily or m ore often); and (b) a Reaction rating, rep resen tin g the degree to which individual behaviors b o th er o r upset the caregiver (0 = n o t at all, 1 = a little; 2 = moderately; 3 = very m uch; 4 = extrem ely). Thus, this in stru m en t also provides inform ation on caregiver stress. Teri et al. derived th ree subscales based on factor analysis: Depression, Disruption (e.g., includes verbal aggression, destroying property, arguing, etc.), and Memory-Related TABLE 11.1

Normative Inform ation on the Present Functioning Questionnaire Scales Personality Scale Everyday Tasks Scale Language Skills Scale Memory Scale Self-care Scalc Total Score

M

SD

1.40 .17 .16 .80 .09 2.60

1.91 .59 .65 1.62 .44 3.68

Not*. Collaborative inform ants who reported on 70 normal elderly volunteers from senior’s activity groups. Data from Crockett, D., Tuokko, H., Koch, W., & Parks, R. (1989). Copyright © 1989 by T he Haworth Press Inc., Binghamton, NY. Clinical Gerontologist, The assessment of everyday functioning using the Present Functioning Questionnaire and the Functional Rating Scalc in elderly samples 8(3), p. 15. Adapted with permission.

208

CHAPTER 11

Problems. They also presented m eans and standard deviations o b tain ed by collaborative inform ants who reported on elderly persons, m ost o f whom were diagnosed with dem entia. T he internal consistency was .75 an d .87 for F requency and Reaction, respectively, an d validity evidence (in term s o f relation to o th e r scales) has been presented (Teri et al., 1992). This scale may be useful when it comes to assessing the course an d severity of geriatric disorders. Its specificity an d sensitivity in differentiating norm al elders from those with dem entia, however, requires fu rth e r investigation. In th at sense, the PFQ m ight prove to be a m ore useful instrum ent than the RMBPC as norm s on norm al elders are available. T he RMBPC was published in its entirety in T eri et al. (1992, appendix). Jo rm and Jacom b (1989) presented general population norm s as well as psychom etric inform ation on the Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE). T he instrum ent was specifically developed to assess decline from a prem orbid level. T he caregiver reports on degree o f change over a 10-year period. T he m easure consists o f 26 items. It contains questions p ertin en t to memory, language abilities, and everyday functioning. Responses arc provided on 5-point form at ranging from Much B etter (I) to Much Worse (5). T he questionnaire has high in tern al con­ sistency and was found to discrim inate between inform ants derived from the general population and inform ants reporting on persons with d em en ­ tia. T he m ean correlation with the MMSE is .59 and correlations with education have been near zero (Jorm , 1996). T he lack of correlation with education is a m ajor strength o f this instrum ent. Jorm and Jacom b (1989) included the questionnaire in its entirety in an appendix. G eneral p o p u ­ lation norm s were derived from a volunteer Australian sam ple (Jorm & Jacom b, 1989). T he persons, whose nam es were taken from the electoral role for the Australian Capital Territory, were contacted and asked if they knew o f an elderly person. O f the 815 persons who responded an d indi­ cated they knew an elderly person, 613 filled o u t the questionnaire. T he norm s are based on these individuals and are presented as percentages. A nother instrum ent com m only used with caregivers is the Blessed Demen­ tia Rating Scale (BDS; Blessed et al., 1968). T he second p art of the BDS was developed for adm inistration to the care recipient and was discussed in ch ap ter 3. T he first part o f the instrum ent is adm inistered to the caregiver and assesses changes in daily living, self-care, cognitive capacity, and personality dom ains (Blessed et al., 1968). H igher scores are indicative o f m ore severe deficits. It consists o f 22 items and is adm inistered to a proxy using an interview form at. T hree dom ains are assessed: changes in p erform ance o f everyday activities (e.g., inability to find way ab o u t familiar streets); changes in habits (e.g., eating); an d changes in personality, in­ terests, drive (e.g., hobbies relinquished). Scores on the in stru m en t relate to neuropsychological testing and can differentiate between different de­

THE ROLE OF THE CAREGIVER

209

grees of dem entia (Erkinjuntti, H okkanen, Sulkava, & Palo, 1988). Erkinju n tti, H okkanen, e t al. (1988) discussed a revised version o f the BDS (consisting o f the sum o f Items 1-11) th a t achieved h ig h er sensitivity and specificity than the unrevised version. Specifically, a cutoff score o f 4 in the unrevised version gave a sensitivity o f 90% a n d specificity o f 84%. In the revised version (RDS), a cutoff score o f 1.5 gave a sensitivity o f 93% and a specificity of 97%. T he average BDS score for com m unity dwelling elderly persons (ra = 105) was 2.23 (+ /-.2 3 ). R ecent research has shown th a t the in te rra te r reliability betw een two raters who, w orking in d e­ pendently, interviewed each o f 47 caregivers o f persons with dem entia, was low (Cole, 1990). It would be interesting for future research to d eter­ m ine w hether sim ilar intcrrater reliability problem s arc also fo und with respect to o th e r instrum ents adm inistered to caregivers. Williams (1991a) discussed the Cognitive Behavior Rating Scales, which arc m easures that can be adm inistered to relatives o r o th e r caregivers of persons with dem entia. This 117-item questionnaire covers the following areas of functioning: Language Deficits, Apraxia, D isorientation, Agitation, N eed for Routine, Depression, H igher Cognitive Deficits, M emory Disor­ der, and Dem entia. Satisfactory reliability an d validity data o n the instru­ m en t are available (Williams, 1991a). T he m anual o f the scale discusses small age effects that appear for the subscales focusing on A gitation, De­ pression, an d D em entia. It also presents norm s for persons up to age 89. T hese persons were recruited through advertism ents an d were screened for neurological an d psychiatric disorders. Responses were provided by family m em bers who lived with the care recipient. Williams (1991a) p ointed o u t th a t the persons he studied for the developm ent o f the m anual for these scales did n o t include a sufficient n um ber o f m ild cases. Relatives may be less able to provide valid responses for subjects with mild im pair­ m ent. This criticism may hold for m any o th e r proxy-adm inistered instru­ m ents reviewed here and m ore research is n eeded to address this issue. C hapter 3 o f this volume indicated that the Cambridge Mental Disorders of the Elderly Examination (CAMDEX; Roth c t al., 1986) includes an inform ­ an t interview that addresses the areas o f orientation, memory, general intellectual functioning, behavior, personality, m ood, an d activities of daily living of the care recipient. Each item is coded on a 3-point scalc with a m axim um score o f 63. Some validity evidence for the scale exists (O ’Con­ n o r, Pollitt, Brook, & Reiss, 1989). For instance, the inform ant m em ory an d orientation scores correlate with m em ory an d orientation tests. O ’C on­ n o r et al. (1989) presented m eans b u t n o t standard deviations o f total scores derived from inform ants reporting abo u t persons with mild, m od­ erate, and severe dem entia. Spouse interviews resulted in m ean scores of 11.92, 25.95, an d 36.43 for m ild (n = 24), m oderate ( n = 19), an d severe cases ( n = 7), respectively. Interview scores derived from nurse an d child

210

CHAPTER 11

inform ants did n o t differ significantly from the scores o btained during interviews with spouses. D uring the last several years, instrum ents have been developed to diag­ nose dem entia based solely on inform ant data (see Jo rm , 1996). These instrum ents may prove useful in unusual situations d u rin g which it is im­ possible to exam ine and assess the person (e.g., postm ortem ). As this area o f instrum ent developm ent is relatively new, however, m ore research is n eed ed before the clinical utility o f such instrum ents can be established. F or exam p le, th e Informant Based Questionnaire (B arber, S now don, & C raufurd, 1995) was designed to differentiate A lzheim er’s disease from frontotem poral dem entia. It takes approxim ately 1 h o u r an d 20 m inutes to adm inister an d includes a series o f questions focusing o n language, m em ory /o rien tatio n , personality, spatial skills, and o th e r related dom ains. B arber c t al. were able to devise a scoring system that separated perfecdy 20 early-onset A lzheim er’s disease cases and 18 confirm ed cases o f frontotem poral dem entia. Clearly, however, cross validation an d norm ative research is need ed on this instrum ent. Measures Assessing Activities o f Daily Living A lthough som e o f the caregiver strain scales reviewed later in this ch ap ter as well as the PFQ, allow for an assessment o f activities o f daily living (ADL), m ore specialized instrum ents exist ADL assessment instrum ents are often adm inistered to caregivers w hen patients are unable to com plete them. An ADL m easure could be self-report, proxy-adm inistered, o r based on observation. Generally, the validity o f ADL scales can be com prom ised w hen they are adm inistered to persons o f dim inished capacity. O ne o f th e m ost widely used measures is the Index of ADL (Katz, Downs, Cash, & Grotz, 1970). Six ADL functions are rated in a dichotom ous fash­ ion: bathing, dressing, going to the toilet, transfering, continence, and feeding. O ne point is given for each item for which the person requires assistance. Satisfactory validity and reliability evidence for the scalc can be fo und in the literature. For exam ple, Katz et al. (1970) found th at the scale relates in the predicted direction to indices o f adaptive capacity and Sherwood, Morris, Morr, and G utkin (1977) found th at the m easure has high reproducibility. T he original in te n t o f the scale was to be adm inistered to knowledgeable inform ants who would rep o rt on the p erso n ’s status. Law an d Letts (1989) wrote a review of ADL scales tapping basic activities such as eating, dressing, and groom ing. O f the scales they reviewed, they concluded that the Barthel Index (M ahoney & Barthel, 1965), the Index o f ADL (Katz, Ford, Moskowitz, Jackson, & Jaffee, 1963), the Level of Rehabili­ tation Scale (LORS-II; Carey & Posavac, 1982), and the Physical Self-Maintenance Scale (PSMS; Lawton & Brody, 1969) have satisfactory psychometric

THE ROLE OF THE CAREGIVER

211

properties and are short. They also concluded that these scales would m ake good diagnostic measures. Law and Letts (1989) hypothesized th a t the best potential for responsibly m easuring change in ADL function is re­ flected on the Barthel Index (M ahoney & Barthel, 1965), the Kenny Self-Care Evaluation (Schoening et al., 1965), an d the Klein-Bell ADL scale (Klein & Bell, 1982). N onetheless, they stressed that this potential rem ains to be evaluated. More reliability evidence is needed, for exam ple, for both the Kenny Self-Care Evaluation and the Klein-Bell ADL scale. O th er reviewers have argued that the Katz Index is adequate for most purposes an d is used widely (A. R. Kane & L. R. Kane, 1981). For hospital-based clinicians, w orking with patients with chronic long care needs an d m ultiple disabilities, m ore detailed scales such as the Barthel Index (M ahoney & Barthel, 1965) may be particularly useful (A. R. Kane & L. R. Kane, 1981). Instrum ental Activities o f Daily Living (IADL) m easures tap m ore com ­ plex activities associated with daily living. T hese were reviewed in detail by A. R. Kane and L. R. Kane (1981). Many were developed for adm inistration to the person and som e are proxy adm inistered. However, it is com m on for clinicians to adm inister scales to caregivers that were originally devel­ o p ed as self-report measures. T he Functional Health Status Test (Rosow & Breslow, 1966) is a relatively p u re IADL m easure an d contains 25 questions. It is a G uttman-type hier­ archically organized scale. T he items range from questions concerning ability to go to a church m eeting an d the movies to the ability to engage in strenuous physical work such as shoveling snow. M ore research is n eed ed to establish the psychometric properties o f the scale. T h e Philadelphia Geri­ atric Centre Instrumental Activities of Daily Living (Lawton, 1972) also has a G uttm an form at an d taps ability to use the telephone, shopping, food p reparation, housekeeping, laundry, public transportation, taking m edica­ tions, and handling finances. Reproduceability was high b u t the investiga­ tors com m ented that it may be m ore suitable for w om en an d it does n o t tap the full range o f ADL. T he Older Americans Resources Service (OARS): IADL scale (Fillenbaum , 1988) was adm inistered to caregivers who participated in the Canadian Study o f H ealth an d Aging. T he items range from ability to eat and dress to ability to m anage money. In o th e r words, it includes b o th ADL and IADL items (see Tables 11.2 and 11.3). Reliability and lim ited validity evidence exist for this instrum ent (Duke University C entre for the Study o f Aging and H um an D evelopm ent, 1978), although studies arc need ed to d eterm ine the extent to which reliability an d validity o f the instrum ent are affected d epending on w hether the instrum ent is adm inistered to a p atien t o r a caregiver. Items are rated “w ithout help," “with h elp ,” or “unable to perform .” T he scale results in a global score of ADL and IADL. An overall 1 to 6 rating can also be m ade by a rater (1 = excellent and 6

212

CHAPTER 11 TABLE 11.2 C an ad ian Study o n H ealth a n d Aging N orm s on the OARS A D L /IA D L Scale Based o n C aregivers o f P ersons W ho S cored G reater T h an 77 o n th e 3MS (n = 484) Number o f Persons Able to Perftrrm With No Help (% )

Items

Can Can Can Can Can Can Can Can Can Can Can Can Can Can

subject subject subject subject subject subject subject subject subject subject subject subject subject subject

eat? dress and undress? take care of appearance? walk? get in and out of bed? take a bath or shower? use the bathroom or toilet? use the telephone? get to distant places? go shopping? prepare own meals? do housework? take own medicine? m anage own money?

477 471 472 455 477 426 481 451 395 362 405 284 454 443

(98.6) (97.3) (97.5) (94.0) (98.6) (88.0) (99.4) (93.2) (81.6) (74.8) (83.7) (58.7) (93.8) (91.5)

Number o f Persons Able to Perform With Sams Help (% )

6 12 10 29 7 48 2 29 76 85 52 149 26 33

(1.2) (2.5) (2.1) (6.0) (1.4) (9.9) (.4) (6.0) (15.7) (17.6) (10.7) (30.8) (5.4) (6.8)

Number of Perstms Un/ible to 1)0 This (% )

1 (.2) 1 (.2) 2 (.4) — — 9 (1.9) —

4 9 33 25 48 3 8

(.8) (1.9) (6.8) (5.2) (9.9) (.6) (1-7)

Note. T he items from the OARS come from Multidimensumal b'unctitrrud Assessment o f Older Adults (pp. 143-145) by G. G. Fillenbaum, 1988, Hillsdale, N|: Lawrence Erlbaum Associates.

Copyright © 1975 by the Duke University Center for the Study of Aging and Development. Adapted with permission.

= totally im paired) based on the subject’s responses to the various items. Frequencies o f the responses on individual items o f caregivers rep o rtin g on people who obtained scores o f 78 o r greater on the 3MS (Teng & Chui, 1987) appear on T able 11.2. Inform ation on the responses o f caregivers o f persons with dem entia ap p ear on T able 11.3. Frequencies o f overall ADL ratings obtained in die CSHA study appear on Table 11.4. Earlier in this chapter, findings concerning patient-proxy ag reem en t in symptom reporting were discussed. Some research o f this kind focused specifically on ADL/IADL. R ubenstein, Schairer, Wieland, and Kane (1984) fo und that scores on ADL and IADL scales may vary d ep en d in g o n who provides the inform ation about the perso n ’s functioning. These investiga­ tors adm inistered (to a group o f hospitalized patients overage 65) the Index o f ADL (Katz ct al., 1970), the PSMS (Lawton & Brody, 1969), an d the IADL scale developed at the Philadelphia G eriatric C entre (see Rubenstein et al., 1984). T he investigators concluded that inform ation sources ab o u t function­ ing can n o t be used interchangeably. This is consistent with the findings of Z im m erm an and M agaziner (1994). A lthough proxies’ and self reports of functioning can often be similar, the similarities may n o t be strong enough

TABLE 11.3 Canadian Study on Health and Aging Norms on the OARS ADL/IADL Scale Based on Caregivers of Persons With Dementia (« = 353) Number of Persons Able to Perform With No Help (% )

Items Can subject eat? Can subject dress and undress? Can subjcct take care of appearance? Can subject walk? Can subjcct get in and out of bed? Can subject take a bath or shower? Can subject use the bathroom or toilet? Can subject use the telephone? Can subject get to distant places? Can subject go shopping? Can subject prepare own meals? Can subject do housework? Can subject take own medicine? Can subject manage own money?

308 243 265 274 312 160 294 180 84 73 90 58 129 82

(87.3) (68.8) (75.1) (77.6) (88.4) (45.3) (83.3) (51.0) (23.8) (20.7) (25.5) (16.4) (36.5) (23.2)

Number of Persons Able to Perform With Some Help (%) 37 73 53 55 20 120 31 39 167 110 105 116 115 89

(10.5) (20.7) (15.0) (15.6) (5.7) (34.0) (8.8) (26.3) (47.3) (31.2) (29.7) (32.6) (32.6) (25.2)

Number of Persons Unable to Do This (% ) 8 36 35 24 21 72 27 74 95 164 151 170 97 176

(2.3) (10.2) (9.9) (6.8) (5.9) (20.4) (7.6) (21.0) (26.9) (46.5) (42.8) (48.2) (27.5) (49.9)

Note. The items from the OARS come from Multidimensional Functional Assessment o f Older AduUs (pp. 143-145) by G. G. Fillenbaum, 1988, Hillsdale, NJ: Lawrence Erlbaum Associates. Copyright © 1975 by the Duke University Center for the Study of Aging and Development. Adapted with permission.

TABLE 11.4 Frequencies of Overall ADL Ratings Based on the Responses of Caregivers of Persons With Dementia and Caregivers of Persons Who Obtained Greater than 77 on die 3MS (Canadian Study on Health and Aging)

Level of Impairment No impairment (2) Mild Im pairm ent (3) Moderate Im pairment (4) Severe Im pairment (5) Total Im pairment (6)

Caregivers of Persons With Dementia (% )

Caregivers of Persons With score >77 on 3MS (% )

5.6 13.0 26.9 21.5 32.9

32.5 20.8 20.7 11.0 14.9

213

214

CHAPTER 11

to consider the two types o f reports equivalent. Generally, self-reported scores tended to be indicative o f higher functioning than scores derived from proxies. Zim m erm an and M agaziner speculated on the reasons for this discrepancy. Specifically, it may be due to denial on the p art o f the individual o r to the b u rdened caregiver’s tendency to underestim ate the activities th at the individuals are capable o f perform ing on their own. In the case o f hospitalized patients, the hospitalized status could function as a stereotype th a t could lead caregivers to underestim ate the patients’ capabilities. W einberger et al. (1992) com pared self and proxy perceptions using a m odified version o f Index o f ADL and Instrum ental Activities o f Daily Living in a sample o f medical patients seen in a geriatric evaluation and treatm en t clinic. They concluded that p atient and proxy ratings were con­ co rd an t w hen it cam e to ability to perform ADL tasks. W'ith respect to IADL, concordancc was high for patients whose MMSE scorc was 24 or h ig h er an d relatively po o r for patients with MMSE scores lower than 24. W hen disagreem ent occurred, proxies rated the patients as m ore im paired th an the patients rated themselves. T hese findings seem to suggest that, in m any instances, patients may overrate their ability to carry o u t instru­ m ental activities o f daily living. O f the instrum ents reviewed here, the PFQ (Crockett et al., 1989) has the advantage o f covering several areas o f patient functioning (i.e., cogni­ tive, psychological, ADL, and IADL) and norm s for norm al individuals are available. In addition, satisfactory reliability and validity evidence exists on this instrum ent. T h e IQCODE could also prove to be a very useful index o f cognitive decline, as is suggested by the extensive available validity and reliability evidence (Jorm , 1996). T h e interested read er is referred to jo rm (1996) for a m ore detailed discussion o f inform ant instrum ents th at could be used for the assessment o f cognitive decline. W hen m ore elaborate inform ation on ADL and IADL is needed, o th e r m ore focused scales (e.g., the OARS; Fillenbaum , 1988) may prove useful. N onetheless, th ere is a n eed for additional norm ative a n d /o r validity inform ation for ADL and IADL measures.

ASSESSING THE PSYCHOLOGICAL CONSEQUENCES OF CAREGIVING C aring for a person with dem entia is a very dem anding task with potentially serious psychological consequences for the caregiver. In a 2-year longitu­ dinal study o f A lzheim er’s caregivers, Schultz and Williamson (1991) found th at levels o f depression am ong continuous caregivers were m ore serious th an they were for noncaregivers. F urtherm ore, fem ale caregivers were significantly m ore distressed than m ale caregivers. N onetheless, m ale

215

THE ROLE OF THE CAREGIVER

caregivers showed significant increases in depression symptoms over time. In another study of psychological distress am ong caregivers of dem entia patients, Anthony-Bergstone, S. H. Zarit, and Gatz (1988) adm inistered the Brief Symptom Inventory. Com pared to age-matched norms, caregivers were elevated on the hostility subscale. Both younger and older women scored higher than the norm s on the anxiety subscale,and older women were also elevated significantly on three other subscales. Means derived from this study are presented in Table 11.5. The standard deviations were no t reported. S. H. Zarit (1990) wrote that caregiver outcomes such as decreased well-being are determ ined by an interplay of factors,including patient deficits, appraisals of strain, social support, oth er stressors in the caregivers’ lives, appraisals of coping by carcgivers, and contextual factors such as the relationship of the caregiver with a patient. Some support for Zarit’s assertions exists. C antor (1983) found, for instance, that caregiver burden varies depending on the nature of the relationship of the caregiver with the person with dem entia (e.g., child vs. parent vs. friend). Hadjistavropoulos et al. (1994) found that caregiver burden was affected substantially by the caregiver’s perception of the care recipient’s symptoms and only indirectly by actual symptoms. In addition, caregiver perceptions o f the degree to which the care recipient’s mood was dysphoric was m ore likely to contribute to caregiver burden than perceptions concerning other areas of functioning (e.g., memory and language skills). Such findings suggest that when wellTABLE 11.5 Brief Symptom Inventory Subscale Scores o f Caregivers

Suhsrale

Somatization Obsessive-Compulsive Interpersonal Sensitivity Depression Anxiety Hostility Phobic Anxiety Paranoid Ideation Psychoticism

OUler Women"

Older Men''

Younger Wommr

Younger Men'1

.70 1.20 .61 .92 1.02 .90 .30 .53 .51

.44 .87 .40 .44 .53 .62 .19 .29 .24

.43 .88 .42 .61 .78 .72 .19 .44 .31

.21 .75 .50 .53 .74 .85 .13 .72 .32

Note. From “Symptoms of Psychological Distress Among Caregivers of Dementia Patients," by C. R. Anthony-Rergstone, S. II. Zarit, and M. Galz, 1988, Psychology and Aging, 3, pp. 246-247. Copyright © 1988 by the American Psychological Association. Adapted with permission. an = 77; age 60 or older. hn = 47; age 60 or older. 'n = 47; age 59 or younger. rtn = 13; age 59 or younger.

216

CHAPTER 11

being of caregivers is assessed, it is im portant for the clinician to obtain inform ation on as many o f the dimensions discussed by Zarit as possible (i.e., social supports, nature o f the caregiver-patient relationship, etc.). Caregiver burden has been conceptualized within a stress an d coping framework (S. H. Zarit, 1990). Two processes (coping and cognitive ap­ praisal) have been identified as im portant mediators between a stressful situation such as caregiving and the psychological consequences for the individual (Lazarus & Folkman, 1984). Through appraisal, individuals evalu­ ate w hether they have anything at stake in any particular encounter and w hether anything can be done to prevent or overcome harm or improve the chances for b en efit Coping refers to cognitive or behavioral efforts to m anage dem ands that are appraised as taxing or as exceeding the person’s resources (Lazarus & Folkman, 1984). Secondary appraisal refers to how well people evaluate the adequacy of their resources for coping with threats posed by stressors (Lazarus & Folkman, 1984). The relevant primary stressors arc the stressors or dem ands that result from the eld er’s illness and disability. S. H. Zarit (1990) pointed out that these may have both objective (i.e., specific and necessary care tasks) and subjective com­ ponents (i.e., the degree to which a caregiver feels distress as a result of the disability). Secondary stressors refer to the consequences o f providing care in other areas of the caregiver’s life. Primary and secondary stressors are believed to make contributions to caregiving outcomes such as de­ creased well-being. Furtherm ore, their effects are believed to be m oderated by mediators such as coping, social support, and appraisals (S. H. Zarit, 1990). Such dimensions could also be evaluated during the interview with the caregiver. For the assessment o f the psychological consequences o f caregiving, a brief interview with focused screening questions (e.g., How do you cope with having to care for your relative? Do you manage okay?) is recom ­ m ended. Establishing adequate rapport with the caregiver is im portant in o rder to get sincere answers to such questions. Caregiver functioning can be assessed m ore systematically through the adm inistration of general measures of psychological functioning (e.g., the Brief Symptom Inventory). Caregiver norm s on some psychopathology scales are presented in Tables 11.5 and 11.6. Specifically, Table 11.5 presents norms on the Brief Symptom Inventory based on samples of caregivers of persons with dem entia. All were living in the community. The data are broken down by caregiver sex and age. Table 11.6 presents data on the (Center for Epidemiologic Studies Depression scalc. These data were collected through the Canadian Study on H ealth and Aging. These instruments are discussed in m ore detail in chapi­ ter 9. In some instances, the clinician may be interested in com paring the caregiver’s functioning with that of persons in the general population. G eneral population norms would have to be used for that purpose. In

217

THE ROLE OF THE CAREGIVER TABLE 11.6 C en ter o f E pidem iologic Studies D epression Scale Scores O b ta in e d by Caregivers o f P eople W ith D em entia (C anadian Study o n H ealth an d Aging) Age 65-69 70-74 75-79 80-84 85*

Gender

n

M

SD

M F M F M F M F M F

30 77 19 67 14 51 16 30 9 6

6.00 8.61 7.32 8.18 11.29 10.67 12.75 10.53 6.33 15.50

9.13 9.01 8.31 9.32 12.66 9.59 8.18 8.51 5.59 8.12

ch ap ter 9, population norm s for elderly persons are p resented for several instrum ents developed to assess psychological symptoms. In addition to interview an d general m easures o f functioning, specialized instrum ents that assess caregiver b urden exist. Because the goal o f such instrum ents is to determ ine the degree to which a caregiver is b u rd en ed relative to o th e r caregivers o f dem entia victims, it seems that appropriate norm ative groups would be caregivers o f dem entia victims o r caregivers o f p eople with suspected dem entia. Vitaliano, Young, an d Russo (1991) published a m ajor review o f a variety o f instrum ents developed specifically for the assessment o f caregiver b u r­ d en (e.g., G reene, R. Smith, G ardiner, & Timbury, 1982; Kinney 8c ParrisStephens, 1989; Lawton, Kleban, Moss, Rovine, & Glicksman, 1989; R. J. V. M ontgom ery, Gonyea, & H ooym an, 1985; Novak 8c Guest, 1989; Pearlin 8c Schooler, 1978; Poulshock & D eim ling, 1984; Rabins, Mace, & Lucas, 1982; Vitaliano, Russo, Young, Becker, & Maiuro, 1991; S. H. Zarit, Reever, 8c Bach-Petersen, 1980). Some instrum ents separately assess objective care­ giver experiences as well as th e caregiver’s own response to such experi­ ences. O th er instrum ents confound subjective with objective dim ensions. A lthough this section focuses on obtaining inform ation ab o u t caregiver functioning, some o f the instrum ents developed for that purpose (e.g., Poulshock & Deimling, 1984) also include questions th a t relate to the care recip ien t’s functioning in areas such as activities o f daily living (ADL) and thus, overlap with som e of the ADL instrum ents reviewed in the previous section. T he Burden Interview (BI; S. H. Zarit & J. M. Zarit, 1982; W hitlatch, S. H. Zarit, & von Eye, 1991) is probably the m ost widely used in stru m en t in this area. This 22-item scale (S. H. Zarit, O rr, & J. M. Zarit, 1985) assesses

218

CHAPTER 11

the extent to which caregivers view their responsibilities as having an ad­ verse impact on their social life, health, em otional well-being, and finances. The caregivers respond to the items (e.g., Do you feel stressed between caring for your relative and trying to m eet other responsibilities for your family or work?) along 5-point scales anchored by the polar opposites “Never” and “Nearly always.” The scale has satisfactory reliability and validity (Whitlatch et al., 1991). The BI (S. H. Zarit & J. M. Zarit, 1990) was adm inistered to caregivers in the Canadian Study on H ealth and Aging. Relevant norm s are presented on Table 11.7. In terms of stress and coping conceptualizations, the scale taps appraisal of secondary stressors and sec­ ondary appraisal (S. H. Zarit, 1990). Lawton et al. (1989) proposed five dimensions of burden based on a factor analysis o f items they drew from a variety of scales. They proposed that caregiver appraisals of the caregiving process arc multifaceted and include caregiver burden. The following dimensions were derived: impact of caregiving burden (analogous to secondary appraisal), impact of caregiv­ ing (analogous to secondary stressors), caregiving mastery, caregiving sat­ isfaction, and cognitive reappraisal (a coping strategy) (S. H. Zarit, 1990). Although Lawton et al. derived some support for these dimensions, the appraisal dimensions were not conclusively confirmed. Additional research on this instrum ent is needed to determ ine its clinical utility. However, the main contribution of Lawton and colleagues was the expansion of the m easurem ent of the caregiving experience to include several dimensions of appraisal. Poulshock and Deimling’s (1984) approach involves indices o f burden th at correspond to dependency and mental impairment. Dependency refers to activities of daily living. Specifically, for each ADL (bathing, dressing, toileting, mobility, incontinence, and eating), caregivers indicate along a TABLE 11.7 B urden Interview (22-item version; S. H. Zarit & J. M. Zarit, 1990) Scores o f Caregivers o f Patients With Dem entia (C anadian Study o f H ealth and Aging) Agt 66-69 70-74 76-79 80-84 85+

Gentler

n

M

SD

M F M F M F M F M F

28 78 19 65 14 52 17 29 9 4

13.29 18.45 14.16 17.17 16.36 17.59 18.35 15.14 16.11 35.25

10.08 15.42 12.78 17.89 16.29 15.17 14.64 11.37 12.87 19.45

THE ROLE OF THE CAREGIVER

219

0 to 3 scale w hether they find the assistance tiring, difficult, o r upsetting. For m ental im pairm ent, they indicate along a 0 to 3 scale how upset they feel as a result of the patien t’s level sociability (8 item s), disruptive behavior (7 item s), and cognitive incapacity (8 items). To m easure caregiving im­ pact, items similar to Zarit’s B urden Interview are used. Im pact is op era­ tionalized in terms o f both negative changes in the eld er/careg iv er family relationships (11 items) an d in term s o f restrictions in the caregiver’s activities (8 items). Reliability and construct validity evidence is also pre­ sented. A strength o f this approach is that the caregiving experience can he m easured in relation to specific symptoms. A nother interesting approach was discussed by Novak an d G uest (1989). They developed the Caregiver Burden Inventory (CBI) that involves five factor analytically derived scales (Tim e-D cpcndence Burden, D evelopm ental Bur­ den, Physical B urden, Social Burden, and Em otional B urden). T h e CBI consists o f 24 items. Adjusted raw scores derived from these scales can be used to plot caregiver burden profiles that graphically rep resen t the degree to which each o f the five areas m easured is problem atic relative to the rest. M ore evidence on the construct validity o f this m easure is needed. G reene et al. (1982) described the developm ent o f two scales: a Behavior and Memory Disturbance Scale (BMDS) and a Relative’s Stress Scale (RSS). T he first scale contains 34 items tapping behavioral problem s that the care recipients may be displaying (e.g., talk aloud to themselves) and the latter consists of 15 items and relates to negative feelings that caregiving may lead the caregiver to experience (e.g., Do you ever get cross and angry with . . . ? ) . T he responses are given along 5-point scales corresponding to each item. The scales were factor analyzed using a very small sample (N = 39) and, thus, the factor analytically derived subscales (within each o f the two m ain scales) may reflect an unstable solution. A lthough both objective and subjective aspects o f burden are tapped by the scales, there is no d irect correspondence between the items of the BMDS scale tapping objective burden and the RSS tapping subjective burden. Some validity evidence was presented but m ore research is needed to d eterm in e the utility o f this scale. Kinney and Parris-Stephens (1989) developed a Caregiver Hassles Scale. This 42-item instrum ent differs from others reviewed h ere in th at it focuses on the day-to-day experience o f caregiving rath er than caregiving events o r responsibilities occurring over longer periods of time. T h e five scales o f the instrum ent are Basic Activities o f Daily Living, Instrum ental Activities o f Daily Living, Cognitive Status, Patient Behavior, an d the Caregiver’s Social Network. The caregiver indicates w hether o r no t each event occurred over the past week and then every occurrence is rated with respect to w h ether or not it represented a hassle along a 4-point scale. C onstruct validity and some reliability evidence was also presented.

220

CHAPTER 11

R. J. V. M ontgom ery et al. (1985) developed two measures; the first consists o f nine items designed to assess objective burden. T h e questions assess consequences o f caregiving on the caregiver’s life (e.g., am o u n t of privacy at hom e, am ount o f time you have to yourself) and are scored along 5-point scales anchored by the polar opposites “a lot m o re” and “a lot less.” T he second scale consists o f 13 items and aims to assess subjective burden (e.g., “I feel it is painful to watch my [relative] age”; “I feel useful in my relationship with my relative”). Some reliability and validity evidence was presented. A criticism o f this approach is that the item s in the two scales do n o t show a conceptual correspondence. Only one o f the instrum ents (Screen for Caregiver Burden, Vitaliano et al., 1991) reviewed by Vitaliano, Young, and Rousso (1991) m et all th ree o f the following criteria: exam ined prevalence o f caregiver dem ands (Objec­ tive B urden), exam ined caregiver response to these specific dem ands (Sub­ jective B urden), and has been studied with respect to criterion validation as well as sensitivity to change. This m easure com bines features o f prim ary appraisal an d subjective evaluation o f secondary stressors. All items in the scale are included in Vitaliano, Russo, et al. (1991). Some norm ative in­ form ation on the subjective burden portion o f the scale was given by Vitaliano c t al. (1991). T he sam ple from which these norm s were developed consisted o f 79 spouse caregivers who were living with a person with pos­ sib le/p ro b ab le Primary D egenerative Dem entia. T he norm ative inform a­ tion is presented in quartiles. Caregivers o f less than 65 years o f age with a score higher than 42.2 would be at th e top 25%, w hereas they would fall at the bottom 25% if their score were less than 31. T he m eans and stan dard deviations repo rted by Vitaliano and colleagues were 9.01 (SD = 4.30) for objective burden and 35.54 (SD = 8.02) for subjective b u rd en (average caregiver age = 67.4 years). Vitaliano et al. p resen ted data col­ lected at two different points in time. T he data for people over age 65 collected at Tim e 2 appeared som ew hat elevated com pared to those at Tim e 1. T he inform ation presented here is based on Tim e 1 data. At this p o in t there does n o t seem to be a consensus as to w hat is the best way to m easure caregiver burden. M ore research is n eed ed to d eter­ m ine the relative utility o f global versus specific measures. Based on the previous review and keeping in m ind the lim itations discussed, clinicians may select the m easure that taps best the area they are in terested in as­ sessing. T h e BI is probably the scale on which the greatest am o u n t o f inform ation is available. O th e r scales, however, tap inform ation n o t as­ sessed by the BI. F urtherm ore, in com paring the data to norm s, it is im p o rtan t to note the com position o f the caregiver norm ative groups be­ cause caregiver’s response to b urden could be affected by factors such as dem ographic characteristics.

THE ROLE OF THE CAREGIVER

221

CONCLUSIONS In m ost circumstances, a knowledgeable proxy should be interviewed in o rd er to provide inform ation about the care recipien t’s symptomatology. Standardized m easures developed for such a purpose could supplem ent the clinical interview. T h e interview could also include some screening questions regarding the degree to which the caregiver/proxy is b u rd en ed by the p atien t’s condition. Knowledge about caregiver b u rd en is im p o rtan t n o t only because it may lead the caregiver to experience symptoms o f psychological distress but also because caregiver burden could potentially interfere with effective p atient care. Many clinicians may wish to pursue with the caregiver the possibility o f an assessment o f bu rd en an d caregiver functions through the use of standardized assessment tools and m ore elabo­ rate interviewing. T he purpose o f the assessment should d eterm in e the m easure of choice.

Chapter

Ethics in the Assessment of Elderly Persons

Most ethical issues pertaining to the assessment of elderly persons are com mon to the assessment of all populations. For the purposes o f this discussion, several references are made to psychologists and codes of ethics developed by psychologists. However, the principles articulated here are relevant to the practices of many types of health professionals who work with elders. It is recom m ended that all practitioners using psychological and cognitive tests become familiar with the Standards for Educational and Psychological Testing (American Psychological Association, 1985), the General Guidelines for Providers of Psychological Services (American Psychological Association, 1987) a n d /o r the Practice Guidelines for Providers o f Psychological Services (Canadian Psychological Association, 1989), Specialty Guidelines for the Delivery of Services by Clinical Psychologists (American Psychological Association, 1981), as well as with the Ethical Principles and Code of Conduct (American Psychological Association, 1992) and the Canadian Code of Ethics for Psychologists (Canadian Psychological Association, 1991). Additional ethical guidelines and standards of practice are available for physicians (e.g., Council on Ethical and Judicial Affairs, 1994), social workers (e.g., National Association of Social Workers, 1993), and other health professionals. Several im portant general issues need to be considered when conducting any psychological assessment (e.g., Keith-Spiegel & Koocher, 1985). First, to be com petent, practitioners must have received specialized training in the type of population with whom they work. In the case o f geriatric neuropsychologists, extensive supervision and training in both general and geriatric neuropsychology is required. Second, practitioners should be fa­ 222

ETHICS

223

miliar with m easurem ent theory and issues of test reliability and validity. This allows for an evaluation of the adequacy o f the psychometric instru­ m ents they employ. They should also be familiar with the factors that could bias the results of psychological testing. Such factors were discussed in chapters 1 and 2 and include cultural and language issues. In addition, the content of psychological tests must be kept secure because test validity could be com prom ised if test security is violated and members of the public become familiar with the content of psychological tests. Test validity is affected when individuals become exposed to the items of a test before they arc adm inistered that test A nother im portant concern is the use of com puterized scoring and interpretation services as discussed by Matarazzo (1986). Any reports derived from com puterized scoring systems should be viewed as means of generating hypotheses about the person’s deficits. They should not be accepted at face value because they may be of limited validity. The com puter program may fail to present more than one hypotheses for the clinician’s consideration and to integrate interpretations with background history (Matarazzo, 1986). Com puterized test interpretations arc often simulations of clinicians’ deci­ sion-making rules and are subject to erro r (Matarazzo, 1986). Thus, they should be checked against other sources of data (e.g., other psychological tests, background inform ation). It is imperative that com puterized packages provide sufficient evidence for the validity of their interpretations and the reports they generate. Such reports should never be used by individuals who do n ot have good knowledge of psychological assessment procedures and the limitations of such procedures. Keith-Spiegel and Koocher (1985) stressed the im portance of inform ed consent and the client’s right to know the purpose of the assessment and the potential use of assessment results. These issues are of particular im­ portance in geriatric assessment because the capacity to provide inform ed consent may be com prom ised by dem enting illness and, in some institu­ tional a n d /o r family contexts, the voluntariness of inform ed consent may also be compromised. The issue of inform ed consent is especially salient when it comes to the assessment of elderly persons and is addressed in more detail later. It is recom m ended that test scores be retained in a client’s file only as long as they serve a valid and useful purpose (Keith-Spicgcl & Koocher, 1985). In the cognitive assessment of elderly persons, test results are of special im portance with respect to potential future use because they provide a baseline for determ ining the presence or absence of cognitive deterioration. Questions about third-party access to psychological test data and its effects can often arise. In the case of elderly persons with dem entia, there may be the complicating factor of family members wishing access to such data. W hen a client is cognitively impaired, a legal guardian may be entitled

224

CHAPTER 12

to legal access (Keith-Spiegel & K oocher, 1985). Keith-Spiegel an d K oocher recom m ended that from the beginning o f the working relationship, parties should be inform ed o f the limits of confidentiality an d an early discussion o f the types o f inform ation to be shared should take place. O th er questions p ertaining to third-party access apply to m ost assessment situations and are discussed in detail elsew here (see Keith-Spiegel & K oocher, 1985; Ogloff, 1995). Geriatric neuropsychologists, by virtue o f their role, will be dealing with a disproportionate n um ber of elderly persons who suffer cognitive im pair­ m ents an d are, therefore, vulnerable. Thus, it is im perative th at they be aware o f the ethical issues involved in the assessment o f such ciders. Al­ though the n u m b er o f ethical issues that may arise is potentially high, two arc discussed in m ore detail because they are o f special im portance in geriatric assessments. Specifically, issues pertaining to inform ed consent and assessment feedback are exam ined.

INFORMED CONSENT Obtaining Consent A substantial portion o f elders are victims o f dem entia (Canadian Study o f H ealth an d Aging W ork G roup, 1994), so consent concerning health care decisions involving dem entia victims is often obtained from proxy decision m akers (e.g., family m em bers). Rozovsky (1990) wrote that th ree elem ents m ust be present for consent to be legally and ethically valid: voluntariness, mental capacity, and adequate information. In the case o f cogni­ tively intact elderly persons, addressing the issue of com petence is n o t terribly com plicated. W hen, on the o th e r hand, there is evidence that m ental capacity is significantly com prom ised, consent for the assessment m ust be sought from an appropriate proxy (e.g., legal guardian). R ecent research evidence suggests that practitioners may have to be m ore conservative in their determ inations o f com petence to provide con­ sent. Abramovitch, Finstad, and Silberfeld (1993) conducted a study to investigate the issue o f inform ed consent for m ental capacity assessments. T hey concluded that it was the assessor’s im pression th at everyone who p articipated in the capacity assessment provided inform ed consent at least to som e degree. However, at the en d o f the assessment, it was fo und th at eig h t clients (47% o f the total group) were found to have only a general sense o f being tested o r showed no understanding at all for the reason of the assessment. Four o f these clients showed no understanding o f why they h ad com e to the clinic. Abramovitch et al. recom m ended th a t m ore at­ tention be paid when it comes to the determ ination of capacity to give

ETHICS

225

consent and suggested an informal assessment of that capacity with some interview questions. Clients could be asked with appropriate probes about why they came to the clinic and about their expectations. Tymchuk and O uslander (1990) recom m ended an initial basic assess­ m ent of the person’s ability to hear and see (i.e., can the person see, discriminate, and label words?) followed by a brief assessment of mental capacity to determ ine w hether there is a need for proxy consent. They also supported assessment of reading capacity, presentation of consent m aterial developed in a format and at difficulty level to match the person’s ability, assessment of com prehension of consent material, and a follow-up to determ ine w hether com prehension of the inform ation remains at cri­ terion level. These suggestions imply that a brief screening tool such as the Mini-Mental State exam (M. L. Folstein et al., 1975) com bined with a test such as the Wide Range Achievement Test-Revised (Jastak & Wilkin­ son, 1984) could be used as quick assessment tools for capacity to consent in the case of some individuals (chapter 3 includes additional discussion o f com petency assessments). Although a problem with this suggestion is that persons would be given psychological tests in order to assess their capacity to agree to be given more psychological tests, the procedure is probably justified on ethical grounds because consent could be potentially obtained for a variety of oth er purposes (e.g., im plem entation of assessment recom m endations, release of records). Perhaps, in the future, a videotaped form at could be developed with information that could facilitate the con­ sent process— especially in individuals of relatively lower educational at­ tainm ent (Tymchuk & Ouslander, 1990). Although proxy consent is often obtained when working with the elderly person, assent should be sought in all instances. U nder no circumstances may a person be forced to participate in neuropsychological testing or research. It has been suggested that a professional activity may proceed without assent only if the service is of direct benefit to the person of diminished capacity (Canadian Psychological Association, 1991). Lack of assent and cooperation in neuropsychological assessment would typically jeopardize the validity of psychological test results and testing without assent is unlikely to be significantly beneficial to the person. In some instances, it may benefit the person to derive assessment inform ation from a caregiver even if the individual refuses to participate in neuropsychologi­ cal testing. In most cases, obtaining assent is not difficult. In a supportive environm ent, most people are cooperative with the assessment process. A lthough at first glance the process of proxy consent may appear straightforward, Kapp (1991) discussed several caveats pertaining to proxy consent. First, elderly persons must be protected against family coercion masquerading in the guise of shared decision making. Furtherm ore, po­ tential conflicts of interest between the person and the family may occur

226

CHAPTER 12

an d service providers should be alert to these issues. For exam ple, the family that refuses consent for routine beneficial care a n d /o r assessment in o rd er to save m oney and conserve future estate is n o t acting in the p erso n ’s best interest. In extrem e cases, it may be necessary to seek form al legal protection for the person. Limits o f Confidentiality In general, practitioners m ust inform their clients o f the limits to confi­ dentiality before any assessment is u ndertaken an d inform ed consent is obtained (American Psychological Association, 1992). W here the individ­ u a l’s m ental abilities are com prom ised, it would be appropriate to inform th eir legal guardian o f these limits. These limits may vary from jurisdiction to jurisdiction but generally involve situations such as suspicion o f child abuse, danger to a person (and, in some cases, risk o f substantial dam age to property), and co urt o rd er (see Ogloff, 1995). Some jurisdictions have ad opted legislation that makes the reporting of eld er abuse m andatory (Ogloff, 1995). Psychologists m ust be well aware o f the relevant laws o f their jurisdiction an d be p repared to act accordingly. In all instances, appropriate support should be offered to the person a n d /o r the family. A com m on situation in the assessment of elders involves suspicion th at the individual is unable to drive. Many jurisdictions require h ealth profes­ sionals to inform the authorities if a person continues to drive despite im paired capacity. T he act o f driving is very im portant for many elders and rem oving their privilege to drive may be tantam oun t to rem oving their ind ependence. Thus, it is very im portant for the geriatric neuropsychologist to be sensitive to these types o f issues and support individuals in their attem p t to ad ap t to a new lifestyle through consultation a n d /o r appropriate referral.

GIVING FEEDBACK T he role o f the neuropsychologist has been redefined to encom pass, in addition to assessment, a m ore active involvement with the client and treatm en t team o f health professionals (Gass & Brown, 1992). As part of such involvement, neuropsychologists provide assessment feedback to the family, the client, and o th e r professionals. From an ethical p o in t o f view, it is im p o rtan t to recognize the perso n ’s right to self-determ ination and, as such, feedback to the person m ust be a concern to the ncuropsychologist. Clinically, feedback can be a vehicle for providing persons with decision­ m aking guidelines. For exam ple, provincial and state laws outline specific rules with respect to com petence to drive. Feedback pertain in g to com pe­ tence in this and o th e r areas could be discussed with the patient. Involve­

ETHICS

227

m ent of family members is necessary when the person’s ability to participate in the feedback process is compromised. During the process of feedback, the neuropsychologist can provide the person and family with support as the realization of the impact of neurological conditions often has negative em otional consequences. In addition, during the provision of feedback, recom m endations can be m ade that could benefit the person and the family. Gass and Brown (1992) provided some guidelines for the provision of assessment feedback to persons with brain dysfunction. Specifically, they recognized that no single approach to feedback can be used with all individuals and suggested the following steps: review the purpose of testing (this is especially im portant when it comes to persons with memory prob­ lems) , define the tests as behavior samples assessing behavioral skills related to the functional integrity of the brain, explain test results and behavior in a m anner the person can understand, describe strengths and weaknesses, address diagnostic and prognostic issues (it is often prudent to defer the diagnosis to a physician who can assess the psychologist’s report in relation to medical diagnostic data), and make recommendations. Pope (1992) discussed the im portance of ensuring that the client understands the limitations of psychological assessment procedures. Gass and Brown (1992) also stressed that the way the feedback is provided is important. Specifically, potentially stigmatizing terminology (e.g., dem ented, retarded) should be avoided and every effort should be m ade for feedback to be given in a way that is understandable to the clients. The em otional needs of the person may affect the timing of the feedback and the way this is given. Gass and Brown (1992) suggested, for instance, that some persons may require a supportive approach in which their relative strengths are stressed. Although denial on the part of the person and the family is a com mon initial response, and some degree of denial may be psychologically adaptive, the occurrence of pro­ longed denial may cause m ajor problems (Gass & Brown, 1992). Such denial could prevent the person and the family from preparing for the future and for im plem enting treatm ent recommendations. Given such potential diffi­ culties, the role of the geriatric neuropsychologist in the feedback process is delicate and professionals must consider carefully, and with a high degree of sensitivity, the issus pertaining to neuropsychological assessment feedback o f each individual person.

NEUROPSYCHOLOGICAL RESEARCH WITH GERIATRIC POPULATIONS T he ethical concerns relating to inform ed consent and vulnerability be­ come especially potent when engaging in testing for research as opposed to clinical purposes. Although testing for clinical purposes is intended to

228

CHAPTER 12

directly benefit the person, this is n o t necessarily the case w hen it comes to research. Typically, in research studies, w ritten consent m ust be ob­ tained. A lthough consent form s are m eant to be w ritten in lay language, they may be technical and difficult to understand w ithout a high level o f education a n d /o r technical background. In addition, m any elderly persons have visual, auditory, a n d /o r o th e r im pairm ents that could interfere with th eir ability to read consent forms. A daptations for elderly persons may involve the use o f large p rin t and easy-to-understand language (Sachs, Rhymes, & Cassel, 1994). W here the person is unable to provide consent, it would be appropriate to obtain proxy consent. But, testing for the purposes o f research should n o t be attem pted w ithout assent o f the participant. If the subject— no m atter how im paired—objects to being in the study eith er verbally o r behaviorally, then participation should be discontinued (Sachs et al., 1994). It has been suggested that persons with dim inished capabilities should n o t be sought as research participants unless the research questions can n o t be answered with an o th er group o f individuals who are m ore able to co n sent (C anadian Psychological Association, 1991). Sachs et al. (1994) en d orsed the concept o f advance directives for research with elders. Spe­ cifically, they suggested that institutionalized elderly persons o r those with­ o u t dem entia may be given the opportunity to instruct others o n their wishes regarding participation in research if they ever lose th eir decision­ m aking capability.

ETHICS CODES AND THE RESOLUTION OF ETHICAL DILEMMAS T h e codes o f ethics o f the A m erican Psychological Association (APA; Am eri­ can Psychological Association, 1992) an d C anadian Psychological Associa­ tion (CPA; C anadian Psychological Association, 1991) recognize the special care th a t should be taken in the assessment o f vulnerable persons. T he ethical principles o f psychologists published by the APA are listed in Table 12 . 1.

Principles A to E apply directly to the types o f dilem m as en co u n tered by the geriatric neuropsychologist and Principle F ( Social Responsibility) applies indirectly. First, the practitioner m ust consider issues pertaining to Competence. Are practitioners qualified to draw the conclusions they have reached both with respect to the perso n ’s cognitive function an d psycho­ logical state? Have they assessed the person adequately? T h e Principle o f Integrity is also im p o rtan t in so far as health professionals m ust consider the n eed to provide services with honesty and respect toward th eir clients. T h e principle o f Professional a n d Scientific Responsibility m ust be considered

229

ETHICS TABLE 12.1 Ethical Principles American Psychological Association (1992) Principle Principle Principle Principle Principle Principle

A: Competence B: Integrity C: Professional and Scientific Responsibility D: Respect for People’s Rights and Dignity E: Concern for O ther's Welfare F: Social Responsibility

Canadian Psychohtgictd Asstn:uUion (1991) Principle Principle Principle Principle

I: Respect for the Dignity of Persons II: Responsible Caring III: Integrity in Relationships IV: Responsibility to Society

as well. Specifically, a decision should be made concerning collaboration with other professionals (e.g., referrals and consulations). Moreover, prac­ titioners should be prepared to accept responsibility for any professional decision they make. Appropriate respect for the person’s dignity, as speci­ fied in Respect fo r People’s Rights a n d Dignity , must also be shown. Clinicians m ust consider the welfare of the client and the consequences that any action they m ight take could have for the client ( Concern fo r Others’ Welfare). Finally, clinicians should work for the benefit of their community and show respect for the law (Social Responsibility). In addition to its ethical principles, APA’s code oudines a variety of standards that must be m aintained by psychologists. The section o f stand­ ards referring specifically to assessment includes a discussion of the im­ portance o f a professional context, com petence, and test security. The appropriate construction, use, and interpretation of tests is also discussed. The code specifically stresses in the section “Use of Assessment with Special Populations” that psychologists must identify situations in which particular interventions or assessment techniques or norms may n o t be applicable becausc of factors such as the individual’s ethnicity, language, and other dem ographic characteristics. In addition, the im portance o f providing feed­ back that is reasonably understandable to the person a n d /o r guardian is also stressed. The Canadian code comprises four ethical principles that arc hierar­ chically listed in order of im portance (see Table 12.1). These principles apply in a m anner similar to those of the APA code. Specifically, the three most relevant CPA principles are: Respect fo r the Dignity o f Persons, Responsible Caring, and Integrity in Relationships. Like the APA’s corresponding principle,

230

CHAPTER 12

the fo u rth principle o f the CPA code (Responsibility to Society) applies indi­ rectly. Regardless o f w hether psychologists consult the APA code, the CPA code, or both, they should also review m ore specific standards o u tlined in the codes that may be p ertin en t to each dilem m a with which they are dealing. Both the APA and CPA codes oud in e standards referring broadly to issues such as inform ed consent, confidentiality, and com petence. T he geriatric neuropsychologist will often en co u n ter ethical dilemmas. T hese m ust be considered carefully and any decision taken m ust be able to w ithstand scrutiny. T he decision-making approach reco m m en d ed by the CPA’s ethics code is particularly useful in the resolution o f ethical dilemmas. Specifically, the Canadian Code o f Etiiics for Psychologists o u t­ lines seven steps for resolving ethical dilemmas. Psychologists m ust first identify the ethically relevant issues and practices. In o th e r words, they should consider how each ethical principle o f the codes applies to th eir situation. Second, they should consider alternative courses o f action. Third, they should analyze all likely risks and benefits for each course o f action. Fourth, practitioners m ust select an appropriate course o f action after conscientious application o f existing principles, standards, and m oral val­ ues. Fifth, they should take appropriate action with a com m itm ent to assume responsibility for the consequences of the selected course o f action. Sixth, they m ust evaluate carefully the consequences o f the course o f action. Finally, practitioners m ust assume responsibility for consequences o f their action. This could involve die correction o f any negative consequences or th e reengaging o f the decision-m aking process if the ethical p ro b lem is n o t resolved. Sometimes as practitioners consider alternative courses o f action, they may realize th a t there will be tim es w hen ethical principles are in conflict. T h ere may be a situation, for instance, where the n eed to give m axim al pro tection to clients’ welfare has to be balanced with the n eed to show concern for their dignity. T he need to protect clients’ dignity is m aximized w hen they are given full and com plete inform ation ab o u t assessment con­ clusions O n the o th e r hand, the disclosure o f such inform ation could som etim es have devastating em otional consequences for clients and, in rare occasions, it could lead them to cease their cooperation with the health service provider. In such a case, the ethical principles co n cern ed with protection o f dignity are in conflict with ethical principles co ncerned with protection o f welfare. A health care provider could, therefore, becom e tem pted to proceed with appropriate referral while being som ew hat evasive ab o u t the assessment conclusions. T he CPA code deals with such situations by having its principles organized in a hierarchical fashion. Principle I (Respect for the Dignity of Persons) is considered to be the m ost im portant. Thus, in situations where principles are in conflict, the hierarchical or­ ganization o f the code could lead to resolution. In contrast, the principles

ETHICS

231

of the APA code are not organized hierarchically, making the code less helpful for guiding actions. It is im portant to point out, however, that the in tent of the hierarchical organization of the CPA code is not to underm ine the im portance of showing responsible caring and protecting client welfare. In the scenario discussed, a psychologist should take all reasonable steps to protect client welfare and minimize and correct any negative conse­ quences that a selected course of action could have. The psychologist should ensure, for instance, that appropriate follow-up a n d /o r referral is arranged if the assessment results have negative emotional consequences for a client. Ethical dilemmas usually do not have perfect solutions and the CPA code points out that there are exceptions to the hierarchical organization of its principles. It states, for instance, that Respect for the Dignity of Persons should be given the highest weight except in situations where there is clear and im m inent danger to the physical safety of any individual. Naturally, psychologists and oth er practitioners should also be familiar with laws and precedence-setting cases affecting their jurisdiction and take these into account in their chosen course of action. As stated in the APA code of ethics, “w hether or not a psychologist has violated the ethics code does n ot by itself determ ine w hether he or she is legally liable in any court action” (American Psychological Association, 1992, p. 1598). It is im portant to consult with colleagues when in doubt about any situation involving the resolution of ethical dilemmas in practice or research.

References

Abikoff, H., Alvir, J., H ong, G., Sukoff, R., Orazio, J., Solom on, S., 8c Saravay, S. (1987). Logical memory subtest of the W echsler Memory Scale: Age an d education norm s and alternate-form reliability of two scoring systems. Journal o f CUinical and Experimental Neuropsychology, 9, 435-448. A braham , I. L., M anning, C., Boyd, M. R., Neese, J. B., Newman, M. C., Plowfield, L. A., 8c Reel, S. (1993). Cognitive screening o f nursing hom e residents: Factor structure o f the Modified Mini Mental State Exam ination. International Journal of Geriatric Psychiatry, 8, 133-138. Abramovitch, R., Finstad, M., 8c Silberfeld, M. (1993). Preliminary report on m ental capacity assessments. Canadian Journal on Aging, 12, 373-381. A bsher, R., 8c Cum mings, J. L. (1993). Noncognitive behavioural alterations in dem entia syndromes. In F. Boiler 8c J. Grafman (Eds.), Handbook o f neuropsychology (Vol. 8 , pp. 315-338). A msterdam: Elsevier Science. Adams, R. L., Boake, C., 8c Crain, C. (1982). Bias in a neuropsychological test classification related to education, age, and ethnicity. Journal o f Consulting and Clinical Psychology, 50, 143-145. Adams, R. L., Smigielski, J., & Jenkins, R. L. (1984). D evelopm ent o f a Satz-M ogel short form of the WAIS-R. Journal of Consulting and Clinical Psychology, 52, 908. Agnew, J., Bolla-Wilson, K., Kawas, C. H., 8c Bleecker, M. L. (1988). P urdue Pegboard age and sex norm s for people 40 years old and older. Developmental Neuropsychology, 4, 29-35. Albert, M., 8c C ohen, C. (1992). The test for severe im pairm ent: An instrum ent for the assessment of patients with severe cognitive dysfunction. Jinirrwl of the American Geruitrics Society, 40, 449-453. Albert, M. L. (1978). Subcortical dem entia. In R. Katzman, R. D. Terry, 8c K. L. Bick (Eds.), Alzheimer’s disease: Senile dementia and related disorders (pp. 175-180). New York: Raven. Albert, M. L., 8c Knoefel, J. E. (Eds.). (1994). Clinical neurology o f aging (2nd cd.). New York: Oxi'ord University Press. A lbert, M. S. (1981). Geriatric neuropsychology. Journal of Consulting and Clinical Psychology, 49, 835-850. Albert, M. S. (1988). Cognitive function. In M. S. Albert 8c M. B. Moss (Eds.), Geriatric neuropsychology (pp. 33-56). New York: Guilford.

232

REFERENCES

233

Albert, M. S., Heller, H. S., 8c M ilberg, W. (1988). Changes in nam ing ability with age. Psychology and Aging, 3, 173-178. A lbert, M. S., 8c Moss, M. B. (1988). Geriatric neuropsychology. New York: Guilford. A lekoum bides, A., Charter, R. A., Adkins, T. G., 8c Seacat, G. F. (1987). T h e diagnosis of brain dam age by the WAIS, WMS, and Reitan Battery utilizing standardized scores corrected for age and education. International Journal o f Clinical Neuropsychology, 9, 11-28. Alexopoulos, G. S. (1989). Late-life depression and neurological brain disease. International Journal of Geriatric Psychiatry, 4, 187-190. Allen-Burge, R., Storandt, M., Kinscherf, D. A., 8c Rubin, E. H. (1994). Sex differences in the sensitivity o f two self-report depression scales in older depressed inpatients. Psychology and Aging, 9, 443-445. Almeida, O. P., Howard, R. J., Forstl, II., 8c Levy, R. (1992). Late paraphrenia: A review. International Journal of Geriatric Psychiatry, 7, 543-548. Almeida, O. P., Howard, R. ]., Levy, R., 8c David, A. S. (1995a). Psychotic states arising in late life (late paraphrenia): Psychopathology and nosology. British Journal o f Psychiatry, 166, 205-214. Almeida, O. P., Howard, R. J., Levy, R., 8c David, A. S. (1995b). Psychotic states arising in late life (late paraphrenia): T he role o f risk factors. BritishJournal o f Psychiatry, 166, 215-228. A m erican Psychiatric Association. (1987). Diagnostic and statistical manual of mental disorders (3rd, rev. ed.). W ashington, DC: Author. Am erican Psychiatric Association. (1994). Diagnostic and statistical manual o f mental disorders (4th ed.). W ashington, DC: A uthor. Am erican Psychological Association. (1981). Specialty guidelines for the delivery o f services by clinical psychologists. American Psychologist, 36, 640-651. Am erican Psychological Association. (1985). Standards for educational and psychological testing. W ashington, DC: A uthor. A m erican Psychological Association. (1987). G eneral guidelines for providers of psychological services. American Psychologist, 42, 712-723. A m erican Psychological Association. (1992). Ethical principles of psychologists an d code of conduct. American Psychologist, 47, 1597-1611. A m m ons, R. B., 8c Ammons, C. H. (1962). T he Quick Test (QT): Provisional manual. Psychological Reports, 11, 111-161. Anastasi, A. (1988). Psychologiud testing (6th ed.). U p p er Saddle River, NJ: Prentice-Hall. Angoff, W. H. (1971). Scales, norm s and equivalent scores. In R. L. T h o rn d ik e (Ed.), Educatiorud measurement (2nd ed., pp. 508-600). W ashington, DC: American Council on Education. Angoff, W. H., 8c Robertson, G. R. (1987). A procedure for standardizing individually adm inistered tests, norm ed by age or grade level. Applied Psychological Measurement, 11, 33-46. Anthony, J. C., Le Resche, L., Niaz, L., Von Korff, M. R., & Folstein, M. F. (1982). Limits o f the Mini-mental state as a screening test for dem entia and delirium am ong hospital patients. Psychological Medicine, 12, 397-408. Anthony-Bergstone, C. R., Zarit, S. H., & Gatz, G. (1988). Symptoms o f psychological distress am ong caregivers of dem entia patients. Psychology and Aging, 3, 245-248. A ronson, II. (1985). M anual o f administration and scoring: Aronson Shopping List, Form I and 11. U npublished m anuscript. Axelrod, B. N., 8c Henry, R. R. (1992). Age-related perform ance on the W isconsin Card Sorting, Similarities, and Controlled O ral W ord Association Tests. Clinical Neuropsychologist, 6, 16-26. A xelrod, B. N., Jiron, C. C., 8c Henry, R. R. (1993). Perform ance of adults ages 20 to 90 on the A bbreviated Wisconsin Card Sorting Test. Clinical Neuropsychologist, 7, 205-209.

234

REFERENCES

Bak, J. S., Sc Greene, R. L. (1980). Changes in neuropsychological functioning in an aging population. Journal of Consulting and Clinical Psychology, 48, 395-399. Banks, P. G., Dickson, A. L., 8c Plasay, M. T. (1987). The Verbal Selective Reminding Test: Preliminary data for healthy elderly. Experimental Aging Research, 13, 203-207. Barber, R., Snowdon, J. S., Sc Craufurd, D. (1995). Frontotemporal dem entia and Alzheimer’s disease: Retrospective differentiation using information from informants. Journal of Neurology, Neurosurgery and Psychiatry, 59, 61-79. Barer, M. L., Evans, R. G., &: Hertzman, C. (1995). Avalanche or glacier? Health care and the dem ographic rhetoric. Canadian Journal on Aging, 14, 193-224. Barona, A., 8c Chastain, R. (1986). An inproved estimate of prem orbid IQ for blacks and whites on the WAIS-R. International Journal of Clinical Neuropsychology, 8, 169-173. Barona, A., Reynolds, C. R., 8c Chastain, R. (1984). A demographically based index of prem orbid intelligence for the WAIS-R. Journal of Consulting and Clinical Psychology, 52, 885-887. Basset, S. S., Magaziner, J., 8c Hebe, J. R. (1990). Reliability of proxy response on mental health indices for aged, community-dwelling women. Psychology and Aging, 5, 127-132. Bayles, K. A , Boone, D. R., Tomoeda, C. K., Slauson, T. J., 8c Kaszniak, A. W. (1989). Differentiating Alzheimer’s patients from the normal elderly and stroke patients with aphasia. Journal of Speech and Hearing Disorders, 54, 74—87. Bayles, K. A., 8c Kaszniak, A. W. (1987). Communication and cognition in normal aging and dementia. Boston: Little, Brown. Bayles, K. A., 8c Tomoeda, C. (1991). Arizona Battery for Communication Disorders of Dementia. Gaylord, MI: National Rehabilitation Services. Bcck, A. T., 8c Beck, R. W. (1972). Screening depressed patients in family practice: A rapid technique. Postgraduate Medicine, 52, 81-85. Beck, A. T., Ward, C. H., Mendelson, M., Mock, J., 8c Erbaugh, J. (1961). An inventory for measuring depression. Archives of General Psychiatry, 4, 561-571. Beland, R., 8c Lecours, A. R. (1990). The M T-86p Aphasia Battery: A subset o f normative data in relation to age and level of school education. Aphasiology, 4, 439-462. Beland, R., Lecours, A. R., Giroux, F., 8c Bois, M. (1993). The MT-860 Aphasia Battery: A subset o f normative data in relation to age and level of school education: II. Aphasiology, 7, 359-382. Bellack, A. S., Sc Hersen, M. (1988). Behavioral gy, II, 423-428. Snow don, J., 8c Lane, F. (1994). A longitudinal study of age-associated m em ory im pairm ent. International Journal of Geriatric Psychiatry, 9, 779-787. Sobel, L. C., T oneatto, T., 8c Sobell, M. B. (1994). Behavioral assessment an d treatm ent planning for alcohol, tobacco, and o ther drug problem s: C u rren t status with an em phasis on clinical applications. Dehavun Therapy, 25, 533-580. Speirs, P. A. (1981). Have they com e to praise Luria o r to bury him? T h e Luria-N cbraska Battery controversy. Journal of Consulting and Clinical Psychology, 49, 331-341. Spielberger, C. D., Gorsuch, R. L., 8c Lushene, R. E. (1970). M anual for the State-Trail Anxiety InvenUrry. Palo Alto, CA: Consulting Psychologists Press.

REFERENCES

257

Spielberger, C. D., Gorsuch, R. L., Lushene, R. E., M ontouri, J., & Platsek, D. (1973). STAIC: Preliminary manual for the State-Trait Anxiety for Children. Palo Alto, CA: Consulting Psychologists Press. Spielberger, C. D., Gorsuch, R. L., Lushene, R. E., Vagg, P. R., & Jacobs, G. A. (1983). Manual for the State-Trait Anxiety Inventory (form Y). Palo Alto, CA: Consulting Psychologists Press. Spielman, A. (1986). Assessment of insomnia. Clinical Psychology Review, 6, 11-25. Spitzer, R. L., & Endicott.J. (1978). NIMH clinical research collaborative fprogram on the psychobiology of defrression: Schedule for affective disorders and schizophrenia (SADS). New York: New York State Psychiatric Institute, Biometrics Research Division. Spitzcr, R. L., E ndicott.J., 8c Robins, E. (1978). Research diagnostic criteria: Rationale and reliability. Archives of General Psychiatry, 35, 773-782. Spitzform, M. (1982). Normative data in the elderly on the Luria-Nebraska Neuropsychologi­ cal Battery. Clinical Neuropsychology, 4, 103-105. Spreen, O., 8c Benton, A. L. (1969). Neurosensory Centre Comprehensive Examination for Aphasia. Victoria, Canada: University of Victoria. Spreen, O., & Strauss, E. (1991). A compendium of neurofpsychological tests: Administration, norms and commenUiry. New York: Oxford University Press. Squire, L. R. (1987). Memory and brain. New York: Oxford University Press. Stambrook, M., Cardoso, E. R., Ilawryluk, G. A., Erikson, P., Piated, D., 8c Sicz, G. (1988). Neuropsychological changes following the surgical treatment of normal pressure hydrocephalus. Archives of Clinical Neurofpsychology, 3, 323-330. Stambrook,. M., Gill, D. D., Cardoso, E. R., 8c Moore, A. D. (1993). Communicating (normal-pressure) hydrocephalus. In R. W. Parks, R. F. Zees, & R. S. Wilson (Eds.), Neuropsychology of Alzheimer's disease and other dementias (pp. 283-307). New York: Oxford University Press. Stanton, B. A., Jenkins, C. D., Savageau, J. A., Zyzanski, S. J., 8c Aucoin, R. (1984). Age and educational differences on the Trail Making Test and Wechsler Memory Scales. Percefptual and Motor Skills, 58, 311-318. Stebbins, G. T., Gilley, D. W., Wilson, R. S., Bernard, B. A., 8c Fox, J. H. (1990). Effects of language disturbances on prem orbid estimates of IQ in mild dementia. Clinical Neurofpsychologist, 4, 64-68. Stebbins, G. T., Wilson, R. S., Gilley, D. W., Bernard, B. A., 8c Fox, J. H. (1990). Use of the National Adult Reading Test to estimate prem orbid IQ in dementia. Clinical Neuropsycholo­ gist, 4, 18-24. Straker, M. (1982). Adjustment and personality disorders in the aged. Psychiatric Clinics of Ninth America, 5, 121-129. Streiner, D. L., 8c Norman, G. R. (1995). Health measurement scales: A practical guide to their development and use (2nd ed.). New York: Oxford University Press. Stroop, J. R. (1935). Studies of interference in serial verbal reactions. Journal of Experiment/il Psychology, 18, 643-662. Strub, R. C., 8c Black, F. W. (1988). Neurobehavuntd disorders: A clinical approach. Philadelphia: Davis. Sunderland, T., Hill, J. L., Mellow, A M., Lawlor, B. A., Gundersheimer, J., Newhouse, P. A., & Grafman, J. H. (1989). Clock drawing in Alzheimer’s disease: A novel measure of dem entia severity. Jourrud of the American Geriatrics Society, 3, 725-729. Sweet, J. J., Mober, P. J., 8c Tovian, S. M. (1990). Evaluation of Wechsler Adult Intelligence Scale-Revised premorbid IQ formulas in clinical populations. Psychological Assessment, 2, 41-44. Talland, G. A (1965). Three estimates of word span and their stability over the adult years. Quarterly Journal of Experinumtal Psychology, 17, 301-307. Tamkin, A. S., & Jacobsen, R. (1984). Age-related norms for the H ooper Visual Organization Test. J(mmal of Clinical Psychology, 40, 1459-1463.

258

REFERENCES

T eng, E. L., 8c Chui, H. C. (1987). T he modified Mini-Mental State (3MS) Exam ination. Journal of Clinical Psychiatry, 48, 314-318. T eri, L., T ruax, P., Logsdon, R., Vomsto, J., Zarit, S., 8c Vitaliano, P. (1992). Assessment o f behavioral problem s in dem entia: T he revised m em ory and behavior problem s checklist. Psychology and Aging, 7, 622-629. T eri, L., & W agner, A. (1992). A lzheim er’s disease an d depression. Journal of Consulting and Clinical Psychology, 60, 379-391. T hom pson, M. G., Heller, K-, & Rody, C. (1994). R ecruitm ent challenges in studying late-life depression: D o com m unity samples adequately represent depressed older adults? Psychology and Aging, 9, 121-125. T horp, T. R., 8c M ahrer, A. R. (1959). Predicting potential intelligence. Journal of Clinical Psychology, 15, 286-288. Tiffin, J. (1968). Purdue Pegboard Examiner manual Chicago: Science Research Associates. T om baugh, T. (1989). An itemized scoring system for the Taylor Complex Figure. U npublished research, C arleton University. T om baugh, T. N., McDowell, I., Kristjansson, B., 8c Hublcy, A. M. (1996). Mini-Mental State Exam ination (MMSE) and the M odified MMSE (3MS): A psychometric com parison and norm ative data. Psychological Assessment, 8, 48-59. T om baugh, T. N., 8c McIntyre, N. j. (1992). T he Mini-Mental State Exam ination: A com prehensive review. Journal of the American Geriatrics Society, 40, 922-935. T om baugh, T. N., 8c Schmidt, J. P. (1992). T he learning and memory battery (LAMB): D evelopm ent and standardization. Psychological Assessment, 4, 193-206. T om baugh, T. N., Schmidt, J. P., 8c Faulkner, P. (1992). A new procedure for adm inistering the Taylor Com plex Figure: Normative data over a 60-year age span. Clinical Neuropsycholth gist, 6, 63-79. T rah an , D. E., G oethe, K. E., 8c Larrabee, G. J. (1989). An exam ination o f verbal supraspan in norm al adults and patients with head traum a or unilateral cerebrovascular accident. Neuropsychology, 3, 81-90. T rah an , D. E., 8c Larrabee, G. J. (1993). Clinical and m ethodological issues in m easuring rate o f forgetting with the Verbal Selective R em inding Test. Psychological Assessment, 5, 67-71. Tsang, M. H., A ronson, H., 8c Hayslip, B. (1991). Standardization o f a learning an d retention task with com m unity reading older adults. Clinical Neuropsychologist, 5, 67-77. Tuokko, H. (1993). Psychosocial evaluation and m anagem ent o f the A lzheim er’s patient. In R. W. Parks, R. F. Zee, 8c R. S. Wilson (Eds.), Neuropsychology o f Alzheimer’s disease and other dementias (pp. 565-588). New York: O xford University Press. T uokko, H., 8c Crockett, D. (1987). Central cholinergic deficiency WAIS profiles in a non-dem ented aged sample. Jourrud of Clinic/d and Experimented Neurofjsychology, 9, 225-227. T uokko, H., 8c Crockett, D. (1989). Cued recall and m em ory disorders in dem entia. Journal o f Clinic/d and Experimental Neuropsychology, 11, 278-294. T uokko, H., Crockett, D., Holliday, S., 8c Coval, M. (1987). T he relationship between perform ance on the Multi-focus Assessment Scale an d functional status. Canadian Journal on Aging, 6, 33-45. T uokko, H., 8c Crockett, D. J. (1991). Assessment of everyday functioning in norm al and m alignant memory disordered elderly. In D. E. T u p p er 8c K. D. Cicerone (Eds.), The neuropsychology of everyday life: Issues in deveUtpment and rehabilitation (pp. 135-181). Boston: Kluwer. T uokko, H., 8c Hadjistavropoulos, T. (1994). A comparison of the clinical utility of alternative approaches to the scoring of clock drawing. U npublished m anuscript. T uokko, II., Hadjistavropoulos, T., Miller, J. A., 8c Beattie, B. L. (1992). T h e Clock Test: A sensitive m easure to differentiate norm al elderly from those with A lzheim er’s disease. Journal oj the American Geriatrics Society, 40, 579-584.

REFERENCES

259

Tuokko, H., Hadjistavropoulos, T., Miller, J. A., H orton, A., 8c Beattie, B. L. (1995). The Clock Test: Administration and scoring mannaL T oronto: Multi-Health Systems. T uokko, H., Hadjistavropoulos, T., Rae, S., 8c O ’Rourke, N. P. A comparison o f the clinical utility o f alternative, afpproaches to the scoring o f clock drawing. M anuscript subm itted for publication. T uokko, H., Krisyansson, E., 8c Miller, J. (1995). T he neuropsychological detection of dem entia: An overview of the neuropsychological com ponent o f the C anadian Study of H ealth an d Aging. Journal o f Clinical and Experimental Neurofpsychology, 17, 352-373. T uokko, H ., 8c Purves, B. (1993). Understanding dementia: A jproblem afpproach for caregivers [Videotape]. Available from the A lzheimer Society o f BC, 20-601 West Cordova Street, Vancouver, British Columbia, Canada V6T 1G1. T uokko, H., Vcmon-W ilkinson, R., Weir, J., 8c Beattie, B. L. (1991). Cued recall an d early identification o f dem entia. Journal of Clinical and Experimented Neuntfpsychology, 13, 871-879. Tuokko, H., 8c Woodward, T. (1996). D evelopment an d validation o f a dem ographic correction system for the neuropsychological measures used in the C anadian Study of H ealth a n d Aging. Journal of Clinic/d and Experimental Neurofpsychology, 18, 479-616. Tym chuk, A., 8c O uslander, J . G. (1990). O ptim izing the inform ed consent process with elderly people. Educational Gerontology, 16, 245-257. U.S. Bureau o f the Census. (1950). U.S. census of pofpulation: 1950 (C urrent population reports, series P-23, No. 128). W ashington, DC: U.S. G overnm ent Printing Office. U.S. Bureau o f the Census. (1983). America in transition: A n aging society (C urrent population reports, series P-23, No. 128). W ashington, DC: U.S. G overnm ent Printing Office. U hlm ann, R. F., Teri, L., Rees, T. S., Mozlowski, K. J., 8c Larson, E. B. (1989). Im pact of mild to m oderate hearing loss on m ental status testing: Comparability o f standard an d written Mini-Mental State Exam inations. Journal of the American Geriatrics Society, 37, 223-228. Ulatowska, H. K., Hayashi, M. M., Cannito, P., 8c Fleming, S. G. (1986). Disruption o f reference in aging. Brain and L