Comprehensive geriatic assessment in the outpatient setting: Population characteristics and factors influencing outcome

Experimental Gerontology, Vol. 28, pp. 447--z,57,1993 Printed in the USA. All rights reserved.

0531-5565/93$6.00 + .00 Copyright© 1993PergamonPress Ltd.

COMPREHENSIVE GERIATRIC ASSESSMENT IN THE OUTPATIENT SETTING: POPULATION CHARACTERISTICS AND FACTORS INFLUENCING OUTCOME

JANE F. POTTER Department of Internal Medicine, Section of Geriatrics and Gerontology, University of Nebraska Medical Center, 600 South 42nd Street, Omaha, Nebraska 68198-5620 AbstractmThe chronic diseases of old age reduce the "active life expectancy" of

older persons and result in significant health care expenditures. Yet, among disabled persons, the degree of disability does not explain who actually accesses long-term care (LTC) services. Comprehensive geriatric assessment (CGA) has been proposed as the technology of geriatric medicine which will reduce both excess disability and health care costs. Outpatient CGA programs may be the ideal setting in which to study older disabled persons who are at risk for increasing their use of LTC services. A prospective study of our CGA population suggests that understanding the factors which influence outcome will be critical to designing interventions and improving outcome for this population. CGA patients suffer from a limited number of functionally disabling problems, each of which has a broad array of etiologies. Use of LTC services seems most closely .tied to social factors, including the feelings of burden experienced by caregivers. Caregiver burden is a better predictor of service use than any measure of the older person's mental or physical health or functional abilities. Reducing health care costs including nursing facility placement will likely require interventions which reduce caregiver burden. Further definition of the problems seen in the CGA population and delineation of the factors influencing outcome will identify interventions to achieve the desired health care outcomes. Key Words: geriatrics, health services, comprehensive geriatric assessment, disability evaluation, dementia, depression, elderly

INTRODUCTION As CLINICAL director for the Baltimore Longitudinal Study of Aging (BLSA), Reubin Andres (Shock et al., 1984) was influential in the design o f that program. The B L S A and other longitudinal studies are the basis of modern understanding of aging and the differentiation o f aging and disease. The significance of the findings from those studies is shown in part by the important role they play in the day-to-day clinical care and counselling of older persons. More recently, the observation that a portion of aging individuals display a much reduced degree of typical aging changes has led to the concept of " u s u a l " versus " s u c c e s s f u l " aging (Rowe, 1990). The smaller decrement in physiologic function seen in successful aging may be due to better health practices, including appropriate diet and exercise. This suggests that normal old age can be 447

448

J.F. POTTER

consistent with a high level of health and vigor. Remaining limitations to attaining good health in old age are defining the health practices essential to achieve successful aging, educating and convincing the general populace in those principles, and overcoming the chronic diseases of old age. Chronological age, quite naturally, is the best predictor of mortality. Age is also a predictor of both morbidity and disability. While life expectancy is increasing even among older persons, it is not clear how much of that gain is in "active life expectancy" (Katz et al., 1983). It appears that, after age 65, those added years are some mixture of good health and disability (Branch et al., 1991). As a result, the need for long-term care services is expected to increase into the next century. A major societal concern brought on by a projected increase in the number of older disabled persons is the cost of that care. Yet disability alone does not accurately predict who uses those most costly institutional services. For every disabled person in a nursing home there are two equally disabled persons who continue to live at home. Studies are needed to identify why some disabled persons stay at home while others enter LTC facilities. Comprehensive geriatric assessment (CGA) has been described as a multidisciplinary diagnostic process used to quantify an older person's medical, psychosocial, and functional capabilities and problems with the intention of arriving at a plan for long-term follow-up and care (Rubenstein, 1983). As part of this process, patients receive at least short-term management of identified problems. CGA sets as its goal improving the independence of impaired older persons. The target population is made up of persons at risk for institutional care. Geriatric evaluation services came into use in the United States during an era of cost constraint. Therefore, studies of CGA have usually dealt with demonstrating benefits and effectiveness. Indeed, in many, but not all, settings this process has proven itself worthwhile (Rubenstein, 1983; Applegate et al., 1991; Rubenstein et al., 1991). The setting where cost effectiveness and beneficial effects on mortality are most in doubt is outpatient assessment. Perhaps it is ineffective; however, there have been few controlled studies, and the apparent lack of effectiveness may be due to inadequate sample sizes or relatively low mortality rates (Rubenstein et al., 1991). An alternate explanation for the apparent lack of effectiveness is inadequate understanding of the population being served and the factors which influence outcomes in this setting. At UNMC we have not conducted effectiveness studies. Rather, we set out to better understand the patient population that undergoes CGA. The rationale for this approach is that knowledge of the factors that influence outcome is limited, and the overall benefit of CGA depends on how well those factors are understood. The CGA population is of interest because, when properly selected, these are persons who are experiencing functional failure. The breadth and methodical nature of the evaluation that persons undergoing CGA receive is difficult to duplicate in other study sites. This may be the ideal setting for study of persons in need of LTC services. The overall hypothesis of our study is that better understanding of the factors that contribute to dependence and disability would suggest approaches and interventions to improve the clinical care and outcomes for these patients. The UNMC CGA program, described below, has taken a prospective longitudinal view of this population. The program was in operation for 1.5 years before the start of data collection. A development grant from the Enron and Peter Kiewit Foundations for the CGA program included funding for start-up of the research data base.

449

OUTPATIENT GERIATRIC ASSESSMENT

SUBJECTS Subjects or their caregivers who request CGA are screened for perceived problems in the following areas: memory loss, depression, change in behavior, weight loss, falls, incontinence, decline in functional status, failing health, and concern regarding present living arrangements. Only those subjects who meet some of these screening criteria are scheduled for evaluation. Subjects or caregivers provide basic demographic information, an initial list of concerns and problems, and information on the number of days in hospital and or nursing homes in the prior 6 months. STUDY DESIGN Subjects presenting for outpatient CGA are invited to participate in a prospective study of functional outcomes. In addition to a protocol assessment, subjects and caregivers are asked to return to clinic or to respond to phone calls and mailed surveys after 6 and/or 12 months. Initial subject enrollment began in January 1986. Data collection occurs in cycles during which variables are not deleted or modified. Participating investigators have the opportunity to modify variables at the beginning of each new data cycle. Many variables have remained constant since the start of the study (see Table 1). The time frame for cycles of data collection and the methods of follow-up employed in each cycle are given in Table 2. Refusal rate for the study was highest (at 6%) in cycle I, when the protocol required on-site follow-up visits. More recently, refusals have fallen to 2-3%. The proportion of subjects (or caregivers) who complete one-year follow-up interviews is 88%. TABLE 1. VARIABLESINCLUDED Domain

Variable

Status

Social

Caregiver Burden Social Network Index Formal Services Informal Services

Continuous Added 8/3/88 Continuous Added 8/3/88

Cognitive

Mini-Mental State Exam Washington University Brief Neuropsychological Screening Clinical Dementia Rating

Continuous Continuous Continuous Continuous

Affective

Yesavage Geriatric Depression Scale Morale Scale

Continuous and Modified

Functional

Activities of Daily Living (ADL) Instrumental ADL Williams Manual Test

Continuous Continuous Modified 8/3/88

Physical Health

Cumulative Illness Rating Scale Medication Profile Visual Acuity Audiometry Geriatric Syndromes

Continuous Modified 8/3/88 Continuous Added 8/3/88 Modified 8/3/88

Dropped as of 6/6/90

450

J.F. POTTER

TABLE 2. DATA COLLECTION CYCLES

Time

!/86-6/87 7/87-7/88 8/88-8/89 4/90-8/91 9/91-current

Number

Follow-up procedure

146 138 157 195 30

1, 2, 3 3 2, 3 2, 3 2

1 = 6 - and 12-monthon-site visits, 2 = phone interview, 3 = National Death Index.

Most initial data is collected during one morning in the outpatient area of the University Geriatric Center. All subjects undergo a complete history, physical, detailed neurological examination (Zemcov et al., 1984), and review of medical records. Prior to July 1987, subjects with a clinical indication (approximately one-half of all subjects) were psychiatrically interviewed. Since July 1987, all have received a semistructured psychiatric interview. Trained nursing staff conduct additional tests of cognitive functioning and ascertain information on basic and instrumental activities of daily living (ADL and IADL). A single experienced social worker who has been with the program since its inception conducts a family interview and collateral source history, and oversees collection of elements in the social work data base. Medication history from patient, family, and record review is collated by a geriatric pharmacist. Subjects and caregivers are seen for a follow-up conference between 2 and 6 weeks after the initial visit to learn the results of studies and team recommendations. Between visits subjects complete an individually tailored program of diagnostic studies and interventions. At the time of the follow-up conference, subjects are classified with respect to cognitive and affective diagnoses and their degree of physical illness, and often complete special testing when they meet criteria for specific protocols in the current cycle of data collection. STUDY VARIABLES Initial variables

These variables characterize the study population and cover the broad range of parameters which might be expected to influence outcome. The domains assessed have remained constant throughout the study, while the measurements themselves have evolved. The domains are social, cognitive, affective, functional, and physical health (Table 1). Social

The Family Burden Interview developed by Steven Zarit et al. (1980; Zarit, 1982; Zarit et al., 1986) to measure the subjective burden of families in providing care is completed by the patient's primary caregiver. The use of formal and informal services are determined by interview, and information is collected that allows a Social Network

OUTPATIENT GERIATRIC ASSESSMENT

451

Index (Berkman, 1977; Berkman and Syme, 1979), which measures the strength of close social bonds to be calculated. Cognitive/affective

Measures of cognitive function include Folstein and Folstein's (1975) Mini-Mental State Exam (MMSE), the Washington University Brief Neuropsychiatric Screening Instrument (Storandt et al., 1984), and a Clinical Demential Rating scale (CDR) (Hughes et al., 1982). Measures of emotional health include the Philadelphia Geriatric Center Morale Scale (Lawton, 1975) and the Yesavage Geriatric Depression Rating Scale (Yesavage et al., 1983). Subjects with dementing illness are classified into 1 of 18 categories of dementia as outlined by Zemcov et al. (1984). Physical~functional

Measures of the older person's abilities include Activities of Daily Living (ADL) (Lawton and Brody, 1969), Instrumental Activities of Daily Living (IADL) (Berkman and Syme, 1979), timed manual performance (Williams and Mitchell, 1984; Williams, 1987), and a measure of comorbid illness, the Cumulative Illness Rating (CIR) (Linnet al., 1968). Outcome variables

At a specified interval (typically one year after initial evaluation), subjects or their caregivers (as appropriate) are contacted to determine outcomes of interest. These include vital status, the number of days in nursing facilities, and acute care hospitals, ciarrent living arrangements, and use of formal and informal services. In the first cycle of data collection, subjects returning to clinic for a 6- and/or 12-month visit had a series of additional parameters remeasured, including neuropsychiatric, functional, and general health parameters. Survival status of subjects is also tracked using the National Death Index. RESULTS Overview

Sixty-five percent of UNMC CGA subjects are found after evaluation to be suffering from cognitive disorders, a figuresimilar to the 74% reported in a population of 438 outpatients undergoing CGA (Altkorn et al., 1991). The average subject is 78 years of age, female, and unmarried (Table 3). This profile is similar in sex and marital status and slightly older in age (75 vs. 78 years) when compared to that reported from California (Altkorn et al., 1991), and similar in all three parameters to a second population from New York (Williams et al., 1987). The similarities in these three populations suggest some consistency in the population that receives outpatient CGA services. Our CGA subjects are relatively heavy users of both acute and LTC services. Fortyeight percent of subjects spend some time in acute care hospitals and 37% in nursing homes during the year following assessment. The average subject who enters the hospital during the year has 11.2 days of care, while those entering nursing homes experience 162 days of care. Subjects are fairly independent in personal care (e.g., dressing,

452

J.F. POTTER T A B L E 3. D E M O G R A P H I C S O F T H E

664

SUBJECTS IN T H E

STUDY P O P U L A T I O N

Age (years)

78.3

% Female

74

Marital Status Married (%) Widowed/Single/Divorced (%)

34 66

Living Arrangement Alone (%) With spouse/relative (%) Sheltered environment (%)

38 47 15

Annual Income

< $8,000 (%) $8-15,000 (%) > $15,000 (%)

38 36 26

Education: ~<8 years (%) 9-12 years (%) > 12 years (%)

24 51 25

bathing, and toileting) but have substantial limitations in conducting those more complex activities summarized as IADLs (e.g., cooking, cleaning, shopping, and finance). Analysis of the CGA data base has resulted in 10 published reports, 25 abstracts, and a better understanding of the process and where it could be improved. A brief summary of several of the published studies is given below. Caregiver burden In our CGA population, the burden experienced by caregivers is among the most powerful predictors of outcome for these patients. The study that identified this phenomenon involved 146 subjects studied over an 18-month period (Brown et al., 1990); 109 met selection criteria and had identified primary caregivers. The Burden Interview of Zarit et al. (1980), which assesses feelings of burden in providing care, was completed by the primary caregiver. Change in each subject's use of formal services, including change to a more dependent living arrangement, was determined at the end of one year. Using logistic regression, the score on the Burden Interview was a more powerful predictor of service use than any measure of the subject's mental or physical status. In a representative subset of caregivers who completed the Burden Interview at foUow-up, only those whose relatives were placed in nursing homes experienced a decrease in burden. Increasing home services did not significantly reduce the feelings of burden. These findings have resulted in an increased emphasis on caregiver counselling and support during all phases of the CGA process. Further research is needed to identify and test interventions with caregivers that are likely to reduce both burden and the use of institutional services. Assessment of a patient's functional abilities is critical in the evaluation of any chronic condition. Functional abilities are usually assessed by verbal report of the

OUTPATIENT GERIATRIC ASSESSMENT

453

patient or caregiver. Patients often overestimate true abilities, while caregivers underestimate them. An objective or performance-based measure of function, therefore, has appeal. One such measure, the Williams Manual Test (WMT) (Williams and Mitchell, 1984), was evaluated in CGA patients. One hundred and seventeen older subjects completed the WMT and questionnaires assessing current service use (Scholer et al., 1990) at baseline and again at 12 months. The WMT was found to be a significant predictor of additional service use at 12 months (X2 = 6.4, P < .01). The test had high specificity (83%), but relatively low sensitivity (40%). In further analysis, the lack of sensitivity in this study group was best explained by higher levels of burden among caregivers whose relatives received, but were not predicted to need, increases in services. The WMT is the most objective measure of abilities employed in this study. However, a patient's abilities can be maintained and they will still use formal services if the primary caregiver feels heavily burdened. Depression and personality changes A relatively high proportion of outpatient CGA subjects in this and other populations (Altkorn et ai., 1991) suffer from dementing illness. Affective symptoms and behavioral problems are among the most troublesome complications of dementia for caregivers. Accurate screening for depression in the presence of dementia is an important issue in this population. Over a two-year period, 283 patients were seen in the CGA program and screened using the Yesavage Geriatric Depression Rating Scale (GDS) (Linn et al., 1968). Subjects were also given a retrospective psychiatric diagnosis by one of two geropsychiatrists. Cognitively intact subjects (N = 70) were compared to individuals with mild Alzheimer's disease (N = 72). Data were analyzed using receiver-operator characteristic curves. In the intact group, the GDS was both sensitive and specific in the detection of depression, while in mild Alzheimer's its performance was no better than chance (Burke et al., 1989). Additional data has been collected to prospectively assess the utility of the GDS in this population, as well as to develop a caregiver version of the GDS. Family caregivers of demented patients are not only concerned with depression, but also struggle to understand and cope with a whole host of personality changes. Investigation of the changes of personality, however, has been an area of limited study. Items from the Blessed Dementia Scale, "changes in personality, interests, and drives," have been operationalized into a series of 30 questions as one part of an extensive telephone screening instrument for dementia (ALFI) (Fischbach, 1990). This instrument was evaluated in our CGA patients (Wengel et al., submitted). One hundred patients and their collateral sources (CS) were interviewed. Nineteen were cognitively intact, 15 questionably impaired, and 66 were mildly to moderately demented. The total number of Blessed items reported by the CS significantly correlated with dementia severity using the CDR scale (r = .39, p = .0001). The caregivers of demented patients endorsed six items significantly more often than their counterparts. Those were as follows: 1) change bothers subjects more, 2) striking others or throwing things, 3) decreased emotional responsiveness, 4) decreased interest in hobbies, 5) difficulty starting new tasks, and 6) indecisiveness. By self-report, subjects with dementia described two of these changes more often than those without dementia: 1) lack of concern about others and 2) decreased interest in hobbies. These results help expand and clarify the personality

454

J.F. POTTER

changes seen in dementia. The availability of operationalized items should facilitate detection and allow greater reliability in the assessment of such changes (Fischbach, 1990). This same phone interview using the ALFI has been used to validate phone versions of the MMSE (Roccaforte et al., 1992) and the Short-Portable Mental Status Questionnaire (Roccaforte et al., 1990). Valid telephone versions of these instruments will facilitate data collection of cognitive outcomes in this population. Sensory changes and dementia One function of CGA is to uncover sources of excess morbidity. An early analysis of our data (Peters et al., 1988) supported the observation of others that auditory impairment appears to accelerate cognitive losses in individuals with Alzheimer's dementia (AD). As a result, the clinic has implemented audiometry as part of its screening of all patients. The presence of visual impairment does not seem to hold the same sort of negative prognostic significance (Berlinger and Potter, 1988), and the prevalence of significant visual disorders was found to be quite low. Dementia and body weight The high prevalence of dementia in the outpatient CGA population provides an opportunity to study dementing illness in its early stages. Low body weight had been described in demented individuals in nursing homes. In order to determine the association between dementia and low body weight in outpatients, body mass index (BMI) was evaluated prospectively in 346 CGA outpatients (Berlinger and Potter, 1991). Patients were categorized into four groups (cognitively intact, dementia of the Alzheimer's type, other dementia, and patients with depressive symptoms). Differences between groups for various clinical parameters were evaluated using an analysis of variance and Duncan's multiple range test. Patients with dementia, regardless of etiologic type or severity, and patients with depressive symptoms had BMIs >i 10% lower than the cognitively intact patients. BMI was positively correlated with IADL. Low BMI was not associated with increased physical illness. In fact, in the subset of patients with AD, lower BMI correlated with significantly lesser amounts of comorbid physical illness. Compared to cognitively intact outpatients, patients with AD appeared to be physically healthier despite their lower BMI. These results suggest an association between dementia and low BMI. On the other hand, the presence of comorbid physical illness, a common focus of evaluation in these patients, was not more common in those patients with lower BMIs. A separate retrospective longitudinal analysis of body weight change in Alzheimer's patients suggested that weight loss had occurred after the onset of AD symptoms and continued during the year prior to assessment (Goodwin et al., 1991). Weight loss was not explained by the presence of comorbid illness. Survival analysis A meta-analysis of controlled trials of outpatient CGA programs (Rubenstein et al., 1991) failed to show a beneficial effect of CGA on mortality. This is in contrast to inpatient CGA programs, where the results have been impressive. This negative finding has been attributed to either the small sample sizes and/or the relatively low mortality

OUTPATIENTGERIATRICASSESSMENT

455

rates in those series. The one-year mortality rate in the UNMC CGA population is 8.3%, which is in the range reported by others (Williams et al., 1987). Predictors of mortality for the first 181 prospectively studied patients in this population have been reported (Waldman and Potter, 1992). In logistic regression, three variables (ADL, IADL, and CIR) were each predictive of death by one year. In separate logistic regressions, CIR and age predicted hospitalization during the year following assessment, while ADL or IADL predicted a nursing home stay. A life table analysis of 435 cases followed for between 1 and 3 years shows that survival is predicted both by a combined ADL/IADL score and the CIR, but not by age (J. Potter, unpublished data). Using this method, the population can be stratified for risk of death from less than 2% predicted death in the most independent to 14% in the most dependent group. These analyses suggest that improvement in mortality for the CGA population would depend on improvement either in functional status or in the burden of illness. Since functional status and chronic illness are interdependent, reduction of mortality will to some extent be linked to the nature of the underlying illness. Improvement of functional status could also be linked to reduction of excess morbidity from such treatable conditions as depression, drug toxicity, unrecognized rehabilitation needs, and unrecognized or untreated medical illness. SUMMARY A 1989 consensus conference on the "Future of Geriatric Assessment" suggested that research be conducted to define the process of geriatric assessment and management and to determine benefits and effectiveness in various sites. Since 1986, the outpatient CGA program at UNMC has made careful observation, intervention, and follow-ups of a large cohort of these patients. The analyses conducted to date have focused on describing the population that is identified under targeting criteria for multiple functional problems, exploring predictors of mortality, understanding the factors predicting use of health services, and validating instruments for measuring health status and outcomes in this population. Acknowledgments--The author thanks the patients and their families who participated in these studies, the investigators who conducted the analyses, Dr. William J. Burke for review and critique of the manuscript, and Ms. Linda Sovereign for preparation of the typescript.

REFERENCES ALTKORN, D.L., RAMSDELL, J.W., JACKSON, J.E., and RENVALL, M. Recommendations for a change in living situation resulting from an outpatient geriatric assessment: Type, frequency and risk factors. J. Am. Geriatr. Soc. 39, 508-512, 1991. APPLEGATE, W., DEYO, R., KRAMER, A., and MEEHAN, S. Geriatric evaluation and management: Current status and future research directions. J. Am. Geriatr. Soc. 39, 2S-7S, 1991. BERKMAN, L.F. Social networks, host resistance, and mortality: A follow-up study of Alameda County residents. University of California, Berkeley, CA, 1977. Available from: University Microfilms International, Ann Arbor, MI; Publication no. 78-12, 490. Dissertation. BERKMAN, L.F. and SYME, S.L. Social networks host resistance and mortality: A nine-year follow-up study of Alameda County residents. Am. J. Epidemiol. 109, 186-204, 1979. BERLINGER, W.G. and POTTER, J.F. Visual impairment as a predictor of cognitive decline. Paper presented at annual meeting of the American Geriatrics Society. Anaheim, CA, 1988 May 19--22.

456

J.F. POTTER

BERLINGER, W.G. and POTTER, J.F. Low body mass index in demented outpatients. J. Am. Geriatr. Soc. 39, 973-978, 1991. BRANCH, L.G., GURALNIK, J.M., FOLEY, D.J., JOHOUT, F.J., WETLE, T.T., OSTFELD, A., and KATZ, S. Active life expectancy for 10,000 Caucasian men and women in three communities. J. Gerontol. 46, MI45-MI50, 1991. BROWN, L., POTTER, J.F., and FOSTER, B.G. Caregiver burden should be evaluated during geriatric assessment. J. Am. Geriatr. Soc. 38, 455-460, 1990. BURKE, W.J., HOUSTON, M., BOUST, S.J., and ROCCAFORTE, W.H. Use of the Geriatric Depression Scale in dementia of the Alzheimer type. J. Am. Geriatr. Soc. 37:856-860, 1989. FISCHBACH, R. Early identification of demented persons in the community. In: Alzheimer Disease, Current Research in Early Diagnosis, Becker, R.E. and Giacobini, E. (Editors), pp. 49-74, Taylor and Francis, New York, NY, 1990. FOLSTEIN, M.F. and FOLSTEIN, S.E. Mini-Mental State: A practical method for grading the cognitive state of patients for the clinician. J. Psychiatr. Res. 12, 189, 1975. GOODWlN, L.S., BERLINGER, W.G., and ROBERTSON, B.J. Do outpatients with mild Alzheimer's disease lose weight? Poster presented at AGS/AFAR annual meeting. Chicago, IL, 1991 May. HUGHES, C.P., BERG, L., DANZIGER, W.L., COBEN, L.A., and MARTIN, R.L. A new clinical scale for the staging of dementia. Br. J. Psychiatry 145, 566, 1982. KATZ, S., BRANCH, L.G., BRANSON, M.H., PAPSIDERO, J.A., BECK, J.C., and GREER, D.S. Active life expectancy. N. Engl. J. Med. 309, 1218-1224, 1983. LAWTON, M.P. The Philadelphia Geriatric Center Morale Scale: A revision. J. Gerontol. 30, 85, 1975. LAWTON, M.P. and BRODY, E.M. Assessment of older people: Self-maintaining and instrumental activities of daily living. Gerontology 9, 179, 1969. LINN, B.S., L1NN, M.W., and GUREL, L. Cumulative illness rating scale. J. Am. Geriatr. Soc. 16, 622, 1968. PETERS, C.A., POTTER, J.F., and SCHOLER, S.G. Hearing impairment as a predictor of cognitive decline in dementia. J. Am. Geriatr. Soc. 36, 981-986, 1988. ROCCAFORTE, W.H., BURKE, W.J., and BAYER, B.L. Telephone utility of SPMSQ and a shortened MMSE. Paper presented to the 143rd annual meeting of the American Psychiatric Association. 1990 May; New York. ROCCAFORTE, W.H., BURKE, W.J., BAYER, B.L., and WENGEL, S.P. The reliability of a phone version of the Mini-Mental State Examination. J. Am. Geriatr. Soc., in press. ROWE, J.W. Toward successful aging: Limitation of the morbidity associated with normal aging. In: Principles o f Geriatric Medicine and Gerontology, Hazzard, W.R., Andres, R., Bierman, E.L., and Bluff, J.P. (Editors), pp. 138-141, McGraw-Hill, New York, NY, 1990. RUBENSTEIN, L.Z. Geriatric assessment: An overview of its impacts. Clin. Geriatr. Med. 3(1): 1-16, 1983. RUBENSTEIN, L.Z., STOCK, A.E., SIU, A., and WIELAND, D. Impacts of geriatric evaluation and management programs on defined outcomes: Overview of the evidence. J. Am. Geriatr. Soc. 39, 85-165, 1991. SCHOLER, S.G., POTTER, J.F., and BURKE, W.J. Does the Williams Manual Test predict service use among community dwelling elderly. J. Am. Geriatr. Soc. 38, 767-772, 1990. SHOCK, N.W., GREULICH, R.L., ANDRES, R., ARENBURG, D., COSTA, P.T., LAKATTA, E.G., and TOBIN, J.D. (Editors). Normal Human Aging: The Baltimore Longitudinal Study o f Aging. NIH Publication No. 84-2450. U.S. Government Printing Office, Washington, DC, 1984. STORANDT, M., BOTWlNICK, J., DANZIGER, W.L., BERG, L., and HUGHES, C.P. Psychometric differentiation of mild senile dementia of the Alzheimer's type. Arch. Neurol. 41,497, 1984. WALDMAN, E. and POTTER, J.F. A prospective evaluation of the Cumulative Illness Rating Scale. Aging Clin. Exp. Res. 4, 171-178, 1992. WILLIAMS, M.E. Identifying the older person likely to require long-term care services. J. Am. Geriatr. Soc. 35, 761, 1987. WILLIAMS, M.E. and MITCHELL, C.A. Timed Manual Performance in Older Persons: A Handbook and Training Guide. Authors, Chapel Hill, NC, 1984. WILLIAMS, M.E., WILLIAMS, F., ZIMMER, J.G., HALL, W.J., and PODGARSKI, C.A. How does the

OUTPATIENT GERIATRIC ASSESSMENT

457

team approach to outpatient geriatric evaluation compare with traditional care: A report of a randomized controlled trial. J. Am. Geriatr. Soc. 35, 1071-1078, 1987. YESAVAGE, J.A., BRINK, J.L., ROSE, T.L., and LUM, O. Development and validation of a geriatric depression screening scale: A preliminary report. J. Psychiatr. Res. 17, 37, 1983. ZARIT, J.M. Family role, social supports, and their relation to caregivers' burden. Paper presented at meeting of the Western Psychological Association. Sacramento, CA, 1982 October. ZARIT, S., ZARIT, J., and TODD, P. Subjective burden of husbands and wives as caregivers: A longitudinal study. Gerontology 26, 260, 1986. ZARIT, S.H., REEVER, K., and BACH-PETERSON, J. Relatives of the impaired elderly: Correlates of feeling of burden. Gerontology 20, 649, 1980. ZEMCOV, A., BARCLAY, L.L., BRUSH, D., and BLASS, J.P. Computerized data base for evaluation and follow-up of demented outpatients. J. Am. Geriatr. Soc. 32, 801, 1984.