Factors Associated with Symptoms of Depression Among Informal Caregivers of Demented Elders in the Community

Transcription

1 Copyright 1998 by The Cerontological Society of America The Cerontologist Vol. 38, No. 2, This study assessed variables associated with increased severity of depressive symptoms among 321 informal caregivers of demented subjects living in the community identified by a Canadian national population-based survey. Measures included patient and caregiver demographic variables; patient behavioral disturbance, activities of daily living, and cognitive impairment; living arrangement, and duration of disease. Depressive symptoms were measured with the Center for Epidemiologic Studies Depression Scale (CES-D). Multiple regression analyses indicated that a higher CES-D score was significantly associated with three caregiver characteristics (being a spouse or child of the patient, self-identified ethnicity other than English and French Canadian, and lower education), and two patient characteristics (greater behavioral disturbance and moderate to severe functional impairment). Key Words: Canada, Health, Characteristics, Aging, Regression Factors Associated with Symptoms of Depression Among Informal Caregivers of Demented Elders in the Community Garbis Meshefedjian, MPH, MSc, 12 Jane McCusker, MD, DrPH, 13 Francois Bellavance, PhD, 13 and Mona Baumgarten, PhD 4 A substantial body of research has documented a variety of mental and physical health problems among informal caregivers of persons with dementia (Baumgarten, 1989). One such problem, depression, has been associated with caregiving for persons with dementia (Baumgarten et al v 1992; Dura, Haywood- Niler, & Kiecolt-Glaser, 1990; Kiecolt-Glaser, Dura, Speicher, Trask, & Claser, 1991; Schulz & Williamson, 1991). Depression can be a serious problem that interferes with the ability to function normally. Even symptoms of depression that do not meet conventional criteria for depressive disorders may have significant effects on functional status (Broadhead, Blazer, George, & Tse, 1990; Wells et al., 1989). Furthermore, depression is often treatable using pharmacological and/or psychosocial approaches (Paykel & Priest, 1992). However, physicians often fail to recognize depressive disorders in general medical settings, hence these problems are often not treated. The identification and treatment of depression in caregivers has the potential to improve their functioning and ability to continue to provide care, thus reducing the need for institutional care for those with dementia. 'Department of Epidemiology and Biostatistics, McCill University, Montreal, Quebec, Canada. 'Department of Public Health, Montreal General Hospital, Montreal, Quebec, Canada. 3 St. Mary's Hospital Center, Montreal, Quebec, Canada. Address correspondence to Dr. Jane McCusker, Department of Clinical Epidemiology and Community Studies, St. Mary's Hospital Center, 3830 Lacombe Ave., Montreal (QC) H3B1M5, Canada. 'Department of Biostatistics and Epidemiology, University of Pennsylvania, Philadelphia, PA. In order to plan appropriate services for informal caregivers, it is important to know the factors that are associated with increased severity of symptoms of depression. A number of previous studies have addressed this question, but there are methodological limitations and the results are not entirely consistent (Baumgarten, 1989). Perhaps most consistent has been the association between increased behavioral disturbance in the demented patient and increased psychological distress in the caregiver (Deimling & Bass, 1986; Haley, Brown, & Levine, 1987; O'Connor, Pollitt, Roth, Brook, & Reiss, 1990). Other measures of cognitive and functional deterioration have also been linked to caregiver distress (Gallagher-Thompson, Brooks, Bliwise, Leader, & Yesavage, 1992; Hadjistavropoulos, Taylor, Tuokko, & Beattie, 1994; Kurtz, Kurtz, Given, & Given, 1995). Increased clinical deterioration tends to be associated with older patient age and a longer duration of disease. Characteristics of caregivers that have been found to be associated with depression, such as relationship to the demented individual and living arrangement, may be interrelated (George & Gwyther, 1986). Similarly, older caregivers are less likely to be working and less likely to be well educated. As the majority of the research on the risk factors for caregiver depression and psychological distress has been based on nonrepresentative samples (Baumgarten, 1989), the generalizability of the identified risk factors is not known. To our knowledge, only one previous study by O'Connor and colleagues (1990) from the United Kingdom investi- Vol. 38, No. 2,

2 gated the caregivers of a community sample of demented elders. In that study, the frequency and severity of problems faced by caregivers and a measure of strain increased with the degree of dementia. Wives reported more problems and strain than husbands, and coresident children were under greater strain than children who lived independently. Since a specific measure of depression was not used, it is not possible to make a direct comparison with our study results. A second methodological issue raised by previous research on caregivers is the interrelationship of many patient and caregiver risk factors. To better focus prevention and management interventions, it would be helpful to separate caregiver risk factors (age, relationship to patients, etc.) from patient risk factors (severity and duration of dementia). Simultaneous investigation of these risk factors is also needed to determine the relative importance of patient and caregiver variables. These concerns led us to use data collected in a large, population-based, national survey of dementia, the Canadian Study of Health and Aging (CSHA), to investigate associations between patient and caregiver variables and the severity of depressive symptoms among informal caregivers. The CSHA recently estimated that the prevalence of dementia was 8% in the Canadian population aged 65 years and older, with markedly increased rates at older ages (Canadian Study of Health and Aging Working Group, 1994). In the community-based sample of the CSHA, the prevalence of depression (using the conventional threshold of 16 on the Center for Epidemiologic Studies Depression Scale [CES-D]) was 25.9% in the caregivers of patients with dementia versus 13.8% among caregivers of nondemented patients (Canadian Study of Health and Aging, 1994). These rates are higher than those reported from older primary care samples (Callahan et al., 1994; Stephenson-Cino, Steiner, Krames, Ryan, & Huxley, 1992) but lower than that reported in a caregiver sample (Baumgarten et al., 1992). Methodology Sample Selection The study methods used in the CSHA are described in detail elsewhere (Canadian Study of Health and Aging Working Group, 1994). Briefly, the study sampled 9,008 community-residing people aged 65 and older in 36 urban centers and surrounding rural areas across Canada in The community samples were drawn from the computerized records of the provincial health insurance plans, except in Ontario where the Enumeration Composite Record (an aggregated list based on election and other municipal records) was used. Random samples were drawn in the age strata 65-74, 75-84, and 85 and older; the sampling fractions in the two latter age strata were 2 and 2.5 times, respectively, that of the 65 to 74 age group. Subjects in the community were screened for cognitive impairment, using the Modified Mini- Mental State Exam (3MS) (Teng & Chui, 1987). Those who screened positive (below the cut-off point of 78) and a sample of those who screened negative were referred for clinical examinations that consisted of neuropsychological test batteries (for subjects with a score of 50 or more on the 3MS exam) and physician assessments of mental, physical, and neurological status. The final consensus diagnosis of dementia was made at a case conference. The primary caregiver was identified through an interview with the elderly subject and, in most instances, through discussions with other family members. The primary caregiver was defined as the person who was ordinarily the most responsible for the day-to-day decision making and care of the index subject. Primary caregivers could be formal (paid) or informal (unpaid). When both formal and informal caregivers were involved, the informal caregiver was interviewed if he or she was in contact with the index subject at least once a week. Interviews were conducted with the caregiver, usually in his or her home, to assess the patient's ability to perform activities of daily living, the presence of behavioral problems in the patient, and depressive symptoms in the caregiver. The CSHA identified 395 subjects with dementia in the community; seven had no primary caregiver who met the selection criteria, 39 caregivers could not be contacted or declined to participate, and 22 subjects had formal caregivers only. Among the sample of 327 informal caregivers interviewed, six were excluded because their CES-D score was missing. Thus, the sample for this report comprised 321 caregivers. Study Variables Dependent Variable. Depressive symptoms were assessed by the CES-D (Radloff, 1977; Radloff & Teri, 1986). The CES-D comprises 20 items, each rated on a 4-point response scale corresponding to the frequency of the symptom in the preceding week. We analyzed the CES-D score as a continuous variable as we were interested in identifying risk factors for greater severity of depressive symptoms over the entire range of severity. The possible range of CES-D scores is 0-60, with a higher score indicating more severe symptoms. Independent Variables.We considered two categories of independent variables: caregiver demographic characteristics and patient variables. Caregiver demographic characteristics included caregiver age, gender, years of education, marital status, ethnic background, current employment, relationship to the patient, and living with the patient. Ethnicity was measured by the response to an open-ended question about the caregiver's ethnic or cultural background. Patient variables included patient age, gender, years of education, duration of disease, and the three health status measures described below. We did not analyze patient ethnicity separately because 87% of patients had the same ethnicity as their caregiver. 248 The Gerontologist

3 The Activities of Daily Living (ADL) questionnaire was developed for the Older American Resources and Services Project (Fillenbaum, 1988). The caregiver rated the patient's ability to perform 14 tasks with or without assistance. An algorithm was used to arrive at a rating on a five-point ordinal scale: no, mild, moderate, severe, or complete impairment. This scale was analyzed in three categories: no/mild impairment; moderate impairment; and severe/complete impairment. The Dementia Behavior Disturbance (DBD) Scale, (Baumgarten, Becker, & Cauthier, 1990), administered by an interviewer to the caregiver, evaluates behavioral disturbances in persons with dementia. It includes 28 items, such as passivity, agitation, aggressiveness, sexual misdemeanor, and disturbances in eating and diurnal rhythm. Each item is rated on a 5- point scale ranging from "never" (score of 0) to "all the time" (score of 4) during the preceding week. The theoretical scale range is 0 to 112, with a higher score indicating more behavioral disturbance. The Modified Mini-Mental Status (3MS) Scale (Teng & Chui, 1987) ranges from 0 to 100, with higher scores indicating better cognitive performance. In this study, scores were missing for 28 subjects either beimpairment (such as deafness), cause of sensory mental retardation, or current illness (such as stroke). Statistical Methods We examined the univariate associations between the CES-D score and the independent variables using one-way analysis of variance. For these analyses we categorized age and education into conventional groupings; tertiles were used for the 3MS and DBD scores, and for duration of disease. Multiple regression models were then developed separately for the caregiver and patient characteristics and also considering the characteristics of both together. The following interaction terms were investigated for potential inclusion in the model: caregiver relationship with caregiver age; caregiver relationship with caregiver gender; and patient age and patient gender with each of MMS, DBD and ADL. Using the "best subset" method (Neter, Wasserman, & Kutner, 1990), we identified three or four potential models with the lowest values on Mallow's C P and highest adjusted R 2 values. The choice of the final model in each category was based on conceptual coherence and parsimony. Finally, a multiple regression model was built combining the caregiver and patient characteristics. Although unequal sampling ratios were used in the three age strata, we did not perform differential weighting of the age strata because the oversampling was applied to the original sample of elders and not to the caregivers who are the subject of this report. The oversampling limits the generalizability of some of our results, because they are based on a sample of caregivers caring for patients who are, on average, somewhat older than those in the target population. However, it does not affect the main results of this research, i.e., the evaluation of associations between the independent variables and the CES-D score. All statistical analyses were performed using Statistical Analysis System (SA) software (SAS Institute Inc., 1987). Results were considered to be statistically significant for p values of less than Results Our study sample comprised 321 informal caregivers of demented older adults who lived in the community. The majority of caregivers were aged 50 or older, female, and married; identified themselves as English Canadian; were the spouse or child of the patient; resided with the patient; and were not employed (Table 1). Most of the patients in the sample were aged 75 and older, female, had less than a secondary education, and had an average disease duration of 31.6 months. Clinically, half of the patients were severely or completely impaired; their mean DBD and 3MS scores were 16.7 and 57.7, respectively. The mean CES-D score for our sample of caregivers was (Table 1). CES-D scores were significantly associated with caregiver characteristics (age, education, ethnicity, relation to patient, occupational status, and living arrangement) and patient characteristics (ADL, DBD, and 3MS). CES-D scores were higher among caregivers who were older, had fewer years of education, were spouses of the index patient, were not currently employed, resided with the demented patient, and identified themselves as neither English nor French Canadian. Caregivers were also more depressed if they cared for patients who had greater behavioral disturbance and greater functional or cognitive impairment. Multivariate Analysis Regression analyses of caregiver characteristics resulted in a model that included gender, relationship to the patient, and ethnicity (Table 2, Model 1). Caregiver CES-D scores were higher for female care providers who were either the spouse or child of the patient and had self-identified ethnicity other than English or French Canadian. Caregiver age, education, employment status, and residence were not significantly associated with the CES-D score in the presence of these three variables, and inclusion of these variables did not improve the fit of the model. As regards patient characteristics, the caregiver CES-D score was significantly associated with age of patient, ADL status, duration of disease, and DBD score (Table 2, Model 2). CES-D scores were higher if patients were younger, had moderate to severe ADL impairment, disease duration between one to three years, and greater behavioral disturbance. In the model that combined caregiver and patient characteristics, the regression coefficients for caregiver gender and for patient age and disease duration were smaller and no longer significantly associated with the CES-D score (Table 2, Model 3). However, caregiver ethnicity, relationship to the patient, and education, as well as patient ADL status and DBD score, were significantly associated with the CES-D score. This model explained 29.4% of the Vol. 38, No. 2,

4 Table 1. Distribution of Caregiver and Patient Characteristics With Mean and Standard Deviation (SD) CES-D Scores (N= 321) Variable N M Total Caregiver Characteristics Age (years) Mean ±SD = 61.5 ±13.3 Gender Male Female Education Level (years) Primary school or less (0-8) Secondary school or less (9-12) More than secondary (13-24) Mean ±SD = 11.1 ±3.9 Marital Status Never married Married Divorced/separated Widowed Ethnic Background English Canadian French Canadian Other Relationship to the Patient Spouse Child Other Occupational Status Not working Working Living Arrangement Not with patient With patient Patient Characteristics Age (years) Mean ±SD = 82.3 ± 6.2 Gender Male Female Education (years) Less or equal to primary (0-8) Less or equal to secondary (9-12) More than secondary (13-20) Mean ±SD = 8.0 ± 4.1 Disease Duration (months) or more Mean ±SD = 31.6 ±50.3 Activities of Daily Living No/mild impairment Moderate impairment Severe/complete impairment CES-D' SD p value b (continued on next page) 250 The Gerontologist

5 Table 1. Distribution of Caregiver and Patient Characteristics With Mean and Standard Deviation (SD) CES-D Scores (N= 321) (Continued) CES-D" Variable N M SD p value" Patient Characteristics (continued) Dementia Behavior Disturbance Score Mean ±SD = 16.7 ±13.8 Modified Mini-Mental Status Score Mean ±SD = 57.7 ± "Center for Epidemiologic Studies Depression Scale. "By one-way analysis of variance, excluding missing values. Table 2. Regression Coefficients (and Standard Errors) for Regression Models of Caregiver and Patient Variables of Caregiver CES-D* Scores Model 1 R 1 = 12.9% (n = 285) Model 2 A? 2 = 16.6% (n = 285) Model 3 R 2 = 29.4% (n = 283) Caregiver Characteristics Gender Male" Female Ethnicity English French Other" Relationship Spouse Child Other" Education (years) 2.55(1.26)* -2.38(1.34)*** -1.57(1.51) 7.80(1.35)** 4.49(1.28)** (1.26)* -2.03(1.40) 6.77(1.30)** 3.39(1.17)** (0.13)* Patient Characteristics Age Activities of Daily Living Mild/no impairment Moderate impairment Severe/complete impairment Disease Duration (months) 0-12" or more Dementia Behavior Disturbance (DBD) (0.08)* 3.36(1.44)* 2.58(1.41)*** 2.52(1.27)* 0.51 (1.37) 0.21 (0.04)** 3.59 (1.34)** 2.84(1.30)* 2.32(1.19)*** 1.20(1.28) 0.22 (0.04)** Center for Epidemiologic Studies Depression Scale. "Reference category for the corresponding indicator variable. Age and DBD are considered on continuous scales. *p < 0.05; **p < 0.01; ***0.05 p variance in caregiver depression. None of the interaction terms of interest were significant in any of the three models presented. Discussion The main strength of this study lies in our use of data from a large, population-based survey. This data provided an opportunity to examine the joint effects of several risk factors for depression that have previously been studied in smaller, more selected samples. Another strength is the inclusion of a wide range of pertinent patient and caregiver variables. The principal study limitation is its cross-sectional design; causality cannot be inferred from the associations we examined. Secondly, because of the way the informal caregiver was identified, in some cases a formal caregiver may have been more responsible than the informal caregiver for the dayto-day care of the index subject. Third, absence of a Vol.38, No. 2,

6 measure of the physical health status of the caregiver in our data base may have introduced some confounding. The results of our multivariate analyses suggest that the best subset of variables associated with increased symptoms of depression in informal caregivers of demented patients in the community comprises two patient characteristics (greater behavioral disturbance and functional impairment) and three caregiver characteristics (being the spouse or child of the patient, self-identified ethnicity other than French or English Canadian, and lower level of education). Previous research has reported that behavioral disturbance is a key predictor of caregiver distress (Deimling & Bass, 1986; Haley et al., 1987; O'Connor et al., 1990) and our results confirm this association. This association may be due to the greater burden of care or to the inability of the caregiver to predict or control these behaviors. Greater impairment in activities of daily living in the patient was also associated with a higher CES-D score in the caregiver in our study. A longitudinal study of cancer patients and their caregivers found that higher levels of caregiver depression were associated with declines in patient mobility (Kurtz et al., 1995). Decreased functional independence of patients increases the physical burden of care that, in turn, can reduce the psychological well-being of the caregiver. Another explanation for these associations is a reporting bias whereby depressed caregivers are more likely to report abnormal behaviors or disability in patients. The caregiver's relationship to the patient was an important correlate of depression; spouses had the highest average CES-D scores, followed by children. Several authors have reported more symptoms of stress or depression in spouse caregivers compared to children or other informal caregivers (Cantor, 1983; George & Gwyther, 1986; Baumgarten et al., 1992). These relationships have been attributed to the concept of centrality (Cantor, 1983): the closer the bond, the more stressful the caregiver role. On the other hand, some investigators have found no significant difference in measures of caregiving burden between child and spouse caregivers (Robinson, 1983; Zarit, Reever, & Bach-Peterson, 1980), or have reported that children caring for their parents suffer more distress than spouses (Johnson & Catalano, 1983). These discrepancies may be due to methodological factors, such as differences in the sample selection, in the choice of the outcome variable (depression, burden, distress, strain, stress), or the influence of other caregiver or patient characteristics, such as living arrangement (George & Gwyther, 1986). Our finding of a significant association between ethnicity and depression is intriguing, and, to our knowledge has not been reported previously. However, the interpretation is complicated, and the study design does not allow us to explore this finding in detail. The structure of informal caregiver networks has been found to differ by ethnocultural factors (Thornton, White-Means, & Choi, 1993) and may affect, for example, the availability of informal support for the caregiver and opportunities for the sharing of caregiving duties. Since fluency in either English or French was a criterion for study eligibility, our study sample is limited to those members of other language groups who have a working understanding of either of these languages. Therefore, ethnicity of the patient is another concern that influences interpretation of caregiver depression. For instance, if the patient, like the caregiver, is neither English nor French Canadian, the observed association might be explained by barriers to care encountered by ethnic minorities, and by comprehension problems that affect the responses of those whose native tongue was neither French nor English (Callahan & Wolinsky, 1994). There may also be ethnocultural differences in the way stress is experienced and expressed. For example, if the patient is identified as English or French Canadian and is of a different ethnic background than the caregiver, the caregiver's depression might be explained by cultural differences and communication problems with the patient. Ethnic groups may differ in the way they respond to the CES-D, as Callahan & Wolinsky demonstrated in their study of differences between Black and White elderly primary care patients in the United States. In contrast, Roberts (1980) reported that the quality and reliability of CES-D results were similar in samples of Anglo, Black, and Mexican- American subjects in the U.S. Clearly, the role of ethnocultural factors in the informal care of elderly patients and in psychological distress among caregivers deserves further research. In one of our multivariate models (Model 3), we found that lower caregiver education levels were significantly associated with increased depression. Such a finding has been reported in other studies of depression in older adults (e.g., Stephenson-Cino et al., 1992; Kurlowicz, 1993), but not, to our knowledge, in previous studies of caregivers. Possible explanations of this association may include lack of monetary resources (Kurlowicz, 1993). Interventions for informal caregivers of people with dementia should include efforts to prevent, detect, and treat depression. Interventions that reduce the psychological and physical burdens of caregiving may also be needed, such as home care services, respite care, and day care. Caregivers at risk for more severe depressive symptoms include spouses and children of the demented person and those caring for patients with severe behavioral disturbance and functional impairment. More research is needed to determine more discriminating risk factors; a substantial amount of variance in caregiver depression was not explained by the risk factors in our model. At present, clinicians caring for individuals with these characteristics should be alert to depressive symptoms to determine whether they interfere with caregiving. Although the effectiveness of interventions to alleviate caregiver distress has not been demonstrated consistently, respite interventions and individual psychosocial interventions appear to be moderately effective 252 The Gerontologist

7 (Knight, Lutzky, & Macofsky-Urban, 1993). In the future, it may be reasonable to target high-risk caregivers for these services and to evaluate the effectiveness of interventions in this population. References Baumgarten, M. (1989). The health of persons giving care to the demented elderly: a critical review of the literature. Journal of Clinical Epidemiology, 42, Baumgarten, M., Battista, R. N., Infante-Rivard, C, Hanley, J. A., Becker, R., & Cauthier, S. (1992). The psychological and physical health of family members caring for an elderly person with dementia. Journal of Clinical Epidemiology, 45, Baumgarten, M., Becker, R., & Gauthier. S. (1990). Validity and reliability of the Dementia Behaviour Disturbance scale. Journal of the American Geriatrics Society, 38, Broadhead, E. W., Blazer, D. C, George, L K., & Tse, C. K. (1990). Depression, disability days, and days lost from work in a prospective epidemiologic survey. Journal of the American Medical Association, 264, Callahan, C. M., & Wolinsky, F. D. (1994). The effect of gender and race on the measurement properties of the CES-D in older adults. Medical Care, 32, Callahan, C. M., Hendrie, H. C, Dittus, R. S., Brater, C. D., Hui, S. L, Tierney, W. M. (1994) Depression in late life: The use of clinical characteristics to focus screening efforts. Journal of Gerontology, 49, M9-M14. Canadian Study of Health and Aging. (1994). Patterns of caring for people with dementia in Canada. Canadian Journal on Aging, 13, Canadian Study of Health and Aging Working Group. (1994). The Canadian Study of Health and Aging: Study methods and prevalence of dementia. Canadian Medical Association Journal, 150, Cantor, M. H. (1983). Strain among caregivers: A study of experience in the United States. The Gerontologist, 23, Deimling, G. T., & Bass, D. M. (1986). Symptoms of mental impairment among elderly adults and their effects on family caregivers. Journal of Gerontology, 41, Dura, J. R., Haywood-Niler, E., & Kiecolt-Glaser, J. K. (1990). Spousal caregivers of persons with Alzheimer's and Parkinson's disease dementia: A preliminary comparison. The Gerontologist, 30, Fillenbaum, G. G. (1988). Multidimensional functional assessment of older adults: The Duke Older Americans Resources and Services procedures. Hillsdale, NJ: Lawrence Erlbaum Associates. Gallagher-Thompson, D., Brooks, J. O. Ill, Bliwise, D., Leader, J., & Yesavage, J. A. (1992). The relations among caregiver stress, "sundowning" symptoms, and cognitive decline in Alzheimer's disease. Journal of the American Geriatrics Society, 40, George, L. K., & Gwyther, L. P. (1986). Caregiver well-being: A multidimensional examination of family caregivers of demented adults. The Gerontologist, 26, Hadjistavropoulos, T., Taylor, S., Tuokko, H., & Beattie, B. L (1994). Neuropsychological deficits, caregivers' perception of deficits and caregiver burden. Journal of the American Geriatrics Society, 42, Haley, W. E., Brown, S. L., & Levine, E. G. (1987). Experimental evaluation of the effectiveness of group intervention for dementia caregivers. The Gerontologist, 27, Johnson, C. L., & Catalano, D. J. (1983). A longitudinal study of family supports to impaired elderly. The Gerontologist, 23, Kiecolt-Glaser, J. K., Dura, J. R., Speicher, C. E., Trask, O. J., & Glaser, R. (1991). Spousal caregivers of dementia victims: Longitudinal changes in immunity and health. Psychosomatic Medicine, 53, Knight, B. G., Lutzky, S. M., & Macofsky-Urban F. (1993). A meta-analytic review of interventions for caregiver distress: Recommendations for future research. The Gerontologist, 33, Kurlowicz, L. H. (1993). Social factors and depression in late life. Archives of Psychiatric Nursing, 7, Kurtz, M. E., Kurtz, J. C, Given, C. W., & Given, B. (1995). Relationship of caregiver reactions and depression to cancer patients' symptoms, functional states and depressiona longitudinal view. Social Sciences and Medicine, 40, Neter, J., Wasserman, W., & Kutner, M. H. (1990). Applied linear statistical models: Regression, analysis of variance, and experimental designs (3rd ed.). Boston: Richard D. Irwin. O'Connor, D. W., Pollitt, P. A., Roth, M., Brook, C. P., & Reiss, B. B. (1990). Problems reported by relatives in a community study of dementia. British Journal of Psychiatry, 156, Paykel, E. S., & Priest, R. G. (1992). Recognition and management of depression in general practice: Consensus statement. British Medical Journal, 305, Radloff, L. S. (1977). The CES-D scale: A self-report depression scale for research in the general population. Applied Psychological Measurement, 7, Radloff, L S., & Teri, L. (1986). Use of the Center for Epidemiologic Studies- Depression scale with older adults. In T. L. Brink (Ed.), Clinical gerontology: A guide to assessment and intervention (pp ). New York: Haworth Press. Roberts, R. E. (1980). Reliability of the CES-D scale in different ethnic contexts. Psychiatric Research, 2, Robinson, B. C. (1983). Validation of a caregiver strain index. Journal of Gerontology, 38, SAS Institute Inc. (1987). SAS/STAT guide for personal computers (Version 6.04 ed.). Cary, NC: SAS Institute Inc. Schulz, R., & Williamson, G. M. (1991). A 2-year longitudinal study of depression among Alzheimer's caregivers. Psychology and Aging, 6, Stephenson-Cino, P., Steiner, M., Krames, L., Ryan, E. B., & Huxley, G. (1992). Depression in elderly persons and its correlates in family practice: A Canadian study. Psychological Reports, 70, Teng, E. L, & Chui, H. C. (1987). The Modified Mini-Mental State (3MS) Examination. Journal of Clinical Psychiatry, 48, Thornton, M. C, White-Means, S. I., & Choi, H. K. (1993). Sociodemographic correlates of the size and composition of informal caregiver networks among frail ethnic elderly. Journal of Comparative Family Studies, 24, Wells, K. B., Stewart, A., Hays R. D., Brunam, A. M., Rogers, W., Daniels, M., Berry, S., Greenfield, S., & Ware, J. (1989). The functioning and wellbeing of depressed patients: Results from the Medical Outcomes Study. Journal of the American Medical Association, 262, Zarit, S. H., Reever, K. K., & Bach-Peterson, J. (1980). Relatives of the impaired elderly: Correlates of feelings of burden. The Gerontologist, 20, Received July 30,1996 Accepted September 19,1997 Vol. 38, No. 2,