| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
Correspondence: Ian McDowell, Department of Epidemiology & Community Medicine, University of Ottawa, 451 Smyth Rd., Ottawa, Ontario K1H 8M5, Canada. E-mail: mcdowell{at}uottawa.ca.
Decision Editor: Laurence G. Branch, PhD
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionReferences |
|---|
Key Words: Alzheimer's disease • Caregiver • Home care • Epidemiology • Canada
The aging of populations will result in growing numbers of people whose health prevents them from caring for themselves. To plan systems of community support services, we must understand the types of care that can realistically be provided for persons with chronic conditions by their families and friends, and we must anticipate the human cost of such care. Intuitively, caring for a person with dementia at home must place the caregiver at high risk of adverse health consequences, for dementia caregiving often requires 24-hr surveillance over a period of years and represents a situation of chronic strain. There is a consensus that caregiving commonly has adverse health consequences (Chappell and Penning 1996
; Schulz and Beach 1999; Schulz, O'Brien, Bookwala, and Fleissner 1995; Schultz, Vistainer, & Williamson, 1990), but it also appears that these are frequently not as extreme as might be supposed, and there is considerable variability in the response (Taylor, Ford, and Dunbar 1995; Walker and Pomeroy 1996). Although an association has consistently been found between caregiving and depression, the impact on the caregiver's physical health is less clear (Baumgarten et al. 1994; Grafstrom and Winblad 1995; Schulz et al. 1995; Shaw et al. 1997).
Research has therefore addressed moderating variables in the association between caregiving and health of the caregiver (Vitaliano, Schultz, Kiecolt-Glaser, & Grant, 1997). The results suggest that factors such as the duration of providing care or the amount of care provided appear to be only modestly related to strain (Schulz et al. 1995; Townsend, Noelker, Deimling, and Bass 1989); there is no simple cause-and-effect model that links care provided and strain experienced. Instead, caregiving is a dynamic situation involving adaptive mechanisms. For example, stressed caregivers may be expected to seek alternative caregiving solutions; families may be able to choose who assumes the caregiver role, and social support can be martialed to lessen the association between caregiving and personal strain. The personality and resiliency of caregivers are also relevant (Monahan and Hooker 1995), as is their relationship to the care recipient (Meshefedjian, McCusker, Bellavance, and Baumgarten 1998). Situational factors may also mitigate the adverse impact of caregiving; studies have examined institutional placement (Matsuda, Hasebe, Ikehara, Futatsuya, and Akahane 1997) and the death of the care recipient (Mullan 1992). There remains considerable uncertainty; for example, is strain higher among long-term caregivers of people with dementia or among caregivers of those recently diagnosed with dementia, owing to the adaptation this requires? Does admission to institutional care or the death of the care recipient relieve the caregiver, or are such changes in themselves stressful?
Since 1991, we have undertaken a longitudinal study of the epidemiology of dementia, including Alzheimer's disease, in Canada (Canadian Study of Health and Aging Working Group 1994a, Canadian Study of Health and Aging Working Group 1994b; McDowell, Hill, and Lindsay 2001). The Canadian Study of Health and Aging (CSHA) cohort offers the opportunity to examine the impact of providing care on the caregiver's health in a population-based, representative sample. Here we use longitudinal data over a 5-year period to analyze the impact on caregivers of the changing health status of their care recipients, classified into four main groups. The first group included elderly people who had dementia or developed it over the 5-year period; the second group included those who were cognitively normal but physically frail; a third group included those who remained cognitively and physically healthy; and a fourth group included those who died during the follow-up period.
This report has three objectives: (a) to describe the typical evolution in the cognitive and physical health status of these groups of elderly people over 5 years, summarizing their needs for care; (b) to link changes in the health status and residence of care recipients with the amount of assistance provided by unpaid and paid caregivers; and (c) finally, to examine the extent to which changes in care recipients' health were reflected in adverse changes in their caregivers' feelings of strain and health. This analysis examines the modifying effect of situational changes, including admitting the care recipient to an institution and the care recipient's eventual death.
|
Methods |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
). In 1991, the baseline study (CSHA-1) drew nationally representative samples of Canadians aged 65 or older living in the community (n = 9,008) and in institutions (n = 1,255; Canadian Study of Health and Aging Working Group 1994a; McDowell, Aylesworth, Stewart, Hill, and Lindsay 2001). Participants were screened for cognitive problems using the Modified Mini-Mental State Examination (3MS; Teng and Chui 1987), and those who screened positive underwent a detailed clinical and neuropsychological examination (Graham et al. 1996; McDowell, Stewart, et al. 2001); diagnoses of dementia were based on established criteria (Graham et al. 1996). The primary caregivers of those diagnosed with dementia were then interviewed (n = 1,048), as were people responsible for providing care to a comparison group of 638 people without dementia randomly selected from study participants who screened negatively (3MS scores > 77). Of the 1,686 caregivers at CSHA-1, 87%, or 1,472, were family members or friends. Formal, or paid, caregivers (n = 214) were interviewed when a family member or close friend could not be identified as a primary caregiver. The objectives were to describe patterns of care provided to people with dementia in Canada and to perform cross-sectional analyses of the association between caregiving and caregiver health. The 1996 follow-up study (CSHA-2) repeated the same screening and clinical assessments. In addition, we sought to recontact the same 1,472 informal caregivers in order to record the care they were providing and their current health status. Caregivers were included even if the care recipient had died, but no new caregivers were added to the sample. The diagnostic methods and interviews with the caregivers were closely comparable at CSHA-1 and CSHA-2.
Measures
Care Recipients
Care recipients were grouped into those with dementia and those who were cognitively normal, based on the clinical and neuropsychological examinations; for this report, all subtypes of dementia were grouped together. The physician rated the severity of dementia using the criteria given in the Diagnostic and Statistical Manual of Mental Disorders (3rd ed., Rev.; American Psychiatric Association 1987). The classification of cognitive normality was based on a person's having screened negative (scores of 78 or above on the 3MS test). For this presentation, 42 people who were diagnosed at the CSHA-2 clinical examination as falling into an intermediate category of cognitive impairment that did not reach the criteria for dementia were discarded from the analysis. Their mean 3MS score was 72 (range = 39–88; SD = 11.0).
The interview with the caregiver collected information on the care recipient's need for assistance in basic and instrumental activities of daily living (ADLs and IADLs, respectively), using the 14-item scale developed for the Older Americans Resources and Services project (Fillenbaum 1988; Fillenbaum and Smyer 1981). Results are presented as the number of areas (range = 0–14) in which each person required assistance or was completely unable to perform the activity. Using this information, we classified the cognitively normal people at each study wave as being frail if they required assistance for two or more ADL or IADL activities. This threshold was derived empirically as including 25% of participants at CSHA-1; use of a more exacting criterion would have yielded too few cases for analysis. Note that, although almost all of those with dementia would have met the criterion for frailty, the frail group in this report refers only to those who were cognitively normal. We asked caregivers of those who died before the follow-up interview to rate their functional status as of 3 months before death. For those in the community sample who required assistance, we asked caregivers to identify everyone who assisted and how many hours each helper spent per month in assisting with each ADL or IADL activity; this included paid caregivers. Caregivers of people with dementia completed the Dementia Behavior Disturbance (DBD) scale to record the frequency of behavior problems. This scale includes 28 items, and scores range from 0 to 112, with higher scores indicating more problems. It has a reported alpha coefficient of .84 (Baumgarten, Becker, and Gauthier 1990).
From the information on cognition and on disability, recipients of care were classified into eight health status groups for analysis. Those who had dementia were subdivided into (a) those who had dementia at both waves (prevalent dementia), (b) those newly diagnosed with dementia at CSHA-2 (incident dementia), and (c) those who had dementia at CSHA-1 and died. Those who did not have dementia were subdivided into (a) those who were frail at both waves (prevalent frailty), (b) those who became frail (incident frailty), (c) those who were frail and died, (d) those who were not frail but died, and (e) those who were cognitively and physically healthy at both study waves.
Care recipients were also grouped into five residential categories: (a) those who remained in the community, (b) those who were admitted to institutional care during the study period, (c) those who were in a long-term care institution for the entire study, (d) those who were initially in the community and died, and (e) those in an institution who died. No one moved from an institution to the community.
Measures of Caregivers' Health
Physical health was assessed using a list of 12 self-reported chronic health conditions and symptoms (such as heart problems, high blood pressure, or asthma) derived from previous surveys. An overall score counted the number of symptoms or conditions reported by the caregiver. Self-rated health was recorded using responses to the question "How would you say your health is these days?" on a 5-point response scale (Fillenbaum 1988). Depression was evaluated by means of the Center for Epidemiologic Studies–Depression (CES-D) scale (Radloff 1977). This includes 20 items graded on the frequency of occurrence of each symptom in the past week (ranging from 0 = none of the time to 3 = most or all of the time). Possible scores range from 0 to 60, and scores of 16 or above were taken to indicate depression. For those caring for a person with dementia, feelings of distress were recorded using Zarit's 22-item Burden Interview (Zarit, Reever, and Bach-Peterson 1980); scores range from 0 to 84, with high scores indicating more burden.
Analyses
All analyses used the study participant's cognitive and residential classifications as the independent variables. For the descriptive analyses of Objectives 1 and 2, characteristics of caregivers (age, burden, health outcomes) and care recipients (disability, behavior disturbance scores) were summarized as means. Where distributions were nonnormal, a logarithmic transformation was used, and analyses of variance tested differences among the cognitive and residential groups. For hours of care provided, however, the log transformation failed to normalize the distribution, and significance of differences was tested using the Kruskal-Wallis analysis. Comparison of other characteristics was based on differences in proportions, calculated according to Fleiss 1981. We do not present significance figures for the contrast between groups in 3MS or ADL/IADL scores, as these were used as the basis for defining the groups.
For Objective 3 (impact of providing care on caregiver health), we used a sequential approach. First, the proportions of caregivers with health problems at CSHA-2 were calculated for three health indicators (number of chronic health problems, self-rated health, and depression scores) for each combination of the care recipient's cognitive and residential status. Although these analyses summarize the caregivers' health, they do not permit direct comparisons across groups as caregivers' ages differed between the residential groups. For example, spouse caregivers predominated in the community, whereas daughters were often the caregivers for those in institutions. To standardize for caregiver's age, we used generalized linear modeling (SAS Proc GLM; SAS Institute 1989); the results are shown in the final table as the CSHA-2 score for each health measure. The next step in the analysis considered change in caregivers' health over time, controlling both for their age and their score at CSHA-1. This analysis used residualized gain scores (Cronbach and Furby 1970), as described by Streiner and Norman 1995. This involves first using a regression analysis to estimate each caregiver's expected health scores at CSHA-2 from their CSHA-1 scores and then calculating the difference between the expected and the actual CSHA-2 scores. This removed from consideration that component of the CSHA-2 scores that could be predicted from the baseline health score. The resulting difference score was used as the dependent variable in Proc GLM to compare average residual change in caregiver health for each combination of cognitive and residential change. Caregiver age was used as a covariate in all of these analyses. The resulting scores have been oriented such that a positive value indicates more health problems than would be predicted from the baseline score and age, and negative scores indicate better outcomes (fewer problems) than predicted.
|
Results |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
Objective 1: Characteristics of Care Recipients and Caregivers and Needs for Care
Details of the cognitive and residential status of care recipients are shown in Table 1 . This was a very elderly sample. The mean age of surviving care recipients at CSHA-2 was 85.6 years; those in the dementia groups were slightly older (87.8 years) than those in the nondementia comparison groups (81.6 years for those who were well and 85.6 years for those who were frail). Those in the dementia groups were slightly more likely to be women (71% of the prevalent dementia cases) than those in the nondementia groups (68%). The mean 3MS score was 37.8 among prevalent dementia cases and 56.5 among incident cases; the mean was 32.1 for those with severe dementia as compared with 46.8 for those rated as moderate and 63.0 for those with mild dementia. The mean 3MS score was 88.7 among those without dementia who were frail, the same as in the healthy sample. People with dementia were more likely to move into an institution than were those without. As would be expected (Ostbye, Hill, and Steenhuis 1999), mortality was higher among those with dementia at CSHA-1 (72%) than among those without dementia (42%), and death rates were highest among those in institutions.
|
Indicators of needs for care included ADL and IADL disabilities and also the DBD scale. Table 2 shows mean numbers of ADL and IADL problems at CSHA-1 and CSHA-2 (or as of 3 months before death for those who died) by cognitive status and residence. By CSHA-2, those with dementia had severe ADL/IADL problems, with a mean of 11.7 problems (out of a possible 14) for prevalent cases, 10.1 for incident cases, and 12.3 among people with dementia who died. The mean scores for those with frailty appear to be intermediate, and those who were healthy at both study waves had a mean score of only 0.2. For most groups, people in institutions had the highest mean number of ADL/IADL problems; there were especially large increases in disability from CSHA-1 to CSHA-2 for both community and institutional groups who had incident dementia.
|
Objective 2: Assistance Provided by Caregivers
Table 3 shows the number of ADL or IADL tasks for which community caregivers provided assistance and the time this required. There was a clear gradient by cognition and functional status. Those with dementia received the most assistance, those with frailty required intermediate assistance, and the healthy sample required very little assistance. There was also a clear gradient in the amount of assistance provided by type of caregiver, such that primary caregivers helped with twice as many tasks as did paid care providers, who helped with twice as many tasks as did other relatives or friends. We also analyzed the patterns of help by change in health status, comparing incident with prevalent cases. In terms of the number of tasks, for both dementia and frailty groups, the prevalent and incident cases did not vary significantly from the results shown in Table 3 . In terms of number of hours of assistance, however, the pattern was more complex, in that people with incident dementia received significantly more hours of help than those with prevalent dementia, albeit for the same number of tasks. Conversely, those with incident frailty received significantly fewer hours of assistance than those with long-term frailty, F(4,190) = 26, p < .0001.
|
Objective 3: The Health Impact of Caregiving
Zarit's caregiver burden scores showed significant contrasts between residential groups. Caregivers of people with dementia who remained in the community had significantly higher burden scores at CSHA-2 (mean score = 21.9) than did caregivers of people with dementia who moved into an institution (M = 13.6) or of those whose recipients remained in an institution (M = 11.7), F(2,197) = 7.4, p < .001. These contrasts remained after adjusting for the caregiver's age and baseline scores: Residualized gain scores increased by an average of 6.3 points (i.e., more burden) for caregivers whose care recipients with dementia stayed in the community; there was a decrease of 2.8 points for caregivers of people who moved from the community to an institution, and a decrease of 3.0 points among caregivers of those who were in an institution at both times, F(2,170) = 10.3, p < .0001.
Table 4 and Table 5 summarize the health outcomes implied in Objective 3. To begin, Table 4 shows raw results (uncorrected for age or baseline health levels) of cross-sectional analyses of the caregiver health indicators at CSHA-2. Caregivers for people with dementia were significantly more likely to report three or more chronic conditions (53% for incident dementia, 49% for prevalent dementia) than caregivers of healthy elders (33%; z = 2.7, p < .01). The level of chronic conditions among caregivers of frail people was not significantly different from that of caregivers in the dementia groups. In terms of self-rated health, most caregivers reported being in good or excellent health. However, 11% of primary caregivers of people with prevalent dementia, and 16% of caregivers of people with incident dementia, reported fair or poor self-rated health. This compares with 6% of caregivers of people without dementia or frailty. Only the difference between the latter two figures was significant (z = 2.1, p < .05).
|
|
As the caregivers may not have been in comparable health at baseline, Table 5 shows the caregiver's health scores at CSHA-2 and the change in health scores, controlling for age and baseline values. It also breaks out the analysis by residential status, where numbers permit. The results tend to confirm those seen in Table 4 : With the exception of caregivers of healthy people, there was remarkably little difference in either the CSHA-2 scores or the change scores between the groups. The exceptions are that the change scores for caregivers of the healthy group were frequently negative, indicating fewer adverse effects than expected, whereas the CES-D scores were elevated for caregivers of those with incident dementia.
For the three groups in which the care recipient had died, the mean delay between death and the CSHA-2 interview was 34.2 months (SD = 17.8). Because time since death may have influenced the caregiver's health outcomes, we repeated the analyses shown in Table 5 , adjusting for the time since death. The effect was to increase the CSHA-2 depression and self-rated health outcome scores; the CES-D scores at CSHA-2 rose, for example, by an average of 1.2 points compared with those shown in Table 5 , the risk of depression and of poor self-rated health increased with more recent deaths.
|
Discussion |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
At the same time, several limitations were imposed by the study's design. First, at follow-up only 32 people were newly diagnosed with dementia, owing to the limited number of people showing early stages of cognitive impairment in the baseline sample. Second, as would be expected after a 5-year delay, few of those with dementia at baseline survived to the follow-up study. Mortality was especially high in the institutional group, so there were only 80 people with dementia in institutions at both study waves. Third, it bears repeating that this is a very old sample (the mean age of care recipients at follow-up was 85.6 years). Finally, caregiving patterns and health were measured at only two points, 5 years apart. We missed fluctuations in the caregiving process and their impact on the health of caregivers during this interval, and relevant variables such as the caregiver's coping skills were not measured.
Many of our results confirm those of other studies. For example, the association between residence and type of caregiver suggests that spouses, and in particular wives, delay institutionalization of their spouses by providing care over a prolonged period; institutional admission commonly occurs for those who do not have a spouse to care for them. Among those with dementia, the subgroup who moved into an institution had the highest behavioral disturbance scores and the largest loss of ADL abilities. Both are markers of the burden of care and similar results have been observed in other studies (Baumgarten et al. 1994; Grafstrom and Winblad 1995; Kiecolt-Glaser, Duva, Speicher, Trask, & Glaser, 1991). The results in Table 2 also suggest that caregivers of people with dementia in the community face a task, in terms of their care recipients' level of disability, as challenging as that of caring for people without dementia who are in an institution. However, in addition to their physical dependency, the challenge of caring for a person with dementia is made qualitatively more demanding by their cognitive problems. It is initially surprising, therefore, that the gradient in caregiving tasks between the dementia, frail, and healthy groups did not translate across the board into corresponding contrasts in health impacts on the caregivers.
Within the limitations of a nonexperimental study, several conclusions are clear. In the main, the caregivers appeared to be surprisingly healthy. Although roughly half reported three or more chronic conditions (most of them minor), our mean CES-D scores were markedly lower than those in other studies (e.g., Schulz et al. 1995). Only between 5% and 16% of caregivers reported being in less than good health (Table 4 ), and there were few consistent contrasts between the groups; in effect, these caregivers did not tend to complain of poor health.
The clearest link between caregiving load and health was that caregivers in the healthy group showed fewer problems on all indicators. Beyond this, the impression from the change scores in Table 5 (most of which had positive coefficients, indicating deterioration) is that the depression scores showed clearer variation between diagnostic and residential groups than did self-rated health or chronic conditions. Caregivers in the dementia groups showed both the highest depression scores at CSHA-2 and the largest increases in depression. Depression persisted among caregivers for people with dementia who had died, and controlling for time since death suggested that more recent bereavement was associated with higher depression scores. Previous studies of depression among dementia caregivers have reported the continuation of depressive symptoms through time, although these studies were shorter (less than 30 months compared with the CSHA's 60 months; Grafstrom and Winblad 1995; Kiecolt-Glaser et al. 1991; Rabins, Fitting, Eastham, and Fetting 1990). Examining these associations further by residential status, caregivers of people with incident dementia demonstrated a significant increase in depression, principally for those in the community at both waves of the study. The finding that the number of hours of assistance provided was higher for incident than for prevalent dementia may suggest an adjustment phase of intensive caregiving, perhaps accompanied by apprehension for the future, that could be associated with depression. Caregivers of people with long-standing dementia had the highest depression scores, but, curiously, mainly when the care recipient moved to an institution. Arguably, the depression may have preceded, even precipitated, institutional admission and was not resolved by it.
Perhaps because of this study's relatively long follow-up compared with other studies, we did find some evidence of an exacerbation of physical health problems and a modest gradient of effect. First, the chronic condition change scores in Table 5 tended to show deterioration compared with the expected scores, with five statistically significant positive coefficients and no significant negative coefficients. Second, the coefficients representing deterioration clustered in the prevalent and incident dementia groups, and also in the caregivers of cognitively normal people who died. Third, improvement scores (negative signs) were most frequent in the healthy group and for those with dementia who died (four of seven coefficients were negative).
Overall, the care recipient's death did not appear to have a consistent health impact: Coefficients showing improvement in health measures were balanced by others showing deterioration. However, among community caregivers whose care recipient died, six coefficients showed an improvement, and only one showed deterioration across the three health indicators. The impact of adjusting for time since death suggests that time does, indeed, heal—but only for reactive conditions: Self-rated health and depression showed gradients by time since death, but chronic conditions did not. Admission to an institution also showed an inconsistent association with caregiver health. Among the incident and prevalent frail groups, five of the six coefficients shown in Table 5 for those who moved to an institution were negative, suggesting improvement. By contrast, all of the coefficients were positive for the dementia groups (although only two were significant). There are several possible reasons why institutional admission might reduce caregiver strain only for those whose care recipients did not have dementia. Those with dementia are sicker, experiencing both ADL and cognitive problems that may make care qualitatively more difficult. It is also possible that the processes involved in institutional admission exert their toll. Zarit and Whitlatch 1992 reported that institutionalization was followed by the caregivers' experiencing relief from primary effects of caregiving (e.g., feelings of overload and tension) and improved well-being, and other indicators of stress remained unchanged. Methodologically, although our analysis adjusted for caregiver age, the results may mix two groups: those who chose institutional admission and those for whom it became an undesired necessity, the latter often being spouses. Hence, although we tried in these analyses to define homogeneous groups of care recipients in terms of residence and type of disability, there was considerable variability in the health of caregivers within each group. The link between the caregiving task and caregiver health is complex and not fully explained by the variables we recorded here.
More broadly, several reasons may explain our failure to find a straightforward association between caregiver load and health status. The first concerns our sample. The mean age of caregivers was 64, and most had been providing care before our study commenced; some had already developed chronic health problems, and so we may not have begun from a true baseline. Working against this observation, however, is the finding of the relatively low levels of CES-D depression scores: Very few of these caregivers were depressed. Compared with studies based on clinical samples, our population-based study found comparatively healthy caregivers. Relating to this, a second, methodological reason may be that our use of residual gain analysis for the change scores in Table 5 overcontrolled for health status by removing the effect of the caregiver's initial health level, limiting our chances of seeing further deterioration. Certainly, in comparing our results with those of other studies, the unadjusted results shown in Table 4 should be used. A third explanation returns to the theme of self- and family selection of the caregivers. Depending on the size and closeness of a family, there may be chances for caregiver substitution, so that a caregiver who suffers significant strain may be replaced by another family member, leading to a healthy caregiver effect. We failed to recontact 26% of our original caregivers (386 of 1,472), including many who died, so our follow-up sample may be biased toward healthier caregivers. Selection into the caregiving role and the resulting healthy caregiver effect are likely to complicate such analyses, especially for a population-based sample. Caregiving is a dynamic and adaptive process of the type described by complexity theorists, and the regression-based linear analyses we have undertaken here may not adequately reflect the complexity of the underlying processes.
In summary, and within the limitations of an observational study, these results point to a complex relation between the cognitive and residential status of elderly people and the health of their caregivers. Recent pharmacologic treatments may delay dementia but will not reverse it, so we shall still be faced with providing care to a growing older population of people with dementia. Learning about the caregiver's role and the health effects of caregiving thus remains a high priority for families and for society alike, and the mechanisms appear complex and elusive. Whatever the mechanisms involved, however, studies such as the CSHA can assist in choosing subsets of vulnerable target groups to be evaluated in trials of supportive care interventions.
The Forum
Book Reviews
Practice Concepts
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received for publication December 21, 2001. Accepted for publication April 22, 2002.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  S. Jaglal, R. Cockerill, L. Lemieux-Charles, L. W. Chambers, K. Brazil, and C. Cohen Perceptions of the Process of Care Among Caregivers and Care Recipients in Dementia Care Networks American Journal of Alzheimer's Disease and Other Dementias, April 1, 2007; 22(2): 103 - 111. [Abstract] [PDF]
|
|
|
|
 |
|
 R. Schulz, S. H. Belle, S. J. Czaja, K. A. McGinnis, A. Stevens, and S. Zhang Long-term Care Placement of Dementia Patients and Caregiver Health and Well-being JAMA, August 25, 2004; 292(8): 961 - 967. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Berg-Weger, S. M. Rauch, D. M. Rubio, and S. S. Tebb Assessing the health of adult daughter former caregivers for elders with Alzheimer's disease American Journal of Alzheimer's Disease and Other Dementias, July 1, 2003; 18(4): 231 - 239. [Abstract] [PDF]
|
|
|
|
 |
|
 N. O'Rourke and H. A. Tuokko Psychometric Properties of an Abridged Version of the Zarit Burden Interview Within a Representative Canadian Caregiver Sample Gerontologist, February 1, 2003; 43(1): 121 - 127. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
