| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
RESEARCH ARTICLE |
1 Social Activity and Care Research Group, Tokyo Metropolitan Institute of Gerontology, Tokyo, Japan.
2 Gerontology Program, Obirin University Graduate School, Tokyo, Japan.
3 Department of Social Welfare, Japan Women's University, Tokyo, Japan.
4 Section of Geriatrics, The University of Chicago, Chicago, IL.
Address correspondence to Yoko Sugihara, PhD, Division of Human Sciences, Social Activity and Care Research Group, Tokyo Metropolitan Institute of Gerontology, 35-2 Sakae-cho, Itabashi-ku, Tokyo 173-0015, Japan. E-mail: sugihara{at}tmig.or.jp
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionReferences |
|---|
In recent years, longitudinal studies on caregiver well-being have increased. Townsend, Noelker, Deimling, and Bass (1989) summarized two leading hypotheses about the course of caregiver well-being. One hypothesis proposes that the longer caregiving is sustained, the more deterioration occurs in caregiver well-being. In contrast, another hypothesis emphasizes the ability of caregivers to adapt to stressful situations over time. Some longitudinal studies support the deterioration pattern (Gold, Reis, Markiewicz, & Andres, 1995
; Haug, Ford, Stange, Noelker, & Gaines, 1999; Pot, Deeg, & Van Dyck, 1997), whereas others support the improvement pattern (Kramer, 2000; Townsend et al., 1989). However, most longitudinal studies on caregiver well-being show relative stability over time (e.g., Gaugler, Davey, Pearlin, & Zarit, 2000; Goode, Haley, Roth, & Ford, 1998; RIS MRC CFAS, 2000; Seltzer & Li, 2000). Although this stability tends to be interpreted overall as supporting an adaptation hypothesis, methodological improvements and additional specification are needed before we can reject the deterioration hypothesis.
A first limitation of past research is that most work has been done in Western cultures, and little is known about change in caregiver well-being in Confucian cultures such as Japan. There are some important differences in the caregiving environments between Japan and Western societies such as the United States. The Japanese, for example, tend to provide family caregiving while living together in the same household, whereas North Americans are more likely to provide it while living separately (Campbell & Ingersoll-Dayton, 2000). In the United States, parents generally do not live with their adult children, but in Japan, they often do. Many studies show that caregivers who live with the care recipient experience higher levels of burden than those who do not (for a review, see Schulz, O'Brien, Bookwala, & Fleissner, 1995). Therefore, changes in the well-being of Japanese caregivers may differ from those experienced by Western caregivers.
In Japan, longitudinal studies on caregiver well-being are limited. These studies (Arai, Zarit, Sugiura, & Washio, 2002; Matsuda, Hasebe, Ikehara, Futatsuya, & Akahane, 1997; Sugisawa, Nakamura, Nakano, & Sugisawa, 1992) suggest that changes in the well-being of Japanese caregivers may be similar to those reported in Western studies (i.e., stable or even improved outcomes). However, these Japanese studies are generally based on small convenience samples and may fail to capture the real impact of caregiving.
A second limitation is that much of the previous research on change in caregiver well-being has been restricted to the analyses of average change, whereas there might actually be considerable individual variability in their trajectories (Schulz & Williamson, 1991; Townsend et al., 1989; Walker, Acock, Bowman, & Li, 1996). Recent advances in statistical methodology such as latent growth curve modeling (LGM) allow for the testing of complex models regarding longitudinal trends from both interindividual and intraindividual perspectives. Examples of the application of LGM in caregiving research can be found in Walker and colleagues (1996), Gaugler and colleagues (2000), and Roth, Haley, Owen, Clay, and Good (2001).
These studies suggest that changes in caregiver well-being may be different depending on the type (general vs. caregiving specific) of outcome measures being used. Gaugler and associates (2000), for example, suggest that caregivers experienced slight deterioration in caregiving-specific outcomes such as role overload and role captivity, but relative stability in general psychological well-being measures such as depression. They also reported that only role overload demonstrated significant individual variability in rate of change. Although using general well-being measures makes it possible to compare caregivers with noncaregivers (George & Gwyther, 1986), general measures may be less sensitive than caregiving-specific measures (Stull, Kosloski, & Kercher, 1994). Some studies suggest assessing both caregiving-specific and general well-being types of measure dimensions (Braithwaite, 2000; Lawton, Moss, Hoffman, & Perkinson, 2000).
Variations Across Kin Relationships
In Japan, daughters-in-law (usually the wives of the eldest sons) have historically been more involved in caring for elders than their counterparts in the West (Campbell & Ingersoll-Dayton, 2000). However, little is known about the differential impact of caregiving on the well-being of daughters-in-law relative to other kin.
Several cross-sectional studies in the West have found that, the longer caregiving is sustained, wife caregivers tend to score worse on physical or mental well-being measures than husband (Bookwala & Schulz, 2000; Pruchno & Potashnik, 1989) or daughter caregivers (Hoyert & Seltzer, 1992; Li, Seltzer, & Greenberg, 1997; Stone, Cafferata, & Sangl, 1987). Few studies have examined how the longitudinal impacts of caregiving differ by kin type, and the findings are inconsistent (Seltzer & Li, 2000; Zarit, Todd, & Zarit, 1986; Zarit & Whitlatch, 1992). Among the few studies that have addressed differences between caring for a biological parent and caring for a parent-in-law, the relationship status had no main effect on caregiver stress among adult children (Ingersoll-Dayton, Starrels, & Dowler, 1996).
Moreover, even in Japan, research on the well-being of daughters-in-law is limited. Some cross-sectional studies have reported that wives tend to express higher levels of burden (Arai & Sugiura, 2000; Ohyama, Suzuki, & Yamada, 2001) and a more negative appraisal of the caregiving role (Yamamoto-Mitani et al., 2002) than other kin. Other studies have reported that daughters-in-law tended to express a weaker desire to continue care at home (Saito, Kunizaki, & Kanagawa, 2001) and a less positive appraisal (Yamamoto-Mitani et al., 2002) than other kin.
Reasons for Variations by Kin Relationship
Regarding possible reasons why changes in caregiver well-being might differ by kinship, we can consider the mechanism on the basis of the stress process model (Pearlin, Mullan, Semple, & Skaff, 1990). For example, differences in levels of primary stressors, such as the care recipient's disability, may be related to variations in caregiver stress by kin relationship. Past research suggests that wives provide more assistance to impaired elderly persons than daughters (Stone et al., 1987), and they report higher levels of cognitive impairment or behavioral problems in the care recipient than do husbands (Bookwala et al., 2000; Miller, 1990). Daughters were more likely to provide personal care or support to their parents than sons (Horowitz, 1985; Ingersoll-Dayton et al., 1996). Although few empirical studies examine whether the primary stressors to which daughters-in-law are exposed differ relative to other kin, under the traditional values of the Japanese family, daughters-in-law have been the mainstays of the family labor force, and therefore they disproportionately bear such exposure (Kasuga, 2001).
Further, according to the stress process model, social support, such as formal service utilization, has also been considered as one of the important factors explaining variability in stress outcomes. Whereas some research suggests that kin type is not related to level of service use (Bass & Noelker, 1987), qualitative research in Japan suggests that the relatively low authority level of daughters-in-law within the family hierarchy leads to a reduced likelihood of their being able to access and use formal caregiving services (Yamamoto & Wallhagen, 1998).
The Present Study
The purpose of this study was to examine longitudinal changes in the well-being of Japanese caregivers, variations in well-being by kin relationship, and reasons why caregiver well-being might differ by kin type. On the basis of the literature reviewed herein and the Japanese caregiving context, we proposed the following hypotheses.
Hypothesis 1: On average, caregivers who continued to provide in-home care during the study period would exhibit significant deterioration in our caregiving-specific well-being measure (caregiver's emotional exhaustion) as well as in our general measure of well-being (depression). However, only the caregiving-specific measure would demonstrate significant individual variability over time because of the insensitivities inherent in the more general measure.
Hypothesis 2: Changes in caregiver well-being would differ by kin relationship. Daughters-in-law would experience a more negative trajectory of well-being than other family caregivers, although the well-being of wives would also worsen.
Hypothesis 3: The reasons why changes in caregiver well-being differ by kin relationship would be explained by differences in primary stressors, such as activity of daily living (ADL) dependencies and problem behaviors in the care recipient, and formal service utilization by kin type. Because daughters-in-law might have higher levels of primary stressors and less use of formal services, they would exhibit the worst pattern of change in well-being.
|
METHODS |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
). The average per capita income was higher than the national average, and 76% of working persons were engaged in the white collar and service industries (Japanese Statistics Bureau, 2003).
We conducted a census-like mail survey or face-to-face interview with family members of all residents aged 65 and older (21,567 persons) in that city in February–March of 1996 (response rate = 95.5%), and we asked about the current disability status of those elderly residents. If the elders needed any help with ADLs (walking, eating, dressing, bathing, toileting, or getting in and out of bed), or if they had any memory or behavioral problems, the elders were considered care recipients (Nakano et al., 1996). Through this screening process, we identified 1,379 elders as community-dwelling frail elderly persons and asked who provided the most help to them in order to identify primary caregivers.
After the initial screening process, we conducted face-to-face interviews (usually at home) with 941 of these primary caregivers in April–May of 1996 (Time 1). We conducted the first follow-up interviews about 1 year later in March–April 1997 (Time 2), and the second follow-up interviews in August–September 1998 (Time 3), about 1.5 years after Time 2.
Of the 941 participants who completed the Time 1 assessment, 807 caregivers met criteria for the present analyses: (a) the caregivers identified themselves as the family members who were primarily responsible for the impaired elder's care, (b) the caregivers were wives, husbands, daughters, sons, or daughters-in-law of the elders, and (c) the caregivers answered all important study questions at Time 1.
Of these 807 caregivers, 350 (43.4%) provided complete data at all assessments, 176 (21.8%) provided data at Time 1 and Time 2, 12 (1.5%) provided data at Time 1 and Time 3, and 269 (33.3%) caregivers provided data at Time 1 only. Of the 457 caregivers who did not provide complete data, 256 were bereaved of their care recipients during the follow-up period, 120 placed their care recipients in institutional settings, and 81 refused or had other reasons.
At Time 1, the mean age of these 807 caregivers was 62.2 (range 32–92). Kin relationship type was wife (29.9%), daughter (29.9%), daughter-in-law (20.3%), husband (12.1%), and son (7.8%). Caregiving duration was <1 year (14.7%); 1 to <3 years (29.5%); 3 to <5 years (22.4%); 5 to <10 years (21.6%); and 10 years or more (11.8%). The vast majority (96.4%) lived with care recipients.
Measures
Emotional exhaustion
We measured caregiver emotional exhaustion by using a subscale of the Family Caregiver's Burnout Scale developed by Nakatani (1992, 1996). He translated and backtranslated the Maslach Burnout Inventory (MBI; Maslach & Jackson, 1981), a widely used approach for measuring burnout among human service professionals, and then slightly revised each item to apply to family caregivers of frail elderly persons. On the basis of the MBI, the original Family Caregiver's Burnout scale was designed to consist of the following three subscales: Emotional exhaustion, low personal accomplishment, and a cynical and uncaring reaction toward care recipients (depersonalization). However, Nakatani (1992) demonstrated that internal consistencies and construct validities were slightly low for the personal accomplishment and depersonalization subscales. Therefore, we used only the emotional exhaustion subscale as the caregiving-specific well-being measure.
The subscale consists of seven items measuring the intensity of perceived caregiver exhaustion. Examples of items are "I feel burned out from caregiving," "I feel frustrated by caregiving," and "I feel like I'm at the end of my rope from caregiving." The content of this measure is similar to measures of overload (e.g., Pearlin et al., 1990). Response categories ranged from 0 (very mild) to 4 (very strong). Total scores could range from 0 to 28, with higher scores indicating greater emotional exhaustion. Internal consistency for this measure at each time point was good (
>.84).
Depression
The Center for Epidemiological Studies–Depression scale (CES-D; Radloff, 1977) is a measure of depressive symptoms that has commonly been used in the caregiving and stress literature as a general psychological well-being measure. We used the Japanese version of the CES-D as translated and validated by Shima, Shikano, Kitamura, and Asai (1985). The CES-D is a 20-item self-report measure used to examine the frequency of mood and behavioral symptoms that occurred during the previous week. Frequency of behavior is rated on 4-point scales ranging from 0 (rarely or none of the time) to 3 (most of the time). Positive items are reverse coded. Example items include "I felt depressed," "I felt that people dislike me," and "I felt hopeful about the future." Total scores could range from 0 to 60, with higher scores indicating higher depressive symptoms. Internal consistency for this measure at each time point was also good ( >.85).
Kin relationship
We indicated kin relationship of the caregiver by dummy coding. We specified daughter-in-law as the reference group, and we compared wife, daughter, husband, and son categories with the daughter-in-law category, respectively.
ADL dependencies
We measured the ADL dependency level by the elder's ability to do six ADLs (walking, eating, dressing, bathing, toileting, and getting in and out of bed; Shimizu, 1996). Each item was rated from 0 (not at all) to 3 (completely) according to the degree of dependency on the caregiver. Total scores could range from 0 to 18, with higher scores indicating higher dependency. Internal consistency for this measure at each time point was excellent ( >.89).
Memory and behavior problems
We used the Checklist for Dementia Rating in Japanese community residents developed by Homma (1996) to assess the care recipient's memory losses, recognition failures, annoying behaviors, and socially inappropriate behaviors. Response categories were dichotomized (0 = absence, 1 = presence), and the potential range on this 16-item scale was 0 to 16, with higher scores indicating more memory and behavioral problems. Examples of items are "difficulty remembering his–her age," "difficulty recognizing relatives living with him–her," and "agitation at night." Internal consistency for this measure at each time point was good ( >.87).
Service utilization
We measured formal service utilization by the number (0–6) of community long-term care services that care recipients or caregivers had received: short-term nursing home respite, homemaker service, adult day care service, bathing service, visiting nurse, and home-delivered meals.
Caregiving duration
In pilot testing, caregivers were often unable to report the duration of caregiving exactly, so we chose to use an ordinal categorization: <1 year (0.5); 1 to <3 years (2); 3 to <5 years (4); 5 to <10 years (7.5); and 10 or more years ago (10).
Statistical Analysis
We used LGM, a statistical technique that describes both the average change in populations and individual differences in trajectories of change (Duncan, Duncan, Strycker, Li, & Alpert, 1999; McArdle & Anderson, 1990; Meredith & Tisak, 1990; Willet & Sayer, 1994), to examine changes in psychological well-being of caregivers. Latent factors representing intercept (initial status) and slope (rate of change) components are extracted from the three observations across time, that is, baseline (V1), 1-year follow-up (V2), and 2.5-year follow-up (V3) for well-being, stressors, and service use variables. Factor loadings of the intercept component to all three observations were fixed to 1, and we defined the linear slope component by fixing those parameters to 0, 1, and 2.5 for V1, V2, and V3, respectively. We estimated separate measurement errors for each time point. We evaluated goodness-of-fit indices to determine the fit of models. As a general rule, an acceptable fit is indicated by a chi-square probability value of p >.05, a comparative fit index (CFI) >.95, and a root-mean-square error of approximation (RMSEA) <.05.
Next, to examine if the intercept and slope components for each well-being measure differ by kin group, we conducted simultaneous analyses of multiple samples (Duncan et al., 1999). In the multiple-sample analyses, we divided the sample of 807 caregivers into five separate subsamples by kin group and then simultaneously estimated the intercept and slope components of well-being across all five subsamples. We first estimated a baseline model in which all parameters were constrained to be equal across all five kin groups. Next, we relaxed equality constraints hierarchically to probe the differences across kin groups, and we used nested 
2 tests to test the adequacy of the constraints.
Finally, we estimated the path model that examined whether the contribution of kin relationship to the intercept and slope components of well-being measure was explained by the intercept and slope components of ADL dependency, behavioral problems, and service utilization. We included caregiving duration as a covariate in the model. We implemented all LGM procedures in the structural equation modeling program AMOS 4.0 (Arbuckle & Wothke, 1999).
Typical approaches to analyses with incomplete data would use listwise deletion methods, using only data that are complete at all observations. However, this procedure necessitates the reduction of the sample size available for analysis, and the likelihood of a large sample bias if the missing data are not missing completely at random (MCAR). Some structural equation modeling programs offer an approach to the analysis of missing data, called full information maximum likelihood (FIML) estimation, which computes parameter estimates on the basis of all available data, including the incomplete cases. FIML estimates tend to be less biased than either listwise or pairwise deletion methods when the missing data process is not MCAR but only MAR (missing at random; Arbuckle, 1996; Wothke, 2000). Compared with all 807 cases, the 350 cases with complete data exhibited a slight tendency to exclude daughters-in-law and more stressful caregivers (i.e., caregivers who had the elders with heavy disabilities and those who experienced more psychological distress). Thus, our analyses were based on a sample size of 807, using the FIML procedure. However, when we ran the same analyses using only the 350 cases, the parameter estimates were comparable.
|
RESULTS |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
|
|
2 (1, N = 807) = 11.21, p <.01, CFI =.99, RMSEA =.08. However, in large samples, 2 tends to often reject most models (Bentler & Bonett, 1980), and a value of about 0.08 or less for the RMSEA would indicate a reasonable error of approximation (Browne & Cudeck, 1993). In addition, previous research has established the trajectory of ADL dependencies as roughly linear (Gaugler et al., 2000; Walker et al., 1996); therefore, we considered the slope of ADL dependencies in this sample to be approximately linear for these analyses.
|
For emotional exhaustion, the average intercept was 9.04, and the variance (24.04, p <.001) indicated significant individual differences at baseline. The average slope of the emotional exhaustion showed a significant increase (0.49, p <.001), similar to the CES-D. However, unlike the CES-D, there were significant individual differences in this linear trend (2.92, p <.001).
Comparison of Changes in Well-Being Across Kin Groups
To determine whether the intercept and slope of caregiver well-being were different by kin type (Hypothesis 2), we conducted simultaneous analyses of multiple samples for each well-being measure. For the CES-D, the model without an equality constraint on the intercept mean indicated a significant improvement in model fit over the baseline model in which all parameters were constrained to be equal across all five kin groups, 
2 (4, N = 807) = 11.24, p =.02, and it provided the best fit indices (CFI =.97, RMSEA =.02) among all competing models. This finding suggests that there were significant differences in the initial levels of depression by kin group, but the slopes of depression were similar across all kin groups. At baseline, wives and daughters-in-law tended to experience more depressive symptoms than other kin.
For emotional exhaustion, the model, which hypothesized cross-group differences for both the intercept and slope, demonstrated a significant improvement over the baseline model, 2 (8, N = 807) = 42.23, p <.001, and provided the best fit to the data (CFI =.99, RMSEA =.01) among all competing models. At baseline, daughters-in-law, wives, and daughters reported more severe emotional exhaustion than husbands and sons. As for the slopes, only wives and daughters-in-law showed significant increase (M = 0.99, SE = 0.21, p <.001; M = 0.51, SE = 0.26, p <.05; respectively). These findings on emotional exhaustion generally supported Hypothesis 2, which suggested that the patterns of change in caregiver well-being would differ by kin group. Although wives tended to experience the most negative change in emotional exhaustion, daughters-in-law also tended to demonstrate a similar negative trajectory.
To examine the reasons why daughters-in-law tended to show more negative trajectories than husbands, sons, and daughters, we estimated the path model including the intercept and slope components of ADL dependency, behavioral problems, and service utilization as the intermediate factors. Because the CES-D did not demonstrate significant individual differences in the slope components as shown in Table 3, we could not use the CES-D as a well-being measure for this analysis.
As shown in Table 4, the model on emotional exhaustion provided an acceptable fit to the data, except for the 2 statistic (CFI =.99, RMSEA =.03), and the results partially supported Hypothesis 3. As hypothesized, daughters-in-law experienced more behavioral problems in their elders at baseline than other kin (ß = –.27, p <.001, wives; ß = –.09, p <.10, daughters; ß = –.21, p <.001, husbands; ß = –.10, p <.05, sons), and many behavioral problems, in turn, tended to increase emotional exhaustion at baseline (ß =.32, p <.001). For ADL dependencies, daughters-in-law experienced a greater increase in ADL dependencies in their elders than husbands (ß = –.28, p <.001), and increases in ADL dependencies, in turn, was associated with an increase in emotional exhaustion (ß =.26, p <.001). As for service use, contrary to our hypothesis, daughters-in-law did not necessarily have limited availability to formal services relative to other kin. In addition, increases in service use were associated with increases in emotional exhaustion (ß =.23, p <.10).
|
Total effects in Table 5 suggested that daughters, husbands, and sons tended to experience less emotional exhaustion than daughters-in-law at both initial status (ß = –.05, ß = –.16, ß = –.12, respectively) and rate of change (ß = –.05, ß = –.12, ß = –.03, respectively), whereas only total effects of husbands and sons on the intercept component showed statistically significant differences. However, for the reason why daughters-in-law tended to show worse average trajectories than husbands, sons, and daughters, we found that our Hypothesis 3 was only partly confirmed because the indirect effects of husbands, sons, and daughters on the intercept and slope components of emotional exhaustion were smaller relative to the direct effects.
|
|
DISCUSSION |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
These small rates of changes in well-being may suggest that Japanese caregivers also adapt over time, or, the negative impact of caregiving may accumulate gradually rather than appear rapidly. Chronic stressors such as caregiving may deplete personal resources very gradually. Therefore, in studies with small samples or short-term follow-up, we should be cautious about interpreting any stable change patterns in caregiver well-being as indicating adaptation processes are operative.
One reason why our analyses demonstrated significant deterioration in caregiver well-being on average may be related to the significant increase in average levels of primary stressors such as ADL dependencies and problem behaviors in their elders, as shown in Table 3. Although the number of services used also increased significantly over time, these services might not meet caregivers' needs. Optimal tailoring of service provision to the needs of caregivers remains an area for further study.
Another reason for the significant deterioration in psychological well-being may be attributable to the context of Japanese caregivers. In Japan, and in our sample, most caregivers live with their impaired relatives. Many studies show that caregivers who live with the care recipient experience higher levels of burden than those who do not. We tested this hypothesis by conducting the same analyses but excluding caregivers who lived in a separate residence, and we confirmed that the findings remained essentially the same.
Interestingly, concerning the variance in the estimates of slopes for individuals, emotional exhaustion exhibited significant variation but the CES-D did not. Other caregiver studies using LGM or hierarchical linear modeling similarly find that depression does not demonstrate significant variance in rates of change (Gaugler et al., 2000; Redinbaugh, MacCallum, & Kiecolt-Glaser, 1995). These results suggest that general well-being measures such as depression may be inadequate to detect variability of change in caregivers over time and that caregiving-specific well-being measures such as caregiver burden or role strain may be more suitable.
This study focused on kin relationship, because the composition of primary caregivers' kinship in Japan is very different from the West, especially North America. Consistent with our expectation, daughters-in-law as well as wives were more likely to experience more negative trajectories of emotional exhaustion compared with other relatives.
Regarding the causal reasons behind the differences in trajectories of well-being by kin group, our initial hypothesis was that the differences in primary stressors and accessibility to formal services by kin group would be related to the differences in trajectories of well-being. These hypotheses were partially confirmed. Daughters-in-law tended to care for elders who exhibit more behavioral problems at baseline and a greater increase in ADL dependencies than other kin. Problem behaviors at baseline, in turn, tended to increase emotional exhaustion at baseline, and increases in ADL dependencies were related to increases in emotional exhaustion.
However, our findings suggest that there are still other factors to be considered. For example, attachment to care recipients and past relationship with care recipient may be associated with differences in the well-being trajectories between daughters-in-law and other kin.
Initially, we hypothesized that availability of formal services would be limited for daughters-in-law compared with other family caregivers, but the results did not support this hypothesis. However, because there was little use of formal services overall, significant differences across kin groups might be hard to detect. In addition, increase in service use was paradoxically associated with a deterioration in emotional exhaustion.
There are several interpretations for these last findings. First, using only the number of formal services that caregivers used may not adequately reflect the protective effect of service utilization. The timing, duration of use, or type of services used may be better measures by which to observe any protective effects. Alternately, our findings may suggest that the formal services available during the study period were not able to meet caregiver needs. Because this study was fielded before the establishment of a new long-term care insurance system in Japan (Ministry of Health and Welfare, 1999), the formal services available at the time might have been insufficient to confer much benefit.
There are several limitations to this study. First, our study cannot explain all the reasons why the trajectories of caregiver well-being might vary by kin type. Other factors such as emotional attachment to the care recipient, past relationship, and rewards from caregiving should be examined in future work. Second, this study was fielded before the establishment of a new long-term care insurance system in Japan, and the supply system and functions of formal service provision have been changing recently. Whether or not these newly implemented formal community services can reduce the risk for caregiver distress is a topic for future research.
|
Acknowledgments |
|---|
We gratefully acknowledge Harumi Sato, Taro Fukaya, Hideki Okabayashi, Kaoru Takanashi, Katsuhide Kawamoto, and Hye-kyung Kim for assistance with data collection and management, the research participants and staff members of the related organizations for their contributions, and the editor (Margie E. Lachman) and anonymous reviewers for extensive and helpful suggestions.
|
Footnotes |
|---|
Received for publication November 21, 2002. Accepted for publication March 1, 2004.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  Y. Sugihara, H. Sugisawa, H. Shibata, and K. Harada Productive Roles, Gender, and Depressive Symptoms: Evidence From a National Longitudinal Study of Late-Middle-Aged Japanese J. Gerontol. B. Psychol. Sci. Soc. Sci., July 1, 2008; 63(4): P227 - P234. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
