| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
a Waisman Center and School of Social Work, University of Wisconsin—Madison, Madison, WI
Correspondence: Marsha Mailick Seltzer, PhD, Waisman Center and School of Social Work, University of Wisconsin—Madison, 1500 Highland Avenue, Madison, WI 53705. Email: mseltzer@waisman.wisc.edu
Decision Editor: Vernon L. Greene, PhD
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionConclusionReferences |
|---|
Key Words: Family caregiving • Kinship relationships • Role salience
Two decades of research on the effects of family caregiving have documented that this is a stressful role with deleterious consequences for the person providing care (Aneshensel, Pearlin, Mullan, Zarit, and Whitlatch 1995
; Haley, Levine, Brown, Berry, and Hughes 1987; Lawton, Moss, Kleban, Glicksman, and Rovine 1991; Pruchno, Peters, and Burant 1995). Nevertheless, there is evidence of considerable heterogeneity in the effects of caregiving, with some caregivers showing patterns of adaptation while occupying this role, whereas others are characterized by wear-and-tear effects (Townsend, Noelker, Deimling, and Bass 1989).
Whereas most research on caregiving is cross-sectional in design, more recent longitudinal studies have tracked changes in the well-being of caregivers (Aneshensel et al. 1995; Goode, Haley, Roth, and Ford 1998; Li, Seltzer, and Greenberg 1999; Schulz and Williamson 1991; Townsend et al. 1989). Consistent with this new emphasis on prospective investigations has been a reconceptualization of caregiving as a career. Pearlin 1992 noted that the role of caregiver has some features of a career and that, during the period of family caregiving, the demands of the role change (Dwyer, Henretta, Coward, and Barton 1992). Consequently, the effects on the well-being of the caregiver are not static. The purpose of this article is to examine the prevalence and consequences of transitions into, within, and out of the caregiving career during a 3-year period among wives and daughters who had an elderly husband or parent.
Pearlin and Aneshensel 1994 clarified that, unlike most careers, caregiving is unplanned, unexpected, and not entered into by choice. Rather, family members are "drawn into it insidiously" (p. 376), often functioning as caregivers even before they self-define in this role. The ambiguity of the onset of the caregiving role might be more characteristic of spouses than adult child caregivers (Montgomery and Kosloski 1994), because spousal caregiving grows naturally out of the prior patterns of support and assistance that are exchanged between husband and wife, whereas the character of the exchange of support and resources between parent and adult child is more likely to be altered when caregiving begins (Seltzer and Li 1996). The structure of the family also influences when a particular family member will enter into the caregiving role. There is good evidence supporting the existence of a hierarchy of caregiving (Cantor and Little 1985; George 1988), with spouses being the first-line caregivers and daughters likely to enter the role when the spouse is deceased or unable to function as caregiver (Dwyer et al. 1992). Thus, daughter caregivers tend to care for older care recipients than do wife caregivers (Lawton et al. 1991; Montgomery and Kosloski 1994).
Past studies of the transitions of caregiving have had little to say about the effects of entry into this role because virtually all samples consist of individuals who had become caregivers before the beginning of the study. Research comparing caregivers with non-caregiving peers has shown that caregiving takes a toll on the psychological well-being of the caregivers (Dura, Stukenberg, and Kiecolt-Glaser 1990; Kiecolt-Glaser, Dura, Speicher, Trask, and Glaser 1991; Russo, Vitaliano, Brewer, Katon, and Becker 1995; Strawbridge, Wallhagen, Shema, and Kaplan 1997) and is associated with reduced social involvement (Haley et al. 1987) and family support (Creasey, Myers, Epperson, and Taylor 1989). In particular, spouse caregivers fare worse than adult child caregivers in comparison with their non-caregiving counterparts (George and Gwyther 1986; Hoyert and Seltzer 1992; Kinney and Stephens 1989; Pruchno and Potashnik 1989; Stone, Cafferata, and Sangl 1987). Although these studies provide indirect evidence that entry into the caregiving role may result in a negative change in the well-being of caregivers, the cross-sectional structure of most of these studies makes it difficult to know whether the less advantaged profiles observed for caregivers existed before caregiving actually began. To examine the changes that accompany entry into the caregiving role, we needed a prospective design to assess non-caregivers before and after caregiving begins.
In addition to entry into the caregiving role, there are two other transitions of caregiving: institutional placement of the relative, which is experienced by a minority of caregivers (Montgomery and Kosloski 1994; Pruchno, Michaels, and Potashnik 1990; Scott, Edwards, Davis, Cornman, and Macera 1997); and bereavement, which is the ultimate end of the caregiving career. Regarding institutional placement, past research has suggested that adult children are more likely to place their parents in a nursing home than are spouses (McFall and Miller 1992; Montgomery and Kosloski 1994; Morycz 1985; Scott et al. 1997; Zarit and Whitlatch 1993) and that high levels of subjective burden and stress also increase the risk of nursing home placement (Colerick and George 1986; McFall and Miller 1992; Pruchno et al. 1990; Zarit, Reever, and Bach-Peterson 1980). Although some studies have shown a reduction in social strain and subjective burden after placement (Skaff, Pearlin, and Mullan 1996; Stull, Cosbey, Bowman, and McNutt 1997; Zarit and Whitlatch 1993), continuing patterns of elevated psychological distress have also been reported (Aneshensel et al. 1995; Brody, Dempsey, and Pruchno 1990; Collins, Stommel, Wang, and Given 1994; Lichtenberg and Barth 1989; Stull et al. 1997; Zarit, Todd, and Zarit 1986; Zarit and Whitlatch 1993). There is some evidence that these patterns are more pronounced in wives than daughters (Harper and Lund 1990, Skaff et al. 1996). As Zarit and Whitlatch 1992 concluded, "The careers of caregivers do not stop at the institution's door" (p. 672) even though placement may alleviate some of the negative effects of direct caregiving. Thus, the net effects of caregiving on well-being may remain unchanged after institutionalization of the care recipient because new stressors are substituted for those previously experienced.
Past research on the effects of the end of the caregiving career due to bereavement have yielded inconsistent findings. Whereas George and Gwyther 1984 reported increasing social participation and satisfaction with social relationships in former caregivers following bereavement, Bodnar and Kiecolt-Glaser 1994 found no change in social support from the period of caregiving to the post-bereavement stage. Similarly, although George and Gwyther found reduced psychological distress in caregivers following bereavement, other studies reported stability in depression before, during, and after this transition (Bodnar and Kiecolt-Glaser 1994; Norris and Murrell 1987; Umberson and Chen 1994; Wright 1994). However, Mullan 1992 found a curvilinear relationship, with depression decreasing within the first 6 months after bereavement and then increasing, with no net change in depressive symptoms from before to a year after bereavement. Collins et al. 1994 found that women who provided care showed a pattern of decreasing depression following their relative's death, whereas men showed the opposite pattern. A large longitudinal study of caregivers for persons with dementia reported that bereaved caregivers increased in their sense of mastery as compared with their pre-bereaved state (Aneshensel et al. 1995; Mullan 1992; Skaff et al. 1996). Mastery has also been shown to be higher among widows who were caregivers before their husband's death as compared with widows who had not provided care (O'Bryant, Straw, and Meddaugh 1990), possibly because the caregivers had taken on responsibility for financial and household management before their husband's death and therefore felt more confident to handle life alone. Other research has suggested that the degree of strain experienced during caregiving is predictive of the extent of strain after bereavement (Bass and Bowman 1990). Thus, past research has shown that the end of caregiving brings both losses and gains in the caregiver's social and psychological well-being, but the pattern of changes varies from study to study.
One explanation for the inconsistency of findings from past research regarding the transitions of caregiving is that different subgroups of caregivers experience the transitions differentially. For example, past research indicates that women are affected more negatively by the caregiving role than men (Young and Kahana 1989), in part because women have traditionally been the kinkeepers and may feel the losses associated with mental or physical illness in their relatives more acutely than men. In addition, as we noted, spouses are affected more negatively by caregiving than adult children (George and Gwyther 1986; Harper and Lund 1990; Hoyert and Seltzer 1992; Seltzer and Li 1996), possibly as a result of the greater saliency of the spousal role (Richards, Bengtson, and Miller 1989; Rossi and Rossi 1990; Skaff and Pearlin 1992) and the differential timing (old age vs midlife) of the transitions of caregiving in the life course of spouses as compared with adult children (Moen and Wethington 1999). Caregiving transitions that occur during midlife are experienced in the context of the demands of the multiple roles occupied by women at this stage, whereas in old age women have fewer other roles and resources. Thus, subgroup differences might be obscured in research that does not separate out the effect of transitions on male and female caregivers, or on spouses and adult-child caregivers.
To contribute to the emerging body of research on the prevalence and consequences of the transitions of caregiving, the present study tracks a probability sample of caregivers and non-caregivers during a 3-year period. During this time, some of the non-caregivers entered the caregiving role, some of the caregivers experienced the institutionalization of their relative, and other caregivers became bereaved. To control for possible gender effects, only women were included in the sample and, to identify possible kinship effects, wives were contrasted with daughters.
We posed two research questions. First, we investigated the distribution of caregiving transitions (entry, institutionalization, and bereavement) experienced during a 3-year period by wives and daughters of older persons. Second, we asked whether the transitions of the caregiving career occur in tandem with changes in caregivers' social involvement, family support, and psychological well-being.
Specifically, we had the following hypotheses:
|
Methods |
|---|
|
TOPAbstractMethodsResultsDiscussionConclusionReferences |
|---|
). Our approach to sample identification led to the inclusion of wives and daughters who were in the beginning of the caregiving role as well as those who were further along in this process. This generated a group of care recipients who were extremely heterogeneous in their reason for needing care, including dementia, heart disease, stroke, arthritis, and diabetes. We also included two comparison groups of non-caregivers: wives who had a husband aged 60 or older, and daughters who had a parent aged 60 or older who were not in need of any care, using the above definition. For daughters who had two living and healthy parents, the mother was designated as the focal parent for this study. When our study began, there were 541 respondents in four groups: wife caregivers (n = 129), daughter caregivers (n = 207), wife non-caregivers (n = 119), and daughter non-caregivers (n = 86), representing a participation rate of 70.2% of the women who were screened and who met the eligibility criteria for these four groups.
The data for this analysis were taken from Wave 1 and Wave 3 of our three-wave longitudinal study. Fully 88% of the initial participants (n = 476) continued in this study through Wave 3. Among them, 108 were wife caregivers, 187 were daughter caregivers, 106 were wife non-caregivers, and 75 were daughter non-caregivers at Wave 1. The remaining 12% of the initial sample of respondents were lost to follow-up because of death (n = 12), severe illness (n = 5), or refusal to participate at Wave 2 or 3 (n = 48). We note that virtually all of the sample was White (97.9%), a limitation of the study but reflective of the racial composition of Wisconsin's elderly population.
We conducted a series of comparisons between the sample members who continued throughout the study period and those who refused to participate at Wave 2 or 3. The participants and nonparticipants were similar at Wave 1 except for education of the caregiver and health of the care recipient. Participating wives and daughters were significantly (p < .05) better educated (43.9% of participants vs 26.2% of nonparticipants had post-high school education), and their husbands or parents were in better health than the husbands or parents of those who did not participate (55.2% vs 34.9% were in good or excellent health).
The characteristics of the longitudinal sample (n = 476) at Wave 1 were as follows. The wife caregivers averaged 70.2 years of age. Nearly one third (31.5%) continued their education beyond high school. According to our sampling criteria, all were married (i.e., wives caring for their husband). One fourth (25.0%) of the wife caregivers were employed at Wave 1 and their family income averaged $27,010. Although 67.6% of the wife caregivers considered their health to be good or excellent, fully one third (32.4%) rated their health as only fair or poor. When the study began, wives had been in the caregiving role for an average of 6.4 years. The husbands to whom they provided care averaged 75.2 years of age. Nearly all (92.6%) lived with their wives at Wave 1, with the remaining few living in nursing homes or group homes (7.4%). More than half (58.9%) were rated by their wives as being in fair or poor health.
The wife non-caregivers averaged 68.2 years of age at Wave 1. About half (49.1%) had education beyond high school. More than one fifth (21.7%) were employed at Wave 1 and their family income averaged $34,382. The majority (84.0%) considered themselves to be in good or excellent health. Their husbands averaged 71.4 years of age. All lived with their wives and only 17.9% were considered to be in fair or poor health at Wave 1.
The daughter caregivers averaged 57.9 years of age at Wave 1. Nearly half (48.7%) had some post-high school education. More than half (54.0%) were employed and their family income in 1993–1994 averaged $41,946. Most (89.3%) were married and in good or excellent health (81.8%) at Wave 1. When the study began, these daughters had been in the caregiving role for an average of 8.4 years. The parents for whom they provided care averaged 84.6 years of age. Only 17.4% were male (i.e., the father of the caregiver) and fewer than 1 in 5 (18.7%) were married. These parent care recipients lived in a variety of settings, with only about one tenth (10.2%) living with the daughter caregiver, one fifth (21.3%) living in a nursing or group home, over half (58.3%) living in their own home, and few either in housing for the elderly (7.0%) or with other relatives (3.2%). More than half (56.1%) were rated by their daughters as being in fair or poor health.
The daughter non-caregivers averaged 54.8 years of age at Wave 1. Fully 42.7% had post-high school education. Nearly all (90.7%) were married and fully three fifths (61.3%) were employed. Their average family income was $41,146. Most (85.3%) considered themselves in good or excellent health at Wave 1. Their parents averaged 79.6 years of age. Only 14.7% were male and about one third (36.7%) were married. Fully 89.3% lived in their own home, with few living either in housing for the elderly (8.0%) or with other relatives (2.7%). At Wave 1, about one third (34.7%) were considered to be in fair or poor health.
Data Collection and Measures
We collected data by interviewing the wives and daughters, primarily in their homes. Respondents also completed a set of self-administered measures. The variables used in this analysis can be classified into three domains: social involvement, family support, and psychological well-being.
Social Involvement
Past research has shown that social involvement is sensitive to changes that accompany the transitions of caregiving (George and Gwyther 1984; Stull et al. 1997). We conceptualized social involvement as including social participation and participation in leisure activities. Social participation was indicated by the frequency of social activities engaged in by the respondent. Using a modified version of a measure developed for the National Survey of Families and Households (Sweet and Bumpass 1987), respondents rated the frequency of participating in five types of social activities from 0 (never) to 4 (more than once a week): spending social time with relatives, with people with whom the respondent works, or with friends and neighbors; attending a social event at church or synagogue; and participating in a group recreational activity. The sum of the ratings of the five types of social activities was used to indicate social participation. Participation in leisure activities was indicated by the frequency of engaging in three types of activities, rated from 0 (never) to 4 (more than once a week): working on a hobby, taking trips, and engaging in physical exercise. The sum of the ratings was used to indicate the level of participation in leisure activities.
Family Support
Past research has shown that family support may change in the context of caregiving transitions (Creasey et al. 1989; Kramer 1993). Family support was conceptualized as including family relations and marital satisfaction. Family relations was measured using the Family Relations Index (FRI) of the Family Environment Scale (Moos and Moos 1986) and the Marital Satisfaction Scale of the Marital Satisfaction Questionnaire for Older Persons (Haynes et al. 1992). The FRI measures the extent to which the family is perceived by the respondent to be cohesive, expressive, and conflictual (which was reverse coded). Each of three subscales consists of nine items, rated as 1 (true) or 0 (false) for the respondent's family. The FRI is the sum of the three subscales, with an alpha reliability of .60. The Marital Satisfaction Scale measures domains of marital satisfaction that are relevant for older couples. Seventeen items address specific areas of marital satisfaction, such as communication, decision making, handling conflicts, and companionship, each rated on a 6-point scale (1 = very dissatisfied to 6 = very satisfied) and summed, with an alpha reliability of .96.
Psychological Well-being
We measured both positive and negative aspects of psychological well-being. We decided to include measures of positive well-being as well as psychological distress in response to the criticism that past research has perpetuated too pessimistic a picture of caregiving outcomes (Kramer 1997; Matthews 1988). The two measures of positive well-being we selected were Ryff 1989 scales of Mastery and Personal Growth, which have been used extensively with midlife and aging samples. Each scale consists of five items, rated from 1 (strongly disagree) to 6 (strongly agree), with adequate levels of internal consistency (.67 for Mastery; .62 for Personal Growth). Mastery refers to a feeling of control in managing the environment, whereas personal growth refers to a feeling of continued development and growth. We included the Mastery scale for the present analysis because previous research has shown the significance of efficacy and control in regulating the psychological well-being of caregivers (Li et al. 1999; Skaff et al. 1996) and because mastery has been shown to be sensitive to the effects of bereavement following caregiving (O'Bryant et al. 1990; Skaff et al. 1996). We selected the Personal Growth scale because there is evidence to suggest that caregiving may provide a complex challenge, giving a sense of satisfaction, competence, and expansion of the self (Lawton et al. 1991; Wells and Kendig 1997) as well as resulting in stress and burden.
Negative aspects of psychological well-being were assessed with measures of depressive symptoms and subjective caregiving burden. Depressive symptomatology was measured by the Center for Epidemiology Scale for Depression (CES-D; Radloff 1977). The scale consists of 20 items reflecting symptoms of depression that might have been manifested during the past week, each rated 0 (rarely or none of the time) to 3 (most or all of the time), with a possible range of 0 to 60. A score of 16 or more indicates clinical depression (Radloff 1977). Alpha reliability for the CES-D scale was .87. Subjective burden was measured by the Zarit Burden Interview (Zarit et al. 1980) which assessed problems frequently faced by caregivers, including caregiver's health, finances, social life, and the relationship between the caregiver and the care recipient. Alpha reliability for the burden scale was .87 in this sample.
Data Analysis
For our first research question, we tracked the number of sample members who experienced each caregiving transition between Wave 1 and Wave 3 (entry, institutionalization, or bereavement) or continued in their initial status (i.e., as either caregivers or non-caregivers) throughout the study period. We did not separate out those transitions that occurred between Waves 1 and 2 from those that occurred between Waves 2 and 3 because of the small number of caregiving transitions that occurred at each of these intervals. Using chi-square statistics, we contrasted wives and daughters with respect to the frequency of each caregiving transition.
For our second research question, we conducted a series of two-way repeated measures analyses of variance (ANOVAs). The repeated measures were taken at Waves 1 and 3 and reflected sample members' social involvement, family support, and psychological well-being. We conducted separate analyses for wives and for daughters. Each analysis contrasted a specific pair of subgroups in the sample. First, women who entered the caregiving role during the study period were contrasted with those who remained non-caregivers (new caregivers vs continuing non-caregivers). Second, caregivers whose care recipient entered a nursing home or group home during the study period were contrasted with caregivers whose care recipient continued to receive care in the community (nursing or group home caregivers vs community caregivers). Third, women who left the caregiving role because of the death of the care recipient during the study period were contrasted with women who continued as caregivers (bereaved caregivers vs continuing caregivers).
Our rationale for contrasting these pairs of study groups was that both groups in each pair occupied the same caregiving status at Wave 1, with one group subsequently experiencing a caregiving transition before Wave 3 and the other group remaining in the initial caregiving status throughout the study period. Finally, as a point of comparison, women who continued in the caregiving role throughout the study period were contrasted with those who were not caregivers at any point in the study (continuing caregivers vs continuing non-caregivers). (Note that women who placed their care recipient in a nursing home during the study period were not included in the continuing caregiver group because the continuing caregivers were conceptualized as a contrast group who did not experience a caregiving transition.)
These two-way repeated measures ANOVAs tested three types of effects: the main effect of caregiving status, the main effect of time, and the interaction effect of Caregiving Status x Time. Our greatest interest was in the interaction effect, which indicated whether there was a different pattern of change from Wave 1 to Wave 3 manifested by those who experienced a transition in caregiving status as compared with those who did not change in caregiving status. We refer to this as a caregiving transition "effect," mindful that causal conclusions are not warranted on the basis of this research design. When the interaction effect is significant (or at a trend level), we do not interpret or comment on the main effects.
Along these lines, to strengthen the validity of the comparisons, we conducted a set of t tests to determine if the women experiencing each of these transitions differed from their respective comparison groups at Wave 1 in age, level of education, income, and employment status. For all but one of the comparisons, there were no differences in these variables between those who later experienced the transition and those who did not. The one exception was that daughter caregivers whose parent died during the study were significantly older than their counterparts who continued as caregivers (60.66 vs 56.00), t (163) = 3.19, p = .002). Therefore, age was controlled in the analyses contrasting bereaved daughter caregivers and continuing daughter caregivers. Thus, with this one exception, the analyses of transitions that are presented here were made between pairs of groups of women who were very similar in sociodemographic profiles.
As noted earlier, as a point of comparison, we also contrasted the two groups of women who did not experience a caregiving transition during the study period (i.e., continuing caregivers vs continuing non-caregivers). Among the wives, the continuing caregivers and the continuing non-caregivers were similar at Wave 1 in age and employment status, but they differed in education and income, with the caregivers less likely to have graduated from high school (33.8% vs 51.7%), t (157) = -2.32, p = .02) and having lower incomes ($27,784 vs $35,376), t (159) = 2.91, p = .004, than the non-caregivers. Therefore, these variables were controlled in the repeated measures analyses that contrasted continuing wife caregivers and continuing wife non-caregivers. Among daughters, those who were caregivers throughout the study and those who remained non-caregivers did not differ at Wave 1 in age, income, education, or employment status.
It is possible that the duration of time that a caregiver occupied the caregiving role may have influenced the pattern of change from Wave 1 to Wave 3. To investigate this possibility, in preliminary analyses we added a covariate for duration in the caregiving role before Wave 1. This covariate did not alter the pattern of findings reported here (data available from the authors) and hence this covariate was not included in the final analyses.
Missing data were handled by mean substitution in instances in which fewer than one third of the respondent's items for a given scale were missing. For the analyses reported in this article, given the small number of cases in each of the subgroups, the p < .05 level of significance was used, with trends reported up to p < .10.
|
Results |
|---|
|
TOPAbstractMethodsResultsDiscussionConclusionReferences |
|---|
|
2(1, n = 169) = 14.01, p < .001.
Institutionalization
None of the wife caregivers placed their husband in a nursing home or a group home between Wave 1 and Wave 3, whereas 20 (10.7%) of the daughter caregivers whose parent initially lived in the community placed their parent in a nursing home during the 3-year study period. This difference between wife and daughter caregivers in propensity to place their relative in a nursing home or a group home was statistically significant, 2(1, n = 180) = 13.99, p < .001, and consistent with our hypothesis and past research.
Bereavement
Counter to our hypothesis, wives and daughters did not differ in their likelihood of exiting the caregiving role during the study period due to the death of the care recipient, 2(1, n = 273) = 1.77, p = .18. Of the 108 wives who provided care to their husband at the beginning of the study, 34 (31.5%) were no longer caregivers by the end of the study due to their husband's death. About the same percentage of caregiving daughters (34.8%, or 65 out of 187) were bereaved during the study period as a result of their parent's death. Two daughters exited the role because their parent fully recovered health and independence. Due to the small number of daughters experiencing this type of change, they were not included in the analysis presented here of the changes accompanying caregiving transitions. A small number of women who were non-caregivers at Wave 1 experienced the death of either their husband (5.7%, n = 6) or parent (8.0%, n = 6) during the study period. They, too, were dropped from our subsequent analyses because of the small number of non-caregivers who experienced this transition.
Continuing Caregivers and Non-caregivers
Of the wives who were caregivers at Wave 1, 74 (68.5%) were still in the caregiving role 3 years later. Of the daughter caregivers at Wave 1, 120 (64.2%) were still caregivers at Wave 3, but 20 of these had placed their parent in a nursing home during the study period. Thus, 100 (53.5%) of the Wave 1 daughter caregivers were still in the same caregiving status at Wave 3. Of the women who were non-caregivers at Wave 1, 82.1% of the wives (n = 87) but only 57.3% of the daughters (n = 43) were still in this status when the study ended because their husbands or parents still did not need care for even one ADL or IADL task at Wave 3. The difference between wives and daughters in their likelihood of remaining in their initial caregiving or non-caregiving status was statistically significant, 2(1, n = 304) = 17.77, p < .001.
Summary
To summarize the caregiving transitions that occurred between Wave 1 and Wave 3, a little more than one tenth of the non-caregiving wives at Wave 1 and one third of the non-caregiving daughters at Wave 1 had moved into the caregiving role by Wave 3. A transition only experienced by daughter caregivers during the 3-year study period was the move of the parent care recipient from the community into a nursing home, which occurred by Wave 3 for about 10% of the daughter caregivers. The death of the care recipient between Wave 1 and Wave 3 ended the caregiving role for about one third of the wives and daughters who were caregivers at Wave 1. In total, a change in caregiving status during the 3-year study period was experienced by one third (35.7%) of the women in the sample, whereas stability in either their initial caregiving or non-caregiving status was characteristic of the remaining two thirds (63.9%).
Changes in Family, Social, and Psychological Well-Being
Our second research question asked, "Do the transitions of the caregiving career occur in tandem with changes in caregivers' social involvement, family support, and psychological well-being?" For this question, we conducted two-way repeated measures ANOVAs and ANCOVAs (analyses of covariance), as described earlier. The repeated measures were the dependent variables at Waves 1 and 3 (3 years apart).
First, we describe women who experienced the transitions of caregiving: entry, institutionalization, and exit. Then, as a point of comparison, we describe women who did not change caregiving status during the study period: Women who provided care from Wave 1 through Wave 3 were contrasted with women who did not provide care at any point in the study.
Entry into Caregiving
All sample members included in the analysis of entry into the caregiving role (see Table 2 ) were non-caregivers at Wave 1. Some of these women became caregivers between Wave 1 and Wave 3 (the new caregivers), whereas others continued in their non-caregiving status (the continuing non-caregivers). Table 2 presents data about wives and daughters at two points in time: at Wave 1, when none of the women were caregivers; and at Wave 3, which for some women was after the transition into the caregiving role occurred.
|
For daughters, there were two significant interaction effects of Caregiving Status x Time (i.e., entry into caregiving effects). The new daughter caregivers experienced an increase in social participation, whereas the continuing non-caregiving daughters declined in this dimension, F(1, 63) = 4.67, p < .05. Moreover, the new daughter caregivers experienced a decrease in their sense of mastery, whereas the continuing non-caregivers increased in this dimension of psychological well-being during the study period, F(1, 59) = 4.75, p < .05. In addition, there were two time effects. Both the daughters who entered the caregiving role and the daughters who continued as non-caregivers perceived increasingly positive family relations during the study period, F(1, 65) = 6.46, p < .05, and showed a trend for declining levels of depressive symptoms, F(1, 66) = 3.62, p = .06.
Institutionalization
All sample members included in the institutionalization analysis were providing care to their parent in the community at Wave 1. The parents of some of these daughters moved to a nursing home or group home during the study period (the placed group), whereas others continued to reside and receive care in the community. Table 3 contrasts these two groups at Wave 1 and Wave 3. As shown in Table 3 , there were two interaction effects of Caregiving Status x Time. There was a trend for daughter caregivers whose parent moved to a nursing home or group home during the study period to increase in social participation following placement of their parent in a nursing or a group home, whereas daughter caregivers whose parent remained in the community throughout the study period did not change in this dimension, F(1, 107) = 3.14, p = .08. Moreover, daughters who placed their parent showed a sharp decrease in subjective burden, whereas daughter caregivers whose parent lived in the community throughout the study period were stable in their level of subjective burden, F(1, 107) = 7.59, p < .01.
|
Bereavement
All of the sample members included in the analysis of bereavement effects were caregivers at Wave 1. Some of these wives and daughters experienced the death of the care recipient between Wave 1 and Wave 3 (bereaved caregivers), whereas the others continued to provide care. Table 4 contrasts these two groups at Wave 1 and Wave 3. For wives, there were three significant interaction effects of Caregiving Status x Time (i.e., bereavement effects). The bereaved wives increased in their level of social participation, F(1, 104) = 9.83, p < .01; participation in leisure activities, F(1, 104) = 4.81, p < .05; and feelings of personal growth, F(1, 89) = 4.72, p < .05, following the death of their husband, whereas wives who continued in the caregiving role were stable or declined in these dimensions during the study period. In addition, wives (both bereaved and continuing caregiving) showed a trend for increasingly positive family relationships between Wave 1 and Wave 3, F(1, 105) = 3.55, p = .06, and increasing mastery over the environment during the study period, F(1, 91) = 7.44, p < .01.
|
Differences Between Continuing Caregivers and Continuing Non-caregivers
Table 5 contrasts women who were caregivers throughout the study period and women who were non-caregivers during this entire period of time (i.e., women who did not change in caregiving status during the 3-year study). As shown in Table 5 , consistent with past research, wife caregivers showed a significantly more compromised pattern of well-being than wives who did not provide care to their husbands (i.e., main effects for caregiving status). Specifically, wife caregivers had significantly lower levels of social participation, F(1, 153) = 17.08, p < .001, and participation in leisure activities, F(1, 154) = 16.22, p < .001, than non-caregivers; more troubled family relations, F(1, 156) = 10.56, p < .001, less marital satisfaction, F(1, 155) = 33.74, p < .001, a trend for less of a feeling of mastery over their environment, F(1, 136) = 3.48, p = .06, and a greater number of depressive symptoms, F(1, 153) = 16.58, p < .001. These differences were evident both at Wave 1 and Wave 3. In none of the comparisons did the wife caregivers show a more positive well-being profile than the non-caregiving wives. There were no significant interaction effects (i.e., no differential pattern of change between wife caregivers and wife non-caregivers), but there was a trend for decreasing participation in leisure activities for both groups of wives, F(1, 154) = 3.50, p = .06.
|
|
Discussion |
|---|
|
TOPAbstractMethodsResultsDiscussionConclusionReferences |
|---|
; Scott et al. 1997; Zarit and Whitlatch 1993).
Of the women who were not caregivers when the study began, the probability of becoming a caregiver loomed large on the horizon, with about one third of the daughters but just over 10% of the wives entering into the caregiving role before the study period was over. In our sample, the parents of the daughter non-caregivers were older and in poorer health at Wave 1 than the husbands of the wife non-caregivers, which may explain why the former group was more likely to enter into the caregiving role during the 3-year study period. The window of time during which the probability of entering the caregiving role was assessed in the present study was relatively short—3 years. Another study of a somewhat shorter duration reported that, of adult children who did not provide assistance to their elderly parent at Time 1, 5% provided ADL assistance and 13% provided IADL assistance two years later (Dwyer et al. 1992). In a study of much longer duration, Moen 1997 found that 53% of the women in her sample became caregivers during the 30-year study period. These data underscore Brody 1985 observation that caregiving is a normative family role for women, with the probability of becoming a caregiver increasing across the life course.
Our data show that the transitions of caregiving are marked by changes in social involvement, family support, and psychological well-being. This dynamic quality of the effects of caregiving transitions is consistent with Pearlin 1992 conceptualization that the demands of the caregiving role change over time, and consequently its effects change. In general, we found that entering the caregiving role has negative effects, whereas exiting this role is followed by upturns in well-being. However, the effects are dependent on the type of kinship relationship between caregiver and care recipient.
Specifically, we found that daughters experience comparatively few changes in their social, familial, and psychological well-being following entry into or exit from the caregiving role. In contrast, wives show downturns in well-being after entering the caregiving role and upturns after exiting it. These findings are consistent with, and an expansion on, past research that showed more negative effects of caregiving for wives than daughters (e.g., George and Gwyther 1986; Quayhagen and Quayhagen 1988). The present study extends this past research by showing that this different pattern between wives and daughters in caregiving impact becomes evident shortly after entry into the caregiving role and persists after exiting from it.
There are several reasons why entry into and exiting from caregiving might have more profound effects on wives than on daughters. First, the higher level of saliency of the role of wife than the role of daughter (Richards et al. 1989; Rossi and Rossi 1990) may be one reason why transitions in the caregiving role are experienced differently for these two groups. Studies have found that the importance an individual attaches to a life domain qualifies its effects on well-being (Gecas and Seff 1990; Kling, Ryff, and Essex 1997; Krause 1994; Ryff and Essex 1992). As Thoits 1992 argued, "The salience of an identity should indicate its meaningfulness to the individual and thus its potential ... impact" (p. 237). We found that entry into caregiving is disruptive to a wife's leisure pursuits, family relations, and marital relationship because the wife's social and marital partner is now compromised, whereas daughters' primary family relationships with husband and children remain unchanged.
A second and related reason why daughter caregivers may be less likely to experience social and psychological effects of entry into the caregiving role is the timing of this transition in their life course. Daughters tend to provide care for their parent during midlife, a time replete with other roles, when an adult daughter must manage multiple family demands from her children and husband as well as from aging parents, and when career and social commitments are substantial (Moen and Wethington 1999). Indeed, one of the few caregiving transition effects that daughters experienced was a decline in mastery after entering this role, possibly because the addition of caregiving responsibilities strained an already full complement of relationships and time demands and caused the daughter to feel less in control over her environment. However, midlife is also a time rich in personal and social resources (Lachman and James 1997). Hence, at this stage of life, there are other countervailing roles and resources that may buffer the stresses of entering the caregiving role on the daughter's social, familial, and psychological well-being. In contrast, wives tend to care for their husbands when they are in old age, a time when retirement rather than active engagement in work is the norm, when social support from peers is less available, and when their own health problems may threaten coping capacities (Moen 1998). Older wives thus have fewer resources to buffer the stresses of caregiving; thus, it is possible that entry into this role brings greater strains and exit brings greater relief than if it occurred in midlife. Our findings are consistent with Moen 1997 conclusion that taking on new roles, or leaving others, has "different impacts at different stages of the life course and in different contexts" (p. 136).
The findings of this study can be interpreted in the context of life-span development theory (e.g., Featherman and Lerner 1985; Hagestad 1990), which posits both continuities and changes in the well-being of individuals throughout life. Yet, the effects of the changes are not uniform. In some contexts, life transitions become turning points (Clausen 1995; Wheaton and Gotlib 1997), whereas in other contexts such transitions in status do not have a pervasive effect on the course of a person's life trajectory. Our data suggest that wives are likely to experience caregiving transitions as turning points as they experience losses in marital and family relationships after they enter the caregiving role; less social involvement, poorer family relationships, and more compromised psychological well-being while they occupy this role; and then gains in social and leisure participation and personal growth after transitioning out of the role. Whether these gains persist after the initial post-caregiving period warrants future research, as it is possible that the physical and psychological relief that follows the end of caregiving might be replaced in the longer term by feelings of loneliness and loss. There are other indications in the literature, however, that also point to the resilience of bereaved wives who had formerly provided care to their husband. Bodnar and Kiecolt-Glaser 1994 found that bereaved caregivers did not have an elevation in depression as compared with continuing caregivers, measured an average of 20 months after bereavement. Other studies have shown that bereaved wives who had provided care to their husband before his death felt a stronger sense of mastery and competence after they became widowed (O'Bryant et al. 1990; Skaff et al. 1996) and were less likely to be depressed (Wells and Kendig 1997) than those whose husband died suddenly. Thus, caregiving might position a wife well for coping with the demands of old age.
In contrast, daughters appear to weather the transitions of caregiving with fewer effects in social, family, or psychological well-being. The exception is the transition from community caregiver to nursing home caregiver, which we found to bring a marked decline in caregiving burden and a modest increase in frequency of social participation, a pattern that parallels the data reported by Zarit and Whitlatch 1993. We found no general negative effects of nursing home placement on daughters' personal relationships or feelings of well-being. The transitions into and out of the caregiving role may be appraised by daughters as normative (although stressful) at their stage of life and hence do not alter the continuity of their life course trajectories (Wheaton and Gotlib 1997).
Our study had a number of limitations. For instance, the sample was almost entirely Euro-American, reflecting the racial composition of Wisconsin's elderly population. Therefore, generalizations to older persons of color are not warranted. Second, the 3-year study period does not span the full caregiving career, and our research design did not allow for following a full cohort of people at risk for caregiving as they enter the caregiving role, experience it, and exit the role because of the death of the care recipient. Thus, although the study design was longitudinal, it was necessary to piece together the transition effects by making multiple comparisons at Wave 1 versus Wave 3 for different subgroups of the sample. Nevertheless, few studies of caregiving offer this long a view of caregiving transitions. Future research is needed to track full cohorts through this important social role of later life. Finally, the sizes of the subgroups who experienced the various transitions of caregiving were quite small, which limited our ability to conduct within-groups analysis and examine the factors that may affect adjustment and adaptation after caregiving transitions. Small subsample size might also have limited our power to detect the effects of the transitions of caregiving and might account for the several trend-level differences we reported. Although our initial sample was quite large (n = 476), either a larger base sample or a longer duration is needed in future research to have sufficient cases who make each of the transitions.
Nevertheless, the prospective design and probability sample are strengths of this study. In addition, the disentanglement of gender and kinship relationship helps to reveal the different patterns evident in different subgroups. Overall, support was found for most of the hypothesized patterns of effects following the transitions of caregiving. There is, nevertheless, a need for validation of the findings in future research.
|
Conclusion |
|---|
|
TOPAbstractMethodsResultsDiscussionConclusionReferences |
|---|
|
Acknowledgments |
|---|
Received for publication May 21, 1999. Accepted for publication October 5, 1999.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionConclusionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  A. E. Benjamin, R. E. Matthias, K. Kietzman, and W. Furman Retention of Paid Related Caregivers: Who Stays and Who Leaves Home Care Careers? Gerontologist, July 1, 2008; 48(suppl_1): 104 - 113. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. F. Marks, J. D. Lambert, H. Jun, and Jieun Song Psychosocial Moderators of the Effects of Transitioning Into Filial Caregiving on Mental and Physical Health Research on Aging, May 1, 2008; 30(3): 358 - 389. [Abstract] [PDF]
|
|
|
|
 |
|
 J. Sims-Gould, A. Martin-Matthews, and M. A. M. Gignac Episodic Crises in the Provision of Care to Elderly Relatives Journal of Applied Gerontology, April 1, 2008; 27(2): 123 - 140. [Abstract] [PDF]
|
|
|
|
|
|
M. E. Szinovacz and A. Davey Changes in Adult Child Caregiver Networks Gerontologist, June 1, 2007; 47(3): 280 - 295. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S.-T. Cheng and A. C. M. Chan Relationship With Others and Life Satisfaction in Later Life: Do Gender and Widowhood Make a Difference? J. Gerontol. B. Psychol. Sci. Soc. Sci., January 1, 2006; 61(1): P46 - P53. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. W. Li From Caregiving to Bereavement: Trajectories of Depressive Symptoms Among Wife and Daughter Caregivers J. Gerontol. B. Psychol. Sci. Soc. Sci., July 1, 2005; 60(4): P190 - P198. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. J. McCann, L. E. Hebert, J. L. Bienias, M. C. Morris, and D. A. Evans Predictors of Beginning and Ending Caregiving During a 3-Year Period in a Biracial Community Population of Older Adults Am J Public Health, October 1, 2004; 94(10): 1800 - 1806. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Sugihara, H. Sugisawa, Y. Nakatani, and G. W. Hougham Longitudinal Changes in the Well-Being of Japanese Caregivers: Variations Across Kin Relationships J. Gerontol. B. Psychol. Sci. Soc. Sci., July 1, 2004; 59(4): P177 - P184. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. Leiter, M. W. Krauss, B. Anderson, and N. Wells The Consequences of Caring: Effects of Mothering a Child with Special Needs Journal of Family Issues, April 1, 2004; 25(3): 379 - 403. [Abstract] [PDF]
|
|
|
|
|
|
V. A. Freedman, H. Aykan, D. A. Wolf, and J. E. Marcotte Disability and Home Care Dynamics Among Older Unmarried Americans J. Gerontol. B. Psychol. Sci. Soc. Sci., January 1, 2004; 59(1): S25 - 33. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. C. Burton, B. Zdaniuk, R. Schulz, S. Jackson, and C. Hirsch Transitions in Spousal Caregiving Gerontologist, April 1, 2003; 43(2): 230 - 241. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. E. Gaugler, S. B. Wackerbarth, M. Mendiondo, F. A. Schmitt, and C. D. Smith The characteristics of dementia caregiving onset American Journal of Alzheimer's Disease and Other Dementias, March 1, 2003; 18(2): 97 - 104. [Abstract] [PDF]
|
|
|
|
|
|
M. Pinquart and S. Sorensen Associations of Stressors and Uplifts of Caregiving With Caregiver Burden and Depressive Mood: A Meta-Analysis J. Gerontol. B. Psychol. Sci. Soc. Sci., March 1, 2003; 58(2): P112 - 128. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M Hirst Transitions to informal care in Great Britain during the 1990s J Epidemiol Community Health, August 1, 2002; 56(8): 579 - 587. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. J. Karlin, P. A. Bell, and J. L. Noah Long-term consequences of the Alzheimer's caregiver role: A qualitative analysis American Journal of Alzheimer's Disease and Other Dementias, May 1, 2001; 16(3): 177 - 182. [Abstract] [PDF]
|
|
|
|
|
|
D. M. Rubio, M. Berg-Weger, S. S. Tebb, and L. A. Parnell Comparing the well-being of post-caregivers and noncaregivers American Journal of Alzheimer's Disease and Other Dementias, March 1, 2001; 16(2): 97 - 101. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
