This Article Abstract Full Text (PDF) Services Download to citation manager Citing Articles Citing Articles via HighWire PubMed PubMed Citation

Social Support for Spouse Caregivers of Persons With Dementia

^a Mandel School of Applied Social Sciences, Case Western Reserve University, Cleveland, Ohio

Baila Miller, Mandel School of Applied Social Sciences, Case Western Reserve University, 11235 Bellflower Road, Cleveland, Ohio 44106-7164 E-mail: bhm4{at}po.cwru.edu.

	Abstract

TOP Abstract Availability and Distribution of... Relationship Between Support... Methods Results Discussion References

 
Objectives. This study analyzes the value of simultaneously taking helper and caregiver characteristics into account when examining social support received by spouse caregivers of persons with dementia. This study questions whether sources and types of received support vary by the gender and relationship of the helper and/or by the caregiver's race and gender.

Methods. Drawing on a sample of 210 spouse caregivers of persons with dementia, the Generalized Estimating Equation (GEE) method was used to estimate a population-average logistic regression model to address cross-level helper and caregiver interactions.

Results. Results indicate that helper and caregiver characteristics interact in complex ways. White male caregivers are more likely to receive emotional support from adult children and to receive practical assistance from formal sources compared with other types of support and to other race–gender caregiver groups.

Discussion. Description of social status characteristics of caregivers as determinants of their social network may be less fruitful in understanding the benefits of social support than a focus on which helpers assist with what kinds of task for which groups of caregivers.

THE level of support that caregivers give to impaired relatives may depend on the support they receive from those in their networks; conversely, the level of support that caregivers receive from their networks may depend on how much care is required by the health status of their impaired relative (Pearlin, Aneshensel, Mullan, and Whitlatch 1995). These interrelated dimensions of support have not been examined with equal fervor by gerontologists. Much research on caregiver support describes the instrumental support provided by secondary caregivers for the frail elderly rather than support provided to the caregiver per se (Miller and McFall 1991b; Stoller and Pugliesi 1991; Stone, Cafferata, and Sangl 1987). Spouses are likely to be the primary caregiver (when available) and least likely to have secondary caregivers involved in care (Stone, Cafferata, and Sangl 1987; Tennstedt, McKinlay and Sullivan 1989). Yet, little is known about the characteristics of persons who are drawn on by spouse caregivers for support and about what kinds of support are received.

Social support can best be described as a metaconstruct, comprised of several distinguishable theoretical constructs (Vaux 1988). Three dimensions are consistently identified (Barrera 1986; House, Umberson, and Landis 1988; Vaux 1988): (a) network structure, that is, support network resources or social embeddedness; (b) functional support, that is, supportive behaviors or enacted support; and (c) perceived social support, that is, subjective appraisals of support and satisfaction with support. Satisfaction with support has been suggested as a more effective component of support than number and types of helpers in moderating the effect of caregiving stressors on negative caregiver emotional and physical health outcomes (George 1996; Lubben 1988; Turner and Marino 1994). Support does not exist in a vacuum but is part of the "social fund" that people draw from when handling stressors (Thoits 1995). The structure of the support network indicates the availability of individuals who can respond to a person's need for support in specific situations.

This study simultaneously examines characteristics of helpers of spouse caregivers of persons with dementia (i.e., those providing support to the caregiver) and attributes of spouse caregivers. We address two interrelated questions, which pose the issue of availability and distribution of support resources for spouse caregivers: (a) Do the types of support provided to caregivers vary by the gender and relationship of the helper? (b) Do the sources of help and types of help provided vary by the spouse caregiver's race and gender and/or the impaired spouse's limitations?

Addressing similar issues of the relationship between helper and caregiver characteristics, Suitor and Pillemer 1993 suggested that these questions can be addressed either by aggregating information about persons in the caregiver's helper networks to the caregiver level or by examining which characteristics differentiate individual network members who provide help (Wellman and Wortly 1990). A major limitation to these approaches is that the support process represents an indivisible association between provision and receipt of assistance. Helpers are not randomly assigned to caregivers, violating key assumptions of traditional multivariate approaches. Thus, simultaneous analysis of the characteristics of the members of a caregiver's support network and the characteristics of the caregiver is preferable. The present research performs an analysis of the joint influence of caregiver and helper characteristics that takes into account the nonrandom association of helpers with caregivers.

	Availability and Distribution of Support for Spouse Caregivers

TOP Abstract Availability and Distribution of... Relationship Between Support... Methods Results Discussion References

 
Support for spouse caregivers occurs within the web of surrounding social relationships and thus is subject to the influence of personal, economic, and social forces that link the individual to larger social systems (Pearlin et al. 1995; Seeman and Berkman 1988). Issues of gender permeate the support process for spouse caregivers at all levels because the majority of caregivers for frail older persons are women (Dwyer and Coward 1992; Miller and Cafasso 1992; Walker 1992). Theories of filial responsibility, hierarchical compensation, and task specificity detail the importance of helper relationship in the availability and selection of who provides help (Cantor 1991; Hanson and Sauer 1985; Litwak 1985). Caregiving and social support research also suggest patterned differences in the extensiveness and nature of social support between African American and White caregivers (Silverstein and Waite 1993; Taylor and Chatters 1991). In this study, we focus on two linking attributes of the helper, gender and relationship to the caregiver, and two linking attributes of the caregiver, gender and race.

Gender
Sociological, psychological, and feminist theories of gender differences in support provide many reasons for the greater likelihood of women being helpers (Stoller 1990; Walker 1992). These include the following: Providing care may be central to women's identity; daughters have a special sense of attachment to their mothers; societal support of women as helpers is reinforced by social ideologies, devaluing of women's paid work, and inadequate government support. The support literature suggests that older men and women have personal networks of different size and composition in which male and female network members typically provide different types of support (Akiyama, Elliot and Antonucci 1996; Antonucci 1990). These differences have been mirrored in caregiving studies that note that men and women provide different types of caregiving assistance (Horowitz 1985; Matthews 1996; Miller and Cafasso 1992; Stoller 1990). In addition, there is some evidence that adult children may be more likely to provide care to a parent of the same gender (Lee, Dwyer, and Coward 1993). Thus, support relationships may be influenced by the gender of the caregiver and the gender of the helper and also by the combination of gender of caregiver and helper (Akiyama, Elliot, and Antonucci 1996).

Relationship of helper
The hierarchical compensatory model of support suggests a hierarchy of helpers in order of familial closeness and intimacy to the frail older person: spouses, when available, followed by adult children, other family members, and friends, and lastly, formal services (Cantor and Little 1985; Chatters, Taylor, and Jackson 1986; George 1988). This ordering does not specifically predict the role of adult children as secondary caregivers in networks of spouse caregivers because adult children are most likely to be caregivers of widowed parents. In older families, adult children are the most frequently mentioned members of older persons' social networks (Antonucci and Akiyama 1987; Mitchell and Register 1984; Mutran 1985; Taylor 1986). Given norms of filial responsibility in American society (Hamon and Blieszner 1990; Hanson and Sauer 1985), we may expect that adult children will be the most important sources of support to their parents, the spouse caregivers. These caregiver selection models do not address the question of what kinds of help each type of helper provides.

Race
Race is a marker variable that cannot be separated from socioeconomic, cultural, and political factors (LaViest 1994; Miller et al. 1996). Evidence is equivocal if race differences in support remain when socioeconomic class is taken into account (Markides and Mindel 1987; Smerglia, Deimling, and Barresi 1988; see also Angel and Tienda 1982; Krause and Borawski-Clark 1994). Information about helpers of African American spouse caregivers of persons with dementia is lacking in the research literature. This caregiving configuration is relatively rare as older African Americans are significantly less likely to be married than older Whites (Angel and Hogan 1991). Older African Americans reportedly have larger extended families and more multigenerational contacts than older Whites (Markides and Mindel 1987; Taylor 1986). This premise, however, has been challenged in recent years by research that focuses on variations within race/ethnic groups, changing family structures, and socioeconomic conditions (Burton and Dilworth-Anderson 1991; Haley et al. 1995; Minkler, Roe, and Price 1992; Silverstein and Waite 1993; Smerglia et al. 1988). More similarities than differences in caregiving patterns by race have been observed, but few of these studies focus on social support (Lawton, Rajagopal, Brody, and Kleban 1992; Miller, Campbell, Farran, Kaufman, and Davis 1995). Thus, traditional assumptions about the expansive nature of social support in the African American extended family structure and the role of adult children in such networks are in transition (Burton et al. 1995; Silverstein and Waite 1993).

	Relationship Between Support Networks and Support Functions

TOP Abstract Availability and Distribution of... Relationship Between Support... Methods Results Discussion References

 
The empirical findings that satisfaction with support is an important buffer of caregiver distress neglect to take into account the relationship between sources of support and support functions (Barrera 1986). Support functions are the types of support, such as instrumental, emotional, or informational assistance, that an individual or organization can provide (House et al. 1988). The assumed relationship between these two dimensions of support is that network structure influences access to received functional support (Lin and Westcott 1991; Thoits 1995).

Applying this model to spousal caregiving, we argue that a crucial dynamic is the specificity of the relationship between structure and function. In other words, the benefits of social support are regulated by the forms of support, its sources, and the fit with the types of problems confronting individuals (Jackson and Antonucci 1992; Pearlin et al. 1995). A variety of specificity hypotheses in the social support and caregiving literature highlight the relevance of matching between different support components. These include matching between type of support and type of stressor (Clipp and George 1990; Cohen and McKay 1984; Cutrona and Russell 1990; Li, Seltzer, and Greenberg 1997), salience of social roles and type of distress (Krause and Borawski-Clark 1994; Wethington and Kessler 1987), and type of caregiver burden and structure of group providing support (Litwak 1985; Messeri, Silverstein, & Litvak, 1993; Thompson, Futterman, Gallagher-Thompson, Rose, and Lovett 1993).

The task-specificity model of Litwak 1985 clarifies the links between support networks and support functions. Members of a support network may provide different functions of support, and substitution of tasks theoretically occurs among groups whose structure most closely matches the tasks of the other (Litwak 1985). Within informal groups, spouses and other household residents are believed to be best suited for the provision of daily tasks; the extended family is best suited for temporary short-term needs and emotional support; friends are best suited for companionship and for peer-related advice. Formal organizations provide higher levels of technical knowledge and more detailed division of labor typically applied to tasks that the caregiver can no longer perform for reasons of poor health or burden (Penning 1990). Ideally, specific measures of group structure and of support function (e.g., proximity and predictability) should be gathered. Messeri, Silverstein, and Litvak (1993) argued, however, that everyday terms for groups (i.e., children, friends, or neighbors) and for tasks (i.e., emotional support and practical assistance) can be assumed to represent distinct clusters of structural dimensions that represent the model.

In this research, we investigate the "matching" between four sources of support for spouse caregivers (i.e., adult children, other relatives, friends/neighbors, and formal helpers) and two types of support (i.e., emotional support and practical assistance). Our first hypothesis is that the probability of providing emotional support increases when the helper is an adult child (compared with other relatives, friends/neighbors, and formal sources). Emotional support (i.e., the receipt of reassurance and respect from others) may be conveyed in all types of social relationships. We assume that the primacy of the parent–child relationship and the history of emotional attachment associated with parent–child roles will be satisfied most strongly when emotional support is provided by adult children compared with other sources of support (Hamon and Blieszner 1990). Our second hypothesis is that the probability of providing practical assistance (i.e., help with tasks of everyday living) increases when the helper is from a formal resource (compared with adult children, other relatives, and friends/neighbors). Formal assistance is the least common source of support used by caregivers (Miller et al. 1996). Its use is determined by high degrees of need for assistance with activities of daily living (ADLs) or high levels of caregiver burden (Miller and McFall 1991a). Use of formal assistance is also influenced by many parents' wishes to not become a burden on their children or on other informal sources to maintain their sense of social and emotional independence from significant others.

We recognize the assumption that the type of support provided is largely determined by need. Clipp and George 1990 found that need was not a primary predictor of stability of support over time, but cross-sectional research has identified the role of need in mobilizing support. The need of spouse caregivers of persons with dementia is most typically represented by the number of the impaired spouse's functional limitations (Miller and McFall 1991a) and frequency of behavior problems (Haley et al. 1995; Miller, McFall, and Montgomery 1991). By controlling for need, researchers can clarify the joint influence of helper and caregiver sociodemographic characteristics and the association between sources and functions of support.

	Methods

TOP Abstract Availability and Distribution of... Relationship Between Support... Methods Results Discussion References

 
Sample
The data were from a 1991 National Institute of Aging supported study (Miller et al. 1995) of the influence of race and gender on spouse caregivers of persons with dementia. In-home structured interviews were performed, with interviewer and respondent matched by race. Selection criteria for respondents included: English-speaking spouse caregivers of persons aged 60 and older with a diagnosis of some form of dementia and coresidence of the impaired person and the spouse caregiver.

Respondents were recruited by referrals from Alzheimer's clinics, Veteran's Administration hospital programs, community home care programs, and adult day care centers that provided services to persons with presumed medical diagnoses of possible dementia in a large metropolitan area. Our intent was to recruit caregivers who had sought medical or social service help for their impaired spouses, not necessarily for themselves. African Americans and males were oversampled relative to their proportion in the caregiving population at a rate of approximately 1.75 to 1, an approach designed to ensure orthogonality of race and gender and reasonable cell sizes. Our final analytic sample size was 210, composed of 21 African American men, 55 White men, 52 African American women, and 82 White women. Response rates by type of recruitment source, caregiver race, and gender varied but could not be calculated, because in the social agencies, initial screening was carried out by agency staff rather than by research staff.

Differences in source of assistance by race and gender reflect societal processes of health care delivery, service access, and availability. We were concerned with potential confounding of site with severity of illness, diagnosis, race, and gender. We thus examined the influence of source of referral on the demographic composition and functional impairment of the sample. Analyses of a range of demographic and impairment levels by caregiver's race and gender confirmed that race and gender differences by source of referral were not accompanied by differences in cognitive or functional impairment (data not shown).

Sample characteristics were similar to those of other studies of spouse caregivers (Stone, Cafferata, and Sangl 1987). Analysis of variance (ANOVA) comparisons of sample demographic characteristics by gender and race suggest both similarities and differences (data not shown). Not surprisingly, male caregivers were older than female caregivers by an average of 7 years and had been married significantly longer. There were no significant differences in impaired spouse's age, caregiver education or number of children by gender. There was little difference in years of caregiving by race (e.g., an average of 5 years). Race differences occurred in caregiver age, education, years married, religion, and income level. African American caregivers were younger, had less education and income, and had been married fewer years than White caregivers. Race and gender groups differed only in terms of number of years married and education; African American women had the highest proportion of second marriages and the lowest educational levels.

Because of the lower marital rates among African Americans, this sample may be less representative of Black caregivers compared with Whites. As in other studies of caregivers, the sample did not represent an easily definable universe (Lawton et al. 1992). The sample was not representative of either those community-dwelling spouse caregivers who did not define a need for medical assessment or help for an impaired spouse or of those who sought medical help from local physicians rather than well-known hospital programs. In addition to possible differences in mortality, morbidity, and willingness to participate in the study by caregiver gender and race, differential rates of institutionalization by race may also have affected sample representativeness. At comparable levels of disability, African Americans are less likely to be placed in nursing homes than Whites. It is also possible that the social support networks of those who do not seek out assistance for their impaired spouses differ from those who do.

Measures
Helper level measures
Helper level measures were derived from the concentric circle approach to identifying social support networks described by Antonucci and Akiyama 1987. Participants were presented with a set of three concentric circles and asked to identify "all the people who currently help you take care of your spouse in any way." The innermost circle A was for persons who were "most helpful to you, who provide the kind of help to you that you probably couldn't get along without." Circle B was for "people who don't help you quite as much but are still important sources of help," and circle C was for "persons who help only a little bit." There were no restrictions on the number of people who could be nominated.

The 210 spouse caregivers named 1,560 persons as helpers who were identified by their relationship to the caregiver and by their gender. Relationship to the caregiver was a categorical variable coded as adult children, other relatives, friends and neighbors, and formal helpers. Gender of helper was coded Approximately 40 helpers could not be identified by gender (e.g., home care agency or neighbor). In order not to omit cases of the relatively rare occurrence of use of formal help, the 23 helpers who were identified as representing formal agency staff were randomly assigned a value of female or male, using a 80:20 ratio of women to men. This ratio reflects the gender distribution of agency helpers. The remaining 17 helper cases (1% of the nominated helpers) were dropped from the analysis. Analyses of the final models that used imputed and missing versions of gender were similar and we present results dropping the missing gender cases. Approximately two-thirds (943; 63%) of the helpers were women. Adult children were the most common source of help (31%), followed by friends or neighbors (29%), other relatives (25%), and formal sources including clergy (15%).

In follow-up questions about social support, caregivers were asked to identify who they could count on specifically for five types of support: emotional support, practical assistance, advice and guidance, help in making decisions, and socializing. The largest number of nominations of types of support was for emotional support and practical assistance. Specifically, caregivers were asked, "Whom can you count on for emotional support, i.e., sympathy, caring, understanding or reassurance? You may say ‘no one’ or list up to nine people from your circles." A similar set of questions asked whom they could count on for practical assistance, that is, concrete assistance with chores, tasks, or other needs. Each helper was coded by the caregiver nomination of the type of support they provided. Thus, provision of emotional support and practical assistance were coded dichotomously for each nominated helper, that is, The number of other support functions provided by the helper ranged from 1 to 4, excluding the function under study.

Caregiver level measures
Preliminary analyses investigated a range of caregiver and impaired spouse variables believed to be related to type and function of the helper. Caregiver education, years of caregiving, proximity of adult child, caregiver health, income adequacy, depression, family strain, and impaired spouse's cognitive status and number of behavior problems were considered but found to be nonsignificant in all analyses and were thus dropped from the final models. The race–gender group of the caregiver was coded in four categories: African American women, African American men, White men, White women. White women were the reference group.

We assessed functional limitations with questions about whether or not the care receiver needed help with a specific ADL. Responses were categorized as performed without help versus needed help or unable to do. A measure of the number of functional limitations was created by summing 14 ADL limitations: inside mobility, eating, dressing, grooming, bed mobility, bathing, toileting, housework, laundry, preparing meals, travel outside home, grocery shopping, using phone, and managing money (coded 0 or 1). The reliability of this measure was .94 (Cronbach's alpha). Although personal care and instrumental ADLs can be seen as tapping different dimensions of functional ability, research suggests that a hierarchy of limitations exists that points to a progressive trend of increased disability (Spector, Katz, Murphy, and Fulton 1987; Wiener, Hanley, Clark, and Van Nostrand 1990).

Plan of Analysis
The data in this study are hierarchical in nature: responses to questions regarding helpers and their functions are nested within caregivers. As noted earlier, helpers are not randomly assigned to caregivers, creating intraclass correlation. The presence of intraclass correlation violates the basic assumption of independence of sample observations embedded in standard multivariate regression models. Applying conventional models of analysis could produce misleading estimates of statistical significance.

In addition, our substantive interest is in the cross-level interactions between helper and caregiver characteristics. Thus we developed models of the joint interaction between caregiver and helper to determine the probabilities of providing emotional support and practical assistance controlling for functional limitations of the impaired spouse. Identification of a cross-level interaction between caregiver and helper means that the impact of a caregiver's characteristic on the dependent variable (i.e., probability of emotional support or practical assistance) varies by the level of helper characteristics.

Because the dependent variables in this study were dichotomous, we developed population-average logistic regression models (Gibbons and Hedeker 1994; Zeger, Liang, and Albert 1988) using the nonlinear Bernoulli module offered by the HLM computer software package, Version 4 (Bryk, Raudenbush, and Congdon 1996). The population-average logistic regression model is an application of Hierarchical Generalized Linear Modeling.

In this study, we are more interested in the average difference between caregivers in the probability of receiving emotional support or practical assistance than in individual differences. For example, we want to know, other things being equal, how male caregivers differ from female caregivers in the probability of receiving emotional support on average in the population, rather than how sample caregiver John differs from sample caregiver Amy in the probability of interest. We used the Generalized Estimating Equation (GEE) method to estimate the population-average logistic regression model, which provides robust estimation of standard errors (Breslow and Clayton 1993; Diggle, Liang, and Zeger 1995; Zeger, Liang, and Albert 1988). This approach provides a conservative test of our research hypotheses.

In our population-average regression model, the helper is the basic unit of analysis. The log odds of providing a certain type of help (emotional support or practical assistance) is represented by a linear function of helper's characteristics (i.e., gender, relationship to the caregiver, and number of other types of support provided), caregiver's characteristics (i.e., gender, race, and spouse's ADL limitations), and the interaction between the helper's and caregiver's characteristics. All continuous variables were centered around their sample means.

We tested all possible cross-level interaction terms, but only report the significant interactions in the final models. The procedure can be summarized as follows. First, we ran one interaction at a time (i.e., we ran 20 models in total for each dependent variable). Second, we ran one model including all identified interactions that were significant in the individual runs. Third, we reran the model using only interactions that were significant in the preceding step. This process was repeated until all interactions included in the model were significant (data available from the author). This approach resulted in two significant interactions for the emotional-support model and one significant interaction for the practical assistance model.

Because the HLM estimation of nonlinear models is based on a penalized quasi-likelihood rather than maximum likelihood (Bryk, Raudenbush, and Congdon 1996), common measures of goodness of fit, such as a likelihood-ratio test, were not applicable. We carefully compared results between competing models, including the direction, magnitude, and significance of coefficients (data available from the author).

	Results

TOP Abstract Availability and Distribution of... Relationship Between Support... Methods Results Discussion References

 
Context of Caregiver Support
In Table 1 we present a description of the context of caregiver support by caregiver's race and gender. At the caregiver level, the sample composition was 25% African American women, 10% African American men, 39% White women, and 26% White men. Based on the mean number of helpers, the primary source of helpers varied by caregiver group. African American female caregivers nominated adult children and other relatives as the largest categories in their helper networks, whereas African American male caregivers were more likely to rely on friends and neighbors. White men and women reported adult children and friends/neighbors as their largest source of helpers. Overall, White caregivers reported larger helper networks and were more likely to use formal resources than were African Americans. There also was variation in the overall size of helper networks, with White female caregivers having the most helpers, and Black female caregivers the fewest. White caregivers, regardless of gender, were more likely to report receiving emotional support from their helpers, compared with African American caregivers. This pattern was less marked in terms of receipt of practical assistance: White male caregivers reported the highest amount of practical assistance, whereas African American female caregivers reported the least.

The number of other types of support provided by a helper (e.g., socializing, dependency, decision making, advising, and practical assistance or emotional support) were included in the analysis as a control variable representing the extent of a helper's involvement with a caregiver. The results shown in Table 2 indicate that helpers who provide emotional support or practical assistance are more likely to provide other types of support as well. This pattern is especially strong in the provision of practical assistance. Each additional type of support provided increases the likelihood by 3.5 times that the helper will be also providing practical assistance.

Do the Sources of Help and Receipt of Emotional Support and Practical Assistance Vary by the Spouse Caregiver's Race and Gender and Impaired Spouse's Limitations?
The findings representing caregiver's race, gender, and spouse's functional limitations are also found in Table 2 . In addition, we have graphed the significant cross-level interaction effects. Caregiver's race and gender are significant correlates of the receipt of emotional support. In general, caregiver gender is more salient than caregiver race in the receipt of emotional support. African American female caregivers are almost half as likely to receive emotional support as are White females, whereas African American and White males are about a third as likely . This general pattern, however, is conditioned by results of the cross-level interaction between source of help (adult children) and caregiver race–gender group.

Fig. 1 illustrates this interaction using predicted probabilities: White male caregivers are significantly more likely to receive emotional support from adult children compared with other sources of help and the other race–gender caregiver groups. This likelihood is even higher than that for White female caregivers. The data shown in Fig. 1 also suggest that when cross-level interactions are taken into account, White caregivers tend to receive more support from their adult children than do African American caregivers.

View larger version (42K): [in this window] [in a new window]   Figure 1. Source of help, caregiver's race and gender, and probability of emotional support.

View larger version (41K): [in this window] [in a new window]   Figure 2. Source of help, caregiver's race and gender, and probability of practical assistance.

View larger version (14K): [in this window] [in a new window]   Figure 3. Impaired spouses' activities of daily living (ADL), source of help, and probability of emotional support.

	Discussion

TOP Abstract Availability and Distribution of... Relationship Between Support... Methods Results Discussion References

Our results contribute to research focusing on understanding race and gender similarities and differences in support networks. Descriptions of the social status characteristics of caregivers as determinants of their social network may be less fruitful in understanding the benefits of social support than a focus on which helpers assist with what kinds of task for which groups of caregivers. Examination of the interaction patterns presented in Fig. 1 and Fig. 2 provide tantalizing suggestions that require further research for corroboration. For example, Fig. 1 suggests that when cross-level interactions are taken into account, adult children are more likely to provide emotional support to their White parents than are children of African American caregivers. This result adds another piece of data to the ongoing discussion about whether the expansive nature of social support in the African American extended family structure and the role of adult children in such networks are in transition (Silverstein and Waite 1993). White female spouses were likely to receive more emotional support from all sources of help, except in the case of adult children in which White caregiving husbands were advantaged. If we assume that married men's connections to kin networks operate through their wives (Akiyama, Elliot, and Antonucci 1996), we can speculate that when the wife of a male caregiver is cognitively impaired and no longer able to carry out this function, adult children provide emotional support to fill the gap.

Another interpretation is that functional support provided in the context of salient roles has more of an effect than support provided in the context of roles more marginal to a person's identity (Krause and Borawski-Clark 1994). Yet the provision of support by adult children was not uniform across all types of need. Fig. 2 shows that white male spouses were more likely to receive practical assistance from formal sources, compared with their adult children or other informal sources. We did not have the data to determine whether this pattern is a result of greater economic resources of this group of caregivers, of higher likelihood by social agencies to define this group as more helpless in household tasks or of other factors. Although we do not have comparative data on patterns of support for adult child caregivers, if our findings are specific to spouse caregivers, they support Akiyama and colleagues 1996 argument that relationship of the helper may be more salient than gender of helper in specifying support relationships.

This research contributes indirectly to recent research that emphasizes differential effects of social support. The receipt of emotional support and practical or instrumental assistance reflects different support processes that vary by caregiver characteristics (Li, Seltzer, and Greenberg 1997). Li and colleagues 1997 found that emotional support had differential effects depending on type of caregiver relationship, but caregiving support (i.e., their general measure of instrumental support) did not. Provision of practical assistance is a concrete response to a defined need. Once the need is met, its effects may be taken for granted. Thus, we surmise that the more uniform experiences with receipt of practical assistance or instrumental support may reflect the concrete and expected nature of such help. In contrast, emotional support is grounded in social relationships and thus subject to such qualitative elements as past relationship history, degree of relationship intimacy, knowledge of preferred patterns of communication, and so forth. Other studies, such as Thompson and colleagues 1993, further demonstrate the complexity of the relationship between type of support, recipient of support, and type of caregiver burden. In Thompson and colleagues' study, however, neither emotional support nor practical assistance had much effect on different types of caregiver burden. Neither of these studies examined interactions between caregiver and helper characteristics. In addition, because of differences in samples and measures of type of social support, it is difficult to identify theoretical generalizations.

Several limitations must be considered in interpreting the results of this study. First, the sample of spouse caregivers is nonrandom, restricted to dementia caregivers, and cross-sectional. In situations involving noncognitive impairments, patterns of support may vary considerably (Akiyama et al. 1996). Second, although we had illustrative information about helpers as the unit of analysis, our helper information was limited. At the minimum, research based on the social convoy model of social support needs to seek more information about the helpers' current demographic status and caregivers' perceptions of their past and current relationship quality with individual helpers and patterns of reciprocity in support provision. Because these questions can create high caregiver-respondent burden, ideally, studies of support processes should interview the helpers directly. Finally, the relatively small sample size of caregivers creates two methodological limitations. The low numbers of African American male spouse caregivers limits our ability (i.e., power) to detect differences related to this group. In addition, the GEE analytic approach is based on asymptotic theory; thus, estimates derived from such small samples can change when repeated with a larger sample. These methodological concerns further limit the generalizability of our findings.

A contribution of the current research is its application of a form of clustered data analysis to support relationships that address the indivisible association between provision and receipt of assistance. Because helpers are not randomly assigned to caregivers and because we are asking caregivers to report on their helpers, key assumptions of independence of measures in traditional multivariate approaches are violated. Caregiver-level analyses that aggregate helper network information to the caregiver level may satisfy interest in network structural attributes but may lose critical information about the specificity of helpers and variations in the types of support they provide. Simultaneous analyses of characteristics of the members of a caregiver's support network and characteristics of the caregiver are preferable. Thus, analytic strategies for clustered data permit a new range of research questions to be addressed. Addressing such questions can lead to opportunities to develop specificity hypotheses that take into account how the interaction between the "who" and "what" of social support and the attributes of the person receiving support jointly influence the stress process. This knowledge of the simultaneous interrelationship among these different aspects of social support will facilitate the tailoring of interventions for increasing support resources for caregivers. The results of this study suggest that it may be more useful to pay attention to who is providing what kinds of support as well as to the attributes of the caregivers themselves.

	Acknowledgments

 
This study was funded by National Institute of Aging Grant RO1AGO9416.

Received for publication October 9, 1998. Accepted for publication November 8, 1999.

	References

TOP Abstract Availability and Distribution of... Relationship Between Support... Methods Results Discussion References

 

• Akiyama H., Elliot K., Antonucci T. C., 1996. Same-sex and cross-sex relationships. Journal of Gerontology: Psychological Sciences 51B:P374-P382. [Abstract]
• Angel J. L., Hogan D. P., 1991. The demography of minority aging populations. Journal of Family History 17:95-115.
• Angel R. J., Tienda M., 1982. Determinants of extended household structure: Cultural pattern or economic need?. American Journal of Sociology 87:1360-1383.
• Antonucci T. C., 1990. Social supports and social relationships. Binstock R. H., George L. K., , ed.Handbook of aging and the social sciences 3rd ed. 205-226. Academic Press, San Diego, CA.
• Antonucci T. C., Akiyama H., 1987. Social networks in adult life: A preliminary examination of the convoy model. Journal of Gerontology 42:519-527. [Abstract/Free Full Text]
• Barrera M., Jr. 1986. Distinctions between social support concepts, measures, and models. American Journal of Community Psychology 14:413-445.
• Breslow N., Clayton D. G., 1993. Approximate inference in generalized linear mixed models. Journal of the American Statistical Association 88:9-25.
• Bryk A. S., Raudenbush S. W., 1992. Hierarchical linear models: Applications and data analysis methods Sage, Newbury Park, CA.
• Bryk A. S., Raudenbush S. W., Congdon R., 1996. HLM: Hierarchical linear and nonlinear modeling with the HLM/2l and HLM/3l programs Scientific Software International, Chicago, IL.
• Burton L. M., Dilworth-Anderson P., 1991. The intergenerational family roles of aged black Americans. Marriage and Family Review 16:311-330.
• Burton L., Kasper J., Shore A., Cagney K., LaViest T., Cubbin C., German P., 1995. The structure of informal care: Are there differences by race?. The Gerontologist 35:744-752. [Abstract]
• Cantor M. H., 1991. Family and community: Changing roles in an aging society. The Gerontologist 31:337-346. [Abstract]
• Cantor M. H., Little V., 1985. Aging and social care. Binstock R., Shanas E., , ed.Handbook of aging and the social sciences 745-781. Van Nostrand Reinhold, New York.
• Chatters L. M., Taylor R. J., Jackson J. S., 1986. Aged blacks choices for an informal helper network. Journal of Gerontology 41:94-100. [Abstract/Free Full Text]
• Clipp E. C., George L. K., 1990. Caregiver needs and patterns of social support. Journal of Gerontology: Social Sciences 45:S102-S111.
• Cohen S., McKay G., 1984. Social support, stress and the buffering hypothesis: A theoretical analysis. Baum A., Taylor S. E., Singer J. E., , ed.Handbook of psychology and health, (Vol. 4) 253-267. Erlbaum, Hillsdale, NJ.
• Cutrona C. E., Russell D., 1990. Type of social support and specific stress: toward a theory of optimal matching. Sarason B.R., Sarason I. G., Pierce G. R., , ed.Social support: An interactional view 319-366. Wiley, New York.
• Diggle P. J., Liang K. Y., Zeger S. L., 1995. Analysis of longitudinal data Clarendon Press, Oxford, England.
• Dwyer J. W., Coward R. T., 1992. Gender, families and elder care Sage, Newbury Park, CA.
• Franks M. M., Stephens M. A. P., 1996. Social support in the context of caregiving: Husbands provision of support to wives involved in parent care. Journal of Gerontology: Psychological Sciences 51B:P43-P52. [Abstract]
• George L. K., 1988. Social participation in later life: Black–White differences. Jackson J. S., , ed.The Black American elderly: Research on physical and psychological health 99-126. Springer, New York.
• George L. K., 1996. Social factors and illness. George L.K., Binstock R. H., , ed.Handbook on aging and the social sciences 4th ed. 229-252. Academic Press, New York.
• Gibbons R. D., Hedeker D., 1994. Application of random-effects probit regression models. Journal of Consulting and Clinical Psychology 62:285-296. [Medline]
• Haley W. E., West C. A., Wadley V. G., Ford G. R., White F. A., Barrett J. J., Harrell L. E., Roth D. L., 1995. Psychological, social, and health impact of caregiving: A comparison of Black and White dementia family caregivers and non-caregivers. Psychology and Aging 10:540-552. [Medline]
• Hamon R. R., Blieszner R., 1990. Filial responsibility expectations among adult child-older parent pairs. Journal of Gerontology: Psychological Sciences 45:110-112.
• Hanson S. M., Sauer W. J., 1985. Children and their elderly parents. Sauer W. J., Coward R. T., , ed.Social support networks and the care of the elderly 41-66. Springer, New York.
• Horowitz A., 1985. Sons and daughters as caregivers to older parents: Differences in role performance and consequences. The Gerontologist 25:612-617. [Medline]
• House J. S., Umberson D., Landis K. R., 1988. Structures and processes of social support. Annual Review of Sociology 14:293-318.
• Jackson J. J., Antonucci T. C., 1992. Social support processes in health and effective functioning in the elderly. Wykle M., Kahana E., Kowal J., , ed.Stress and health among the elderly 72-95. Springer, New York.
• Krause N., 1990. Perceived health problems, formal/informal support, and life satisfaction among older adults. Journal of Gerontology: Social Sciences 45:S193-S205.
• Krause N., Borawski-Clark E., 1994. Clarifying the functions of social support in later life. Research on Aging 16:251-279. [Abstract]
• Krause N., Liang J., Yatomi N., 1989. Satisfaction with social support and depressive symptoms: A panel analysis. Psychology and Aging 4:88-97. [Medline]
• LaViest T. A., 1994. Beyond dummy variables and sample selection: What health services researchers ought to know about race as a variable. HSR: Health Services Research 29:1-16.
• Lawton M. P., Rajagopal D., Brody E., Kleban M., 1992. The dynamics of caregiving for a demented elder among Black and White families. Journal of Gerontology: Social Sciences 47:S156-S164.
• Lee G. R., Dwyer J. W., Coward R.T., 1993. Gender differences in parent care: Demographic factors and same-gender preferences. Journal of Gerontology: Social Sciences 48:S9-S16.
• Li L. W., Seltzer M. M., Greenberg J. S., 1997. Social support and depressive symptoms: Differential patterns in wife and daughter caregivers. Journal of Gerontology: Social Sciences 52B:S200-S211. [Abstract]
• Lin N., Westcott J., 1991. Marital engagement/disengagement, social network, and mental health. Eckenrode J., Gore S., , ed.The social context of coping 175-204. Plenum, New York.
• Litwak E., 1985. Helping the elderly: The complementary roles of informal networks and formal systems Guilford Press, New York.
• Lubben J. E., 1988. Assessing social networks among elderly populations. Family and Community Health 11:42-52.
• Markides K. S., Mindel C. H., 1987. Aging and ethnicity Sage, Beverly Hills, CA.
• Matthews S. H., 1996. Gender and the division of filial responsibility between lone sisters and their brothers. Journal of Gerontology: Social Sciences 50B:S312-S321.
• Messeri P., Silverstein M., Litwak E., 1993. Choosing optimal support groups: A review and reformulation. Journal of Health and Social Behavior 34:122-137. [Medline]
• Miller B., Cafasso L., 1992. Gender differences in caregiving: Fact or artifact?. The Gerontologist 32:498-507. [Abstract]
• Miller B., Campbell R. T., Davis L., Furner S., Giachello A., Prohaska T., Kaufman J. E., Perez C., 1996. Minority use of community long-term care services: A comparative analysis. Journal of Gerontology: Social Sciences 51B:S70-S81. [Abstract]
• Miller B., Campbell R. T., Farran C. J., Kaufman J. E., Davis L., 1995. Race, control, mastery, and caregiver distress. Journal of Gerontology: Social Sciences 50B:S374-S382. [Abstract]
• Miller B., McFall S., 1991a. The effect of caregiver burden on change in frail older persons' of formal helpers. Journal of Health and Social Behavior 32:165-179. [Medline]
• Miller B., McFall S., 1991b. Stability and change in the informal task support network of frail older persons. The Gerontologist 31:735-745. [Abstract]
• Miller B., McFall S., Montgomery A., 1991. The impact of elder health, caregiver involvement, and global stress on two dimensions of caregiver burden. Journal of Gerontology: Social Sciences 46:S9-S19.
• Minkler M., Roe K. M., Price M., 1992. The physical and emotional health of grandmothers raising grandchildren in the crack cocaine epidemic. The Gerontologist 32:752-761. [Abstract]
• Mitchell J., Register J. C., 1984. An exploration of family interaction with the elderly by race, sex and residence. The Gerontologist 24:48-54. [Medline]
• Mutran E., 1985. Intergenerational family support among Blacks and Whites: Response to culture or to socioeconomic differences. Journal of Gerontology 40:382-389. [Abstract/Free Full Text]
• Pearlin L. I., Aneshensel C. S., Mullan J. T., Whitlatch C. J., 1995. Caregiving and its social support. George L. K., Binstock R. H., , ed.Handbook on aging and the social sciences 4th ed. 283-302. Academic Press, New York.
• Penning M. J., 1990. Receipt of assistance by elderly people: Hierarchical selection and task specificity. The Gerontologist 30:220-227. [Abstract]
• Seeman T. E., Berkman L. F., 1988. Structural characteristics of social networks and their relationship with social support in the elderly: Who provides support. Social Science and Medicine 26:737-749.
• Silverstein M., Litwak E., 1993. A task-specific typology of intergenerational family structure in later life. The Gerontologist 33:258-264. [Abstract]
• Silverstein M., Waite L. J., 1993. Are Blacks more likely than Whites to receive and provide social support in middle and old age? Yes, no, and maybe so. Journal of Gerontology: Social Sciences 48:S212-S222.
• Smerglia V. L., Deimling G. T., Barresi C. M., 1988. Black/White family comparisons in helping and decision making networks of impaired elderly. Family Relations 37:305-309.
• Spector W. D., Katz S., Murphy J. B., Fulton J. P., 1987. The hierarchical relationship between activities of daily living and instrumental activities of daily living. Journal of Chronic Disease 40:481-489. [Medline]
• Stoller E. P., 1990. Males as helpers: The role of sons, relatives, and friends. The Gerontologist 30:228-235. [Abstract]
• Stoller E. P., Pugliesi K. L., 1991. Size and effectiveness of informal helping networks: A panel study of older people in the community. Journal of Health and Social Behavior 32:180-191. [Medline]
• Stone R., Cafferata G., Sangl J., 1987. Caregivers of the frail elderly: A national profile. The Gerontologist 27:616-626. [Medline]
• Suitor J., Pillemer K., 1993. Support and interpersonal stress in the social networks of married daughters caring for parents with dementia. Journal of Gerontology: Social Sciences 48:S1-S8.
• Taylor R. J., 1986. Receipt of support from family among Black Americans: Demographic and familial differences. Journal of Marriage and the Family 48:67-77.
• Taylor R. J., Chatters L. M., 1991. Extended family networks of older Black adults. Journal of Gerontology: Social Sciences 46:S210-S217.
• Tennstedt S. L., McKinlay J. B., Sullivan L. M., 1989. Informal care for frail elders: The role of secondary caregivers. The Gerontologist 29:677-683. [Abstract]
• Thoits P. A., 1995. Stress, coping, and social support processes: Where are we? What next?. Journal of Health and Social Behavior 36: (extra issue) 53-79.
• Thompson E. H., Futterman A. M., Gallagher-Thompson D., Rose J. D., Lovett S. B., 1993. Social support and caregiving burden in family caregivers of frail elders. Journal of Gerontology: Social Sciences 48:S245-S254.
• Turner R. J., Marino F., 1994. Social support and social structure: A descriptive epidemiology. Journal of Health and Social Welfare 35:193-212.
• Vaux A., 1988. Social support: Theory, research, and intervention Praeger, New York.
• Walker A. J., 1992. Conceptual perspectives on gender and family caregiving. Dwyer J. W., Coward R. T., , ed.Gender, families and elder care 34-48. Sage, Newbury Park, CA.
• Wellman B., Wortly S., 1990. Different strokes from different folks: Community ties and social support. American Journal of Sociology 96:558-588.
• Wethington E., Kessler R., 1987. Perceived support, received support, and adjustment to stressful life events. Journal of Health and Social Behavior 27:78-89.
• Wiener J. M., Hanley R. J., Clark R., Van Nostrand J. F., 1990. Measuring the activities of daily living: Comparisons across national surveys. Journal of Gerontology: Social Sciences 45:S229-S237.
• Zeger S., Liang K., Albert P., 1988. Models for longitudinal data: A likelihood approach. Biometrics 44:1049-1060. [Medline]

This article has been cited by other articles:

				M. E. Szinovacz and A. Davey Retirement Transitions and Spouse Disability: Effects on Depressive Symptoms J. Gerontol. B. Psychol. Sci. Soc. Sci., November 1, 2004; 59(6): S333 - S342. [Abstract] [Full Text] [PDF]

				D. Gallagher-Thompson, D. W. Coon, N. Solano, C. Ambler, Y. Rabinowitz, and L. W. Thompson Change in Indices of Distress Among Latino and Anglo Female Caregivers of Elderly Relatives With Dementia: Site-Specific Results From the REACH National Collaborative Study Gerontologist, August 1, 2003; 43(4): 580 - 591. [Abstract] [Full Text] [PDF]

				P. Dilworth-Anderson, I. C. Williams, and B. E. Gibson Issues of Race, Ethnicity, and Culture in Caregiving Research: A 20-Year Review (1980-2000) Gerontologist, April 1, 2002; 42(2): 237 - 272. [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]

This Article Abstract Full Text (PDF) Services Download to citation manager Citing Articles Citing Articles via HighWire PubMed PubMed Citation