| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||
Correspondence: Address correspondence to Lydia Li, School of Social Work, University of Michigan, 1080 S. University, Ann Arbor, MI 48109-1106. Email: lydiali{at}umich.edu
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionReferences |
|---|
Key Words: Frail elders • Informal care networks • Caregiving • Home- and community-based services
Informal Care Structure and Function: Racial Differences
Caregiving research related to race has proliferated and taken different directions in the past few decades (Dilworth-Anderson et al., 2002). Our literature review focused on empirical studies that had addressed racial differences in informal care structure (e.g., caregiving network size) or function (e.g., amount of informal care provided or received). We noted two limitations in this literature: inconsistent findings and limited scope.
Previous studies have focused on caregiving network composition, size, and amount of care for racial comparison. Regarding composition, it is generally agreed that Black elders are more likely than Whites to be cared for by extended family members (Burton et al., 1995; Thornton, White-Means, & Choi, 1993). Research findings related to network size and amount of care are equivocal, however. Some studies found that Blacks had more informal helpers than comparable Whites (Miller & McFall, 1991), and Black caregivers were less likely to be the sole care provider than White caregivers (Stommel, Given, & Given, 1998). Other studies found no significant racial differences in caregiving network size (Burton et al., 1995; Thornton et al., 1993). Likewise, some studies reported that Black caregivers provided or Black elders received more hours of informal care than Whites (Fredman, Daly, & Lazur, 1995; Navaie-Waliser et al., 2001; Tennstedt & Chang, 1998), whereas other studies found minimal racial differences in the extent of help provided to or received by disabled elders (Lawton, Rajagopal, Brody, & Kleban, 1992; Miner, 1995).
The reasons for the inconsistent findings include the commonly known methodological issues, such as variations in samples and measures. Study variation in the use of control variables also contributes to inconsistency. Broadly speaking, previous studies can be classified into two categories, based on their approach to examine racial differences. The first category aimed to examine differences on the dependent variables between racial groups (see, e.g., Fredman et al., 1995; Navaie-Waliser et al., 2001). The second category aimed to isolate the effect of race from socioeconomic status (SES) and other demographic variables (see, e.g., Peek, Coward, & Peek, 2000; Tennstedt & Chang, 1998; Thornton et al., 1993). The selection of control variables can be quite different between studies in these two categories, which is partly responsible for the different findings regarding race.
In addition, some aspects of informal care have received relatively little attention. For instance, few studies have directly examined racial differences in living arrangements among dependent elderly persons. Some researchers implied that older Blacks are more likely to receive intrahousehold care than Whites because they are more likely to be living in extended, multigenerational households (Angel, Angel, & Himes, 1992). Furthermore, less is known about racial differences in types of informal help received. Although multigenerational household arrangements may result in a situation in which Blacks receive more help with personal care and household chores, the lower SES of Black caregivers may decrease their likelihood of providing certain assistance, such as that which requires car ownership (e.g., transportation or shopping). This is not known. In order to understand racial differences in the informal home care received comprehensively, we include the less-studied components of informal care as outcomes.
Study Population, Outcomes, and Conceptualization
We focus on the population of older persons that relies on informal home care the most: community-dwelling elders with low incomes and relatively severe disabilities. These frail elders have been underrepresented in previous studies of informal care because only a small percentage of older persons in the community manifest severe functional limitations (Wiener, Hanley, Clark, & Nostrand, 1990). The study sample came from elderly applicants to Michigan's Medicaid Waiver program, which provides home- and community-based services to low-income individuals at risk of nursing home placement. This sample offers an opportunity for us to examine the informal home care received by a group of vulnerable older persons: low-income frail elders. However, this sample is not comparable with nationally representative samples, as it came from a single Midwestern state with a predominately urban Black population, and the elders in the sample were more financially deprived and disabled than older people in general.
As we alluded to earlier, most previous studies have focused on a limited number of informal care components for racial comparison. These frameworks may have missed the overall picture of informal home care, as well as how different components vary by race. In this study we build on earlier work to examine racial variation in multiple structural and functional components of elders' informal care networks. The structural components are the vehicle by which informal care is provided or received, such as source of care and living arrangement. The functional components are those that are actually provided or received, such as amount and types of assistance. Across the analyses of multiple informal care components, we examine the effect of race independent of other sociodemographic characteristics (SES, marital status, gender, age, and residential area). We selected these factors because they have been known to influence the receipt of informal care and they covary with race (Coward, Horne, & Dwyer, 1992; Coward, Lee, Netzer, Cutler, & Danigelis, 1996; Miner, 1995).
What is most important is that this study conceptualizes disability as the mediating mechanism contributing to racial differences seen in the informal home care received. Two perspectives have been used to explain racial differences in helping behaviors in the literature: socioeconomic differences and ethnic subculture (Mutran, 1985). In this study, we extend both perspectives and argue that the different level of disability between older Blacks and Whites is the primary reason for the differences in informal care received by these individuals.
Research has documented that older Blacks have more debilitating health problems and functional disability than Whites (Geronimus et al., 2001; Kelly-Moore & Ferraro, 2004). This discrepancy is related to the political, social, and economic disadvantages that many Blacks have accumulated over their life course (Barnes, Mendes de Leon, Wilson, et al., 2004; House et al., 1992; Hummer, 1996). In addition, Blacks are less likely to use nursing homes than Whites (Wallace, Levy-Storms, Kingston, & Andersen, 1998), which translates into an overrepresentation of Blacks with severe disability in the community. The lower rate of institutionalization of Blacks may be related to cultural values and family structure (Cagney & Agree, 1999; Sudha & Mutran, 1999), compounded with discrimination and barriers to service use (Cagney & Agree).
The higher disability levels of Blacks generate a greater need for care. The care need can drive how the informal care structure is organized, and how much and what type of care is provided. For instance, studies have shown that functional limitations of older persons are positively correlated with their caregiving network size, and that informal caregivers respond to their care recipients' health declines by increasing the amount (Miller & McFall, 1991; Tennstedt & Chang, 1998) and scope of assistance (Stoller & Pugliesi, 1991). In other words, older Blacks' higher disability levels would require their informal care networks to be structured and function in a way that differs from that of Whites' networks.
Research Questions and Hypotheses
On the basis of prior research, we ask two questions: Are there racial differences in the structural and functional dimensions of informal home care after we control for sociodemographic characteristics including SES, age, gender, marital status, and population density? Does disability account for the remaining differences in the informal home care received by older Blacks and Whites?
In this study we examine three structural (primary source of care, living arrangement, and sole caregiver) and three functional (amount of care, scope of assistance, and types of assistance) informal care components (definitions are given in the Variables and Measures section). We test two major hypotheses. First, Black and White elders will differ in all structural and functional components of informal home care, even after we control for sociodemographic characteristics. Second, disability accounts for racial differences in all structural and functional components of informal home care.
|
Methods |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
Applicants to the waiver program are telephone screened for eligibility, on the basis of two criteria. The first is financial: Individuals must have income at or below 300% of Supplemental Security Income and have assets not exceeding the limit set for Medicaid. The second is functional: Individuals must be medically appropriate for nursing home placement (Tilly & Kasten, 2001). Those deemed eligible then receive a comprehensive in-person assessment, conducted by a nurse and a social worker. The assessment instrument used—the Minimum Data Set for Home Care (MDS-HC)—evaluates multiple aspects of the applicants' lives, including basic demographics, physical and cognitive functioning, and informal support.
Sample
We base our present analysis on the in-person assessment data of waiver applicants who were interviewed between July 1, 1999, and May 31, 2001. We selected only Blacks and Whites aged 65 years or more, for a total of 6,613 persons. We excluded 279 elders who did not have a caregiver. There was no racial difference in the likelihood of having a caregiver.
All study variables except age, in-home chores, and out-of-home chores had missing data. The percentage of missing data varied across variables, ranging from 0.7% (gender) to 14.2% (amount of care). We used listwise deletion in this analysis (Allison, 2002), resulting in 4,118 study participants. Compared with the study sample, the excluded members had a higher percentage of Whites, were more educated, and had fewer limitations in instrumental activities of daily living (IADLs). We ran additional analyses with imputed missing value for all independent variables, using conditional and marginal mean replacement. The pattern of results of these additional analyses was similar to what we reported by using listwise deletion (results available from L. Li).
Variables and Measures
Table 1 presents all study variables and their coding systems. As we noted earlier, dependent variables included three structural (primary source of care, living arrangement, and sole caregiver) and three functional (amount of care, scope of assistance, and types of assistance) measures of informal care.
|
Amount of care represented the number of hours per week that the elder received help with daily life activities from all informal caregivers. We used a log transformation to normalize this variable because of a skewed distribution. We defined scope of assistance as the number of task areas (personal care, meal preparation, housekeeping, medication, shopping, transportation, and personal finances) with which the elder received informal help. Types of assistance consisted of three dummy coded measures for personal care, in-home chores, and out-of-home chores (see Table 1 for dummy codes).
The predictors constituted three categories: race, disability, and sociodemographic covariates. We coded race, the primary variable of note in the study, to represent Black (1) versus White (0). Disability, which we conceived as a mediator, had three indicators: limitations in ADL, limitations in IADL, and limitations in cognitive functioning. Limitations in ADL were the sum of eight basic ADLs (e.g., dressing, eating, and toilet use) rated on a 5-point subscale from 0 = independent to 4 = totally dependent. (In addition, we collapsed a rating of "activity did not occur" with total dependence, assuming that nonoccurrence of the activity was due to complete dependence.) Similarly, limitations in IADL were the sum of seven IADLs (e.g., meal preparation, housework, and managing finances) rated on a 3-point subscale (0 = no difficulty to 2 = great difficulty). Cronbach's alpha reliabilities of the ADL and IADL scales in this study were.92 and.79, respectively. We measured cognitive limitations by using an adaptation of the MDS Cognitive Performance Scale (CPS), which is highly predictive of Folstein's Mini-Mental State Exam scores (Morris, Fries, Mehr, Hawes, & Phillips, 1994). The CPS ranged from intact (0) to severely cognitively impaired (6).
Sociodemographic covariates included SES, age, gender, marital status, and population density. To measure SES, we used three separate variables: education (dummy coded; less than or equal to eighth grade, which was the reference category, ninth grade to high school, and more than high school), home ownership (1 = own home; 0 = other), and adequate income (1 = income adequate to meet needed expenses; 0 = inadequate). We coded age in years. We coded gender as 1 = female, 0 = male. Marital status had four categories (dummy coded; married, which was the reference category, widowed, separated or divorced, and never married). We obtained the population density by merging individual-level data with Census 2000 data by zip code. We measured the variable as population per square mile, and we log transformed it because of a skewed distribution. Population density was an indicator of the degree of urbanization of the elder's residential area.
Analysis
We first conducted a bivariate analysis to examine unadjusted mean differences between the two racial groups on the six components of informal care, as well as their degree of disability. We then conducted a multivariate analysis. We used binary logistic regression for dependent variables that were dichotomous (living arrangement, sole caregiver, and types of assistance). For primary source of care, in which the dependent variable had four categories, we used multinomial logistic regression. We conducted ordinary least squares (OLS) regression for amount of care and scope of assistance.
The multivariate models were similar across all dependent variables. We modeled each component of informal care as a function of race, age, gender, SES, marital status, population density, and disability. We entered independent variables in two successive blocks, with disability added to the second model. On the basis of the research of Baron and Kenny (1986), we deemed that disability completely accounted for racial differences in an outcome if (a) racial differences in disability were significant; (b) disability significantly predicted the outcome; and (c) the effect of race on the outcome variable changed from statistically significant to insignificant before and after we controlled for disability.
|
Results |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
Multivariate Analysis
The results of multivariate analyses are presented in Tables 2–5. To save space, we discuss only the results pertaining to the effects of race and disability on each dependent variable here. Other significant predictors are indicated in the tables.
|
Disability had significant effects on the source of care. Specifically, more limitations in ADLs and IADLs were related to an increased likelihood of being cared for by a spouse than other relatives. Limitations in IADLs increased the probability of receiving care from adult children, whereas cognitive limitations decreased the likelihood of being cared for by friends compared with other relatives.
Living Arrangement
Our analysis of living arrangement (Table 3), controlling for sociodemographic covariates, suggests that older Blacks were more likely than Whites to live with their caregivers. The effect of race remained significant after we added disability into the model, suggesting that racial differences in living arrangement were not explained by disability. All indicators of disability (ADL, IADL, and cognitive limitations) positively predicted the likelihood of coresidence.
|
Amount of Care
For amount of care, Table 4 shows that older Blacks received more log hours of informal care than Whites, when we controlled only for sociodemographic characteristics (Model 1; ß =.07, p <.001). The effect of race became statistically insignificant after we added disability indicators into the model (Model 2; ß =.01, p = ns). All indicators of disability significantly and positively predicted amount of care. Therefore, this set of findings supports our hypothesis that disability accounts for racial differences in the amount of informal care received.
|
Types of Assistance
Table 5 shows that the influence of race on the types of assistance varied, depending on the specific type of help. Controlling for sociodemographic characteristics only, we found that older Blacks were more likely than Whites to receive personal care (odds ratio or OR = 1.30; 95% confidence interval or CI = 1.10–1.54, p <.01). After adding disability to the model, we found that the effect of race became insignificant (OR =.99, 95% CI = 0.82–1.21, p = ns), implying that the higher disability levels of older Blacks account for their greater likelihood of receiving personal care. Racial differences in the likelihood of receiving help with in-home chores, however, were not attributed to disability. Older Blacks were more likely than Whites to be assisted with in-home chores, both when we controlled only for sociodemographic covariates as well as after we added disability indicators as controls. We did not find a racial difference in being helped with out-of-home chores—the effect of race was not statistically significant in both Models 1 and 2. Thus, disability accounts for racial differences in personal care but not in-home chores. Out-of-home chores are associated with disability but not race.
|
|
Discussion |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
We found that race had significant effects on most structural and functional components of informal home care, independent of sociodemographic characteristics, including SES, marital status, population density, age, and gender. Specifically, Blacks were more likely than Whites to be cared for by other relatives, live with their caregivers, have a sole care provider, receive more and broader assistance, and be helped with personal care and in home-chores.
The differences in most functional components of informal care (amount of care, scope of assistance, and personal care) ceased to be significant after we controlled for disability. These findings indicate that older Blacks' greater need for care than Whites is the reason for racial differences in the functional dimension of informal care. Given the same level of disability, Blacks and Whites receive the same amount and scope of assistance and are similar in their likelihood of receiving help with personal care.
Racial differences in all structural components of informal care (source of care, living arrangement, and sole caregiver), however, continued to be significant after we controlled for disability. This suggests that racial differences in the structural dimension of informal care are accounted for by factors not included in this analysis. These other factors may be related to culture, as our analysis suggests that none of the covariates we had available can account for racial differences in the three structural components. Prior studies have suggested several candidates for these cultural factors, including family networks (Taylor & Chatters, 1991), household structure (Peek et al., 2000), preference for family care (Sudha & Mutran, 1999), and family care value (Lawton et al., 1992; Mutran, 1985). These factors have been suggested to predict the use of informal help and vary by race. This study did not have variables indicating culture and therefore cannot investigate its effects. Research is needed to examine whether and which cultural factors account for racial differences in the structure of informal care networks. Figure 1 presents a conceptual model for future investigations. Research based on this model may help us better understand the cultural dimension of informal care.
|
In addition, we found that older Blacks were more likely than Whites to have only one caregiver, after we controlled for sociodemographic characteristics (and disability). Although this finding is contrary to the literature that suggests Blacks to have extended family networks (Taylor & Chatters, 1991), it is consistent with more recent studies that report Blacks to have smaller social networks than Whites (Ajrouch, Antonucci, & Janevic, 2001; Barnes, Mendes de Leon, Bienias, & Evans, 2004). This finding is also in concert with the argument that extended networks of Black families may not be as prevalent as they were in the past, because of the changing sociodemographics and increasing economic deprivation of Blacks (Roschelle, 1997).
Among the sociodemographic covariates used in this study, marital status shows consistent significant effects on all dependent variables, suggesting that availability of spouse is critical in determining the informal care received by disabled older persons. We are aware that marital status and some of the dependent variables, especially source of care and coresidence, may be confounded. In an additional analysis, we dropped marital status from all the models and the results did not change (available from L. Li).
This study has two practice implications. First, the discrepancy in disability between Blacks and Whites shown in this study deserves more attention. Further analysis shows that their differences in disability are not confounded by age. It seems that Blacks delayed enrolling in the waiver program until their disabilities became quite severe. The reason for the delay is not clear, but additional analysis suggests that it is unlikely to be related to the age requirement of the waiver program. The minimum eligibility age for Michigan's Medicaid Waiver program is 18 years (we selected only elderly waiver applicants as the sample in this study). An analysis including all applicants suggests that Blacks were older (age, M = 72.5 vs. 71.3 for Whites) and more disabled (racial differences in ADL, IADL, and cognitive limitations were significant, p <.001 for all) than Whites at the time of application.
These data convey a clear message to service providers: Greater efforts are needed to enroll Blacks into community-based long-term-care programs early. Yeatts, Crow and Folts (1992) have offered an array of strategies to overcome barriers to service use for low-income minority elderly individuals. Two may be particularly useful in an outreach program to Blacks. The first is to enlist the cooperation of significant individuals in the Black community, such as clergies and physicians to inform Blacks of an available service (e.g., waiver). The second is to increase service availability in Black neighborhoods, particularly if the service is in demand and delivered on a first come, first served basis, such as Michigan's Medicaid Waiver program.
Another implication is that practitioners need to be more aware of the meanings and expressions of "burden" of Black caregivers, and they should provide support promptly. Although the literature suggests that Black caregivers are less burdened than Whites (Lawton et al., 1992), some researchers have questioned whether the current constructs of burden adequately capture the experience of Black caregivers (Fox et al., 1999). As we already discussed, the informal care networks of low-income Blacks may be quite strained and the informal support available to their caregivers may not be as extensive as commonly assumed. Practitioners, therefore, should be more sensitive to the experience of low-income Black caregivers.
Compared with that of prior research, the sample size of this study, particularly the number of Black elders, is large, which has allowed for a more powerful test of racial differences. Nevertheless, the results should be interpreted with caution, particularly regarding the generalization of findings. As we noted, this sample represents a distinct group—community-dwelling elderly persons with low incomes and limited functional ability. The sample does not represent other elders in the community or those in institutions. Moreover, the sample is limited to a single Midwestern state with a predominantly urban Black population. Although we controlled for degree of urbanization by means of population density in the analysis, regional differences in the size of support networks of Blacks have been reported (Chatters, Taylor, & Jackson, 1985). Our sample of Blacks may have fewer social resources than Blacks in general, as they were mostly urban residents with low incomes. Mutran (1985) has suggested that poverty severely limits any voluntary help, especially for Blacks. Therefore, our findings may not be comparable with those based on national probability samples, or generalized to elders in other regions or social class. The quality of the type of data used in this study has also been questioned, because the data were collected primarily for administrative and care planning purposes (Teresi & Holmes, 1992). More recent studies, however, tend to conclude that this type of data is valid and reliable and that it is appropriate for research purposes (Landi et al., 2000; Lawton et al., 1998).
Conclusions
One particularly important aspect of this study is that it investigated racial differences in both structural and functional dimensions of informal care. Although we found racial differences in almost all informal care components, controlling for sociodemographics, the analysis shows that racial differences in most functional components are accounted for by disability, whereas differences in all structural components are not. These findings suggest that older Blacks receive more hours and a broader scope of informal care than Whites, primarily because they have a greater need. The severity of disability among low-income older Blacks in the community therefore is a great concern. However, other factors, possibly cultural, are responsible for racial differences in the structural dimension of informal care. Research is needed to examine whether and which cultural factors can explain racial differences in the structure of informal care networks.
|
Footnotes |
|---|
1 School of Social Work, University of Michigan, Ann Arbor.
2 Institute of Gerontology & School of Public Health, University of Michigan, Ann Arbor.
Decision Editor: Linda S. Noelker, PhD
Received for publication February 20, 2004. Accepted for publication July 6, 2004.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
| ||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
