|Home | About | Journals | Submit | Contact Us | Français|
To investigate whether the interaction between individual race/ethnicity and community racial/ethnic composition is associated with health-related home care use among elderly persons in the United States.
A nationally representative sample of community-dwelling elders aged 65+ from the 2000 to 2006 Medical Expenditure Panel Survey (N=23,792) linked to block group-level racial/ethnic information from the 2000 Decennial Census.
We estimated the likelihood of informal and formal home health care use for four racial/ethnic elderly groups (non-Hispanic [NH] whites, NH-blacks, NH-Asians, and Hispanics) living in communities with different racial/ethnic compositions.
NH-Asian and Hispanic elders living in block groups with ≥25 percent of residents being NH-Asian or Hispanic, respectively, were more likely to use informal home health care than their counterparts in other block groups. No such effect was apparent for formal home health care.
NH-Asian and Hispanic elders are more likely to use informal home care if they live in communities with a higher proportion of residents who share their race/ethnicity. A better understanding of how informal care is provided in different communities may inform policy makers concerned with promoting informal home care, supporting informal caregivers, or providing formal home care as a substitute or supplement to informal care.
As the elderly population (ages 65 and older) continues to grow (U.S. Census Bureau 2008), policy makers, the elderly, and their families are concerned with ways to maximize the likelihood that elders can remain living in their communities and “age in place.” One approach to promote independent living and avoid nursing home placement is ensuring that elders can obtain health-related services in their homes. Understanding home care utilization and factors associated with such usage, therefore, has become an increasingly important health services concern.
In the United States, the primary form of home care for community-dwelling elders is informal (unpaid) care provided by family members or friends, with formal (paid) care supplemented by hired helpers (Kemper 1992; Davey et al. 2005;). Previous studies have found that having informal care is independently associated with lower risk of nursing home admission (Newman et al. 1990; Jette, Tennstedt, and Crawford 1995; Charles and Sevak 2005;), except in the presence of informal caregiver burnout (McFall and Miller 1992; Boaz and Muller 1994; Jette, Tennstedt, and Crawford 1995;). Although the association between formal care use and institutional placement remains unclear (McFall and Miller 1992; Greene, Lovely, and Ondrich 1993; Jette, Tennstedt, and Crawford 1995;), research shows that formal assistance is particularly important when specialized health services are needed or elders' dependence exceeds the capacity of informal caregivers (Stoller and Pugliesi 1988; Edelman and Hughes 1990;).
Most research on caregiving has reported racial/ethnic differences in the use of informal and formal home care, with racial/ethnic minority elders being more likely to receive care from family and friends and less likely to rely on formal care compared with non-Hispanic (NH) whites (Mui and Burnette 1994; Mitchell, Mathews, and Hack 2000; Dilworth-Anderson, Williams, and Gibson 2002;). Racial/ethnic differences in the use of home care have been attributed to psychosocial factors (e.g., cultural values toward caregiving for elders) (Dilworth-Anderson, Williams, and Gibson 2002; Bradley et al. 2004;); health needs (e.g., cognitive impairment and disability) (Liu, Manton, and Aragon 2000; Yeo 2001;); economic factors (e.g., income and insurance availability) (Kemper 1992; Weiss et al. 2005;); and presence of informal caregivers (e.g., household living arrangement) (Burton et al. 1995). In this study, we contribute to existing research by examining how individual race/ethnicity may interact with the racial/ethnic composition of the communities in which elders live to influence the likelihood of using home care.
We hypothesize that elders living in communities with a higher proportion of residents who share their racial/ethnic background may be more likely to obtain home care (both informal and formal) compared with elders living in communities with a lower proportion. One reason for this hypothesis is related to health care supply. Communities with a higher proportion of a specific racial/ethnic group may have a supply of service providers, both volunteer and professional, who specialize in meeting the medical and caregiving needs of elders of that race/ethnicity. Another reason pertains to the ideas from research on social capital and social integration. Some research suggests that informal and formal social networks are more likely to flourish in racially/ethnically homogeneous communities (Alesma and Ferrara 2000). These networks, in turn, may facilitate exchange of health-related information and resources helping elders obtain home care. Further, receiving care in one's home may require a high degree of trust with caregivers. Interpersonal trust and norms of reciprocity, important components of social capital, may be stronger among members of communities with a higher proportion of certain racial/ethnic minority groups (Alesma and Ferrara 2002), potentially making home health services more desirable to elders of that particular racial/ethnic group.
Although the relationship between some community characteristics and home care has been examined, prior studies have focused on socioeconomic compositions (Reeder and Calhoun 2002) and rural-versus-urban settings (Congdon and Magilvy 1998; McAuley, Spector, and Van Nostrand 2009;). Other research suggests that community race/ethnic composition is associated with a variety of health services and outcomes, including preventive care use (Wells and Horm 1998; Benjamins, Kirby, and Bond Huie 2004;), barriers to health services use (Haas et al. 2004), and mortality rates (Hart et al. 1998; Jackson et al. 2000;). However, none to our knowledge has examined the association between community racial/ethnic composition and home care use. This paper therefore fills an important gap in current literature. Our findings will help policy makers understand the complex interaction between individual race/ethnic identity and community racial/ethnic composition, and its association with home care utilization. In turn, home health services can be better targeted for elders, especially those with the least access and highest need.
Data for this study came from the following two sources: individual data from the Medical Expenditure Panel Surveys (MEPS) and community data from the U.S. Decennial Census. Sponsored by the Agency for Healthcare Research and Quality, MEPS is a series of longitudinal surveys based on clustered and stratified samples of households that provide nationally representative estimates of health care use and sociodemographic characteristics for the U.S. noninstitutionalized population (Cohen et al. 1996; Cohen 1997;). To enhance the power of our analysis, we pooled seven years (2000–2006) of MEPS data. We restricted our sample to individuals age 65 and older who identified themselves as NH-whites, NH-blacks, NH-Asians, or Hispanic individuals regardless of race (N=23,792). Elderly individuals of other races and NH elders reporting multiple races were excluded from this study because of limited sample size and heterogeneity (N=427).
Residential addresses of MEPS respondents were geocoded and linked to information from the 2000 U.S. Decennial Census regarding the racial/ethnic composition and socioeconomic characteristics of block groups. Communities of elders' residence, therefore, were operationalized in this study as “block groups,” which were the smallest geographic areas for which community-level statistics were available, containing between 600 and 3,000 people (U.S. Census Bureau 2000). The census data contained the percentages of residents in block groups by race/ethnicity and a variety of socioeconomic characteristics.
The key dependent variables measured informal and formal health-related home care use. MEPS respondents were asked whether they received any skilled or unskilled services in their homes for help with a health problem. Additionally, they were asked to report whether the care was provided by unpaid caregivers not residing with them (e.g., family members, friends, neighbors, or volunteers) or by paid care from formal sources (e.g., home health agencies, hospitals, nursing homes, or self-employed persons). From this information, we constructed two dichotomous variables (yes/no), one indicating whether an elder received any informal home care during the year and the other indicating whether any formal care was received.
The main independent variables measured elders' race/ethnicity and the racial/ethnic composition of the block groups where elders lived. At the individual level, four dichotomous variables identified whether an elder was NH-white, NH-black, NH-Asian, or Hispanic. At the community level, four additional dichotomous variables indicated whether 25 percent or more of the residents in a block group were NH-white, NH-black, NH-Asian, or Hispanic. Note that the community-level dichotomous variables were not mutually exclusive. For example, many block groups that contained more than 25 percent NH-Asians also contained more than 25 percent NH-whites.
While using 25 percent as a cutoff between “higher” and “lower” proportion of a race/ethnicity in a block group was somewhat arbitrary, our results were robust to different cutoffs and operationalizations (sensitivity analysis available upon request). We estimated all models with various cutoffs ranging from 15 to 35 percent, and the results followed a pattern similar to that seen in our final tables. Using continuous measures for block group racial/ethnic composition also yielded a similar pattern of results, although with weaker statistical significance. Another approach we considered was to define cutoffs relative to the percentage distribution of block group racial/ethnic composition (e.g., highest 75th percentile of block groups with respect to the percentage of NH-white, NH-black, Hispanic, and NH-Asian residents). This approach yielded cutoffs that varied widely by racial/ethnic composition, from 4 percent in NH-Asian block groups to 28 percent in Hispanic block groups. Such divergent cutoffs seemed objectionable both theoretically and methodologically.
We controlled for a variety of demographic, socioeconomic, and health variables (Table 1). Elders' demographics included age, sex, marital status, and interview language (English, Spanish, or other language) as a proxy for language ability and preference. Elders' socioeconomic status included highest educational attainment, household income, and poverty status (based on gross annual household income as a percentage of the federal poverty levels for the year of MEPS survey; Cohen et al. 1996), receipt of Supplemental Security Income, and health insurance status (public only [e.g., Medicare, Medicaid], and public plus private supplemental health insurance [i.e., Medigap insurance]). Elders' household living arrangement was categorized as living alone, with spouse only, with adult children and a spouse, with adult children but no spouse, or in other arrangements. To measure health status, we included variables for self-assessed overall health, functional limitations as indicated by the need for help or supervision with activities of daily living (ADLs) and instrumental ADLs, and the number of diagnosed conditions predetermined in MEPS (angina, asthma, coronary heart disease, diabetes, emphysema, hypertension, heart attack, and stroke). We also included a dichotomous variable to identify individuals with cognitive impairment defined as having any of the following: (1) experience with confusion or memory loss, (2) problems with making decisions, or (3) need of supervision for their own safety. Rural versus urban location of residence was indicated by metropolitan statistical area, which was defined as one or more counties that contain a city of 50,000 or greater, or that contain a Census-Bureau-defined urbanized area and have a total population of at least 100,000. Finally, community-level control variables included the percentage of residents in block group living under the federal poverty line, the percentage unemployed, the percentage with no high school diploma or equivalent, and the percentage receiving public assistance.
We began our analysis using bivariate statistics to describe the racial/ethnic differences in the proportion of elders with any informal or formal home care. To examine the extent to which the differences in home care use were due to variations in characteristics, we then estimated multivariate logistic regression models in which home care use (informal or formal) was regressed on elders' race/ethnicity and the control variables. Then, to examine the association between home care use and the interaction between individual race/ethnicity and community racial/ethnic composition, we estimated separate, multivariate logistic regression models stratified by elders' race/ethnicity and included the variables for racial/ethnic composition of block groups (together with all control variables). All estimates were calculated using appropriate sample weights, and standard errors were adjusted for both the complex sample design of MEPS and clustering within block groups (Rabe-Hesketh and Skrondal 2005; StataCorp 2007;).
In addition to the main analysis, supplemental analyses were performed to test the sensitivity of our results to data and coding decisions. First, we investigated the odds of using home care among all elders of any race/ethnicity who lived in a block group with a higher proportion (25 percent or more) of a particular race/ethnicity. If our hypothesis was correct, only elders living in communities with higher proportions of residents who shared their race/ethnicity should be more likely to use home care than elders of other races/ethnicities. For example, home health care use for NH-black elders should not be associated with block groups that have 25 percent or more of Hispanics or NH-Asians. Another sensitivity test was performed to ensure that our findings were not an artifact of pooling 7 years of data. We estimated all models with dichotomous variables for each year and interactions between the year and race/ethnicity (both individual and block group level). None of the coefficients pertaining to the year or year interactions were significant, and our main findings were nearly identical to those reported. All supplemental analyses are available upon request.
Descriptive results (Table 1) indicate that overall, 9.2 percent of elders had some home care use during the year (1.7 percent informal and 8.2 percent formal). Proportionately more NH-black elders used some home care compared with NH-whites, while NH-Asians had the lowest proportion. The percentages of Hispanic and NH-white elders did not differ significantly with respect to any home care utilization. The results further suggested that distinguishing between informal and formal home care would be important, at least when considering differences between NH-Asians and NH-whites. More NH-Asian elders disproportionately used informal home care than NH-whites but fewer used formal care.
Regarding other characteristics in Table 1, NH-black and Hispanic elders tended to have lower incomes, lower educational attainment, worse health, at least one indicator of cognitive impairment, and no private supplemental health insurance compared with NH-whites. Further, compared with NH-whites, more NH-black elders lived alone, while more NH-Asian and Hispanic elders lived with their adult offspring. With respect to many characteristics, NH-Asians were similar to NH-whites, except for differences in functional limitations (i.e., two or more ADLs) and education attainment.
There were also racial/ethnic differences in the characteristics of the block groups where elders resided (Table 1). Compared with NH-whites, more NH-black and Hispanic elders lived in block groups with a high prevalence of poverty, public assistance, and unemployment. Perhaps most relevant to the current study, results pertaining to block group characteristics revealed a fairly strong pattern of residential segregation by race/ethnicity. Elders tended to live in block groups with 25 percent or more of residents who shared their race/ethnicity. For example, among NH-black elders, 74.2 percent lived in block groups with at least 25 percent NH-blacks. In comparison, the percentages of NH-white, Hispanic, and NH-Asian elders residing in such block groups were 6.2, 9.4, and 9.8 percent, respectively. Furthermore, while 27.7 percent of NH-Asian elders lived in block groups with 25 percent or more NH-Asians, this percentage was significantly higher than that of other race/ethnic groups (0.8 percent for NH-whites, 1.6 percent for NH-blacks, and 3.8 percent for Hispanics). These findings are consistent with research on patterns of racial, ethnic, and economic residential segregation in the United States (Massey and Denton 1993; Iceland, Weinberg, and Steinmetz 2002; Iceland and Nelson 2008; Dwyer 2010;).
Controlling for living arrangement, household income, health status, and other characteristics, the differences between NH-black and NH-white elders in the likelihood of using any informal or formal home care were insignificant (Table 2). Differences between NH-white and NH-Asian elders, however, remained statistically significant. After including all control variables, the odds of using informal home care for NH-Asian elders was almost 3.3 times that of NH-whites. In contrast, the odds of having formal home care use for NH-Asian elders was only 25 percent that of NH-whites.
Thus far, our analysis focused on differences in home care utilization across racial/ethnic groups, but using models estimated separately by race/ethnicity, we also found evidence of differences within racial/ethnic groups depending on the racial/ethnic composition of communities where elders resided (Table 3). Hispanic elders who lived in block groups with 25 percent or more of the residents being Hispanic were more likely to use informal home care compared with Hispanic elders in other block groups (odds ratio [OR]=6.94). Similarly, NH-Asian elders residing in block groups where at least 25 percent of the residents were NH-Asian had higher odds of using informal home care than NH-Asian elders in other block groups (OR=3.12). Similar associations, however, were not observed among NH-whites and NH-blacks, nor were any apparent for formal home care.
While these findings lend support to our hypothesis, there are other plausible explanations. One possibility is that all elders who lived in block groups characterized by higher proportions (25 percent or more) of NH-Asians or Hispanics were more likely to use informal home care regardless of their own race/ethnicity. To address this, we estimated models that included all block group-level racial/ethnic composition variables in each logistic regression stratified by elders' race/ethnicity. Results indicated that only Hispanic and NH-Asian elders were more likely to use informal home care when they lived in communities with higher proportions (25 percent or higher) of NH-Asians or Hispanics, respectively; results for NH-whites and NH-blacks were not statistically significant (findings not shown but available upon request).
This study examines racial/ethnic differences in informal and formal health-related home care utilization among a nationally representative sample of U.S. elders. Consistent with previous studies, we find that even after controlling for various individual and community characteristics, NH-Asian elders are significantly more likely to use informal home care and less likely to use formal care compared with NH-white elders (Olson 2001). We extend prior research by exploring how the interaction between individual race/ethnicity and community racial/ethnic composition is associated with home care use. We find that NH-Asian and Hispanic elders who live in block groups with 25 percent or more of residents being NH-Asian or Hispanic, respectively, are more likely to report informal home care utilization than their counterparts in other block groups. No such effect is apparent for formal home care. Furthermore, these findings remain consistent in sensitivity analysis using cutoff points between 15 and 35 percent for block group-level racial/ethnic composition.
Unlike Hispanics and NH-Asians, NH-black elders do not seem to have a higher likelihood of using home care if they live in communities with higher proportions (25 percent or greater) of NH-blacks. One possible explanation for this is that NH-Asian and Hispanic elders may face barriers to obtaining home care that differ from those encountered by other older adults, such as those related to English proficiency, immigration status, or citizenship status. Moreover, these barriers may be ameliorated by concordance between individual race/ethnicity and community racial/ethnic composition. Another possible explanation is that communities with higher proportions of NH-blacks may not offer the same level of social capital that other racial/ethnic communities may provide. Prior studies have found that NH-black elders have limited ties to their residential communities, particularly in areas with high poverty or crime rates (Klinenberg 2002).
Previous studies often exclude household living arrangement as a control variable from their analysis in identifying risk factors of home care utilization because of potential endogeneity (Kemper 1992). However, regardless of household living arrangement, this study finds that community racial/ethnic composition remains significantly associated with informal home health use in Hispanic and NH-Asian elders. This suggests that communities may offer additional resources helpful in obtaining home care above and beyond the availability of live-in family caregivers for Hispanic and NH-Asians elders.
Our findings also indicate that while Hispanic and NH-Asian elders living in communities with higher proportions of their own race/ethnicity are more likely to use informal care, they are not more likely to use formal care. These differences may be a result of racial/ethnic preferences for informal home care over formal home care services, regardless of community racial/ethnic composition (Cagney and Agree 1999). Furthermore, this study examines the use of “any” formal home health services, a global measure captured in the MEPS data. Our findings might have been different if we were able to examine racial/ethnic differences in the use of various formal home-based services ranging from highly skilled nursing assistance to personal care. Prior studies have found racial/ethnic variations in different types of formal home care utilization (Dietz, John, and Roy 1998; Laditka, Laditka, and Drake 2006;). Because formal assistance is important for supplementing care provided by informal caregivers, further research is warranted to examine the supply, competency, and stress management of informal caregivers who provide elder care, particularly in communities with higher proportions of Hispanics or NH-Asians. Ensuring adequate formal resources and respite care is of policy relevance to help educate, support, and relieve informal caregivers.
This study has several limitations that should be considered. First, while we offer plausible explanations for our findings, the underlying causal mechanisms were not directly tested. There may be unobserved characteristics related to both community racial/ethnic composition and the use of home care that may partly explain our findings. For example, prior studies have found that psychosocial factors, economic factors, and presence of informal caregivers may contribute to racial/ethnic differences in the use of home care (Burton et al. 1995; Yeo 2001; Bradley et al. 2004;). They may also contribute to the observed differences within racial/ethnic groups living in communities with different racial/ethnic compositions. In our descriptive analysis, we find that more NH-Asian elders are married, have more education, and have higher income than elders of other race/ethnicity. Other unmeasured characteristics, such as acculturation or knowledge of and attitudes toward home care, that differ across community racial/ethnic composition and individual race/ethnicity may in part explain our findings. Further research is necessary to investigate the extent to which our findings are a result of causation or selection.
Another limitation is that we examine only the likelihood of using any home care, not the amount or intensity of the care used. The underlying need for home care and the satisfaction of such care also are not observed and therefore cannot be examined in this study. Finally, the sample limits our ability to assess the heterogeneity of Hispanic and NH-Asian subgroups. These subgroups encompass different cultural values, history, languages, and acculturation.
Despite these limitations, this study provides evidence that the interaction between individual-level race/ethnicity and community-level racial/ethnic composition is associated with the likelihood of home care use among older adults in the United States. Specifically, NH-Asian and Hispanic elders are more likely to use informal home health care if they live in communities with a higher proportion of residents who share their race/ethnicity. As previous research suggests that reliance on informal health care is associated with a reduced risk of nursing home placement, it is important that future research clarify the causal mechanisms underlying our findings. Such research may help programs draw on resources inherent in communities and social networks to foster the accessibility and desirability of informal home care. A better understanding of how informal care is provided within NH-Asian and Hispanic communities also may be important to policy makers concerned with providing formal home care either as a substitute or supplement to informal home care. These efforts, in turn, could improve the likelihood of “aging in place” for more elderly Americans.
Joint Acknowledgment/Disclosure Statement: There were no funding sources for this project. During this study, D. T. Lau was supported by a K01 research award from the National Institute on Aging (K01AG027295).
The views expressed in this paper are those of the authors, and no official endorsement by the Agency for Healthcare Research and Quality or the Department of Health and Human Services is intended or should be inferred.
Funding source for this project: None.
Prior dissemination: This study will be presented in part at the American Sociological Association 105th Annual Meeting in August 2010.
Human subjects protection: Institutional Review Board approval for this study was obtained at Northwestern University.
Conflicts of interest/Disclosures: None.
Additional supporting information may be found in the online version of this article:
Appendix SA1: Author Matrix.
Please note: Wiley-Blackwell is not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.