Abstract
Moving beyond typical dichotomous rural–urban categorizations, this study examines older adults’ likelihood of receiving home- and community-based services. Data from 1608 individuals aged 60+ who requested assistance from Area Agencies on Aging in Virginia in 2014–2015 were analyzed; 88% of individuals received at least one service. Receiving services was associated with geographic-based factors. Individuals living in completely rural areas were significantly less likely to receive any service compared to individuals in mostly rural (OR = 2.46, p = .003) and mostly urban (OR = 1.97, p = .024) areas. There were subtle but significant geographic-based differences in the likelihood of receiving specific services including food/meal, fresh food, information and referral, in-home care, utilities support, and transportation. Findings provide nuanced insights about geographic-based disparities in the receipt of services and suggest the need for new and modified service delivery strategies that maximize older adults’ ability to live.
Keywords
Introduction
As individuals get older, they often engage in help-seeking behaviors to meet ongoing health and functional (e.g., personal care) challenges. For older individuals with low care needs, and for those who are financially vulnerable, informal caregivers provide the bulk of help. If informal caregivers are unavailable, or care provided is insufficient, older adults may supplement their care arrangements with formal home- and community-based services (HCBS; Davey et al., 2005; Savla et al., 2019). These service options are designed to enhance the feasibility of older adults remaining in their homes and communities and go beyond typical health care services to address critical social determinants of health, including food insecurity, housing, and transportation (Artiga & Hinton, 2018). When individuals have sufficient functional and financial need, they may be eligible to receive Medicaid-funded HCBS. Although states have expanded their offerings to improve access to long-term services and supports (LTSS; Musumeci et al., 2020), there are considerable differences among states in offering HCBS, including the scope of benefits and the target population. For people who are not eligible for Medicaid, paying for LTSS like HCBS can be expensive (Hado & Komisar, 2019). Thus, individuals may turn to alternative, affordable HCBS provided through Area Agencies on Aging (AAAs).
The reauthorization of the Older American’s Act (OAA) in 1973 created AAAs to coordinate social services for older adults in local communities across the United States. Though the goal of the OAA is to prioritize individuals with the greatest economic or social need, there is no means testing for receiving community services and supports. That is, anyone aged 60+ can access and receive OAA services, pending available funding. Thus, eligibility criteria for who can receive OAA services is less restrictive than for other assistance programs (e.g., Medicaid HCBS waiver programs). Services may be limited in certain areas because of challenges to connect rural older adults with needed services (Government Accountability Office, 2019). A 2016 AAA national survey revealed that 41% of AAAs served older adults in predominantly rural areas (National Association of Area Agencies on Aging, 2017). Nonetheless, long-standing geographic disparities persist. The purpose of this study was to examine the likelihood of older adults receiving services requested through AAAs. Specifically, we examine geographic location and individual-level variables associated with receiving services.
Social Determinants of Health
Various social factors contribute to overall health. For example, living alone is often associated with negative outcomes, including increased risk of frailty (Kojima et al., 2020), having unmet needs (Dunatchik et al., 2016), and experiencing isolation and loneliness (Cacioppo et al., 2015). Wealth status helped explain the association between living arrangement and disability; affluent older adults living alone reported less difficulty with daily activities that support independence, compared to less affluent older adults (Henning-Smith et al., 2018). However, individuals who live alone intentionally develop and deploy self-care strategies to maintain health and independence (Hayes, 2006). A recent study by Djundeva and colleagues found that the majority of older adults living alone did not have increased risk of vulnerability compared to older adults who live with others; about one-third of older adults had few sources of support and limited interaction, leaving them vulnerable for reporting worse subjective well-being (Djundeva et al., 2019).
Financial vulnerability negatively influences housing options, health insurance coverage, transportation options (Ludke & Obermiller, 2012), and the likelihood of experiencing adverse health situations (Weaver & Burley, 2020). Having limited financial ability to pay for services is known to increase individuals’ risk of having unmet needs (e.g., Leach & Schoenberg, 2008) and experiencing hospitalization (e.g., Ronksley et al., 2013) compared to their counterparts with greater financial resources. Delaying help may be a byproduct of a general inability among older adults to afford formal services (Casado et al., 2011), an issue common among rural-dwelling older adults (Li, 2006).
Historically, rural communities experience economic disadvantages (Lobao et al., 2016), which permeate all aspects of rural life. Rural populations are often poorer, less healthy, and have more restricted access to services than urban populations (Hash et al., 2015). They have a greater risk of chronic disease and disability compared to their urban counterparts (Coburn et al., 2016). These factors are particularly concerning because a disproportionate share of the aging population growth is occurring in rural areas (Lang, 2013).
There are long-standing assumptions that rural residents’ distrust of formal services prohibits their use (e.g., Bull, 1998; Zanjani & Rowles, 2012). However, older adults in rural Appalachia indicated an openness to receiving formal assistance (Weaver et al., 2018). They worried about burdening family and friends with care responsibilities, which contributed to anticipated reliance on formal care in the future. A further analysis of the data (Savla et al., 2019) found that having LTSS available increased the likelihood of utilization, regardless of beliefs about family expectation or attitudes about community services. These findings offer support for increasing access to health services that target rural, underserved communities to deter or prevent adverse health outcomes.
Availability, awareness, and affordability are persistent barriers to receiving HCBS, which contribute to subsequent adverse health experience (Li, 2006; Siconolfi et al., 2019). In addition, there are persistent systemic barriers that inhibit access to services in rural areas, including limited providers, inadequate transportation services, and workforce recruitment and retention challenges (Siconolfi et al., 2019). A shortage of geographically proximate care options contributes to low (Jonk et al., 2015; VanderWielen, et al., 2015) or more costly service use. Rural-dwelling individuals are not benefitting from increased state spending on Medicaid-funded HCBS (Coburn et al., 2016). They are less likely to use HCBS but more likely to use nursing facility services than their urban-dwelling counterparts (Coburn et al., 2016). Strict eligibility for state-funded Medicaid HCBS waiver assistance often impeded low-income, rural-dwelling individuals from receiving assistance because their needs were not at a nursing home level of care (Weaver & Roberto, 2018). Collectively, these findings demonstrate ongoing challenges facing rural older adults seeking HCBS and supports and reinforce long-standing health inequities.
Definitions of Rurality
Definitions of what constitutes “rural” vary across federal agencies. For example, the Office of Management and Budget designates counties as metropolitan (urban), micropolitan (rural), or neither (rural; Myers et al., 2016). The U.S. Census Bureau definition categorizes counties as completely rural, mostly rural, and mostly urban. The Federal Office of Rural Health Policy pulls from both of these definitions and applies a rural–urban commuting area code to each county in an effort to determine extent of rurality (Health Resources & Services Administration, 2020). Each measurement approach yields a slightly different population-based estimate, revealing challenges with each definition. Researchers typically use a binary urban–rural designation to analyze data (e.g., Coburn et al., 2019), but we argue that using the U.S. Census Bureau designation maintains parsimony yet provides nuance with three categorizations.
The 2010 Census (Ratcliffe et al., 2016) found that the majority of the population (86.3%) lives in “mostly urban” counties that consist of urban clusters (2500–49,999 people) and urbanized areas (≥50,000). Far fewer individuals (11.9%) live in “mostly rural” counties, defined as areas with more than half the population considered rural. Less than 2% of the total U.S. population live in “completely rural” counties, which have no areas identified as urban. Maintaining this nuance of rurality can provide new insights about how geographic location (i.e., completely rural vs mostly rural) is associated with the likelihood of receiving services in late life.
Theoretical Framework and Study Hypotheses
Andersen’s (1995) healthcare utilization model provides a framework for identifying factors that influence help-seeking behaviors related to service use (e.g., Ferris et al., 2016; Weaver & Roberto, 2015). The model proposes the influence of predisposing characteristics (e.g., age, sex, race), enabling resources (e.g., geographic location, living arrangement, financial need), and need-based factors (e.g., personal and instrumental activities of daily living [ADL/IADL] limitations). The influence of these factors operates within a larger societal environment and policy arena that determines access to programs designed to assistance at-risk individuals with accessing health services and supports.
To examine potential rural disparities in HCBS available through AAAs, we used Andersen’s (1995) healthcare utilization model and leverage state agency data to answer the following research question: When controlling for cumulative functional need, what predisposing and enabling factors are associated with service use disparities? Expanding from the dichotomous rural–urban categorization, we hypothesize that individuals living in completely and mostly rural areas are less likely to receive any, as well as specific services; and individuals living alone and/or with financial need are more likely to receive any, as well as specific services.
Methods
Data Source
We used data collected by the Virginia Department of Aging and Rehabilitative Services (DARS, n.d.). DARS works in collaboration with the 25 local AAAs across the Commonwealth and community partners to provide resources and services designed to improve the quality of life and promote independence of older Virginians living in the community, assisted living facilities, or nursing homes (DARS, n.d.). DARS serves as the central point of contact for older, vulnerable adults and their families seeking information about appropriate services to address their healthcare needs.
Study Sample
DARS provided enrollment and service use data on all individuals who contacted their agency during a 2-year period from January 1, 2014 to December 31, 2015. Available data included sociodemographic characteristics, functional ability (i.e., ADL and IADL), services requested, and services received. Individuals who did not request a specific service from the agency were removed from the analytic sample (N = 6443), as were individuals who had missing data on basic demographic variables including age, sex, race/ethnicity, and county of residents. The current analysis was restricted to adults aged 60 and older who requested assistance through AAAs, yielding a study sample of 1608 older adults. The university’s institutional review board approved this study.
Study Measures
Outcome Variables
Our primary outcome of interest was likelihood of receiving services, which we coded to indicate whether individuals did not receive (0) or did receive (1) any type of service through the DARS agency during the study period. Very few individuals in the study sample received more than one service (n = 40); thus, we operationalized service use as a binary variable. We were also interested in the use of specific types of services received (or not received) by individuals including fresh food assistance, transportation assistance, utilities assistance, food/meal assistance, information and referral assistance, health management/social programs, and in-home assistance. See Table 1 for more information.
Characteristics and Service Use Prevalence Among Older Adults (N = 1608).
Note. ADLs = activities of daily living; IADLs = instrumental activities of daily living. 1“Other” includes Hispanic, Hawaiian, Pacific Islander, Asian, and Native/Alaskan. 2Missing data from 78 respondents (4.9%). 3Income at or below Federal Poverty Level; missing data from 215 respondents (13.4%).
Predictor Variables
We identified predisposing, enabling, and need-based variables from the available data. Predisposing variables included age (continuous), sex (0 = male, 1 = female), and race/ethnicity (three dichotomous dummy variables for White, Black, and Other [Other included Hispanic, Hawaiian, Pacific Islander, Asian, and Native/Alaskan]). There were three enabling variables included in the analyses. Using the U.S. Census Bureau definition of rural (Ratcliffe et al., 2016), individuals were categorized as living in completely rural (0), mostly rural (1), and mostly urban (2) counties. Individual living arrangement served as a proxy for social support, which we coded to indicate whether individual did not live alone (0) or did live alone (1). DARS determines financial need using the Federal Poverty Threshold level; individuals’ financial need was operationalized as income below the poverty threshold (0) or above the poverty threshold (1).
DARS relies on professional evaluations of applicants’ ability to conduct ADLs and IADLs with and without mechanical or human assistance to determine the need for services. ADLs included bathing, dressing, toileting, transferring, eating/feeding, bowel control, bladder control, walking, wheeling, stairclimbing, and mobility. IADLs included meal preparation, housekeeping, laundry, money management, transportation, shopping, using the telephone, and home maintenance. We controlled for functional need by including a summed score to depict a need index of ADL (range 0–11) and IADL (range 0–8) limitations.
Data Analysis
To characterize the study sample, we ran descriptive statistics (frequencies, percentages, and means) and examined bivariate correlations between all pairs of predictors to check for multicollinearity. We conducted a series of chi-square and one-way ANOVA analyses to determine geographic-based differences and logistic regression analyses to estimate the independent effects of predisposing characteristics, enabling resources, and need-based variables on the likelihood of receiving services. Due to missingness, logistic regression analyses included 1322 individuals. In order to look at between-rural categorization, “completely rural” was the primary reference group. We determined model fit using the Hosmer and Lemeshow goodness of fit test and modified the model as needed (i.e., living alone was excluded as a predictor for meals assistance and utilities assistance). We used IBM SPSS Version 22.
Results
Demographic Characteristics
The mean age of individuals was 74.5 years (SD = 9.299); the majority were women (n = 1363; 84.8%) and White, non-Hispanic (n = 1219; 75.8%). More than half of the older adults lived alone (n = 881; 54.8%), which is double the national average of older adults living alone (26%) (Ausubel, 2020). Four in ten of the older adults lived at or below the federal poverty level (N = 701; 43.6%), which is considerably higher than projections for adults 60+ living in the community living at or below 250% of the federal poverty line (38.1%) (Bauer, 2016). They had over four ADL and IADL limitations (Table 1). Just over one-third of the older adults resided in nonrural counties (i.e., mostly urban; n = 581; 36.1%), while the majority resided in mostly rural counties (n = 920; 57.1%). The remaining individuals resided in completely rural counties (n = 107; 6.7%). Most older adults received at least one service (n = 1421; 88.4%). Sex, financial need, living arrangement, and number of functional I/ADL limitations did not significantly differ based on location. However, individuals living in completely rural counties were significantly younger than individuals living in urban counties (F = 5.01, p = .007) and more likely to be Black in completely rural counties; White, non-Hispanic in mostly rural counties; and Black or another race in urban counties (X 2 = 336.445, p < .001).
Receiving Services
We controlled for functional need-based factors (I/ADL need indices) to examine the association of predisposing characteristics (age, sex, and race) and enabling resources (living arrangement, financial need, geographic location) with the likelihood of receiving at least one service, as well as each specific services (Table 2 for detailed information). Increasing age was associated with lower odds of receiving any service through AAAs. Compared to individuals living in completely rural areas, individuals living in mostly rural areas were 2.5 times more likely to receive any service, and individuals living in mostly urban areas were nearly twice as likely to receive any service.
Factors Associated With Receiving Any and Specific Types of Service (N = 1322).
Note. We controlled for functional limitations as measured by I/ADL limitation indices. Model fit was acceptable for all analyzes after minor modifications for food/meal and utilities assistance. For in-home care, we included “White, non-Hispanic” as the sole predictor of race because no “other race” individuals received the service.
As for specific services, increasing age was associated with lower odds of receiving fresh food and utilities but higher odds of receiving food/meals, information and referral assistance, and in-home care. Women had higher odds of receiving food/meal assistance but lower odds of fresh food assistance; women were twice as likely as men to receive health management/social programs and thrice as likely to receive information and referral assistance. Compared to all other race/ethnicities, White, non-Hispanic individuals were more likely to receive fresh foods, while Black individuals were less likely to receive fresh foods. White, non-Hispanic, and Black individuals had lower odds of receiving health management/social programs and information and referral, compared to all other race/ethnicities. Living alone decreased the odds of receiving health management/social programs and information and referral assistance.
The odds of receiving each specific service was associated with geographic location, except for receiving health management/social programs. Individuals living in completely rural areas had greater odds of receiving food/meal assistance, compared to individuals living in mostly rural and mostly urban areas. Conversely, individuals living in completely rural areas had lower odds of receiving fresh food assistance than their counterparts living in mostly rural or mostly urban areas.
Individuals in completely rural areas were more likely to receive information and referral assistance and in-home care than were individuals in mostly rural areas. As for receiving utilities assistance, compared to individuals in completely rural areas, individuals in mostly urban areas were five times as likely to receive assistance. Individuals in completely rural areas were more likely to receive transportation assistance than were individuals in mostly rural areas, but less likely than individuals in mostly urban areas.
We ran analyses to test whether interactions with geographic location, financial status, and living arrangement were significantly associated with receiving any, as well as specific, services. There were no significant interactions associated with receiving any services. As for specific services, women living in mostly urban areas were more likely to receive fresh food assistance compared to woman in mostly rural areas. Living alone in mostly rural and mostly urban areas increased the likelihood of receiving fresh food assistance compared to living alone in completely rural areas. The odds of receiving food/meals were higher with each increase in ADL limitation for individuals living below the poverty line compared to individuals living above the poverty line. With each increase in IADL limitation, individuals living below the poverty line had lower odds of receiving fresh food assistance, compared to individuals living above the poverty line. Conversely, with each increase in IADL limitation, individuals living in mostly urban areas were more likely than were individuals living in completely rural areas to receive information and referral assistance. No other interactions were significant.
Discussion
We used the designation of completely rural, mostly rural, and mostly urban put forth by the U.S. Census Bureau to examine geographic disparities in receiving services beyond the typical binary rural–urban measure. Taking this approach allowed us to assess nuances across geographic location about likelihood of receiving any type of service, as well as specific types of service. For example, among this sample of Virginians requesting services through AAAs, when controlling for functional limitations, individuals living in mostly rural and mostly urban areas were twice as likely to receive any service as individuals living in completely rural areas. This suggests a potential barrier to service access, as individuals’ functional need did not differ significantly. While some progress has been made to reduce geographic disparities in service access and use (Coburn et al., 2016; McAuley et al., 2009), study findings suggest the need for ongoing efforts to reduce geographic disparities in service use for individuals living in completely rural areas.
In accordance with the behavioral model of health service use (Andersen, 1995), older adults acknowledged a need for help and sought assistance to address their unmet functional care needs with the expectation of receiving help. Although most individuals received at least one service, study findings revealed complexities regarding disparate likelihood of receiving services. We anticipated that living alone and having financial need would significantly increase the likelihood of receiving services, but this was rarely the case. Individuals living alone were less likely to use health management programs/social programs or information and referral assistance, and individuals living alone in completely rural areas were least likely to receive fresh food assistance. One interpretation of this finding is that a potential caregiver (i.e., individuals living with others) may have intervened and assisted with connecting older adults with beneficial services and supports. Among individuals with financial need, an increase in ADL limitations also increased the odds of receiving food/meals compared to individuals without financial need; however, as IADL limitations increased among individuals with financial need, the odds of receiving fresh food assistance decreased. Because services provided through AAAs are not subject to means testing, our findings reveal a potential shortfall in providing services for individuals with the greatest economic or social need (Jeszeck, 2012). Individuals with the greatest economic or social need include individuals who are living alone and/or with financial vulnerability, which may compromise their ability to receive help with day-to-day tasks that promote functional health and independence.
The type of services available through AAAs that we examined included fresh food assistance, transportation assistance, utilities assistance, food/meal assistance, information and referral assistance, health management/social programs, and in-home assistance. We found that the likelihood of receiving information and referral assistance was higher for individuals living in completely rural areas. This type of assistance is typically a hotline to help individuals identify services and programs; thus, services have not come to fruition. However, interaction effects revealed that each increase in IADL limitation among individuals living in mostly urban areas was associated with greater likelihood of receiving information and referral assistance, compared to individuals living in mostly rural areas. These findings support Chen et al.’s (2018) findings that the degree of disability with I/ADLs was associated with having unmet needs, but likelihood of receiving assistance depended on geographic location. An analysis of the National Health and Aging Trends Study by Henning-Smith et al. (2019) showed similar levels of unmet need among urban and rural older adults. This suggests a system-level barrier that disadvantages rural-dwelling individuals who have unmet needs akin to urban-dwelling individuals. More research is needed to understand the mechanisms contributing to system-level barriers in receiving appropriate services. In turn, implementation and evaluation of strategies that address underlying mechanisms and/or target services to proactively address functional limitations and deter/delay avoidable adverse health outcomes is warranted.
The unexpected finding that in-home care assistance was more prevalent among individuals in completely rural areas than mostly rural areas may indicate progress toward addressing long-standing disparities in access to care and services in more remote geographic areas (Coburn et al., 2016; Nelson & Stover Gingerich, 2010). Policy efforts continue to rebalance LTSS in favor of HCBS over institutional care. However, availability of HCBS does not meet the demand, which is especially problematic given the growing numbers of nursing home closures in rural communities (Tyler & Fennell, 2017) and the lower likelihood of HCBS use among rural versus urban individuals (Coburn et al., 2019). Geographic disparities persist and require greater federal and state efforts to build HCBS capacity (Coburn et al., 2019). Location matters—moving away from the dichotomy of rural versus urban allows for attention to the spatial nuances, complexities, and diversity of location; this supports ongoing conversation about moving away from a “have” versus “have not” dichotomy.
Other HCBS options support efforts to address social determinants of health. Securing adequate and affordable housing is a social determinant of health and well-being that has becoming increasing difficult, especially among older, low-income populations (Stone, 2018). We found that individuals living in completely rural areas were less likely to receive utilities assistance than were individuals living in urban areas. This may reflect higher cost of living in urban areas than in rural areas, including cost of housing (U.S. Bureau of Labor Statis, 2016, October 28).
The associations between geographic location and receiving food-related services revealed a divergent effect. While receiving food/meal assistance was more prevalent among individuals living in completely rural areas, receiving fresh food assistance was less likely, especially so for individuals living alone. Lack of high-quality food (i.e., fresh foods) increases health risks, especially problematic in geographic locations without access to healthy food options (Lee et al., 2010). If individuals are nutritionally vulnerable, they may have increased risk for adverse health outcomes (Buys et al., 2014), including nursing home placement (Thomas & Mor, 2013). Our findings suggest that individuals in completely rural areas are receiving food/meals, but without receiving more nutritious fresh food assistance, the services received may unintentionally reinforce the long-standing health disparities among rural communities. Improving nutritional health needs to be prioritized among completely rural populations. It is important to note that individuals living in mostly rural areas were more likely to receive fresh food assistance than individuals living in mostly urban areas, thus demonstrating why it is critical to assess beyond the binary categorization or rural or nonrural.
Individuals in mostly urban areas were more likely to receive transportation assistance than were individuals in completely rural areas. While this finding supports previous findings about the need for greater transportation assistance in rural areas (Henning-Smith et al., 2017), our use of gradient geographic categories uncovered insightful nuances. Specifically, we found that individuals in completely rural areas were more likely to receive transportation assistance than were individuals in mostly rural areas, which was unexpected. Individuals living in mostly rural areas were the least likely to receive transportation assistance. These new findings highlight an opportunity for improvement in targeting transportation services for individuals in this geographic area.
Limitations and Future Research
The findings of this study shed new light on potential geographic-based disparities in the use of home and community-based LTSS. Several limitations need to be acknowledged and addressed in future research. The prevalence of older adults living alone and reporting financial need among our sample was higher than the national average, suggesting high levels of need among individuals seeking and receiving services through AAA; yet, it may limit generalizability of findings. Model fit was acceptable for all analyses after minor modifications for food/meal and utilities assistance. However, mostly rural individuals receiving fresh food assistance yielded a large odds ratio and wide confidence intervals, despite acceptable model fit; this finding should be interpreted with caution. Using secondary data, we were limited to specific analyses based on available data, which limited the questions we could address and interpretation of findings. For example, because we did not have access to service availability data, we were unable to ascertain whether individuals requested but did not receive services due to regional availability of services, access issues, or declination of services when offered. Rather, we looked at likelihood of receiving services and found geographic disparities, even after controlling for functional need, which should be an indicator of service receipt. Research is needed to understand service availability versus accessibility issues.
Having additional information from agencies about why individuals received a specific service would enable researchers to answer additional questions about allocation of scare resources; this would isolate underlying factors that contribute to potential location-based disparities in receiving services. As these type of data are unlikely to be found in existing databases, qualitative research methods would provide a more appropriate and feasible approach for exploring how agencies make decisions about the services they allocate to help meet the needs of older adults (Phoenix, 2018).
We could not account for individual-level characteristics (e.g., self-reliance, hesitance to receive help from outsiders, cognitive status) that sometimes are used to explain lower service use among rural-dwelling older adults compared to urban-dwelling individuals (Behringer & Friedell, 2006; Hong et al., 2011). These factors likely contribute to service use disparities but are cost and time prohibitive and thus not typically asked during client intake processes. We encourage researchers to develop cross-sector partnerships with state and local service agencies to develop systematic and consistent data collection processes to enhance data-driven decision-making for the benefit of the populations they serve as well as advance the scientific study of service utilization among older adults and support modification and development of new service delivery strategies that target high-risk populations. Collaborative partnerships like this are needed to facilitate opportunities that capture greater context about individuals’ care circumstances and identify modifiable factors amenable to intervention that may deter further adverse health outcomes associated with having unmet needs.
Footnotes
Acknowledgments
The authors would like to acknowledge the Department of Aging and Rehabilitative Services for their support of this project. The content is solely the responsibility of the authors and does not necessarily represent the official views of the Department.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
Author Biographies