Developments in the Market for Assisted Living: Residential Care Availability in 2017

Abstract

Objectives:

Describe how the availability of assisted living (AL) and dementia-specific AL vary across counties and correlate with demographic and socioeconomic characteristics.

Design:

Maps, univariate statistics, and standardized mean differences show the differences between counties with high and low levels of AL market penetration, and between counties with and without dementia-specific AL.

Setting and participants:

Data collected from state agencies on licensed AL communities, capacity, and geographic location, and population characteristics from the Area Health Resource file. We include novel and previously undescribed data on dementia-specific AL licenses in 21 states.

Measures:

AL market penetration is reported as the number of AL units or beds per 1,000 persons over age 65 in a county.

Results:

In comparison to counties with the lowest AL penetration, high-penetration counties had higher high school and college education attainment (mean 25.3% vs. 18.5%) and median annual income ($56,000 vs. $46,800), and lower poverty (12.8% vs. 17.3%) and unemployment rates (3.9% vs. 5.1%). Compared to counties with AL but no dementia-specific care, counties with dementia care had substantially higher college attainment (24.6% vs17.7%) and had higher urbanity index (3.8 vs. 5.6 on a 1–9 scale, 1 most urban). Counties with dementia care also had, on average, 16% more in median household income ($54,200 vs. $46,400) and 40% greater home value ($159,800 vs. $113,600).

Conclusions and implications:

Large socioeconomic disparities persist among counties without any AL or low penetration of AL in their borders in comparison to those with high AL penetration, as well as between counties with and without dementia-specific AL communities. There may be a mismatch in need and availability of residential care options for older adults with ADRD that contributes to the disproportionate share of racial/ethnic minorities with ADRD in nursing homes. Lack of available AL beds in the communities where Medicaid individuals reside could make rebalancing efforts doubly difficult, in that Medicaid enrollees may be reluctant to move out of their neighborhoods.

Brief summary:

Availability of assisted living (AL) varies across counties and correlates with socioeconomic factors. Novel data on dementia-specific AL in some states show variation in this important service.

INTRODUCTION

Assisted living (AL), also known as residential care, has become a preferred option for older adults who cannot live independently. AL provides personal care, such as assistance with activities of daily living (ADLs, e.g., dressing, bathing, toileting), instrumental ADLs (e.g., laundry, housework, meal preparation) health maintenance tasks (e.g., medication management), and 24-hour supervision, but does not provide full-time nursing care as in a nursing home. The vast majority of assisted living residents pay for their care privately, either out of pocket or through long-term care insurance. After a period of rapid growth in the 1990s and early 2000s, the AL sector has continued to grow: from 2012 to 2016, the number of residential care providers increased by 30%, with a 17% increase in beds capacity.^1,2 The aging of the US population, the evolution of people’s preferences toward care in more homelike settings, and the expansion of state Medicaid waivers that cover care in AL have fueled demand for AL beds.³ However, the distribution of AL has not been described at the county level in a decade.⁴

AL residences have become an important care provider for older adults with dementia. The majority of residents in AL have some form of cognitive impairment, and more than one-third have behavioral symptoms; but only a minority of AL residents with dementia reside in an AL with a license or certification to provide dementia-specific care.⁵

Unlike skilled-nursing facilities, AL cannot be directly funded by Medicaid or Medicare, leaving states responsible for its oversight and public financing. This article aims to provide a snapshot of how the availability of AL varies across counties in the U.S. and how it correlates with demographic and socioeconomic characteristics of their populations. Because AL supply is driven by demand from private markets, we hypothesized that counties where households have less resources would also be under-resourced with regard to AL availability. With novel, previously undescribed data on the availability of dementia-specific AL communities in some states, we also analyze variation in the availability of this important service for persons with dementia.

DATA AND METHODS

We compiled a list of AL communities licensed in 2017 from lists provided by state agencies. States cover a wide range of licensure categories for congregate residential facilities (eg, residential care facilities, assisted living facilities, community living arrangements, and personal care homes). Facility licenses intended primarily to serve individuals with mental illness or developmental disabilities were excluded from our definition of AL. The list includes facility name, address, and capacity (number of beds or units--a “unit” may contain more than one bed, as the case for couples). Communities were considered duplicates if they had identical name, address, license, and capacity information. Of the 36 states that license, certify, or classify communities providing dementia care, we were able to obtain site-specific designations for individual facilities from a total of 21 states. Appendix Table S2 shows which states were included in the dementia analysis, and which states had dementia care licenses but no data available.

Because the primary objective of this analysis was to correlate the geographic AL availability with sociodemographic characteristics, we obtained county demographic and economic data from the 2017–2018 Area Health Resource File.⁶ We selected variables describing county demograhics (i.e., median age, and percent of population over ages 65 and 85), educational attainment, economic descriptors (including median household income, unemployment rate, percent of persons in poverty), housing markets (i.e., owner-occupied housing and median home value), and race/ethnicity (i.e., percent White, Black, and Hispanic). To measure rurality, we used a scale of urban index from 1 to 9, with 1=most urban. To examine the relationship of AL supply with other health care resources, we also included the number of home health agencies and number of nursing home beds operating in the county.

For county-level analyses, we organize counties by their AL market penetration, which we measure as AL units per 1,000 adults age 65 and over, i.e., per older adult population. We tested for differences between the highest quartile and each of the other groups, using a paired T-test. To interpret the magnitude of the differences between the highest- and lowest- penetration quartile, and between counties with and without dementia-specific care, we report the standardized mean difference (SMD) between these groups. We characterize differences as large or substantial (greater than 0.8 SMD), moderate (0.5 – 0.8 SMD), small (0.2 – 0.5 SMD), and very small (less than 0.2 SMD).(7) For additional detail about data and analyses, see the online appendix.

RESULTS

Nationally, we identified 42,119 unique ALs with 1,267,214 residential units (averaging 41 units per AL) operating in 2017. In twenty-one states for which we obtained data on the dementia designation of AL communities, we identified 6,961 AL communities designated as dementia-specific (16% of ALs). In Figure 1, map A depicts penetration of AL into markets by county in terms of the number of units per older adult population. Overall, 79% of counties had at least one AL within its borders. Map B shows which counties had at least one dementia designated AL for states where facility-level dementia designation data were available.

Figure 1.

Geographic variation in assisted living supply in the United States, by county (2017). Panal (A) shows assisted living beds per 1,000 community-dwelling residents aged sixty-five and over. Panel (B) shows presence of dementia-care assisted living.

Source/Notes: SOURCE Authors’ analysis of licensed assisted living communities collected from state agencies in 2017.

NOTES Quartile cutoffs for county-level penetration of assisted living units per 1,000 elderly who age 65+ are as follows: quartile 1 (lowest) = 0.95–14.93; quartile 2 = 14.94–25.21; quartile 3 = 25.24–39.70; quartile 4 (highest) = 39.73 and above.

Table 1 shows the county-level difference in socioeconomic indicators among different levels of AL supply. AL supply ranged from 1–15 units per thousand older adult population in the lowest quartile of counties, to over 40 units per thousand in the highest. Generally, the measures increased or decreased monotonically going from the lowest to the highest quartile. Differences across almost all characteristics were statistically significant when we compared the highest to the other quartiles, though because the sample is large, even very small differences were statistically significant. College attainment ranged from 16.6% in no-AL counties to over 25% in high-AL counties. Median incomes averaged $44,000 – 47,000 in no- and low-AL counties, to $56,000 in high-AL counties. Percent of the population that is Black was 11.4% in low-AL and 6% in high-AL counties. Figure 2 describes the magnitudes of the differences in terms of the number of standard deviations the differences represent, from small (<0.2 standardized mean difference, or SMD) to large (>0.8 SMD). Figure 2(A) shows large standardized differences between high- and low-penetration counties with regard to poverty, unemployment rates, and educational attainment. The Black and White percentages, owner-occupied housing, median home value, and percent over age 85 were “medium”-sized differences.

Table 1.

County traits, by assisted living penetration quartile in 2017

			AL penetration quartile
County Characteristics Mean, (s.d.)	No AL	1 (low)	2	3	4 (high)
Number of counties	673	618	619	619	619
Median age	41.7 (5.4)	40.4 (4.7)*	40.3 (4.9)	39.88 (4.9)	39.3 (5.1)
Percent of population age 65+	16.8 (4.5)	15.8 (3.6)**	16.0 (4.2)	15.6 (4.1)	15.2 (4.3)
Percent of population age 85+	2.0 (1.0)	1.9 (0.7)**	2.1 (0.8)**	2.17 (0.9)**	2.33 (1)
Less than high school	17.2 (8.0)**	15.9 (6.3)**	14.2 (5.7)**	12.45 (4.9)	10.92 (5.1)
High school diploma or more	82.8 (8.0)**	84.1 (6.3)**	85.76 (5.7)**	87.55 (4.9)	89.08 (5.1)
College education or higher	16.6 (6.7)**	18.5 (7.8)**	20.53 (8.2)**	23.37 (9.1)**	25.33 (10.7)
Median household income ($1,000s)	44.5 (10.9)**	46.8 (11.4)**	48.9 (12.1)**	51.8 (12.3)**	56.0 (14.4)
Unemployment rate	5.0 (2.2)**	5.1 (1.6)**	4.73 (1.4)**	4.37 (1.3)	3.91 (1.3)
Percent persons in poverty	18.4 (7.7)**	17.4 (6)**	16.07 (5.7)**	14.54 (5.1)	12.84 (4.8)
Percent owner-occupied housing	76.0 (8.2)	74.8 (7.8)	74.06 (6.7)**	72.56 (7.6)	72.08 (8)
Median home value ($1,000s)	104.9 (53.2)**	131.0 (81.72)**	134.9 (66.6)**	149.9 (868.3)*	166.0 (956.6)
Percent White population	82.7 (19.6)**	80.6 (17.8)**	82.48 *	83.22 (15.2)	85.41 (14.6)
Percent Black population	8.5 (16.2)**	11.4 (16.4)**	9.86 (14.6)**	8.65 (12.4)	6.02 (11.7)
Percent Hispanic population	10.2 (18.2)**	8.9 (14.9)**	7.36 (10.8)**	7.53 (9.5)	7.33 (9.5)
Rurality index (1=most urban, 9=most rural)	6.67 (2.5)	4.9 (2.5)	4.78 (2.6)	4.23 (2.6)	4.38 (2.6)
Number of home health agencies+	0.22 (0.4)**	0.22 (0.3)**	0.24 (0.3)**	0.25 (0.3)**	0.29 (0.4)
Certified nursing beds +	5.8 (23.3)**	0.8 (4.4)**	1.42 (8.1)**	1.27 (7.3)**	2.01 (9.7)

SOURCE: Authors’ analysis of data collected on licensed assisted living communities in 2017 merged with census estimates by county.

NOTES: Penetration is defined as assisted living units per 1000 people aged sixty-five and older. Standard deviations reported in parentheses. Quartile cutoffs for county-level penetration of assisted living units per 1,000 elderly who age 65+ are as follows: quartile 1 (lowest) = 0.95–14.93; quartile 2 = 14.94–25.21; quartile 3 = 25.24—39.70; quartile 4 (highest) = 39.73 and above. Statistical significance denotes difference from quartile 4.

^**p<0.01

^*p<0.05. +Per 1,000 persons age 65 and over

Figure 2.

Differences in socioeconomic characteristics of county populations associated with assisted living supply. Panel A shows standardized mean difference (SMD) between highest and lowest quartile of assisted living penetration. Panel (B) shows SMD between counties having at least one AL with dementia beds and counties without dementia beds.

SOURCE Authors’ analysis of data collected on licensed assisted living communities in 2017 merged with census estimates by county.

NOTES: Penetration is defined as assisted living units per 1000 people aged sixty-five and older. Quartile cutoffs for county-level penetration: quartile 1 (lowest) = 0.95–14.93; quartile 4 (highest) = 39.73 and above.

Table 2 compares the characteristics of no-AL counties, counties with AL but no dementia beds, and counties with at least one dementia unit. Following a similar pattern to AL supply, counties with dementia beds had higher rates of college attainment (25% vs. 18%), higher income ($54,200 vs. $46,400) and home value ($160,000 vs. $113,000 and a lower percentage of Black residents (7.8% vs 8.7%), than counties with some AL but no dementia beds. In particular, dementia beds were located in distinctly more urban areas, rating 3.8 on the urban-rural index vs. 5.6 and 6.5 averages among no-dementia-care- and no-AL-counties. The range of magnitudes is depicted in Figure 2(B), showing that the largest differences are in college education and rural/urban index, while incomes and home values show medium-sized differences. Racial composition and poverty level differences are small to very small when comparing counties with dementia care to those that have AL but not dementia care (which partly reflects the large variance overall in counties’ racial composition, with standard deviations of 12 to 15 percentage points).

Table 2.

County Traits, by presence of dementia-specialized care among assisted living communities in 21 states with dementia-license data, 2017.

County traits, mean (SD)	No AL	AL but no dementia beds	At least one AL with dementia beds
Number of counties	290	525	572
Median age	40.8 (5.1)	40.6 (4)**	39.2 (4.9)
Percent of population age 65 plus	16.4 (4.3)**	16.5 (3.4)*	15.0 (3.8)
Percent of population age 85 plus	1.92 (0.8)	2.2 (0.8)*	2.1 (0.7)
Less than high school	18.9 (8.5)**	15.0 (5.8)*	11.7 (5.3)
High school diploma or more	81.2 (8.5)**	85.0 (5.8)*	88.3 (5.3)
College education or higher	15.9 (5.7)**	17.7 (5.9)**	24.6 (9.2)
Median household income ($1,000s)	45.2 (11.0)*	46.4 (9.8)**	54.2 (12)
Unemployment rate	5.0 (2.4)**	4.64 (1.5)**	4.2 (1.2)
Percent persons in poverty	18.3 (7.1)**	16.3 (5.7)**	13.78 (5.1)
Percent owner-occupied housing	75.8 (7.7)	74.9 (5.6)**	72.2 (8.4)
Median home value ($1,000s)	96.4 (40.4)**	113.6 (66.2)**	159.8 (74.9)
Percent White population	79.0 (20.8)**	84.1 (16.6)**	83.7 (14.4)
Percent Black population	9.9 (19.0)**	8.7 (15.3)**	7.8 (12.6)
Percent Hispanic population	16.4 (23.5)**	7.6 (12.0)	8.4 (11.9)
Rurality index (1=most urban, 9=most rural)	6.5 (2.4)	5.6 (2.3)	3.8 (2.3)
Number of home health agencies+	0.28 (0.5)**	0.32 (0.4)**	0.27 (0.3)
Certified nursing beds+	4.82 (22.5)**	1.7 (7.8)**	0.81 (5.3)

DISCUSSION

Our main findings are threefold. First, large socioeconomic disparities persist between counties without any AL or low penetration of AL in their borders and those with high AL penetration. For example, the median home value in the highest-penetration counties is nearly 60% higher than in counties with no AL beds, and the rate of college attainment is 50% greater. Second, the contrast between counties with and without certified dementia-specific care follow similar trends, with higher levels of college education and household income and lower levels of poverty and rurality compared to counties without dementia care. Third, we found statistically significant, though small, disparities among county types in their proportion Black and Hispanic populations, a difference that is likely attenuated by the fact that low-penetration counties are more rural, and therefore tend to have fewer minorities.⁸

This study has limitations. First, the definition of AL, and requirements for dementia-specific licenses, vary among states. Although we included only settings whose license types matched the definition are consistent with the definition of AL used by the National Long Term Care Provider Survey (see online appendix), the populations they serve nonetheless are likely to vary among states.

Second, our data regarding dementia-specific care settings are incomplete and do not include all states. Eleven states do not license, certify, or otherwise designate AL settings as providing care specific to the needs of residents with dementia, there are likely AL communities in these states with expertise in caring for this population. For an additional 19 states, we were not able to identify the individual AL communities designated as dementia-specific for 2017 because that information was not included on facility lists provided by state licensing agencies. It is possible that the distribution of AL markets across counties in the 21 states for which we had dementia designation data were not representative of these additional 19 states. Future work should seek a more complete dataset.

States take a varied approach to licensing assisted living care settings. In this paper, we consider the role of dementia care licensing, referring to licensed settings labeled by states as dementia specific and documented as such in state records. Some states clearly define a dementia-specific licensed type of AL or optional certification, with more restrictive regulations covering all relevant components of AL licensure from staff training to door-locking mechanisms. Conversely, some states have a dementia-specific designation for AL communities, but have no additional regulatory requirements for those communities outside of disclosure to the state. Additionally, there are states that do not describe any dementia-specific licensed settings in their regulations, but are regulated in a way that allows AL providers to create secured units or requires higher levels of staffing, often leading providers to offer this setting for dementia care. Finally, the privileges afforded a facility licensed or certified as ADRD-specific also vary, with some states allowing only dementia care facilities to market themselves as “memory care” and some even allowing a higher reimbursement rate for residents relying on the state Medicaid plan for reimbursement.⁹ This wide variation in licensure of dementia care settings, though beyond the scope of this paper, is an important topic for future investigation.

CONCLUSIONS AND IMPLICATIONS

Despite growth in the AL sector, disparities have persisted in the last ten years in access to AL. Use of nursing homes has decreased among Whites and private pay residents and increased among Blacks over time,¹⁰ illustrating a potential racial and ethnic disparity in access to more desirable home- and community-based services. Our work adds to this illustration by documenting a potential disparity in access for more vulnerable and low-income older adults. AL residents with dementia have higher quality of life when workers have specialized training in dementia care in comparison to residents cared for by workers without specialized training.¹¹ Racial and ethnic minorities experience higher prevalence of dementia, and also tend to live in less affluent neighborhoods than their White counterparts.¹² The intersection of race and socioeconomic status may contribute to the disproportionate share of racial/ethnic minorities with ADRD in nursing homes,¹³ and may also potentially magnify the impact of dementia on health outcomes through disparities in dementia-specific AL care, thereby contributing to structural racism non-White older adults experience. Furthermore, older adults with ADRD who live in rural communities—and already face barriers to effective outpatient care¹⁴—also experience disparities in access to AL dementia care.

Unlike the nursing home industry, where the majority (over 60%) of long-stay patients are supported by Medicaid, only 16.5% of AL residents had Medicaid as a payer source in 2015–16.¹ Thus it is hardly surprising that AL locate preferentially in higher-resourced communities. However, public financing of AL is growing: 48 states have expanded Medicaid coverage for AL¹⁵ as part of efforts to rebalance long-term care away from institutional settings toward home and community-based services.¹⁶ Nearly half of all AL communities are authorized to provide care funded by Medicaid waivers.¹ Most people who need to move into residential care prefer to stay near their home communities, children and social networks, and remain healthier when they do.¹⁷ Lack of available AL beds in the communities where Medicaid individuals reside could make rebalancing efforts doubly difficult, as Medicaid enrollees may be reluctant to move out of their neighborhoods. Furthermore, for older adults with cognitive impairment who do relocate, remoteness from their family and friends could reduce quality of life and increased risk of mistreatment.¹⁸ Arguably, access to AL within older adults’ home communities should be considered as part of the criteria for the “homelike” care that HCBS programs set out to provide.¹⁹,²⁰

Free-market approaches to assisted living supply work only insofar as wealthy, private-paying consumers can “vote with their dollars” to encourage the AL industry to offer beds. But other policy levers are necessary for states to address the mismatch between need for and access to assisted living among and vulnerable and historically under-resourced populations. States have discretion to request Medicaid HCBS waivers to target funds to specific geographic areas. HCBS waivers could be designed to target counties or metro regions where demographic data suggest potential need but low AL supply. As Medicaid pays only for nursing and personal care services many states also provide a supplement to social security income for AL residents to help pay room-and-board fees. Increasing the generosity of this supplement could encourage competition in markets with higher shares of Medicaid recipients. Another approach taken by some states to address affordability is to cap the fees that may be charged for room and board to Medicaid-paid AL residents; however, if Medicaid payments are too low to cover operating margins, this type of policy may do little to encourage operators to opening facilities in disadvantaged markets. Certificate of need (CON) laws that apply to AL requiring justification to renovate or build new assisted living facilities exist in 12 states (AR, CT, IA, KY, LA, GA, MA, MO, NC, NJ, NY and VT).²¹ Admittedly CON laws can have unintended consequences when applied to nursing homes,²² but these laws (in states that already have them) or similar incentive-based public funding for building or reopening AL communities could be used to encourage the proliferation of AL communities and dementia-care units in underserved areas.

In sum, our analyses provide empirical evidence suggesting that assisted living supply, particularly dementia care, is unequally distributed between wealthy and under-resourced counties in the U.S. As states seek guidance for their financing, regulation, and oversight of assisted living, examining geographic variation and disparities can assist efforts to understand how AL policies will affect care and access for vulnerable older adults who need residential care.