Abstract
Objectives
To examine how different care arrangements across the long-term care continuum are associated with experiencing unmet care need consequences (UCNCs), such as skipping meals, going without clean clothes, or taking the wrong medication.
Methods
We include older adults receiving assistance with at least one self-care, mobility, or household activity (for health/functioning reasons) in the 2015 National Health and Aging Trends Study (N = 2,388). We examine the likelihood of experiencing a UCNC across the long-term care continuum: those receiving unpaid community care only, paid community care, and residential care. Cross-sectional logistic and longitudinal multinomial logistic regression models examine if type of care arrangement in 2015 is associated with UCNCs in 2015 and change in UCNCs by 2017.
Results
In adjusted cross-sectional models, paid community care recipients had roughly 2 times greater odds of experiencing a UCNC in 2015 compared to those living in residential care or receiving only unpaid care. In adjusted longitudinal models, the risk of experiencing persistent UCNCs (compared to having needs met in both years) was 4.81 times higher for those receiving paid community care compared to those in residential care and 2.17 times that of those receiving unpaid care only.
Discussion
Older adults receiving paid care face significant and consequential gaps in care, particularly in comparison to those in other care arrangements. More attention is needed to determine how paid care arrangements can be improved and/or expanded to meet the needs of the growing number of older adults receiving paid care in the community.
Keywords: Caregiving, Formal care, Long-term services and supports, Residential care
About 70% of older adults will eventually need ongoing help with activities of daily living (e.g., preparing meals, dressing, or managing medications; U.S. Department of Health and Human Services, 2019). Most older adults receive this help from family and unpaid caregivers and roughly one third also receive care from paid long-term care (LTC) providers either at home in the community or in a residential setting, such as assisted living or a nursing home (Freedman & Spillman, 2014a). Where older adults receive care and who they receive care from are shaped by several factors, including health care policy. Medicaid, the federal-state health insurance program for people with low income, is the largest payer of paid LTC services (Harris-Kojetin et al., 2019). Over time, Medicaid funding has shifted from funding only care in institutions to funding home and community-based services (HCBS) in an effort to “rebalance” LTC (Centers for Medicare & Medicaid Services, 2020). However, access to Medicaid-funded HCBS is limited, and paying for ongoing care is unaffordable for most American families. Nationwide, the waitlists for HCBS continue to grow, with 20 states having waitlists for HCBS and older adults spending over 2 years, on average, waiting for services (Musumeci et al., 2019).
The shift toward community care and limited affordable options for paid LTC have contributed to reliance on family and unpaid caregivers to meet the care needs of the growing aging population. Freedman and Wolff (2020) examined five national estimates of caregiving between 2015 and 2017 and concluded that roughly 40 million family and unpaid caregivers provide assistance with daily activities to an older loved one in a given year. In some cases, family and unpaid care alone may not be adequate to meet all the changing care needs of older adults and without adequate assistance older adults may experience unmet care needs related to activities of daily living. For some older adults, experiencing unmet care needs leads to adverse consequences, such as skipping meals, going without clean clothes, or taking the wrong medication.
Preventing unmet care need consequences (UCNCs) has important public health implications. Evidence suggests that experiencing UCNCs is associated with anxiety, depression, increased risk of emergency room visits, hospitalization, nursing home admission, and premature death (Gaugler et al., 2005; Hass et al., 2017; Xiang et al., 2018; Xu et al., 2012; Zuverink & Xiang, 2020). Older adults with Medicaid, higher care needs, and receiving more hours of care are at increased risk of experiencing UCNCs (Allen et al., 2014; Beach et al., 2020).
Given the evolving LTC landscape, it is important to understand differences in the likelihood of experiencing a UCNC across a variety of care arrangements to inform LTC policy and program improvement. Recent research found that hospitalization rates are higher for Medicaid HCBS recipients compared to their nursing home counterparts, suggesting potential gaps in Medicaid-funded care provided in the community (Konetzka et al., 2020). Very little is known about the health outcomes and unmet care needs of older adults living in nonnursing home residential settings, partially due to the loosely regulated industry and varying definitions of residential care by state (Carder, 2017). While we might expect those with only family and unpaid support in the community to have more unmet care needs than those receiving paid community care, a systematic review of the literature from 2000 to 2020 found some evidence that receiving paid care in the community was associated with worse health and well-being among older adults compared to receiving only unpaid care (Coe et al., 2021). Similarly, evidence suggests that the prevalence of UCNCs is especially high among older adults receiving paid care, compared to those receiving only unpaid help (Freedman & Spillman, 2014a). Despite this previous body of research, we do not know how the likelihood of experiencing UCNCs compares across different care arrangements along the LTC continuum and what specific gaps in care (e.g., medication management, bathing, meal preparation) lead to UCNCs in each type of care arrangement.
The goals of this study are (a) to describe specific UCNCs due to potential gaps in assistance with 12 daily activities among older adults and to compare these outcomes across the LTC continuum (those receiving only family and unpaid care, paid community care, or living in residential care) and (b) to determine if older adults’ care arrangement across the LTC continuum in 2015 is associated with the likelihood of experiencing UCNCs 2 years later (2017).
Conceptual Framework
We draw from the Person–Environment Fit Perspective (Lawton & Nahemow, 1973) and conceptualize experiencing a UCNC as a mismatch between the care needs of the older adult and the care received in their current environment. As shown in Figure 1, in response to gaps in care, older adults may also change their care arrangement to better meet their needs, which in turn could influence if they continue experiencing UCNCs. We also draw from the Andersen Healthcare Utilization Model (Andersen & Newman, 1973) to inform our choice of control variables that influence selection into a specific care arrangement. In our adjusted models, we control for a variety of predisposing characteristics (sociocultural characteristics of an individual), enabling factors (resources that contribute to health/functioning and options for LTC), and need factors (health/functioning that necessitates receipt of LTC). We adjust for several factors that influence case-mix severity across the different care arrangements, because those receiving paid care or living in residential care generally have higher care needs than those receiving unpaid care in the community. However, unmeasured factors, such as valuing independence/autonomy over additional support/safety, are likely more common among older adults living in the community. The impact of this selection bias on our results is considered in the Discussion section.
Figure 1.
Conceptual framework informed by the Person-Environment Fit Perspective (Lawton & Nahemow, 1973) and the Andersen Healthcare Utilization Model (Andersen & Newman, 1973).
In the current study, we are most interested in the bold arrows shown in Figure 1—examining how UCNCs in 2015 and change in UCNCs 2 years later (2017) vary by care arrangement in 2015. Drawing from our conceptual framework, we hypothesize that after adjusting for differences in predisposing, enabling, and need factors, older adults in care arrangements with higher levels of environmental support will be less likely to experience UCNCs compared to older adults in care arrangements with lower levels of environmental support. As such, holding all else constant, we expect older adults living in residential care to be the least likely to experience a UCNC (Hypothesis 1). Similarly, given the known lack of availability or affordability of adequate paid care in the community, we expect older adults receiving only unpaid care in the community to be the most likely to experience a UCNC, after adjusting for critical predisposing, enabling, and need variables (Hypothesis 2). For the longitudinal analysis, we expect the same pattern to emerge over time. Holding all else constant, older adults living in residential care will be the least likely to develop or have persistent UCNCs between 2015 and 2017 (Hypothesis 3), and older adults receiving only unpaid care in the community will be the most likely to develop and have persistent UCNCs over time (Hypothesis 4).
Method
Data
Data are from the 2015 and 2017 waves of the National Health and Aging Trends Study (NHATS). NHATS began collecting data on a nationally representative sample of Medicare beneficiaries 65 and older in 2011. Proxy respondents were interviewed when a participant was unable to respond. Unlike other nationally representative data sets that usually exclude people living in institutional settings, NHATS includes older adults across the LTC continuum and allows for differentiation between those receiving unpaid from paid community care and those living in nursing homes from other residential settings (Kasper & Freedman, 2018). NHATS data are collected annually, and the sample was replenished in 2015 to adjust for attrition and mortality. More information on the NHATS sampling design and methods can be found in detail elsewhere (DeMatteis et al., 2016b).
We first limit our sample to only those receiving help in 2015, for health or functioning reasons, with at least one self-care, mobility, or household activity (n = 2,495). If a participant had missing data on measures of interest, then data from the subsequent round were used when possible (n = 161). If Medicaid status was still not known, then we coded the participant as having Medicaid if their income was below $15,000 in 2015 (n = 69). Remaining cases with missing data were deleted (n = 107). Therefore, our final analytic sample for cross-sectional analyses consisted of 2,388 older adults receiving help across the LTC continuum in 2015. Participants were included in the longitudinal analysis (N = 1,524) unless they died between baseline and follow-up (n = 464), were lost to attrition by 2017 (n = 355), or had missing data on the 2017 UCNC outcome measure (n = 45). Chi-square and t-tests comparing participants in the cross-sectional sample to those in the longitudinal sample show that participants remaining in the longitudinal sample were more likely to be younger, higher income, in better health, and less likely to be a proxy respondent. The longitudinal sample also overrepresents older adults receiving only unpaid care in the community (compared to paid community care or residential care). Although the attrition rate was similar across unpaid care, paid care, and residential care, participants receiving only unpaid care in the community were less likely to die between 2015 and 2017. We further discuss the association between care arrangement and mortality in the Results section (Supplementary Table 1).
Measures
Unmet care need consequence
Two measures of UCNC were used: UCNC in 2015 (hereafter referred to as UCNC) and change in UCNC between 2015 and 2017 (hereafter referred to as UCNC change). UCNC was a dichotomous measure based on a self-report of experiencing at least one adverse consequence related to 12 daily activities in the last month that were difficult or required help for the participant to complete. Consequences related to self-care and mobility-related activities included went without eating, wet or soiled clothes, went without getting dressed, went without bathing, stayed in bed, unable to get around inside the home, and unable to get outside the home. These self-care and mobility-related activities correspond with common measures of activities of daily living. Consequences related to household activities included went without clean laundry, went without groceries or personal items, went without a hot meal, made a mistake in managing medications, or went without handling bills/banking matters. These household activities included in the NHATS correspond with common measures of instrumental activities of daily living and were only assessed if a participant received help for health or functioning reasons. The UCNC change measure consisted of four categories: needs met continues (no UCNCs in 2015 or 2017), resolved UCNCs (UCNCs in 2015, needs met in 2017), developed UCNCs (needs met in 2015, UCNCs in 2017), and persistent UCNCs (UCNCs in 2015 and 2017).
Predisposing characteristics and enabling factors
Self-reported age categorized ordinally from 1 to 6 (1 = 65–69 years old; 6 = 90+ years old), gender (female = 1; male = 0), and race/ethnicity (non-Hispanic White, non-Hispanic Black, or Hispanic; another race/ethnicity) were included as predisposing characteristics. Race/ethnicity was included as a proxy measure of differential exposure to interpersonal and systemic racism that can influence use of LTC options (Jenkins Morales et al., 2021). Living with others (vs. lives alone), Medicaid receipt, number of children (count 1–6+), and imputed annual income (logged) were also included as enabling factors that could influence type of care arrangement and UCNCs. The NHATS publicly available imputed income measure has been described in detail elsewhere (DeMatteis et al., 2016a) and is based on a self-reported estimate of individual monthly income or income for coresiding couples from various sources, including earned income, social security, pensions, retirement account withdrawals, and income from interest/dividends (Kasper & Freedman, 2018).
Need factors
We adjust for a variety of health and functioning measures that indicate a need for care and contribute to care arrangement decisions. Dementia status (probable dementia = 1) was determined by meeting one of the following criteria: (a) the participant or a proxy respondent reported a diagnosis, (b) a score ≥2 on the AD8 Dementia Screening Interview indicating probable dementia, or (c) a score at least 1.5 SDs below the mean on at least two domains of cognitive tests (executive functioning, orientation, and memory; Kasper et al., 2013). We created a dichotomous variable indicating surgery (knee, hip, cataract, back/spine, or heart) or overnight hospital stay in the last year. Other self-reported health/functioning characteristics included physical capacity (0–12 scale [α = 0.88], with higher scores reflecting greater physical capacity on activities such as walking, climbing stairs, lifting, bending over, and opening jars; Freedman et al., 2011), self-rated health (1 = poor; 5 = excellent), and severe limitations related to 12 daily activities (count of “a lot of difficulty” with eating, toileting, dressing, bathing, transferring out of bed, getting around inside the home, getting outside of the home, laundry, grocery shopping, cooking, medication management, or banking matters). Participants who received help (for health or functioning reasons) and never completed a task independently in the last month were also coded as having a severe limitation. We included a count of the number of 12 chronic conditions and symptoms (count of heart disease, high blood pressure, heart attack, arthritis, osteoporosis, diabetes, lung disease, stroke, cancer, a broken or fractured hip, anxiety, or depression). Within that count, current symptoms of anxiety and depression were measured using the previously validated Patient Health Questionnaire-4 (Löwe et al., 2010). We also adjust for if the respondent was a proxy because the proportion of proxy respondents varies by care arrangement (Freedman & Spillman, 2014a) and reports of UCNCs might differ between proxies and participants.
Care arrangement
The care arrangement measure consisted of three mutually exclusive categories: (a) those who live in the community with family and unpaid care only, (b) those who live in the community with paid care, (c) those who live in residential care, excluding nursing home residents. Nearly all older adults receiving care across the LTC continuum also received unpaid care (Freedman & Spillman, 2014a). Care arrangement in 2015 and 2017 was used to create the change in care arrangement measure for the longitudinal analysis. Participants either had the same care arrangement in 2015 and 2017 (referent), received a lower level of care in 2017 compared to 2015 (e.g., had paid care in 2015 and unpaid care only in 2017), or received a higher level of care in 2017 compared to 2015 (e.g., had paid care in 2015 and lived in a residential care setting in 2017).
Analysis
For the cross-sectional analysis, the NHATS 2015 analytic weights were used, and standard errors were corrected to adjust for NHATS’ complex sample design (Kasper & Freedman, 2018). All independent variables were taken at baseline in 2015. For bivariate analyses, we used the χ 2 test to examine if the specific type of UCNCs (e.g., went without eating) varied by care arrangement in 2015. Weighted logistic regression was used to examine how care arrangement in 2015 was associated with 2015 UCNCs related to self-care or mobility activities, UCNCs related to household activities, and any UCNC, before and after adjusting for predisposing characteristics (age, gender, race/ethnicity), enabling factors (live alone, Medicaid, number or children, annual income), and need factors (probable dementia, surgery or hospital stay in last year, severe activity limitations, chronic conditions, self-rated health, physical capacity, and proxy respondent). Weighted logistic regression was also used with each UCNC as the outcome variable to identify specific gaps in care related to daily activities after adjusting for potential confounding factors.
For the longitudinal analysis, we used the NHATS 2017 analytic weights to account for differences in the 2015 and 2017 samples due to mortality and attrition. Weighted multinomial logistic regression was used to examine how care arrangement in 2015 was associated with change in UCNCs 2 years later, before and after adjusting for 2015 covariates, and change in care arrangement between 2015 and 2017 (same arrangement, lower level of care, higher level of care). Because UCNCs can lead to premature death (Gaugler et al., 2005) and it is possible that certain care arrangements could keep people alive longer but sicker, we explicitly examine the association between care arrangement in 2015 and death 2 years later using bivariate and multivariate multinomial logistic regression models (Supplementary Table 1). All analyses were conducted in Stata version 16.0 (Stata Corp., College Station, TX).
Results
Table 1 presents weighted characteristics of the analytic sample by care arrangement in 2015. Most older adults receive only family and unpaid care in the community (60.3%). A quarter of the sample receive paid care in the community (24.9%) and 14.8% receive care in a residential setting. Consistent with prior research (Freedman & Spillman, 2014a), the large majority of older adults receiving paid care in the community (85.3%) or living in residential care (82.4%) also report receiving help from unpaid caregivers (results not shown). On average, older adults receiving care in unpaid, paid community, or residential arrangements identified two unpaid caregivers that helped them with self-care, mobility, or household activities (for health and functioning reasons) in the last month (results not shown). Community-dwelling older adults in paid or unpaid care arrangements received a similar number of unpaid care hours per month (roughly 130 h) and those with paid care received an additional 94 h of paid care each month on average (results not shown).
Table 1.
Weighted Characteristics of Older Adults Receiving Help in 2015 by Care Arrangement in 2015
| Unpaid care (n = 1,383) | Paid care (n = 666) | Residential care (n = 339) | Full sample (N = 2,388) | p | |
|---|---|---|---|---|---|
| Age (1–6) | 3.0 | 3.5 | 4.4 | 3.3 | <.001 |
| Female (%) | 64.5 | 62.1 | 70.0 | 64.7 | .174 |
| Race/ethnicity (%) | <.001 | ||||
| White | 73.0 | 68.9 | 88.7 | 74.3 | |
| Black | 11.6 | 11.8 | 5.5 | 10.7 | |
| Hispanic | 9.2 | 14.0 | 2.1 | 9.3 | |
| Another race/ethnicity | 6.3 | 5.4 | 3.7 | 5.7 | |
| Medicaid (%) | 19.4 | 36.2 | 24.7 | 24.4 | <.001 |
| Lives with others (%) | 83.5 | 61.4 | 19.9 | 68.6 | <.001 |
| Number of children (0–6+) | 3.0 | 2.9 | 2.4 | 2.9 | <.001 |
| Annual income (dollars) | 47,033 | 41,371 | 42,177 | 44,908 | .706 |
| Probable dementia (%) | 20.9 | 32.9 | 41.3 | 26.9 | <.001 |
| Surgery or hospital stay (%) | 45.6 | 53.1 | 49.6 | 48.1 | .047 |
| Severe activity limitations (0–12) | 0.6 | 1.8 | 1.2 | 1.0 | <.001 |
| Chronic conditions (0–12) | 3.6 | 4.0 | 3.4 | 3.7 | .001 |
| Self-rated health (1–5) | 2.6 | 2.5 | 3.0 | 2.6 | <.001 |
| Physical capacity (0–12) | 5.9 | 4.0 | 4.9 | 5.3 | <.001 |
| Proxy respondent (%) | 8.2 | 21.5 | 18.4 | 13.0 | <.001 |
| Unmet care need consequences (UCNCs) | |||||
| Any UCNC (%) | 37.7 | 60.3 | 33.6 | 42.7 | <.001 |
| UCNCs related to self-care activities (%) | 12.3 | 32.7 | 18.6 | 18.3 | <.001 |
| Went without eating | 1.1 | 2.0 | 1.4 | 1.4 | .409 |
| Wet or soiled clothes | 6.0 | 23.9 | 15.2 | 11.8 | <.001 |
| Went without getting dressed | 3.3 | 8.3 | 3.2 | 4.5 | <.001 |
| Went without bathing | 4.9 | 11.8 | 6.0 | 6.8 | <.001 |
| UCNCs related to mobility activities (%) | 24.3 | 42.8 | 20.3 | 28.3 | <.001 |
| Stayed in bed | 4.7 | 10.6 | 5.2 | 6.2 | <.001 |
| Unable to get around inside the home | 12.8 | 25.3 | 10.5 | 15.6 | <.001 |
| Unable to get outside the home | 15.3 | 27.4 | 13.0 | 18.0 | <.001 |
| UCNCs related to household activities (%) | 15.2 | 23.1 | 11.9 | 16.7 | .001 |
| Went without clean laundry | 1.8 | 4.8 | 2.7 | 2.7 | .104 |
| Went without groceries or personal items | 3.8 | 7.2 | 5.6 | 4.9 | .054 |
| Went without hot meal | 6.5 | 10.0 | 2.8 | 6.8 | .005 |
| Made mistake managing medications | 5.8 | 8.9 | 2.1 | 6.0 | .002 |
| Went without handling bills/banking matters | 3.2 | 6.0 | 3.4 | 3.9 | .066 |
| Total sample (%) | 60.3 | 24.9 | 14.8 | 100.0 |
Notes: The mean is presented for continuous variables. p value for design-based F test.
Age significantly increased across the LTC continuum (p < .001). The average age of participants receiving only unpaid care in the community was roughly 77 years old, whereas older adults living in residential care were approximately 84 years old on average. Annual income was similar across care arrangements, but older adults with paid care (36.2%) were more likely to have Medicaid compared to those with only unpaid care (19.4%) or residential care (24.7%; p < .001). Consistent with other national estimates (Harris-Kojetin et al., 2019), non-Hispanic White older adults were overrepresented in residential care, comprising 74.3% of the analytic sample and 88.7% of those living in residential care. The proxy response rate was highest among those receiving paid community care (21.5%) compared to those in residential care (18.4%) or receiving unpaid care in the community (8.2%; p < .001). Older adults with paid care had worse health and functioning than those with only unpaid care or living in residential care, except older adults living in residential care were most likely to have probable dementia (41.3%; p < .001).
Almost half of the sample (42.7%) experienced a UCNC in the last month, with consequences related to mobility activities (28.3%) being more common than consequences related to self-care (18.3%) or household activities (16.7%). Roughly one out of six participants were unable to get around inside the home (15.6%) or unable to get outside the home (18.0%) in the last month. The most common UCNC related to self-care activities across all care arrangements was wet or soiled clothes due to needing assistance with toileting. Overall, those in residential care (33.6%) or unpaid care (37.7%) were less likely to experience a UCNC in the last month compared to those in paid care (60.3%).
Table 2 presents the unadjusted and adjusted odds of experiencing each UCNC by care arrangement in 2015, relative to those living in residential care. In the adjusted model, those with only unpaid care were significantly less likely to experience a UCNC related to toileting compared to those with paid care or living in residential care (odds ratio [OR] = 0.38, SE = 0.11; p < .01), whereas older adults with unpaid care had 2.43 times greater odds (p < .10) and those with paid community care had 3.98 times greater odds (p < .01) of making a mistake managing their medications compared to those in residential care. Community-dwelling older adults with unpaid care (OR = 2.44, SE = 1.15; p < .10) or paid care (OR = 3.78, SE = 1.82; p < .01) were also more likely to go without a hot meal in the last month compared to those in residential care. Holding all else constant, older adults with paid care were the most likely to experience a UCNC related to getting around inside the home and getting outside the home, and those receiving paid care had 1.94 times greater odds of going without bathing in the last month compared to those with only unpaid care (p < .05; results not shown).
Table 2.
Unadjusted and Adjusted Odds Ratios for Experiencing an Unmet Care Need Consequence Among Older Adults Receiving Unpaid or Pair Care in 2015 (Relative to Those Living in Residential Care)
| Unadjusted odds ratios | Adjusted odds ratios | |||
|---|---|---|---|---|
| Unpaid care only | Paid care | Unpaid care only | Paid care | |
| Went without eating | 0.75 | 1.39 | 0.55 | 0.70 |
| Wet or soiled clothes | 0.36*** | 1.76* | 0.38** | 1.17 |
| Went without getting dressed | 1.02 | 2.73* | 1.12 | 1.89 |
| Went without bathing | 0.81 | 2.10* | 0.96 | 1.86 |
| Stayed in bed | 0.91 | 2.17* | 0.81 | 0.97 |
| Unable to get around inside the home | 1.26 | 2.90*** | 1.26 | 2.06* |
| Unable to get outside the home | 1.21 | 2.54*** | 1.29 | 1.95* |
| Went without clean laundry | 0.65 | 1.78 | 0.88 | 1.27 |
| Went without groceries or personal items | 0.66 | 1.31 | 0.89 | 1.28 |
| Went without hot meal | 2.43* | 3.88** | 2.44† | 3.77** |
| Made mistake managing medications | 2.94* | 4.62*** | 2.43† | 3.98** |
| Went without handling bills/banking matters | 0.96 | 1.83 | 1.03 | 1.76 |
Notes: Coefficients are reported as odds ratios; regressions employ weights and adjust for standard errors to account for complex survey design. Referent is residential care. Adjusted models control for age, gender, race/ethnicity, Medicaid receipt, living arrangement, number of children, annual income (logged), probable dementia, surgery or hospital stay in the last year, severe limitations with daily activities, multiple chronic conditions, self-rated health, physical capacity, and proxy respondent. N = 2,388.
† p < .10, *p < .05, **p < .01, ***p < .001.
Care Arrangement and UCNCs in 2015
We hypothesized that after adjusting for differences in predisposing, enabling, and need factors, older adults living in residential care will be the least likely to experience a UCNC (Hypothesis 1) and older adults receiving only unpaid care will be the most likely to experience a UCNC (Hypothesis 2). The results given in Table 3 partially support our first hypothesis. There was not a statistically significant difference in the likelihood of experiencing a UCNC between unpaid care and residential care, but those receiving paid care had 2.42 times greater odds (p < .001) of experiencing UCNC compared to residential care. Contrary to our second hypothesis, older adults receiving paid care were more likely to experience a UCNC in 2015 related to self-care of mobility activities (OR = 1.76, SE = 0.20; p < .001) and household activities (OR = 1.57, SE = 0.32; p < .05) compared to older adults receiving only unpaid care in the community, after adjusting for all predisposing, enabling, and need factors (results not shown).
Table 3.
Logistic Regression Models Predicting Unmet Care Need Consequences (UCNCs) in the Last Month Among Older Adults Receiving Help in 2015
| Any consequence | Self-care or mobility UCNCs | Household activity UCNCs | ||||
|---|---|---|---|---|---|---|
| Unadjusted | Adjusted | Unadjusted | Adjusted | Unadjusted | Adjusted | |
| Care arrangement (ref = residential care) | ||||||
| Unpaid care only | 1.19 | 1.31 | 1.06 | 1.18 | 1.33 | 1.32 |
| Paid care | 2.99*** | 2.42*** | 2.75*** | 2.07** | 2.23** | 2.08** |
| Age (1–6) | 0.84*** | 0.85** | 0.76*** | |||
| Female | 1.01 | 0.96 | 1.11 | |||
| Race (ref = White) | ||||||
| Black | 0.93 | 0.86 | 0.97 | |||
| Hispanic | 0.93 | 0.81 | 1.21 | |||
| Another race/ethnicity | 1.44 | 1.36 | 1.51 | |||
| Medicaid | 0.86 | 0.92 | 0.80 | |||
| Lives with others | 0.76† | 0.89 | 0.49*** | |||
| Number of children (0–6+) | 1.10** | 1.08* | 1.06 | |||
| Annual income (logged) | 1.06 | 1.07 | 1.04 | |||
| Probable dementia | 1.41* | 1.27 | 1.31 | |||
| Surgery or hospital stay last year | 1.16† | 1.18 | 1.01 | |||
| Severe activity limitations (0–12) | 1.19*** | 1.23*** | 1.14* | |||
| Chronic conditions (0–12) | 1.06* | 1.07† | 1.02 | |||
| Self-rated health (1–5) | 0.80*** | 0.80** | 0.83** | |||
| Physical capacity (0–12) | 0.85*** | 0.81*** | 0.97 | |||
| Proxy respondent | 0.56** | 0.67† | 0.29*** | |||
| Constant | 0.51*** | 1.17 | 0.40*** | 0.87 | 0.46 |
Notes: UCNCs = unmet care need consequences. Coefficients are reported as odds ratios; regressions employ weights and adjust for standard errors to account for complex survey design. N = 2,388.
† p < .10, *p < .05, **p < .01, ***p < .001.
Care Arrangement and UCNC Change Between 2015 and 2017
Before we present results related to our hypotheses for the longitudinal analyses, we note that participants receiving only unpaid care in 2015 were less likely to die between baseline and follow-up (12.4%) compared to those with paid care (21.2%) or residential care (22.5%). However, after adjusting for the covariates included in our main analyses, care arrangement in 2015 was not a statistically significant predictor of death 2 years later (Supplementary Table 1). As given in Supplementary Table 1, there also was not a significant difference in mortality between paid care and residential care in both unadjusted and adjusted models. Supplementary Table 2 presents weighted characteristics of older adults included in the longitudinal analysis (N = 1,524) by change in care arrangement between 2015 and 2017. Most participants (68.4%) either continued to have their needs met and did not experience any UCNC in 2015 or 2017 (43.5%) or had persistent UCNC and continued to experience a UCNC in 2015 and 2017 (24.9%). Almost a third of participants (31.6%) either had resolved UCNCs (UCNCs in 2015, needs met in 2017) or had developed UCNCs (needs met in 2015, UCNCs in 2017). Older adults receiving paid care were significantly more likely to experience persistent UCNCs, comprising 23.6% of the sample but 36.0% of those with persistent UCNCs (p < .001).
For the longitudinal analysis, we hypothesized that in the adjusted model, older adults living in residential care will be the least likely to develop or have persistent UCNCs between 2015 and 2017 (Hypothesis 3), and older adults receiving only unpaid care in the community will be the most likely to develop and have persistent UCNCs over time (Hypothesis 4). The results given in Table 4 support our third hypothesis. Compared to older adults living in residential care, those with paid care or unpaid care were more likely to develop or experience persistent UCNCs over time (compared to having their needs met in 2015 and 2017). Holding all else constant, for those with paid care, the risk of experiencing persistent UCNCs was 4.81 times that of older adults in residential care (relative risk [RR]: 4.81, p < .001). Contrary to our fourth hypothesis, the risk of older adults with paid care experiencing persistent UCNCs (compared to needs met) was 2.30 times that of older adults receiving only unpaid care (RR: 2.30, p < .01; results not shown). There was not a statistically significant difference between older adults with paid care and only unpaid care in developing a UCNC by 2017 (results not shown).
Table 4.
Multinomial Logistic Regression Models Predicting Change in Unmet Care Need Consequences (UCNCs) Between 2015 and 2017 Among Older Adults Receiving Help in 2015
| Resolved UCNCs | Developed UCNCs | Persistent UCNCs | ||||
|---|---|---|---|---|---|---|
| Unadjusted | Adjusted | Unadjusted | Adjusted | Unadjusted | Adjusted | |
| Care arrangement (ref = residential care) | ||||||
| Unpaid care only | 1.49 | 0.97 | 1.16 | 1.72† | 1.97** | 2.22** |
| Paid care | 3.77*** | 2.22* | 2.17* | 2.40* | 6.20*** | 4.81*** |
| Change in care arrangement (ref = same arrangement) | ||||||
| Lower level of care | 1.72* | 0.38** | 0.71† | |||
| Higher level of care | 1.74† | 1.74† | 1.76† | |||
| Age (1–6) | 0.83* | 0.98 | 0.81** | |||
| Female | 0.86 | 0.81 | 1.06 | |||
| Race (ref = White) | ||||||
| Black | 0.93 | 0.64 | 0.57* | |||
| Hispanic | 0.71 | 0.92 | 0.79 | |||
| Another race/ethnicity | 1.70 | 0.47 | 1.14 | |||
| Medicaid | 1.14 | 1.10 | 0.93 | |||
| Lives with others | 0.95 | 0.67† | 0.81 | |||
| Number of children (0–6+) | 1.24** | 1.10† | 1.11† | |||
| Annual income (logged) | 1.13 | 1.06 | 1.04 | |||
| Probable dementia | 1.18 | 0.81 | 1.68* | |||
| Surgery or hospital stay last year | 1.25 | 1.02 | 1.18 | |||
| Severe activity limitations (0–12) | 1.07 | 0.98 | 1.31** | |||
| Chronic conditions (0–12) | 1.05 | 1.08 | 1.11† | |||
| Self-rated health (1–5) | 0.81† | 0.93 | 0.73*** | |||
| Physical capacity (0–12) | 0.84*** | 0.88*** | 0.80*** | |||
| Proxy respondent | 0.58 | 1.73 | 0.44* | |||
| Constant | 0.22*** | 0.31 | 0.28*** | 0.33 | 0.24*** | 1.04 |
Notes: UCNCs = unmet care need consequences. Referent needs continue to be met in 2015 and 2017. Coefficients are reported as relative risk ratios; regressions employ weights and adjust for standard errors to account for complex survey design. N = 1,524.
† p < .10, *p < .05, **p < .01, ***p < .001.
Discussion
Consistent with prior research, the results suggest that older adults receiving paid care in the community are at risk of experiencing UCNCs (Freedman & Spillman, 2014a). Building from this work, we also find that even after adjusting for expected differences in predisposing, enabling, and need factors, paid care recipients were still significantly more likely to experience UCNCs, both cross-sectionally and over time, compared to those receiving only family and unpaid care in the community or older adults living in residential care. In the wake of the coronavirus disease 2019 (COVID-19) pandemic, the shift toward home-based care is expected to continue and the results of this study provide new insight on potential gaps in care that LTC policy and programs can work to address—even among those already receiving paid community care. For example, wetting or soiling clothes due to insufficient help or difficulty with using the bathroom was the most common UCNC related to self-care activities across all care arrangements, and going without bathing was particularly prevalent among those receiving paid community care. Recent research suggests that 42% of older adults with bathing or toileting difficulty do not have equipment that could help with these tasks (Lam et al., 2021). It is possible that older adults receiving paid care feel less comfortable receiving certain types of help (e.g., bathing) or may not have that type of paid help available, which may contribute to experiencing UCNCs (Reckrey, Bollens-Lund et al., 2020). Improving access to simple assistive equipment (e.g., grab bars and shower seats) is one example of a policy intervention that could prevent UCNCs among paid care recipients. A recent randomized clinical trial of the Community Aging in Place—Advancing Better Living for Elders program, which addresses personal and environmental factors that contribute to disability, has also shown promise as a multidisciplinary intervention to improve physical functioning among community-dwelling older adults (Szanton et al., 2019).
In addition, the results suggest that the additional environmental supports available in residential care settings (e.g., meals, transportation, housekeeping; Freedman & Spillman, 2014b) may help prevent older adults living in residential care from developing or experiencing persistent UCNCs over time. However, currently residential care is usually only an option for older adults with significant financial resources because the majority privately pay for services, as well as room and board (Harris-Kojetin et al., 2019). Non-Hispanic White older adults are also overrepresented in residential care, and previous research suggests that racial disparities in use of nonnursing home residential care options, such as assisted living, do not operate through differences in financial resources alone (Jenkins Morales & Robert, 2020). Our results demonstrating lower levels of UCNC in these settings provide additional impetus for advocates and policymakers to consider how to promote equitable access to quality residential care options for older adults who might prefer this alternative.
Although previous research has linked experiencing UCNC to statements from older adults indicating the need for more help (Allen & Mor, 1997), we do not know if experiencing a UCNC represents a gap in care, differences in personal choices related to accepting help with a given task, or both. While we controlled for many critical predisposing, enabling, and need factors, there could also be unmeasured differences between older adults who chose to remain at home and receive paid care and those who decide to move to a residential setting that is also correlated with UCNCs. For instance, those who remain at home might be more comfortable with risk and value independence/autonomy over additional support/safety. Based on the observational study design, we should not conclude that use of paid care causes UCNCs, but rather the results demonstrate that older adults receiving paid care in the community are more likely to experience a mismatch between their care needs and the care received in their current environment. Although not examined in the current study, the higher prevalence of UCNC may be due to inadequate assistance (e.g., not receiving enough hours of care based on care needs), receiving poor quality care, receiving care that is not appropriately targeted to individualized care needs (e.g., help with laundry, but not with bathing), or insufficient coordination/communication between unpaid caregivers, paid caregivers, and other healthcare providers (Reckrey, Geduldig et al., 2020).
The results also suggest heterogeneity among paid care recipients in their experiences of UCNCs over time. Although paid care recipients were the most likely to experience persistent UCNCs, compared to unpaid care or residential care, they were also the most likely to develop or have resolved UCNCs over time. Future research should explore this variation to better understand how to prevent UCNC among the growing number of older adults receiving paid care in the community. Examining unmet care needs among nursing home residents also warrants further research.
The COVID-19 pandemic and increased attention on LTC present an opportunity to reimagine how care can be provided across the LTC continuum to meet the variety of needs and preferences of the older population. The results of this study provide new insight on potential gaps in care that LTC policy and programs can work to address and emphasize the need for more research to better understand how to improve and/or expand paid LTC assistance for older adults in the community to support the unpaid care many are already receiving (Konetzka et al., 2020).
Supplementary Material
Acknowledgments
We would like to thank Dr. Vincent Mor for his comments on an earlier version of this manuscript. An earlier version of this paper was presented at the NHATS 10th Anniversary Virtual Conference, May 26–27, 2021. The views expressed are those of the authors alone and do not represent those of their employers or the funding agencies.
Contributor Information
Meghan Jenkins Morales, Sandra Rosenbaum School of Social Work, University of Wisconsin–Madison, Madison, Wisconsin, USA.
Stephanie A Robert, Sandra Rosenbaum School of Social Work, University of Wisconsin–Madison, Madison, Wisconsin, USA.
Funding
This work was supported by the National Institute on Aging (P30AG017266, P30AG012846 and U01AG032947) and the University of Wisconsin-Madison Office of the Vice Chancellor for Research and Graduate Education with funding from the Wisconsin Alumni Research Foundation. This paper was published as part of a supplement sponsored by the University of Michigan and the Johns Hopkins Bloomberg School of Public Health with support from the National Institute on Aging (U01AG032947 and P30AG012846).
Conflict of Interest
None declared.
Author Contributions
M. Jenkins Morales and S. A. Robert planned the study. M. Jenkins Morales performed the statistical analysis in consultation with S. A. Robert. M. Jenkins Morales and S. A. Robert both contributed to writing the article.
References
- Allen, S. M., & Mor, V. (1997). Unmet need in the community: The Springfield Study. In Isaacs S. L. & Knickman J. R. (Eds.), To improve health and health care 1997: The Robert Wood Johnson Foundation anthology (pp. 1–24). Jossey-Bass. https://www.rwjf.org/en/library/research/1997/01/to-improve-health-and-health-care-1997.html [Google Scholar]
- Allen, S. M., Piette, E. R., & Mor, V. (2014). The adverse consequences of unmet need among older persons living in the community: Dual-eligible versus Medicare-only beneficiaries. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 69(Suppl. 1), 51–58. doi: 10.1093/geronb/gbu124 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Andersen, R. M., & Newman, J. F. (1973). Societal and individual determinants of medical care utilisation in the United States. The Milbank Quarterly, 51(1), 95–124. doi: 10.1111/j.1468-0009.2005.00428.x [DOI] [PubMed] [Google Scholar]
- Beach, S. R., Schulz, R., Friedman, E. M., Rodakowski, J., Martsolf, R. G., & James, A. E. (2020). Adverse consequences of unmet needs for care in high-need/high-cost older adults. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 75(2), 459–470. doi: 10.1093/geronb/gby021 [DOI] [PubMed] [Google Scholar]
- Carder, P. C. (2017). State regulatory approaches for dementia care in residential care and assisted living. The Gerontologist, 57(4), 776–786. doi: 10.1093/geront/gnw197 [DOI] [PubMed] [Google Scholar]
- Centers for Medicare & Medicaid Services. (2020). Long-term services and supports rebalancing toolkit. https://www.medicaid.gov/medicaid/long-term-services-supports/balancing-long-term-services-supports/index.html [PubMed]
- Coe, N. B., Konetzka, R. T., Berkowitz, M., Blecker, E., & Van Houtven, C. H. (2021). The effects of home care provider mix on the care recipient: An international, systematic review of articles from 2000 to 2020. Annual Review of Public Health, 42, 483–503. doi: 10.1146/annurev-publhealth-090419-102354 [DOI] [PMC free article] [PubMed] [Google Scholar]
- DeMatteis, J. M., Freedman, V. A., & Kasper, J. D. (2016a). National Health and Aging Trends Study Round 5 income imputation. NHATS Technical Paper #15. www.NHATS.org [Google Scholar]
- DeMatteis, J. M., Freedman, V. A., & Kasper, J. D. (2016b). National Health and Aging Trends Study Round 5 sample design and selection. NHATS Technical Paper #16. www.NHATS.org [Google Scholar]
- Freedman, V. A., Kasper, J. D., Cornman, J. C., Agree, E. M., Bandeen-Roche, K., Mor, V., Spillman, B. C., Wallace, R., & Wolf, D. A. (2011). Validation of new measures of disability and functioning in the National Health and Aging Trends Study. The Journals of Gerontology, Series A: Biological Sciences and Medical Sciences, 66(9), 1013–1021. doi: 10.1093/gerona/glr087 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Freedman, V. A., & Spillman, B. C. (2014a). Disability and care needs among older Americans. The Milbank Quarterly, 92(3), 509–541. doi: 10.1111/1468-0009.12076 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Freedman, V. A., & Spillman, B. C. (2014b). The residential continuum from home to nursing home: Size, characteristics and unmet needs of older adults. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 69(Suppl. 1), 42–50. doi: 10.1093/geronb/gbu120 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Freedman, V. A., & Wolff, J. L. (2020). The changing landscape of family caregiving in the United States. In I. V. Sawhill & B. Stevenson (Eds.), Paid leave for caregiving: Issues and answers (pp. 11–30). The Brookings Institution. https://www.aei.org/research-products/report/paid-leave-for-caregiving-issues-and-answers/ [Google Scholar]
- Gaugler, J. E., Kane, R. L., Kane, R. A., & Newcomer, R. (2005). Unmet care needs and key outcomes in dementia. Journal of the American Geriatrics Society, 53(12), 2098–2105. doi: 10.1111/j.1532-5415.2005.00495.x [DOI] [PubMed] [Google Scholar]
- Harris-Kojetin, L., Sengupta, M., Lendon, J. P., Rome, V., Valverde, R., & Caffrey, C. (2019). Long-term care providers and services users in the United States, 2015–2016. https://www.cdc.gov/nchs/nsltcp/index.htm [PubMed] [Google Scholar]
- Hass, Z., DePalma, G., Craig, B. A., Xu, H., & Sands, L. P. (2017). Unmet need for help with activities of daily living disabilities and emergency department admissions among older Medicare recipients. The Gerontologist, 57(2), 206–210. doi: 10.1093/geront/gnv142 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Jenkins Morales, M., Miller, V. J., & Hamler, T. (2021). Dismantling systemic racism in long-term services and supports: A call to action for social workers. Journal of Gerontological Social Work. Advance online publication. doi: 10.1080/01634372.2021.1942375 [DOI] [PubMed] [Google Scholar]
- Jenkins Morales, M., & Robert, S. A. (2020). Black–White disparities in moves to assisted living and nursing homes among older Medicare beneficiaries. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 75(9), 1972–1982. doi: 10.1093/geronb/gbz141 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kasper, J. D., & Freedman, V. A. (2018). National Health and Aging Trends Study (NHATS) user guide: Rounds 1–7 beta release. www.NHATS.org
- Kasper, J. D., Freedman, V. A., & Spillman, B. C. (2013). Classification of persons by dementia status in the National Health and Aging Trends Study. Technical Paper #5. www.NHATS.org [Google Scholar]
- Konetzka, R. T., Jung, D. H., Gorges, R. J., & Sanghavi, P. (2020). Outcomes of Medicaid home- and community-based long-term services relative to nursing home care among dual eligibles. Health Services Research, 55(6), 973–982. doi: 10.1111/1475-6773.13573 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lam, K., Shi, Y., Boscardin, J., & Covinsky, K. E. (2021). Unmet need for equipment to help with bathing and toileting among older US adults. JAMA Internal Medicine, 181(5), 662–670. doi: 10.1001/jamainternmed.2021.0204 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lawton, M. P., & Nahemow, L. (1973). Ecology and the aging process. In C. Eisdorfer & M. P. Lawton (Eds.), The psychology of adult development and aging (pp. 619–674). American Psychological Association. doi: 10.1037/10044-020 [DOI]
- Löwe, B., Wahl, I., Rose, M., Spitzer, C., Glaesmer, H., Wingenfeld, K., Schneider, A., & Brähler, E. (2010). A 4-item measure of depression and anxiety: Validation and standardization of the Patient Health Questionnaire-4 (PHQ-4) in the general population. Journal of Affective Disorders, 122(1–2), 86–95. doi: 10.1016/j.jad.2009.06.019 [DOI] [PubMed] [Google Scholar]
- Musumeci, M., Chidambaram, P., & O’Malley Watts, M. (2019). Key questions about Medicaid home and community-based services waiver waiting lists. http://files.kff.org/attachment/Issue-Brief-Key-Questions-About-Medicaid-Home-and-Community-Based-Services-Waiver-Waiting-Lists
- Reckrey, J. M., Bollens-Lund, E., & Ornstein, K. A. (2020). Content of home-based dementia care: Adverse consequences of unmet toileting needs. Journal of Applied Gerontology, 40(11), 1596–1600. doi: 10.1177/0733464820952995 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Reckrey, J. M., Geduldig, E. T., Lindquist, L. A., Morrison, R. S., Boerner, K., Federman, A. D., & Brody, A. A. (2020). Paid caregiver communication with homebound older adults, their families, and the health care team. The Gerontologist, 60(4), 745–753. doi: 10.1093/geront/gnz067 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Szanton, S. L., Xue, Q. L., Leff, B., Guralnik, J., Wolff, J. L., Tanner, E. K., Boyd, C., Thorpe, R. J.Jr, Bishai, D., & Gitlin, L. N. (2019). Effect of a biobehavioral environmental approach on disability among low-income older adults: A randomized clinical trial. JAMA Internal Medicine, 179(2), 204–211. doi: 10.1001/jamainternmed.2018.6026 [DOI] [PMC free article] [PubMed] [Google Scholar]
- U.S. Department of Health and Human Services. (2019). What is the lifetime risk of needing and receiving long-term services and supports?https://aspe.hhs.gov/basic-report/what-lifetime-risk-needing-and-receiving-long-term-services-and-supports [DOI] [PubMed]
- Xiang, X., An, R., & Heinemann, A. (2018). Depression and unmet needs for assistance with daily activities among community-dwelling older adults. The Gerontologist, 58(3), 428–437. doi: 10.1093/geront/gnw262 [DOI] [PubMed] [Google Scholar]
- Xu, H., Covinsky, K. E., Stallard, E., Thomas, J.3rd, & Sands, L. P. (2012). Insufficient help for activity of daily living disabilities and risk of all-cause hospitalization. Journal of the American Geriatrics Society, 60(5), 927–933. doi: 10.1111/j.1532-5415.2012.03926.x [DOI] [PMC free article] [PubMed] [Google Scholar]
- Zuverink, A., & Xiang, X. (2020). Anxiety and unmet needs for assistance with daily activities among older adults. Journal of Aging and Health, 32(5–6), 491–500. doi: 10.1177/0898264319830805 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.