Service Utilization and Consumer-Reported Unmet Needs in Medicaid Home and Community-Based Services for Older Adults

Abstract

Home and community-based services (HCBS) are increasingly favored over nursing home care by older consumers and by policymakers. Consumer-reported unmet service needs in HCBS are important service quality and person-centeredness indicators. Yet, we know little about consumer-reported unmet needs among HCBS users. Therefore, we evaluated consumer-reported unmet needs (that the services they receive currently were not meeting their needs and goals) for 9,693 Medicaid HCBS beneficiaries (age ≥65 years) in the National Core Indicators- Aging and Disability survey (2016–2019). Personal care (59.7%) and homemaker (24.4%) were the most utilized HCBS. Prevalence of unmet needs was highest in transportation (12.2%) and homemaker (11.7%) services. Consumers with poorer self-rated health, dementia, or mental illness, living alone, and people of color were more likely to report unmet needs in HCBS such as personal care, caregiver support, adult day, or transportation. Proxy survey respondents were more likely to report unmet needs in caregiver support and personal care services, and less likely to report unmet needs in transportation services. Consumer-reported unmet needs might indicate barriers to accessing HCBS. Our findings indicate differences in predictors of unmet needs by service categories, which should inform future targeted policymaking by state agencies and service providers, to improve HCBS.

Introduction

Section 1915(c) of the Social Security Act enlarged the scope of the Medicaid program to allow for comprehensive long-term services and supports (LTSS) in home and community-based settings as an alternative to institutional care (Mann et al., 2023). Over the past several decades, states have used section 1915(c) waiver programs and federally funded demonstrations and grant programs to develop a broad range of home and community-based services (HCBS) that were not previously covered under Medicaid to provide alternatives to institutionalization for eligible Medicaid beneficiaries (Mann et al., 2023). Indeed, most people prefer to receive LTSS in their home or community (e.g., Adult Day Center) compared to more traditional models of institutional LTSS delivered in nursing homes (Casado et al., 2012, Konetzka et al., 2020).

This consumer preference is mirrored by the rebalancing of state and federal support for such services, wherein Medicaid spending on HCBS now surpasses that of costliernursing home care (Murray et al., 2023; Tyler & Fennel, 2017); even though Medicaid programs aren’t mandated to provide these services and the programs can choose which HCBS to cover. This optional nature of HCBS leads to variations across states. The umbrella of HCBS can includepersonal care, homemaker/ chore, meal delivery, caregiver support, adult day, or transportation services(Wang et al. 2021). Although Medicaid is the largest funder of HCBS, there is considerable heterogeneity between states in eligibility requirements and provision of such services (AHRQ, 2012). In response to the increasing medical and social complexities of HCBS consumers, many Medicaid programs have increased the intensity and/or frequency of allowable services, with some expanding access to services overall (Konetzka et al., 2020). However, it is unknown whether the expansion of services considers the increasing heterogeneity in the demographics, social, and medical needs of HCBS consumers; or whether the expansion is driven by ‘one size fits all’ policies which may exacerbate inequities in the increasingly diversified HCBS consumer population.

The expansion of HCBS is largely viewed as positive for consumers. However, HCBS are often intermittent (e.g. caregiver respite), and variation in the availability of services creates barriers for accessing services and may lead to unmet service needs. Challenges around service availability may also result in an unmet need for services, where an individual is unable to receive a service that they need. Limited research has documented a range of consumer-reported unmet service needs among HCBS consumers and associated adverse consequences including excess mortality, falls, lack of community participation, and increased risk of institutionalization (Casado et al., 2012, Chong et al., 2022). Due to the substantial state-to-state variation of HCBS program design, there is a general lack of data measuring HCBS quality and a paucity of research measuring consumer-reported unmet service needs among consumers of Medicaid HCBS (Chong et al., 2022). Indeed, the type of services provided by HCBS programs vary considerably across states and across subpopulations within a state. This study, however, looks at variability in consumer-reported unmet service needs at the level of type of service, which is impacted by state program design and the availability of providers at the local level. Previous quantitative studies have leveraged administrative data sources which do not sufficiently capture consumer perspectives regarding service satisfaction. Thus, consumer-reported unmet service needs in HCBS remain poorly understood. In this regard, the National Core Indicators - Aging and Disability (NCI-AD) is a valuable, under-utilized resource. The NCI-AD is a survey of consumers of publicly-funded LTSS. The survey is administered in a coordinated effort by participating states, ADvancing States, and Human Services Research Institute. It includes a standardized set of measures to assess participants’ quality of life, health outcomes, and LTSS performance. This survey provides a unique opportunity to evaluate consumer-reported perspectives on individual types of HCBS. Therefore, we leveraged the NCI-AD survey data to evaluate patterns in service utilization and consumer-reported unmet service needs (that the services they receive currently were not meeting their needs and goals) in individual types of HCBS.

Conceptual framework

To identify factors associated with consumer-reported unmet service needs (that the services they receive currently were not meeting their needs and goals) for HCBS users, we grounded our analyses using a framework guided by the Andersen’s Behavioral Model (Andersen and Newman, 1973). This framework has been used to examine use of services, including unmet needs in various groups (Alkhawaldeh et al., 2023). According to this framework, health service utilization is a function of (1) predisposing factors–characteristics of individuals that might lead some to use services more than others, such as age, race, gender, living alone, and marital status, (2) enabling factors—factors that facilitate or hinder service use, such as funding program and metropolitan area of residence as indicators of access, and (3) need factors, which encompass conditions that motivate individuals to seek services, such as health conditions and disability (Alkhawaldeh et al., 2023). This framework helped us identify ana priori list of factors/ covariates to be evaluated as potential predictors of consumer-reported unmet needs in HCBS. For improved validity of our findings and accuracy of interpretations, we restricted our analytic sample to older (65+) adult Medicaid HCBS beneficiaries.

We analyzed the NCI-AD survey data (2016–2019) for two main objectives; to evaluate (1) patterns in service utilization and the prevalence of consumer-reported unmet service needs for different types of HCBS, such as personal care services, homemaker/ chore services, delivered meal services, adult day services, transportation services, and caregiver support services, and (2) differences in consumer-reported unmet service needs for each service type in the broad range of HCBS in a national sample of Medicaid HCBS consumers.

Methods

Study population

This study was approved by the Institutional Review Board at the University of Minnesota. We used data from the NCI-AD, Adult Consumer Survey. NCI-AD is an annual survey of consumers of publicly funded LTSS. The survey includes a standardized set of measures to assess participants’ quality of life, health outcomes and LTSS performance. Measures of performance include questions related to service coordination, access, choice, and safety. Proxies are used for respondents unable to participate in the survey. NCI-AD is administered in partnership by ADvancing States, Health Services Research Institute (HSRI), and state Medicaid, aging, and disability agencies. Survey participation is voluntary, and states may choose which HCBS programs they sample from. Each state must have at least 400 total respondents to participate. In our study, we included data from three annual waves, each wave collected survey responses over a 12-month period: 2016–2017, 2017–2018, and 2018–2019. We combined the waves to pool data from 24,348 survey respondents who were community dwelling and 65 years of age or older [Figure 1]. We defined community-dwelling adults as those who reported living in one of the following: own or family house or apartment, senior living apartment or complex, group home, adult family home, foster, or host home. We were unable to differentiate respondents who lived in assisted living, nursing homes or continuing care retirement communities, and as such, excluded all these individuals for living in institutional care, excluding a total of 7,962 non-community dwelling adults. We further excluded 4,696 non-Medicaid beneficiaries. From the remaining analytic sample, we excluded 1,989 subjects with missing service need/use and all respondents (n=8) from one state because of a very small sample size and a high proportion of missing data on >75% variables. Our final analytic sample included 9,693 community-dwelling older adult Medicaid HCBS beneficiaries [Figure 1].

Figure 1.

Participant selection flow for analytic sample.

OAA- Older Americans Act; PACE- Program of All-Inclusive Care for the Elderly

HCBS use and consumer-reported unmet service needs

We evaluated six types of HCBS received: (1) personal care services, (2) homemaker/ chore services, (3) delivered meal services, (4) adult day services, (5) transportation services, and (6) caregiver support services. The NCI-AD background information (BI) section collects consumer demographic and service-related information from state administrative records; including the primary source of LTSS funding and primary program, services received through the given program, duration of participation, self-directed support, and status of legal guardianship. (more details in online supplemental methods).

We obtained information regarding actual service use from the following question in the BI section [Supplemental Table 1], “What type of paid long-term care supports is the person receiving?”. Next, we identified residents with consumer-reported unmet service needs using responses to the following question, “Do the long-term care services you receive meet your current needs and goals?”. We coded respondents who indicated that their current services do not meet their needs as having any unmet need. In a follow-up question, respondents identified which of the six services (they could select multiple services) would help them meet their needs; “What additional long-term care services might help you meet your needs and goals?”[Supplemental Table 1]. Thus, service utilization and consumer-reported unmet service needs data were obtained for each service type.

Covariates

Covariate selection was informed by Anderson’s Behavioral Model of health service use (Andersen and Newman, 1973). Predisposing factors included demographic covariates such as sex (male/female), marital status (single, divorced, widowed), ZIP code designation (i.e., metropolitan, micropolitan, small town, and rural), whether the respondent lives alone (vs. with family or with non-family)t, and race/ethnicity (Black only, non-Hispanic, White only, non-Hispanic, Hispanic/Latino-only, or multi-ethnic/ multi-racial/ other race/ethnicity). Enabling factors included covariates such as year of survey, state identifier, funding program, and insurance type. Health-related needs included overall health measured on a 5-point Likert scale (poor to excellent), diagnosis of Alzheimer Diseases and Related Dementias (AD/ADRD), developmental disability, physical disability, or brain injury, and diagnosed mental illness, as recorded in the background information section or during survey administration (more details in online supplemental methods).

Analyses

We described the demographic and health characteristics of survey respondents using descriptive statistics. Next, we evaluated the prevalence of service use and prevalence of consumer-reported unmet service needs for each type of HCBS. Furthermore, for each service type, we measured the proportion of service users who reported needing more of the service. Finally, we evaluated predictors of consumer-reported unmet need for each service type using multivariable logistic regression, adjusting for all covariates. The outcome variable for consumers’ self-reported unmet service needs in each of the six service types was a binary variable—consumer reported unmet service need in that service type versus consumer did not report unmet need for that service type. The threshold for significance was two-tailed p-value <0.05. Multiple imputation by chained equations (MICE) was used to account for missingness in the following variables [Supplemental Table 2]: gender, marital status, ZIP code designation, living arrangement (alone, with non-family, with family), race (white only, Black only, Hispanic/Latino-only, multi-ethnic/ multi-race/ other race/ethnicity), overall health, Medicare eligibility, legal guardianship, mental illness diagnosis, AD/ADRD diagnosis, physical disability, brain injury, and developmental disability. MICE included year of survey administration, state of respondent residence, and actual/received and desired service use responses. Results from 20 imputed datasets were pooled to calculate the odds ratios (OR) and standard errors. All analyses were conducted using R version 4.1.3 (R Core Team, 2022).

Results

Among 9693 community-dwelling respondents 65 years and older who use Medicaid HCBS, the majority of respondents were female (72%), White (59%), lived in metropolitan zip codes (72%), and had a physical disability (56%). [Table 1]. In this sample, 20% were living with a diagnosis of AD/ADRD and 22% with mental illness. About 15% of all included survey responses were completed with the help of a proxy.

Table 1.

Characteristics of community-dwelling older adult respondents to the National Core Indicators- Aging and Disability survey (2016- 2019) on Medicaid, averaged across multiple imputations.

Characteristic	N = 9,693
Socio-Demographic Characteristics
Female (vs not Female)*	72%
Age (Years), median (interquartile interval)*	76 (70, 84)
Have Legal Guardian*	12%
Marital Status*
Single	13%
Married/Domestic Partner	20%
Separated/Divorced	28%
Widowed	38%
Living Arrangement*
Alone	51%
Family	43%
Other	6.4%
Race/Ethnicity*
White	59%
Black or African-American	20%
Hispanic or Latino	9.7%
Other/Multiracial/Multiethnic	11%
Dually Eligible (Medicare and Medicaid) Beneficiary**	92%
ZIP Code RUCA Classification**
Metropolitan	72%
Micropolitan	14%
Rural	4.9%
Small town	8.5%
Health Conditions
ADRD***	20%
Physical Disability***	56%
Developmental Disability***	7.0%
Brain Injury***	11%
Mental Illness ***	22%
Multiple Chronic Conditions***	30%
Overall Health***
Poor	21%
Fair	39%
Good	28%
Very Good	9.7%
Excellent	2.6%
Proxy Respondent	15%

AD/ADRD- Alzheimer’s Diseases and Related Dementias; RUCA- Rural-Urban Commuting Area Codes.

^*Predisposing factors

^**Enabling factors

^***Need factors.

Prevalence of service utilization and consumer-reported unmet service needs are summarized in Figure 2. Personal care services were the most used service (59.7%), followed by homemaker/chore services (24.4%), meal delivery (19.3%), transportation (7.7%), caregiver support (6.7%), and adult day (6.4%) services. Among HCBS users who used the same type of service, unmet needs in that service ranged from 0.3% to 6.5% across the six different types of HCBS. [Figure 2, Supplemental Table 3].

Figure 2.

Prevalence of Service Utilization and Unmet Needs in Medicaid Home and Community-Based Services for Community Dwelling Older Adults

Met need indicates that the consumers said that the services are meeting all of their needs and goals, completely. Light blue shows the percentage with “met needs” and service use which indicates that the consumers use the services and said that the services are meeting all of their needs and goals, completely. Dark blue shows the percentage of consumers who use the service but reported unmet needs for that service. Green shows the percentage with unmet needs for a given service in the entire study sample, regardless of service use

Table 2 displays predictors of consumer-reported unmet service needs for each service type. Among predisposing and enabling factors, race/ethnicity, marital status, and living situation predicted consumer-reported unmet service needs in several types of HCBS. Consumer-reported unmet needs in personal care services were significantly more likely to be reported by Black consumers (OR, 1.25, p<0.05) and multiracial/ multiethnic/ other race/ethnicity (OR, 1.45, p<0.01), compared to White consumers. Black consumers, compared to White consumers, were significantly less likely to report unmet service needs in transportation services (OR, 0.78, p<0.05). Compared to living alone, living with family was associated with significantly lower odds of reporting unmet service needs in homemaker/ chore services (OR, 0.81, p<0.05) and transportation (OR, 0.80, p<0.05), but significantly greater odds of reporting unmet needs in caregiver support services (OR, 2.18, p<0.001). Compared to living alone, living with someone other than family was associated with significantly lower odds of reporting unmet needs in personal care (OR, 0.45; p<0.001), homemaker/ chore (OR, 0.42, p<0.001), meal delivery (OR, 0.56, p<0.05), and transportation services (OR, 0.69, p<0.05). Unmet needs in transportation services were significantly less likely to be reported by consumers who are married (OR, 0.69, p<0.01) or widowed (OR, 0.74, p<0.01) compared to consumers who are single. Several health conditions, which can contribute to perceived needs for service use, were important predictors of consumer-reported unmet service needs. Diagnosed mental illness and poor overall self-rated health status were strong predictors of consumer-reported unmet service needs for most service types. A diagnosis of AD/ADRD was associated with significantly greater odds of consumer-reported unmet service needs in adult day services (OR, 1.49, p<0.01) and caregiver support services (OR, 1.57, p<0.001), even after adjusting for proxy (versus self) respondent [Table 2, Supplemental Table 4]. Among all HCBS, personal care services had the highest number of its users who reported unmet needs in services they were already using [Figure 2] and was the only service type which significantly predicted higher odds of consumer-reported unmet service needs among consumers who were currently receiving that service [OR = 1.19, p<.05; Table 2, Supplemental Table 4]. Proxy respondents were significantly more likely to report unmet needs in personal care services (OR, 1.27, p<0.05) and caregiver support services (OR, 2.38, p<0.001). We combined the three annual survey waves to have a sufficiently large sample size. In sensitivity analysis, we compared estimates from each individual wave to the estimates from the pooled sample and did not see meaningful differences in our results.

Table 2.

Adjusted odds ratios for unmet HCBS service needs among 9693 consumers.

	Personal Care	Homemaker/ Chore	Delivered Meals	Adult Day services	Transportation	Caregiver Support
Characteristic	OR	OR	OR	OR	OR	OR
Female (vs not Female)	1.11	1.08	1.17	1.00	1.14	0.89
Race/Ethnicity
White	—	—	—	—	—	—
Black or African-American	1.25*	0.84	1.04	1.18	0.78*	1.17
Hispanic or Latino	1.23	0.94	1.03	1.00	1.08	1.07
Other/Multiracial/Multiethnic	1.45**	1.24	1.01	1.24	1.23	1.35
ADRD	1.11	0.93	1.00	1.49**	0.76*	1.57***
Physical Disability	1.08	0.97	1.14	0.74*	1.10	1.13
Developmental Disability	1.11	0.98	1.06	1.97*	0.98	1.10
Brain Injury	1.11	0.95	0.91	1.07	1.40*	0.95
Mental Illness	1.34***	1.44***	1.33**	1.21	1.39***	0.97
Overall Health
Poor	—	—	—	—	—	—
Fair	0.66***	0.75***	0.85	0.97	0.80**	0.86
Good	0.50***	0.56***	0.66**	0.91	0.67***	0.77
Very Good	0.33***	0.47***	0.81	0.63*	0.60***	0.53**
Excellent	0.35***	0.50**	0.49*	0.77	0.44**	0.93
ZIP Code RUCA Classification
Metropolitan	—	—	—	—	—	—
Micropolitan	0.85	0.95	0.99	0.63*	0.80*	0.99
Rural	0.95	0.97	1.23	0.60	0.99	1.08
Small town	0.88	0.89	0.88	0.53**	0.79	1.08
Living Arrangement
Alone	—	—	—	—	—	—
Living w/Family	0.89	0.81*	1.07	0.87	0.80*	2.18***
Living w/non-family	0.45***	0.42***	0.56*	0.72	0.69*	0.56
Dually Eligible (Medicare and Medicaid) Beneficiary	1.25	1.19	1.11	1.24	0.91	1.26
Have Legal Guardian	0.93	0.92	0.77	1.17	0.81	1.06
Marital Status
Single	—	—	—	—	—	—
Married/Domestic Partner	1.20	1.27	1.06	1.59	0.69**	1.05
Separated/Divorced	0.97	1.05	1.03	1.78**	0.86	0.90
Widowed	1.07	0.97	1.03	1.33	0.74**	1.15
Current Service Use (yes/no)
Personal Care (yes/no)	1.19*	0.91	0.98	1.11	1.11	1.14
Homemaker/Chore (yes/no)	1.14	1.17	1.07	1.33*	1.13	1.08
Delivered Meals (yes/no)	1.06	1.08	1.03	1.25	1.14	0.92
Adult Day service (yes/no)	0.92	0.86	0.96	1.02	1.17	1.12
Transportation (yes/no)	1.04	1.14	0.87	0.73	0.83	0.67*
Caregiver Support (yes/no)	0.69*	0.99	0.86	0.99	0.86	1.06
Proxy (versus Self) Respondent	1.27*	1.02	0.94	0.76	0.70**	2.38***

^*p<0.05

^**p<0.01

^***p<0.001

OR = Odds Ratio (adjusted for all covariates, including fixed effects for state ID and survey year); AD/ADRD- Alzheimer’s Disease and Related Dementias. Regression analyses included all 9,693 participants, using multiple imputation by chained equations for missing data. Models were also adjusted for de-identified state id and survey year fixed effects.

Discussion

We analyzed service utilization and consumer-reported unmet service needs among Medicaid HCBS users in a large, multistate sample of older adults. Based on Andersen’s Behavioral Model, we found that (1) predisposing factors such as race/ethnicity, marital status, and, living alone (versus with someone) and (2) need factors such as having a diagnosis of AD/ADRD, mental illness, and poor self-rated health were significant predictors of consumer-reported unmet service needs. Predictors of consumer-reported unmet service need vary by service category. These differences in predictors of unmet service needs by service categories help identify relevant targets for service-specific policy-interventions to equitably improve access and quality of HCBS. We interpret consumer-reported unmet service needs for a service among those who are not already using the service as indicative of lack of access.

First, we examined the prevalence of service use and consumer-reported unmet service needs. We found that personal care, homemaker/chore, and meal delivery services were the most utilized service type. Prevalence of consumer-reported unmet service needs was highest in transportation (12.2%), homemaker/chore (11.7%), and personal care (10.8%) services, which potentially might be explained by inadequate service access. Previous studies have reported that provision of these HCBS are associated with greater choice and control in daily activities, better quality of life, and lower risk of institutionalization (Sands et al, 2016; Fabius, et al, 2020). Despite potential benefits of HCBS, we found that among consumers of Medicaid HCBS, prevalence of utilization of several types of HCBS were low, similar to previous studies (Wang, et al. 2021; Siconolfi, et al., 2023). Previous literature has reported lack of awareness of HCBS as a potential barrier to access (Casado & Lee, 2012; Waymouth, et al., 2023; Siconolfi, et al., 2023), which may be overcome by implementing dissemination efforts to promote awareness of HCBS among target consumers. Aged and Disability Resource Center and Area Agencies on Aging provide services that help older adults remain in their homes. Services include disseminating information about HCBS and ensuring that services and policies are tailored to local communities (ACL, 2023). However, they also face challenges in reaching targeted populations (Coyle et al., 2016). While these entities continue to promote HCBS, there is room for innovative outreach methods, and tailored communication approaches to better engage eligible consumers and their caregivers about the available services. Our findings highlight the need for interventions and policies that target increasing awareness of HCBS among target populations.

Second, we identified predictors of consumer-reported unmet service needs from an a priori list of covariates using Anderson’s Behavioural Model. As per this framework, a sequence of predisposing, enabling, and need factors influence utilization of services and could lead to unmet needs (Alkhawaldeh et al., 2023). Therefore, we evaluated the role of all three types of factors as predictors of consumer-reported unmet service needs in HCBS. We found that need factors, including AD/ADRD, mental illness, and poor self-rated health, were robust, significant predictors of consumer-reported unmet service needs. Older adult HCBS users with AD/ADRD have a unique set of barriers (for e.g. lack of awareness, stigma, lack of cultural competency) and facilitators (e.g., dementia-attuned services, technology, caregiver support, linguistically and culturally competent education and services) to receiving optimal care which warrants tailored policies for making HCBS more dementia-friendly (Waymouth, et al., 2023; Fabius, et al., 2023). In our study, diagnosed mental illness was a strong predictor of consumer-reported unmet service needs, independent of overall health status, highlighting the need for policy interventions aimed at improving awareness, access, and quality of HCBS for older adults with mental illness (Pepin, et al., 2017; Muramatsu, et al., 2012) as well as incentivizing and improving the capacity of providers to serve clients with a diagnosed mental illness. Compared to White older adults, Black older adults or older adults in otherimulti-ethnic/racial and other racial/ethnic groups were significantly more likely to report unmet service needs in personal care services. Previous reports have attributed such racial disparities in HCBS to linguistic and cultural incompatibilities (Shippee, et al., 2022; Fabius, et al., 2018). Disparities in consumer-reported unmet service needs might partially explain why minoritized racial groups have poor outcomes compared to White HCBS users (Yan, et al., 2022). The HCBS user population is a heterogeneous mix of several marginalized subpopulations in our society, each with a unique profile of challenges and needs (AHRQ, 2012). As HCBS continues to expand, it is crucial to focus future efforts in tailored policy-making with a lens of equity for increased access, improved quality, and better outcomes among HCBS users (AHRQ, 2012).

In our study, we analyzed the NCI-AD survey data to measure prevalence and identify differences in consumer-reported unmet service needs for different types of HCBS. Our study supports the feasibility and utility of such consumer surveys for HCBS providers and programs to monitor unmet needs in their clientele and identify targeted areas for improvement in service access and quality. Consumer-reported unmet service needs are an important measure of person-centeredness and quality in HCBS (CMS, 2022). The Centers for Medicare and Medicaid Services (CMS) has curated a list of indicators called the Quality Measures set, which enlists several indicators to measure HCBS quality (CMS, 2022). This list includes consumer-reported unmet service needs as an indicator; however, it relies on the HCBS CAHPS which covers only a few specific questions for a very limited set of services (CMS, 2022). The NCI-AD is leveraged in the CMS Quality Measures set, but not for consumer-reported unmet service needs. Our study indicates that consumer-reported unmet service needs measured in the NCI-AD might be a feasible, useful addition to the CMS quality Measures set.

Limitations

Findings from this study should be interpreted cautiously considering its limitations. This study analyzes administrative, person-reported data from the large-scale NCI-AD survey of a large sample of older adults who use LTSS. Although the NCI-AD has general criteria that all state samples must comply with, such as required thresholds for sample sizes and margins of error, state participation is voluntary and each state may design its own sampling scheme, which may lead to selection bias and restricted external validity (more details in online supplement). There is also potential for residual and unmeasured confounding in regression analyses. Self-rated health was a robust, significant predictor of consumer-reported unmet service needs in most service types; however, due to the cross-sectional nature of the analyses, it could be a cause or a consequence of unmet service needs (reverse causation) and should be interpreted very cautiously. We interpreted consumer-reported unmet service needs as one domain of HCBS quality; there are several domains as indicators of HCBS quality and we cannot comment on the overall quality of HCBS in our study. Among users reporting unmet needs in services they are already receiving, NCI-AD data does not indicate whether the unmet service need is due to consumers being unhappy with the service or whether they are happy with the service but are not receiving enough of it.

Conclusion

Using a large sample of community-dwelling older adult consumers of Medicaid HCBS, this study highlights service utilization and consumer-reported unmet service needs in HCBS users to inform future policymaking for improvement in access and quality of HCBS. Personal care, meal delivery, and homemaker/ chore services were the types of service which were utilized most frequently. Prevalence of consumer-reported unmet service needs was highest in transportation, homemaker/ chore and personal care services, suggestive of lack of access to these types of HCBS. Compared to older adult consumers without ADRD, those with ADRD were more likely to report significantly higher unmet service needs in adult day services and caregiver support but significantly less likely to report unmet service needs in transportation services, highlighting the need for equitable, dementia-tuned policies that improve access to adult day and caregiver support services. Diagnosed mental illness was a strong predictor for unmet service needs in HCBS, highlighting the need for policy interventions aimed at increasing awareness and improving access to and quality of HCBS for older adults with mental illness. Proxy survey respondents were significantly more likely to report unmet service needs in personal care and caregiver support services, suggesting unmet needs for these services. Appropriate use of consumer-oriented quality surveys can be incorporated into tailoring targeting decisions and redesigning HCBS programs through an equitable lens.

Key Points.

• Extent of consumer-reported unmet need in home & community-based services is unknown

• We analyzed the National Core Indicators- Aging and Disability survey (2016–2019)

• We found consumer-reported unmet service needs suggesting barriers to accessing services

• Providers can use such surveys to track satisfaction/ unmet needs in their clientele

• States can use such data for service evaluation, reforms, and equitable policymaking

Data Availability Statement:

The data use agreement between University of Minnesota and the Health Services Research Institute (HSRI) does not allow sharing of NCI-AD data with the general public.