Inequities in access to VA’S aid and attendance enhanced pension benefit to help Veterans pay for long‐term care

Abstract

Objective

To examine characteristics that are associated with receipt of Aid and Attendance (A&A), an enhanced pension benefit for Veterans who qualify on the basis of needing daily assistance, among Veterans who receive pensions.

Data sources

Secondary data analysis of 2016‐2017 national VA administrative data linked with Medicare claims.

Study design

Observational study examining sociodemographic, medical, and healthcare utilization characteristics associated with receipt of A&A among Veterans receiving pension.

Principal findings

In 2017, 9.7% of Veterans with pension newly received the A&A benefit. The probability of receiving A&A among black and Hispanic pensioners was 4.6 percentage points lower than for white pensioners (95%CI = −0.051, −0.042). Married Veterans receiving pension had a 4.4‐percentage point higher probability of receiving A&A (95%CI = 0.039, 0.048). Most indicators of need for assistance (eg, home health utilization, dementia, stroke) were associated with significantly higher probabilities of receiving A&A, with notable exceptions: pensioners with a diagnosis of Post‐Traumatic Stress Disorder (marginal effect = −0.029 95%CI = −0.037, −0.021) or enrolled in Medicaid (marginal effect = −0.053, 95%CI = −0.057, −0.050) had lower probabilities of receiving A&A. Unadjusted and adjusted rates of receiving A&A among Veterans receiving pension varied by VA medical center.

Conclusions

This study identified potential inequities in receipt of the A&A enhanced pension among a sample of Veterans receiving pension. Increased Veteran outreach, provider education, and VA office coordination can potentially reduce inequities in access to this benefit.

What This Study Adds

• Findings from this study suggest that there are subgroups of Veterans receiving pension who have a lower likelihood of receiving A&A, even after accounting for factors related to Veterans’ need for assistance.

• Results suggest that there is variation in receipt of A&A among pensioners by VA Medical Centers.

• These findings call for targeted efforts to ensure equal access to this VA benefit.

1. INTRODUCTION

The Department of Veterans Affairs (VA), through its Veterans Health Administration (VHA) and Veterans Benefits Administration (VBA), supports a large population of older Veterans who require long‐term care (LTC). The VHA, through its office of Geriatrics and Extended Care (GEC), provides a number of clinical programs to older and functionally impaired Veterans needing assistance with daily activities (eg, home‐based primary care, homemaker/home health aide). In addition, the VBA offers monthly cash benefits to Veterans who qualify on the basis of income and need for assistance through its Aid and Attendance enhanced pension benefit (A&A). Veterans do not have to receive care from VHA to be eligible for A&A. While much attention has been devoted to understanding the distribution of resources allotted and access to GEC’s clinical programs, ¹ , ² , ³ , ⁴ , ⁵ , ⁶ less is known about Veterans’ access to VBA’s cash benefit that may assist Veterans in meeting their LTC needs.

Honorably discharged wartime Veterans who have limited income (<$17 724 for a Veteran with one dependent in 2018) and net worth (<$127 061 in 2018) AND who are age 65 years or older, or who have a total and permanent nonservice‐connected disability, are eligible for a pension benefit (see Figure S1). ⁷ Among these Veterans, those who also require the assistance of another person to attend to their daily living needs are eligible for the VBA’s A&A enhanced pension. ⁸ In 2017, the VA paid over $3.3 billion in pension benefits to 276 570 Veterans, 41% of whom received the A&A enhanced pension. ⁹ The estimated average annual amount paid to Veterans receiving pension in 2017 was $12 103 and the average A&A enhanced pension was $4666 higher ($16 769) in 2017.

There are no restrictions on how the A&A enhanced pension cash benefit is to be spent; rather, Veterans make their own decisions regarding how best to use it to meet their own needs. A 2011 GAO report found that receipt of the A&A enhanced pension benefit enabled Veterans to obtain needed assistance through various mechanisms, including in‐home services (eg, personal care, homemaker and chore services, meal preparation, medication management), transportation, offsetting lost income of family members who provide care, and defraying the cost of an assisted living facility. ¹⁰

Importantly, the A&A enhanced pension benefit offered through VBA is an entitlement for which Congress is obligated under law to approve funds to cover all who are eligible and enrolled. The VA is required to ensure that Veterans are informed about the availability of VBA benefits and, to the maximum extent possible, offer assistance in applying for them. ¹¹ The VA accredits Veterans Services Organizations, attorneys, and claims agents ⁷ , ¹² to help Veterans apply for VA benefits free of charge. Nevertheless, the A&A enhanced pension benefit is underutilized: a 2004 VA study estimated that <1/3 of Veterans who received a pension and were potentially eligible for A&A actually received it. ¹³

Previous research suggests that underutilization of the VBA’s A&A enhanced pension benefit stems from Veterans’ lack of awareness about the benefit. ¹⁰ An investigation by the GAO attributes several reasons for this lack of awareness, including VA’s current outreach activities typically do not include A&A; the printed material distributed to Veterans lacks information on eligibility and application requirements; and there is variability among VBA’s 57 regional offices in outreach activities that target older adults who are most likely to qualify for A&A. ¹⁰ Nevertheless, there have not been any systematic studies that evaluate the factors that promote receipt of the A&A enhanced pension benefit.

The well‐known Behavioral Model of Health Care Utilization ¹⁴ , ¹⁵ , ¹⁶ can be used to identify predictors that may impact a Veteran's access to the A&A enhanced pension benefit. With this model, we hypothesize that receipt of the A&A enhanced pension benefit depends on demographic and social characteristics that predispose Veterans to seek benefits, factors that enable or impede access, and Veterans’ health‐related need for assistance. In this model, predisposing factors are exogenous characteristics that affect a Veteran's propensity to receive A&A. They include factors such as education, race, age, and marital status. Enabling factors include circumstances that empower or facilitate a Veteran or their caregiver to obtain A&A. Examples include living arrangement, receipt of care from the VA, income/assets, living in a rural area, a Veteran's caregiver network, and characteristics of the local VA Medical Center. Need variables include requiring assistance from another individual to perform activities of daily living because of a chronic or disabling condition, including blindness, that often precipitate the need for LTSS and could qualify Veterans to be eligible for A&A.

Our objectives in this study, the first of its kind using a comprehensive set of VA and Medicare administrative data, were to (a) identify Veterans’ predisposing, enabling, and need characteristics that are associated with receipt of A&A among a sample of Veterans receiving pension; and (b) determine how receipt of A&A among pensioners varies by VA Medical Center, suggesting differences in access.

2. METHODS

2.1. Data sources

We merged VA and Medicare administrative data from fiscal years (FY) 2016 and 2017 to complete this retrospective observational study. The FY2016 and FY2017 VHA Assistant Deputy Under Secretary for Health (ADUSH) Enrollment Files were used to gather information about Veterans’ demographic characteristics, date of birth, date of death, VHA healthcare utilization, and enrollment in Veterans Benefits Administration (VBA) benefits. Missing demographic information was supplemented with the Medicare enrollment data. In addition to the flag in the ADUSH that indicates receipt of any VHA inpatient or outpatient care, we used data from VHA inpatient and outpatient claims and billing records to gather information about Veterans’ healthcare utilization. We supplemented the VA data with Medicare claims because more than half of VA‐enrolled Veterans eligible for Medicare receive care outside of VA. ¹⁷ Specifically, we merged the VA data with Medicare Part A and B claims, which include all Medicare inpatient, outpatient, and skilled nursing facility claims. Our study and use of these data were approved by the Providence VA Medical Center's Institutional Review Board.

2.2. Analytic sample

Our analytic sample included any Veteran who was enrolled in the VHA and Medicare, age 66 years and older, and received a pension benefit but did not receive the A&A enhanced pension benefit in FY2016 (See Figure 1). The rationale for excluding Veterans already receiving A&A is that we wanted to identify characteristics that precipitates receipt of A&A rather than those that result from prior receipt of A&A. For instance, some of the healthcare utilization that we consider indicators of need, could in turn be affected by having extra financial support for caregiving. Furthermore, the research question is focused on “take up” of the benefit, rather than differences in eligibility. While we cannot precisely identify a Veteran's A&A eligibility with available data, we can identify characteristics of those who receive the benefit among those who are potentially eligible for receiving A&A (ie, receiving pension) thus enabling us to predict A&A enrollment in FY2017 among those receiving pension. In a sensitivity analysis, we replicate our analysis with a prevalence sample of all Veterans receiving A&A in 2017, including those who had A&A in the year prior to account for any concerns around selection bias (see analysis section and Appendix S1).

FIGURE 1.

Flowchart of patient sample Abbreviations: A&A, Aid and Attendance; ADUSH, Assistant Deputy Under Secretary for Health Enrollment File; FY, Fiscal Year

We excluded Veterans with any Medicare Advantage enrollment in FY2016, because we do not have complete healthcare utilization histories for those Veterans (n = 43 919). We also excluded 4940 Veterans whose preferred VA medical facility was outside of the 50 US states or Washington, DC (eg, Puerto Rico, Manila, Pacific Islands) and 3318 Veterans with missing data. This resulted in a final sample of 89,845 Veterans who received pension benefits in FY2016 and FY2017 but did not receive A&A in FY2016.

2.3. Dependent variable

We identified new receipt of A&A in FY2017 among those receiving pension in the ADUSH enrollment file.

2.4. Independent variables

Predisposing factors included age, gender, race, and marital status. Enabling factors included living in an urban vs. nonurban area, whether or not a Veteran received any inpatient or outpatient care from the VHA in FY2016 or FY2017 (based on our assumption that greater exposure to the VHA would increase opportunities for access), residence in a VA community residential care facility in FY2016 or FY2017 (based on the notion that residents in these settings may receive assistance accessing benefits) and whether or not Veterans had a visit with a VA social worker, a member of a VA Home‐Based Primary Care Team, or were enrolled in one of VHA’s Geriatrics Patient Aligned Care Teams in FY2016 or FY2017 (assuming that the these VA services would identify need and provide assistance in completing the application for benefits). Need variables included VA and Medicare‐paid healthcare utilization in the previous year as indicators of condition severity (ie, inpatient, intensive care unit, skilled nursing facility, psychiatric admission, and home health visits, including both Medicare home health and VA homemaker/home health aide services) and Medicaid enrollment in FY2016. With the VA and Medicare claims, we calculated the Jen Frailty Index (a claims‐based measure of functional impairment and long‐term institutionalization risk) ¹⁸ and the Charlson Comorbidity Index, ¹⁹ and identified mental health diagnoses (ie, nonservice‐connected Post‐Traumatic Stress Disorder, serious mental illness, substance abuse, depression) as well as chronic and disabling conditions (ie, dementia, stroke, hemiplegia, diabetes with end organ damage) that often precipitate the need for LTC and if present, could qualify Veterans to be eligible for A&A.

2.5. Analysis

Among a sample of Veterans receiving pension in FY2016 (89,845), we used a logit model to estimate a pensioner's probability of newly receiving A&A in FY2017, including the aforementioned independent variables and VAMC fixed effects. We report the average marginal effect for each variable, or the expected change in the dependent variable given a one‐unit difference in the covariate, holding the others constant. ²⁰ We conducted a separate model without VAMC fixed effects and aggregated the Veteran‐level predictions from this model to the VAMC level. We calculated the difference between the actual rate of the A&A enhanced pension benefit among Veterans receiving pension in each VAMC with the average of all of the Veteran‐level predicted rates in each VAMC. In an Appendix S1, we present the characteristics of a prevalence sample of all Veterans already receiving A&A in 2017 to show how this sample differs from Veterans newly receiving the A&A benefit in FY2017 (n = 140 975). We also replicated our analysis with data from earlier years (FY2013 and FY2014) to account for possible differences in the measurement of VHA‐purchased care that occurred after the enactment of the Veterans Access, Choice and Accountability Act of 2014 (Choice Act). ²¹ The Choice Act changed the way some community‐provided services were paid for and recorded in health records; therefore, examining earlier years addresses the concern that some VA‐purchased care data were missing in later years. Statistical analyses were performed using SAS 9.4.

3. RESULTS

Among 89 845 Veterans who received a pension but not the A&A enhanced benefit in 2016, a total of 8724 (9.7%) Veterans newly received the A&A enhanced pension in FY2017. Veteran pensioners who received A&A were significantly older and more likely to be white and married than those who did not receive A&A (Table 1). Veterans receiving A&A enhanced pension had higher rates of healthcare utilization, social work visits, and frailty/comorbidity than pensioners without A&A. In contrast, A&A recipients had lower rates of psychiatric hospital admissions, lower use of any VHA services, and lower prevalence of serious mental illness than those receiving pension but not A&A.

TABLE 1.

Descriptive characteristics of Veterans receiving pension in 2017, by receipt of Aid and Attendance (n = 89 845)

	No A&A		Newly Receiving A&A		Total
	n = 81 121		n = 8724		N = 89 845
Predisposing characteristics
	n	%	n	%	N	%
Age
66‐74	56 524	69.68	1652	18.94	58 176	64.75
75‐84	10 100	12.45	1282	14.7	11 382	12.67
85+	14 497	17.87	5790	66.37	20 287	22.58
Male	78 745	97.07	8412	96.42	87 157	97.01
Race
White, non‐hispanic	55 996	69.03	7488	85.83	63 484	70.66
Black	19 901	24.53	923	10.58	20 824	23.18
Hispanic	3270	4.03	220	2.52	3490	3.88
Other	1954	2.41	93	1.07	2047	2.28
Married	14 819	18.27	3472	39.8	18 291	20.36
Enabling characteristics
Residing in an urban area	59 521	73.37	6669	76.44	66 190	73.67
Any VHA Use a	70 580	87.01	6780	77.72	77 360	86.1
Resident in VA community residential care facility a	412	0.51	99	1.13	511	0.57
VA social work visit a	12 262	15.12	2130	24.42	14 392	16.02
Geriatrics patient aligned care team visit a	1251	1.54	276	3.16	1527	1.7
Home‐based primary care visit a	2096	2.58	454	5.2	2550	2.84
Need characteristics
Healthcare utilization in previous year b
Inpatient admission	19 115	23.56	4060	46.54	23 175	25.79
Intensive care unit stay	4332	5.34	1314	15.06	5646	6.28
Nursing home stay	17 181	21.18	4046	46.38	21 227	23.63
Psychiatric hospital admission	1038	1.28	81	0.93	1119	1.25
Home health visit	7853	9.68	3628	41.59	11 481	12.78
Medicaid c	20 452	25.21	1516	17.38	21 968	24.45
JEN frailty index d
Low (0‐3)	30 995	38.21	2004	22.97	32 999	36.73
Moderate (4‐5)	20 051	24.72	2156	24.71	22 207	24.72
High (6‐7)	10 894	13.43	2205	25.28	13 099	14.58
Very High (8+)	19 181	23.64	2359	27.04	21 540	23.97
Charlson comorbidity index d	1.97	2.56	3.85	3.12	2.16	2.67
Mental health conditions d
Post‐traumatic stress disorder	5203	6.41	194	2.22	5397	6.01
Serious mental illness	4042	4.98	259	2.97	4301	4.79
Substance abuse	5444	6.71	231	2.65	5675	6.32
Depression	13 066	16.11	2271	26.03	15 337	17.07
Dementia d	6899	8.5	3389	38.85	10 288	11.45
Stroke d	1911	2.36	886	10.16	2797	3.11
Hemiplegia d	1364	1.68	399	4.57	1763	1.96
Diabetes with end organ damage d	8723	10.75	1694	19.42	10 417	11.59

All differences between groups statistically significant (P<.001) according to Chi‐square tests. JEN Frailty Index is measured on a scale of 0 (no frailty) to 13 (high frailty) and is calculated based on geriatric syndromes, functional deficits, and multimorbid clusters. The Charlson Comorbidity Index is a weighted summary measure on a scale of 0 (no comorbidities) to 33 (high comorbidity burden) based on 17 comorbidities

Abbreviations: A&A, Aid and Attendance; VHA, Veterans Health Administration.

^aUtilization in current or previous year (FY2016 and/or FY2017).

^bUtilization in previous year (FY2016).

^cMedicaid enrolled for at least one month in FY2016 according to the Medicare enrollment record.

^dDiagnoses reported in Medicare claims and/or VA records during the prior year (FY2016).

In the fully adjusted model (Table 2), Veterans in older age groups (74‐84 and 85+) who received pension had higher probabilities of receiving A&A compared with Veterans ages 66‐74 receiving pension (marginal effect (ME) = 0.07 (95%CI = 0.06, 0.08) and ME = 0.16 (95% CI = 0.15, 0.16), respectively). In addition, pensioners who were black, Hispanic, or other nonwhite race had lower probability of receiving the A&A enhanced pension than their white counterparts. Specifically, the percentage of Veterans who were black and receiving A&A enhanced pension was 4.6 percentage points lower than whites (95% CI = −0.5, −0.04), 4.1 percentage points lower for Hispanics compared to white pensioners (95%CI = −0.05, −0.03), and 3.3 percentage points lower for other races compared to white pensioners (95% CI = −0.05, −0.02). Veterans receiving pension who were married were 4.4 percentage points more likely to receive A&A (95%CI = 0.04, 0.05). In addition, pensioners residing in a VA Community Residential Care Facility or who had a VA Social Work visit in the prior year had a higher probability of newly receiving A&A than pensioners who did not (ME = 0.04 (95% CI = 0.02, 0.06) and ME = 0.03 (95% CI = 0.03, 0.04), respectively). Veterans receiving pension who were enrolled in Medicaid were 5.3 percentage points less likely to receive A&A (95% CI = −0.06, −0.05). Pensioners who utilized Medicare home health or VA homemaker/home health aide services in the previous year had a 7‐percentage point higher probability of receiving A&A (95%CI = 0.06, 0.08). Pensioners with diagnoses of dementia and stroke had higher probabilities of receiving A&A. However, pensioners with nonservice‐connected diagnoses of PTSD & SMI had lower probabilities of receiving the A&A enhanced pension.

TABLE 2.

Association of Veteran characteristics with new receipt of the aid and attendance enhanced pension benefit among all Veterans receiving pension in 2017 (n = 89 845)

	Marginal effect	95% Confidence Interval
Predisposing characteristics
Age
75‐84 vs. 66‐74	0.069	0.063	0.075
85 + vs. 66‐74	0.157	0.151	0.163
Male	−0.003	−0.012	0.006
Race
Black vs. White, Non‐hispanic	−0.046	−0.051	−0.042
Hispanic vs. White, Non‐hispanic	−0.041	−0.049	−0.033
Other vs. White, Non‐hispanic	−0.033	−0.046	−0.021
Married	0.044	0.039	0.048
Enabling characteristics
Urban	0.024	0.020	0.028
Any VHA Use a	0.013	0.008	0.018
Resident in VA community residential care facility a	0.041	0.018	0.064
VA social work visit a	0.033	0.027	0.038
Geriatrics patient aligned care team visit a	−0.010	−0.019	−0.001
Home‐based primary care a	−0.023	−0.030	−0.017
Need characteristics
Healthcare utilization in previous year b
Inpatient admission	0.004	0.000	0.009
Intensive care unit stay	0.004	−0.002	0.010
Nursing home stay	0.030	0.024	0.035
Psychiatric Hospital admission	0.029	0.009	0.049
Home health visit	0.069	0.063	0.075
Medicaid c	−0.053	−0.057	−0.050
JEN Frailty Index d
Moderate (4‐5) vs. Low (0‐3)	0.006	0.001	0.010
High (6‐7) vs. Low (0‐3)	0.012	0.006	0.018
Very high (8+) vs. Low (0‐3)	−0.009	−0.014	−0.004
Charlson comorbidity index d	0.080	0.085	0.091
Mental health conditions d
Post‐traumatic stress disorder	−0.029	−0.037	−0.021
Serious mental illness	−0.015	−0.023	−0.006
Substance abuse	−0.009	−0.018	0.000
Depression	0.011	0.006	0.015
Dementia d	0.051	0.045	0.057
Stroke d	0.042	0.032	0.051
Hemiplegia d	−0.001	−0.011	0.009
Diabetes with end organ damage d	0.005	−0.001	0.010

Model includes VA Medical Center fixed effects. JEN Frailty Index is measured on a scale of 0 (no frailty) to 13 (high frailty) and is calculated based on geriatric syndromes, functional deficits, and multimorbid clusters. The Charlson Comorbidity Index is a weighted summary measure on a scale of 0 (no comorbidities) to 33 (high comorbidity burden) based on 17 comorbidities. Model C‐statistic = 0.87 (95% CI = 0.87‐0.88).

Abbreviation: VHA, Veterans Health Administration

^aUtilization in current or previous year (FY2016 and/or FY2017).

^bUtilization in previous year (FY2016).

^cMedicaid enrolled for at least one month in FY2016 according to the Medicare enrollment record.

^dDiagnoses reported in Medicare claims, VA, and fee records during the prior year (FY2016).

Figure 2 presents the actual and predicted percent of Veterans with pension who receive the A&A enhanced pension. The actual rates of enrollment in A&A among Veterans receiving pension ranged from less than 1% to 22% across VAMCs. Using the predicted rates of A&A from our model to derive the expected rate of enrollment in A&A among Veterans receiving pension, we identified 48 out of 129 VAMCs that had significantly lower rates of enrollment than their predicted values (see Table S1 for rates by VAMC).

FIGURE 2.

Actual versus Predicted Rates of Aid & Attendance Enrollment among Veterans Receiving Pension, by VA Medical Center (2017) Notes: Actual rates are the percent of Veterans receiving pension that enroll in A&A in 2017, by VA Medical Center. Predicted rates come from the aggregated estimates produced by a logit model with Veterans’ predisposing, enabling, and need characteristics. See Table S1 for Actual and Predicted Rates of A&A among pensioners, by VA Medical Center. A&A, Aid and Attendance; VAMC, Veterans Affairs Medical Center

Results of our sensitivity analysis among a prevalence sample of Veterans receiving A&A in FY2017 suggest that findings are similar to that of the incidence cohort with a few exceptions related to receipt of services from two of VA’s geriatrics programs (ie, HBPC and GeriPACT) and indicators of Veterans’ need (See Tables S2 and S3). First, among Veterans receiving pension, those who were enrolled in HBPC or GeriPACT during the current or prior year had higher probabilities of receiving A&A. In addition, Veterans who had an inpatient admission or a stay in an intensive care unit in the previous year had lower probability of receiving A&A, and Veterans with a diagnosis of serious mental illness had higher probabilities of receiving A&A than Veterans with pension without an SMI diagnosis. The direction of the effects for the remaining predisposing and enabling characteristics was not different between the incidence and prevalence groups of A&A recipients. Similarly, findings from our analyses with earlier years of data were consistent with those from later years (see Tables S4 and S5).

4. DISCUSSION

Using a cohort of Veterans receiving pension, we conducted a study of Veterans’ predisposing, enabling and need characteristics that predict receipt of A&A, an enhanced pension benefit intended to empower Veterans with functional impairment to meet their daily care needs. The results of this study suggest that among Veterans receiving pension, nonwhite Veterans were significantly less likely to receive A&A than their white Veteran counterparts, adjusting for other characteristics that are likely to predispose, enable or signal need to access this enhanced pension benefit. Additionally, among Veterans receiving pension, those who were older and married were more likely to obtain A&A. Veterans who had a recent visit with a VA social worker and those who resided in urban areas or in VA’s Community Residential Care Facilities were more likely to receive the A&A enhanced pension. Most indicators of need for A&A (eg, utilization, disabling diagnoses, level of frailty, and comorbidity burden) were associated with significantly higher probabilities of receiving A&A among Veterans with pension, with notable exceptions. Specifically, Veterans receiving Medicaid, and Veterans with diagnoses of nonservice‐connected PTSD were less likely than their counterparts to receive the A&A enhanced pension benefit in FY2017. These results were consistent across both our incidence and prevalence sample of Veterans. Finally, we found that new receipt of A&A among pensioners varies considerably by VAMC.

While racial differences and disparities are prevalent in many aspects of financial assistance/social welfare program access outside of VA (eg, Temporary Assistance for Needy Families, Supplemental Nutrition Assistance Program, earned income tax credit), ²² , ²³ , ²⁴ , ²⁵ , ²⁶ , ²⁷ , ²⁸ we are the first to report racial inequities in access to VBA’s A&A benefit among a cohort already receiving pension benefits. In differentiating between differences and disparities in healthcare utilization, the Institute of Medicine considers possible causes and concludes that some causes of racial differences are acceptable, such as health beliefs, preferences, or differences in need. ²⁹ We postulate that we can dismiss such “acceptable” factors when the outcome in question is receipt of a cash entitlement in a low‐income population with LTC needs. Additionally, we controlled for differences in underlying need through diverse indicators associated with health‐related need for assistance, yet racial disparities in access remain in both our main analysis and our sensitivity analyses. It is likely that the reasons behind the racial disparities observed are multifactorial and complex. They may include unequal outreach and education about benefits as previously highlighted by the VA’s Advisory Committee on Minority Veterans, ³⁰ provider bias in referring Veterans to appropriate channels for the A&A application, and stereotypical attitudes about how minority Veterans may decide to spend this cash benefit. ³¹ , ³² These disparities may also be driven by differences in Veterans’ social support to identify and advocate ³³ for benefits as well as trust of healthcare professionals ³⁴ , ³⁵ that may influence reporting the need for assistance. This list is not exhaustive and should be more fully explored in future research.

In addition, we observed that enrollment in A&A among pensioners varied by marital status. Similar observations have been found in access to SNAP benefits, in which older adults who live alone are less likely to be enrolled in SNAP, despite being eligible, than those who live with others. ³⁶ The mechanisms underlying unequal access to A&A among Veterans receiving pension by marital status are likely to be quite different from those underlying racial disparities. Yet, given that 80% of our sample of Veterans receiving pension are not married, possible causes require close scrutiny and remediation. It has long been known that the majority of caregivers to older adults are women, and women fill many roles in this capacity, including health provider, care manager, and advocate. ³⁷ Wives, in particular, provide the majority of care to their husbands. ³⁸ Recognizing that family caregiving is the backbone of support for Veterans, and that caregiving takes a physical and psychological toll, the VHA offers a range of services to caregivers, and local Caregiver Support Coordinators at each VAMC are available to assist family caregivers in identifying benefits and services for which they may be eligible. It is likely that the married Veterans in this study benefit not only from the care provided by wives but also from their advocacy efforts on behalf on their spouses and themselves to secure the necessary services and benefits available to meet their husbands’ needs. Daughters and other family and friends of nonmarried Veterans may provide care but may not be as informed about the services available to Veterans and their caregivers, especially if they live a distance away. Additional work is needed to understand and eliminate the differences observed in receipt of A&A among pensioners by marital status.

In addition, we observed that pensioners enrolled in Medicaid were less likely to receive A&A. Given states’ different requirements as it relates to Medicaid eligibility, including the inclusion/exclusion of VA benefits as countable income, it is possible that the benefit of receiving additional cash is not worth the risk of losing Medicaid coverage. Or, it could be the case that Veterans enrolled in Medicaid have access to the services and supports they need and are therefore not seeking out additional cash to assist in meeting their LTC‐related needs. Further research is needed to understand how states’ Medicaid programs facilitate or impede Veterans’ access to VBA’s entitlements.

Veterans receiving pension with diagnoses of nonservice‐connected PTSD were less likely to receive A&A, controlling for other factors, including facilitating events such as meeting with a VA social worker. It is possible that providers are less aware of the impact of nonservice‐connected PTSD on Veterans’ ability to meet their daily living needs than they are of the impact of physical health conditions, despite substantial reports to the contrary by the caregivers of Veterans with PTSD. ³³ Additionally, Veterans with PTSD may be less likely than other eligible Veterans to have a social network to advocate for receipt of A&A. ³⁹ , ⁴⁰ Provider perceptions about the capacity of Veterans with PTSD to appropriately manage their cash benefits may also contribute to our findings of reduced access in this population. Additional investigation into the differences in receipt of the A&A enhanced pension benefit for these groups is warranted.

Our findings also suggest that access to A&A among Veterans already receiving pension varies substantially by VAMC. This may be a function of the integration of VBA and Veteran Service Organizations (VSOs) with the VAMC. For example, some VAMCs have a VBA office co‐located on the campus and/or have scheduled presence of a VSO at the VAMC during the week. Veterans receiving care at these VAMCs may have increased access to information about this benefit and assistance in applying for it. While currently data are not available to examine this relationship, future work would benefit from testing this hypothesis. In addition, education on the eligibility criteria for A&A and the application process may vary within and between VAMCs. While VA has implemented efforts to assist Veterans in understanding and completing the application process, ⁴¹ , ⁴² national efforts to increase VAMC personnel's understanding of A&A eligibility criteria and process for enrolling in this benefit could alleviate the differences in enrollment among pensioners that we observed across VAMCs in this study.

Given our findings regarding unequal access to A&A among Veterans receiving pension, wider and equal dissemination of information around A&A, including eligibility criteria and enrollment processes are warranted. A number of third‐party agencies and individuals have recognized the lack of Veterans’ awareness benefits and processes to obtain them. ⁴³ , ⁴⁴ The US Special Committee on Aging and the GAO found that there are many organizations who charge substantial fees (some up to $10 000) to prepare benefit applications for Veterans, despite federal law prohibiting charging such fees. Without wider and equal dissemination of information about A&A and Veterans’ access to free assistance in completing applications, there is the potential for predatory practices to evolve, and potential inequities in access to continue.

The findings from this work have important implications for both VHA and VBA. While benefits, in general, fall under the purview of the VBA, the VHA has great potential to identify Veterans receiving pension who may be candidates for A&A given their clinical needs. However, it is likely that not all VHA providers equally understand A&A eligibility criteria or enrollment processes. Closer integration of VBA Regional Offices and VAMCs could allow for sharing of information and resources to educate providers and boost outreach and enrollment efforts. The VA’s Caregiver Support Program could also be leveraged to educate caregivers about the A&A enhanced pension benefit. In addition, VHA data could be used by VBA to proactively target Veterans receiving pensions who may be eligible for the A&A enhanced pension. Targeting can be especially helpful to reach homebound Veterans with pension, many of whom are expected to qualify for A&A, and yet they do not show up to typical activities where information about these benefits would be present (eg, senior centers, meetings of VSOs). In addition, for Veterans receiving pension, there is not information on disabling conditions readily available to VBA as there is for Veterans receiving disability compensation. Therefore, having an algorithm to produce estimates of who is potentially eligible using VHA data is one solution for VA to increase access to A&A for Veterans receiving pension. A&A enhanced pension also has the potential to offset some of the challenges that VHA currently faces in providing LTC to Veterans (eg, Veterans living in rural locations, shortages in VHA staffing, variability in access to GEC’s noninstitutional care programs, contracts with providers). ⁴⁵ Therefore, closer integration of VBA and VHA may increase enrollment in A&A and be seen as mutually beneficial for VHA, VBA, and Veterans alike.

4.1. Limitations

There are limitations to our study. First, our analytic cohort was limited to Veterans who were already receiving the pension benefit. There is no information regarding income and net worth in available VA data sets that is required to determine pension eligibility among Veterans who may not be receiving it; however, data from VHA and Medicare claims and enrollment files allow us to approximate the underlying risk, or need, for the A&A enhanced pension benefit among Veterans already receiving pension. Therefore, our findings are not representative of Veterans who may be eligible for pension. Second, we were only able to identify Veterans who were receiving the A&A enhanced pension and are thus unable to determine who had sought out these benefits but had not yet received them. There is a possibility that we are observing differences in approval status rather than initiation of an application for the A&A enhanced pension. There could be variation among VAMCs in accurate completion of these complex enrollment procedures. Third, we are unable to ascertain from where an application originated. This limits our understanding as to whether or not the variation we observe in terms of VAMC and Veteran characteristics reflects differences in VAMC practices, VSO involvement, or the use of third‐party services in initiating and successfully completing the application. Finally, we are limited to the use of VA and Medicare administrative data and were therefore unable to examine some predisposing, enabling and need characteristics that are likely important predictors of receiving A&A (eg, caregiver availability, education, health literacy). Nonetheless, our study is an important first step in understanding factors that are associated with higher or lower probabilities of enrollment in the A&A enhanced pension benefit among current pensioners.

4.2. Conclusion

While substantial attention has been focused on the receipt of benefits by Veterans who are ineligible for them, with appropriate action taken by VA to reduce fraud, ⁴³ much less attention has been devoted to increasing access to benefits by Veterans who are eligible but are not yet receiving them. Our study has uncovered inequities among Veterans receiving pension in access to the A&A enhanced pension benefit by race, marital status, the presence of mental health conditions, Medicaid enrollment, and VAMC affiliation. Future research incorporating providers’ and Veterans’ perspectives is required to better understand and address the reasons for these inequities. However, while provider education and wider dissemination of information about A&A may help to reduce observed inequities, action is required at the system level that will eliminate the possibility of bias in which some eligible pensioners are able to access this enhanced pension benefit and others are not. As an entitlement, guaranteed access to information and application assistance is VA’s responsibility and should not be at the discretion of individual providers or medical centers. Enhanced interoffice coordination between VBA and VHA may be one strategy to reduce inequities in access to the A&A enhanced pension benefit and further move the VA toward a “One VA” system.

CONFLICT OF INTEREST

The authors have no conflicts to disclose.

Supporting information

Supplementary Material

Appendix S1

Thomas KS, Corneau E, Van Houtven C, et al. Inequities in access to VA’S aid and attendance enhanced pension benefit to help Veterans pay for long‐term care. Health Serv Res. 2021;56:389–399. 10.1111/1475-6773.13636