Characterization of the Population Impacted by the 6-Month Recertification Criterion of the Ryan White HIV/AIDS Program’s AIDS Drug Assistance Program in Washington State, 2017–2019

Abstract

Introduction:

AIDS Drug Assistance Programs (ADAPs) provide financial support for medical care for people living with HIV (PLWH) in the US. Federal policy requires that clients recertify for the program every 6 months, which has been described as a barrier to care access. Our objective was to describe the prevalence and factors associated with ADAP disenrollment in Washington State.

Methods:

Between 2017 and 2019, we categorized ADAP clients by the success of their recertification applications as: 1) continuously enrolled, 2) ruled ineligible, or 3) disenrolled if they failed to recertify. We compared individuals who were disenrolled to those who were continuously enrolled by demographic and socioeconomic characteristics and engagement with case management using data from the Washington State HIV Surveillance and Ryan White data systems.

Results:

From 2017 to 2019, 5480 clients were enrolled in ADAP, of whom 1423 (26%) were disenrolled and 984 (18%) were ruled ineligible at least once. Compared to those who were continuously enrolled, disenrolled PLWH were more likely to be Black (unadjusted prevalence ratio (PR) vs White 1.31, 95% CI 1.17–1.46), uninsured (PR vs private insurance 1.24, 95% CI 1.10–1.40), and younger (PR 25–34 vs 35–44 years 1.23 95% CI 1.08–1.41). The median time to return after disenrollment was 12 months (95% CI 8–19 months).

Conclusions:

Disenrollment after failure to recertify was the most common reason why PLWH lost ADAP coverage in Washington. ADAP recertification procedures disproportionately impact Black, young, and uninsured PLWH and may contribute to disparities in HIV outcomes.

Introduction

Despite previous progress in reducing HIV incidence, transmission rates of HIV in the United States (US) have not changed significantly in the past 5 years.¹ Increasing viral suppression among people living with HIV (PLWH) is a cornerstone of federal efforts to prevent HIV transmission and maintain the health of PLWH.^2–4 Barriers to viral suppression span system, social, and individual factors.^5,6 The Ryan White HIV/AIDS Program’s (RWHAP) AIDS Drug Assistance Program (ADAP) is intended to remove financial barriers to treatment and is tremendously successful in improving access to HIV care.^7,8 Two prior studies demonstrated that ADAP was associated with higher rates of viral suppression and retention in care among women,^9,10 even among patients with other health insurance.¹¹ Conversely, interruptions in ADAP coverage may diminish access to treatment, and ADAP policies that affect continuity of coverage have potential to pose system-level barriers to viral suppression.

Per federal requirements, recipients of ADAP services must provide documentation of their residence, income, and insurance status every 6 months to maintain enrollment in the program.¹² This policy is intended to ensure that the HRSA RWHAP is the payor of last resort by verifying that an individual continues to meet the recipient eligibility requirements.¹² The implementation of this requirement is controlled by the individual ADAPs and practices vary between jurisdictions. Jurisdictions have also been given the ability to allow clients to renew eligibility once a year via patient attestation that their eligibility has not changed.¹²

There is evidence to suggest that the 6-month recertification requirement may be a barrier to HIV care access. Prior qualitative studies have found that maintaining enrollment and accessing benefits is challenging for many people and contributes to ADAP underutilization.^13,14 In a 2018 request for information (RFI) by RWHAP, Ryan White programs, clients, and providers frequently identified the 6-month recertification requirement as a barrier to program effectiveness and a potential disruptor to HIV care for people who are least able to navigate healthcare and payor systems.¹⁵ This suggests that recertification requirements for continuing ADAP coverage may pose a structural barrier to treatment continuity and, to the extent that recertification policies differentially impact subpopulations of PLWH, have the potential to contribute to existing health disparities.

A 12-month recertification requirement has been proposed as an alternative policy, which may decrease the number of eligible PLWH who are disenrolled from ADAP due to failure to complete recertification procedures.¹⁵ The first step to understanding the role of the recertification criterion is to characterize the population who lose coverage due to failure to recertify. To our knowledge, there are no published reports describing how many people leave the ADAP each year and how this population compares to those who remain in ADAP in terms of demographics and program usage. We hypothesized that the populations with lower rates of successful engagement in HIV care, such as racial/ethnic minority populations and people who inject drugs would be more likely to be disenrolled from ADAP.^9,16

In this study we examined the prevalence of disenrollment from the ADAP in Washington State, the length of benefits disruption, and the characteristics of the population that loses coverage. The objectives of this study were to: 1) Estimate the frequency of ADAP disenrollment due to failure to recertify and discontinuation based on ineligibility in Washington State between 2017 and 2019; 2) Estimate the average length of benefits disruption for those who are removed from the program; 3) Describe the population that failed to recertify and compare it demographically to those who remained in the program; and 4) Describe the neighborhood contextual characteristics of the population that failed to recertify and compare them to those who remained in the program.

Methods

Study Population and Setting

We conducted a series of cross-sectional analyses of all PLWH who resided in Washington State and were enrolled in the Washington ADAP between 1/1/2017 and 12/31/2019. We divided this time period into six-month periods to mirror the 6-month recertification cycle. If a client died or was reported to live at an address outside of Washington State during a 6-month analysis period, they were excluded from analysis for that period.

Washington State’s ADAP enrolled 4273 PLWH and served 3569 PLWH in 2017, which represents 34% and 29% of the state population of PLWH respectively.^16,17 In comparison, 20% of PLWH nationwide used an ADAP service in 2017.¹⁷ Washington State is one of 12 states that offers an open formulary for some of its clients, meaning that these clients can use the assistance to purchase drugs of any type. It is one of 39 states that pays for insurance premiums and one of 43 states that pays for other, non-HIV medical costs. The annual cost per client served in 2017 was $3,459, which is below the national average of $8,554.¹⁷

Like all ADAPs, eligibility for Washington’s ADAP is primarily based on income.^17,18 All PLWH who have an income between 139% and 425% of the Federal Poverty Level (FPL) are eligible for the program ($17,608 to $54,230 for a household size of 1 in 2020)¹⁹. PLWH who are below 139% FPL are typically eligible for the state’s expanded Medicaid program which covers healthcare and drug-related costs. If a person’s income is below 139% FPL and they are ineligible for the expanded Medicaid program (e.g. certain non-citizens) or have limited benefits as dual Medicare-Medicaid clients, then they are eligible for Washington ADAP.

As required by RWHAP, enrollment in Washington State’s ADAP requires a person to complete an application and provide proof of Washington residency (driver’s license, voter registration, utility bill, lease, etc.), proof of income, documentation of insurance (or attestation of lack of insurance), and documentation of their HIV diagnoses from a healthcare provider.²⁰ To reenroll, clients must submit the same information except for documentation for their HIV diagnosis every 6 months. If a client fails to recertify, the client will be disenrolled from ADAP. ADAP staff will send letters for two months to try to reengage the client. Clients who have been disenrolled are permitted to reenroll once a year by submitting a form attesting continued eligibility, which does not require additional documentation if the client’s information has not changed.

Data Sources and Definitions

We extracted the number of people enrolled in ADAP semiannually from the Washington State Department of Health’s Ryan White Administrative database from 2017 to 2019. For each ADAP client enrolled, we ascertained the result of their recertification for each 6-month period and classified them as 1) continuously enrolled (i.e. the client successfully recertified), 2) disenrolled (i.e. the client failed to recertify), or 3) ineligible (i.e. the client submitted recertification information but did not meet the requirements for the program) for that 6-month period. Clients who newly enrolled in the program during a 6-month period were not categorized, as they did not have an opportunity to recertify. On the first time a client was categorized as disenrolled or ineligible, we calculated the time between their removal from the program and their reenrollment, if it occurred.

We characterized client demographics using multiple data sources. Age (<25, 25–35, 35–44, 45–54, 55–64, >65)¹⁶, race/ethnicity (Black, Hispanic, White, multiple/other), sex at birth (male, female), HIV acquisition risk (male sex with male, injection drug use, male sex with male and injection drug use, heterosexual contact, and other), time since diagnosis (<1 year, between 1–5 years, >5 years), and region (Western Washington, Eastern Washington or Seattle/King County) were extracted from the Washington State HIV surveillance system. We extracted income 0–135% FPL, 135–250% FPL, 250–425% FPL), insurance type (private, dual Medicare-Medicaid, other public insurance, uninsured), and receipt of case management services in the prior six months from the WA State Ryan White Database. Case managers in Washington State are tasked with helping eligible clients maintain enrollment in ADAP, and ADAP clients who use case management may be less likely to be disenroll. We considered characterizing the population in terms of gender instead of sex at birth, but the number of transgender individuals (N=30 and 3, respectively) presented concerns about the quality of gender data and risk of individual identifiability.

We geocoded the residence of all ADAP participants and described each client’s neighborhood socio-economic status (SES) via census tract measures of income (median household income, median home price, and percent of residents below federal poverty level), educational attainment (percent of residents with less than a high school diploma in census tract of residence) and unemployment (percent of residents unemployed in census tract of residence) using data from the 2015–2019 American Community Survey.^21,22 We chose these particular SES indicators to capture the three common dimensions of SES: education, income, and occupation.²³ SES is related to engagement in HIV care on both an individual and neighborhood level.^24,25 While individual SES has a clear connection to the ability to access care, neighborhood SES has been shown to have an independent effect on health beyond SES on the individual level.^26,27 We also described census tract eviction rate (percent of occupied renting households evicted per year) using data from the Princeton Eviction Lab and classified each client’s census tract of residence as rural or urban based on USDA Rural-Urban Commuting Area Codes (Non-rural 1–3; rural 4–9).^28,21,29 We categorized continuous census-tract measures into quartiles defined by the distribution of census tracts in entire state of Washington. For time-varying characteristics (residence, time since diagnosis, age), we selected the value corresponding to a person’s earliest disenrollment. For individuals continuously enrolled in ADAP, we used the value corresponding to their first 6-month recertification opportunity.

Analysis

We calculated the total number and proportion of clients who were continuously enrolled, disenrolled, or ruled ineligible for each 6-month period and overall. Since a client’s motivation to reenroll may change over the course of a year as insurance benefits change, we compared the proportion who disenrolled in the first half of a year to the proportion who disenrolled during the second half of a year with a rate ratio and 95% confidence interval (CL).

For those who were disenrolled or ruled ineligible, we calculated the proportion who were disenrolled more than once and the proportion who returned to the program within 1 and 3 months after their first removal from the program. We used the Kaplan-Meier method to estimate the median time to return and 95% confidence intervals and compared the median between those who were disenrolled and those who were ruled ineligible using a log-rank test.

We characterized the demographics of clients who were disenrolled by comparing the number and proportion of each attribute between those who were ever disenrolled from the program and those who never left the program (both categories excluded those who were ruled ineligible at any point). For individual demographic variables, we calculated prevalence ratios (PR) and 95% CI using a Poisson model with robust standard error estimates. For census-tract measures, we calculated prevalence ratios using multi-level Poisson models with random effects for census tract adjusted for region. We used likelihood ratio test to assess the significance of each census tract variable in its entirety. Finally, we calculated adjusted prevalence ratios using a multi-level model incorporating both the client demographics and census tract characteristics.

As a sensitivity analysis, we repeated this comparison excluding clients who were disenrolled and did not return within 6 months, as these clients may be more likely to have access to care elsewhere or have moved away from Washington State. All demographic data was complete for clients of ADAP apart from county of residence, which contained a small number of missing values from invalid addresses. We did not perform any correction for missing data. All analyses were performed for program evaluation and received exemption from review from the University of Washington IRB.

Results

Frequency of Disenrollment

From 2017 to 2019, a total of 5480 clients were enrolled in ADAP, of whom 3304 (60%) received premium payment assistance to pay for their insurance. During this time, 1423 (26%) clients were disenrolled due to a failure to recertify and 984 (18%) clients were ruled ineligible, representing averages of 6.8% and 4.7% of clients for each recertification opportunity (range 6.0–8.2% and 3.6–6.4%, Table 1). Of the clients who disenrolled, 15% did so more than once in the 2-year study period. There was no significant difference between the proportion of clients who disenrolled in the second half of the year as compared to the first half (PR: 1.08, 95% CI 0.98–1.18).

Table 1:

Number and Outcome of Recertification Attempts by AIDS Drug Assistance Clients by 6-Month Period, Washington State 2017–2019^a

Beginning of Period	Recertification Attempts	Reenrolled	Disenrolled	Ineligible
1/1/2017	3518	3088 (87.8%)	291 (8.3%)	139 (4.5%)
7/1/2017	3587	3123 (87.1%)	323 (9%)	141 (4.5%)
1/1/2018	3623	3229 (89.1%)	244 (6.7%)	150 (4.6%)
7/1/2018	3808	3308 (86.9%)	270 (7.1%)	230 (7.0%)
1/1/2019	3891	3441 (88.4%)	263 (6.8%)	187 (5.4%)
7/1/2019	4040	3483 (86.2%)	297 (7.4%)	260 (7.5%)
Total	22467	19672 (87.6%)	1688 (7.5%)	1107 (5.6%)

^a.Clients were categorized as enrolled if they successfully recertified, disenrolled if they failed to recertify, and ineligible if they recertified but did not meet the income or insurance requirements for the program.

Time away from Services

Of the 1424 clients who were disenrolled during the study period, 29% (95% CI 26–29%) reenrolled within a month of disenrollment and 38% (95% CI 36–38%) returned within 3 months. The median time to return was 12 months (95% CI 8–19 months). Those who were ruled ineligible stayed out of the program for longer; 10% retuned within a month (95% CI 8–10%) and 21% returned within 3 months (95% CI 19–22%). The median time to return for those ruled ineligible was 38 months (95% Cl lower bound 23 months, upper bound not estimable due to censoring; log-rank p-value compared to those disenrolled <0.01.

Demographic Comparison

Compared to those who were continuously enrolled in ADAP, those who were disenrolled from ADAP at least once were more likely to be Black (PR vs White: 1.31, 95% CI 1.17–1.46), younger, and uninsured (PR vs private insurance: 1.26, 95% CI 1.10–1.40). Clients who were disenrolled were less likely to have recently received case management services (PR 0.63, 95% CI 0.58–0.69). They were less likely to have dual Medicare-Medicaid (PR vs private insurance: 0.48 95% CI 0.38–0.60) or other public insurance (PR vs private insurance 0.58 (0.52–0.65) and have an income below 139% FPL (PR vs >425% FPL: 0.73 95% CI 0.65–0.81, Table 2). After adjustment for all of the individual and neighborhood characteristics, there was attenuation of the prevalence ratio for Black race (PR vs White: 1.14, 95% CI 0.99–1.32). The results of our analyses did not change when we removed clients who were disenrolled for more than 6-months.

Table 2:

Demographic Characteristics of ADAP Clients Who Were Disenrolled Versus Clients Who Were Continuously Enrolled, Washington State 2017–2019^a

Variableb	Value	Disenrolled	Continuously Enrolled	Prevalence Ratio (95% CI)	Adjusted Prevalence Ratio (95% CI)c
Total		1390	3169
Race	Hispanic	275 (20%)	683 (22%)	1.00 (0.88–1.12)	0.80 (0.69–0.92)
	Black	308 (22%)	509 (16%)	1.31 (1.17–1.46)	1.14 (0.99–1.32)
	White	667 (48%)	1646 (52%)	Reference	Reference
	Other/Multiple	141 (10%)	331 (10%)	1.04 (0.89–1.21)	0.94 (0.81–1.08)
HIV Acquisition Risk Category	MSM	837 (60%)	1920 (61%)	Reference	Reference
	IDU	65 (5%)	132 (4%)	1.09 (0.88–1.34)	1.34 (1.06–1.68)
	MSM+IDU	128 (9%)	301 (9%)	0.98 (0.84–1.15)	1.12 (0.97–1.31)
	Heterosexual Contact	127 (9%)	350 (11%)	0.88 (0.75–1.03)	0.95 (0.76–1.19)
	Other	234 (17%)	466 (15%)	1.10 (0.98–1.24)	1.08 (0.91–1.29)
Age	<25	23 (2%)	39 (1%)	1.02 (0.73–1.42)	1.05 (0.73–1.51)
	25–34	220 (16%)	269 (8%)	1.23 (1.08–1.41)	1.18 (1.02–1.37)
	35–44	300 (22%)	522 (16%)	Reference	Reference
	45–54	383 (28%)	813 (26%)	0.88 (0.78–0.99)	0.91 (0.8–1.04)
	55–64	331 (24%)	991 (31%)	0.69 (0.6–0.78)	0.78 (0.68–0.9)
	65+	133 (10%)	535 (17%)	0.55 (0.46–0.65)	0.68 (0.55–0.83)
Birth Sex	Female	220 (16%)	525 (17%)	Reference	Reference
	Male	1170 (84%)	2644 (83%)	1.04 (0.92–1.17)	1.18 (1.00–1.40)
Insurance	Dual Medicare-Medicaid	70 (5%)	331 (10%)	0.48 (0.38–0.60)	0.50 (0.39–0.64)
	Other Public Insurance	330 (24%)	1229 (39%)	0.58 (0.52–0.65)	0.64 (0.55–0.75)
	Private Insurance	778 (56%)	1360 (43%)	Reference	Reference
	Uninsured	212 (15%)	249 (8%)	1.26 (1.1–1.4)	1.31 (1.15–1.49)
Geography	Eastern WA	154 (11%)	480 (15%)	0.87 (0.75–1.02)	0.84 (0.71–0.99)
	King County	769 (57%)	1531 (49%)	1.20 (1.09–1.33)	1.03 (0.89–1.2)
	Western WA (Not King County)	429 (32%)	1114 (36%)	Reference	Reference
Time from Diagnosis	<1 year	143 (10%)	364 (11%)	0.92 (0.79–1.08)	0.59 (0.49–0.73)
	1–5 years	243 (17%)	357 (11%)	1.40 (1.25–1.56)	1.06 (0.93–1.2)
	>5 years	1004 (72%)	2448 (77%)	Reference	Reference
Income	0–135% FPL	461 (33%)	1236 (39%)	0.73 (0.65–0.81)	1.06 (0.92–1.21)
	135–250% FPL	465 (33%)	1131 (36%)	0.79 (0.71–0.88)	0.96 (0.86–1.08)
	250–425% FPL	463 (33%)	802 (25%)	Reference	Reference
Received Case Management in past 6M	Yes	961 (69%)	2591 (82%)	0.63 (0.58–0.69)	0.69 (0.63–0.77)
	No	430 (31%)	579 (18%)	Reference	Reference

Abbreviations: MSM= Male-Male Sexual Contact, IDU = Injection Drug Use, FPL = Federal Poverty Level, CI = Confidence Interval

^a.Clients who failed to recertify one or more times between 2017 to 2019 were categorized as disenrolled. Clients who never failed to recertify were categorized as continuously enrolled. Those who were ruled ineligible at any point were excluded. Time varying variables (age, insurance, geography, income, and time from diagnosis) were measured at a client’s first disenrollment for those who disenrolled and first recertification opportunity for those who were continuously enrolled.

^b.Reference groups were selected to maximize sample size and interpretability.

^c.Adjusted prevalence ratio from a mixed Poisson model with a random effects term for census tract adjusted for all demographic and census-tract variables

Neighborhood Characteristics

Ninety-three percent of clients had addresses that could be assigned to a census tract. After adjustment for region, the census tracts where people who were disenrolled lived were not statistically different from the tracts of people who never disenrolled in any of our measures. The full results, along with adjusted prevalence estimates can be found in Table 3. The results of our analyses did not change when we removed clients who were disenrolled for more than 6-months.

Table 3:

Characteristics of Census Tract of Residence of ADAP Clients Who Were Disenrolled Versus Clients Who Were Continuously Enrolled, Washington State 2017–2019^a, 21, 29

Variableb	Value	DIsenrolled (n=1276)	Continuously Enrolled (n=2975)	Prevalence Ratioc	P-valued	Adjusted Prevalence Ratioc	P-valued
Rural vs Urbane	Urban	1206 (95%)	2808 (94%)	0.89 (0.72–1.1)	0.30	1.07 (0.84–1.38)	0.58
	Rural	70 (5%)	167 (6%)	Reference		Reference
% with <HS Degree	Q1 (≤3.9%)	331 (26%)	648 (22%)	1.05 (0.93–1.18)	0.06	1.05 (0.90–1.23)	0.03
	Q2 (3.9–7.1%)	227 (18%)	601 (20%)	0.89 (0.78–1.02)		0.89 (0.76–1.03)
	Q3 (>7.1–11.3%)	278 (22%)	722 (24%)	0.91 (0.80–1.03)		0.89 (0.78–1.02)
	Q4 (>11.3%)	440 (34%)	1004 (34%)	Reference		Reference
% below Poverty Level	Q1 (≤5.7%)	217 (17%)	438 (15%)	1.11 (0.97–1.27)	0.39	1.00 (0.83–1.20)	0.85
	Q2 (>5.7–9.3%)	279 (22%)	609 (20%)	1.06 (0.94–1.21)		0.97 (0.83–1.12)
	Q3 >9.3–14.7%)	317 (25%)	767 (26%)	1.01 (0.89–1.13)		0.95 (0.83–1.09)
	Q4 (>14.7%)	463 (36%)	1161 (39%)	Reference		Reference
% of Homes Rented	Q1 (≤18.5%)	97 (7.0%)	331 (8.4%)	1.02 (0.85–1.22)	0.44	0.90 (0.73–1.12)	0.81
	Q2 (>18.5–31.6%)	180 (12.9%)	521 (13.3%)	1.06 (0.92–1.22)		0.98 (0.83–1.16)
	Q3 >31.6–48.3%)	332 (23.9%)	915 (23.3%)	1.10 (0.98–1.22)		0.98 (0.87–1.12)
	Q4 (>48.3)	665 (47.8%)	1940 (49.4%)	Reference		Reference
% Unemployed	Q1 (≤1.9%)	292 (23%)	656 (22%)	1.06 (0.93–1.21)	0.21	1.05 (0.92–1.21)	029
	Q2 (>1.9–2.9%)	280 (22%)	573 (19%)	1.14 (1.00–1.3)		1.14 (1.00–1.31)
	Q3 >2.9–4.0%)	344 (27%)	838 (28%)	1.01 (0.90–1.15)		1.03 (0.91–1.17)
	Q4 (>4.0)	360 (28%)	908 (31%)	Reference		Reference
% of Lessees Evicted	Q1 (≤0.2)	284 (23%)	616 (22%)	1.01 (0.86–1.18)	0.43	1.00 (0.82–1.23)	0.81
	Q2 (>0.2–0.61)	369 (30%)	818 (29%)	1.02 (0.89–1.18)		1.02 (0.86–1.22)
	Q3 >0.61–1.15)	319 (26%)	708 (25%)	1.11 (0.96–1.27)		1.09 (0.95–1.25)
	Q4 (>1.15)	263 (21%)	722 (25%)	Reference		Reference
Median Home Value	Q1 (≤232)	225 (18%)	620 (21%)	1.04 (0.87–1.25)	0.91	1.13 (0.87–1.46)	0.70
	Q2 (>233–318)	256 (20%)	635 (22%)	1.04 (0.89–1.21)		1.02 (0.83–1.26)
	Q3 >319–435)	328 (26%)	751 (25%)	0.99 (0.88–1.12)		1.02 (0.87–1.19)
	Q4 (>436)	458 (36%)	941 (32%)	Reference		Reference
Median Household Income	Q1 (≤55)	331 (26%)	937 (31%)	0.84 (0.73–0.97)	0.07	0.83 (0.67–1.02)	0.19
	Q2 (>56–72)	380 (30%)	833 (28%)	0.95 (0.83–1.08)		0.97 (0.82–1.15)
	Q3 >73–94)	302 (24%)	714 (24%)	0.88 (0.77–1.01)		0.98 (0.84–1.13)
	Q4 (>95)	263 (21%)	491 (17%)	Reference		Reference

^a.Clients who failed to recertify one or more times between 2017 to 2019 were categorized as disenrolled. Clients who never failed to recertify were categorized as continuously enrolled. Those who were ruled ineligible at any point were excluded. Census tract of residence was ascertained at a client’s first disenrollment for those who disenrolled and first recertification opportunity for those who were continuously enrolled.

^b.Reference groups were selected to maximize sample size and interpretability.

^c.Prevalence ratios and confidence interval from multi-level Poisson model with random effects for census tract adjusted for region (Eastern Washington, King County, Western Washington other than King County). Adjusted prevalence ratios from a mixed Poisson model with a random effects term for census tract adjusted for all demographic and census-tract variables.

^d.P-value from type 3 test of fixed effects.

^e.Rural census tracts defined as those with a USDA Rural-Urban Commuting Code of 1, 2, or 3. All other tracts were classified as urban

Discussion

Our study found that ADAP disenrollment is common. In our two-year study period, 26% of clients were disenrolled from ADAP due to failure to recertify, which is 44% more clients than were found to be ineligible (26% vs 18%). We also found that the population that was disenrolled was disproportionately young, Black, uninsured, and not engaged with case management services. We did not find any association between disenrollment and census tract SES.

The number of disenrollments suggests that there may be a large proportion of clients who are eligible for ADAP in Washington State but are denied services due to a program policy barrier. This is consistent with research from Alabama and North Carolina, which found that the ADAP recertification requirement was a barrier to ADAP utilization.^13,30 This reduction in ADAP clients serves the 6-month recertification policy’s goal of ensuring that RWHAP’s limited grant resources are spent on eligible clients, but the number of disenrollments relative to ineligibility rulings means that the policy is more often removing clients who do not complete the recertification procedure than clients who are confirmed ineligible for ADAP.

The demographics of PLWH who are disenrolled from ADAP reveal an overlap with populations that have lower viral suppression rates in Washington State and nationwide. Black PLWH (78% viral suppression) and young PLWH (74% viral suppression among those between 25 and 34) have the lowest viral suppression rates in the state (82% viral suppression overall).¹⁶ The disproportionate impact of ADAP disenrollment on Black PLWH is noteworthy and may represent a modifiable barrier for a population that has not benefitted from statewide increases in viral suppression in Washington.¹⁶ The attenuation of this association after adjustment for demographic factors does not diminish the importance of the observed disparity but rather highlights the interconnectedness of race and other social determinants of health.

Our finding that disenrollment was not associated with neighborhood SES may indicate that neighborhood SES is not related to a person’s ability to complete the 6-month recertification, or it may represent limitations in our measurement of neighborhood SES. Our measures are, at best, proximal measures of SES, and census tracts do not define perfect boundaries around a person’s social environment. Associations between neighborhood SES and disenrollment may be too subtle to measure with census tract measures.

Our study has several limitations. It is possible that some clients who do not recertify choose not to do so because they know that they are ineligible or no longer need ADAP services. If the latter were common, however, we might expect to see an increase in disenrollment in the second half of the year when privately insured individuals reach their policies’ deductible, which we did not. We did not have data available to describe homelessness, which may be an important factor driving the length of disenrollment as Washington uses letters to try to reengage clients. Our study also did not examine HIV clinical outcomes in relation to disenrollment. This is a key area for future study; due to the length of benefits disruption we observed, disenrollment could lead to disruptions to antiretroviral medication if former clients are unable to access care through other mechanisms. Finally, this study was conducted in the years just prior to the COVID pandemic, which led to substantial changes in HIV care provision, medication dispensation, and ADAP policies. Many states rapidly implemented more flexible policies for continuing ADAP benefits with the intention of preventing disruptions on coverage. The impact and durability of these policies is not yet fully known.^31,32

The implications of this study for national policy depend on whether disenrollment in Washington is representative of disenrollment in other programs. Washington State is one of 59 different ADAP jurisdictions, and the burden of recertification may stem from Washington’s recertification processes rather than the requirements set by federal policy. However, the standard recertification procedures in Washington State represent only the minimum federal requirements, which is not the case in all other jurisdictions.^12,31 Although Washington State does not offer self-attestation to current ADAP clients, it does not seem likely that recertification in Washington is unusually difficult compared to other ADAP’s; self-attestation still requires engagement on the part of the client, and does not completely eliminate the burden of reenrollment. Regardless of the size of the barrier that recertification represents in each jurisdiction, moving to a 12-month recertification interval, as suggested in response to HRSA’s 2018 RFI, would reduce the frequency of disruption.¹⁵

In summary, our study found that the 6-month ADAP recertification policy in Washington State excludes a large number of people from receiving ADAP services on the basis of their ability or willingness to complete the recertification process. This policy disproportionately affects populations with the lowest probability of viral suppression and may contribute to the disparities in HIV incidence and outcomes in the state. While the 6-month recertification requirement is effective in ensuring that RWHAP’s limited grant resources are spent on eligible clients, the impact of this policy in Washington state is unjust. If this finding is consistent in other areas, jurisdictional approaches to how the recertification requirement is operationalized should be re-examined and alternatives to the 6-month recertification requirement should be considered.