|Home | About | Journals | Submit | Contact Us | Français|
Background.Before implementation of the Patient Protection and Affordable Care Act (ACA) in 2014, 100 000 persons living with human immunodeficiency virus (HIV) (PLWH) lacked healthcare coverage and relied on a safety net of Ryan White HIV/AIDS Program support, local charities, or uncompensated care (RWHAP/Uncomp) to cover visits to HIV providers. We compared HIV provider coverage before (2011–2013) versus after (first half of 2014) ACA implementation among a total of 28 374 PLWH followed up in 4 sites in Medicaid expansion states (California, Oregon, and Maryland), 4 in a state (New York) that expanded Medicaid in 2001, and 2 in nonexpansion states (Texas and Florida).
Methods.Multivariate multinomial logistic models were used to assess changes in RWHAP/Uncomp, Medicaid, and private insurance coverage, using Medicare as a referent.
Results.In expansion state sites, RWHAP/Uncomp coverage decreased (unadjusted, 28% before and 13% after ACA; adjusted relative risk ratio [ARRR], 0.44; 95% confidence interval [CI], .40–.48). Medicaid coverage increased (23% and 38%; ARRR, 1.82; 95% CI, 1.70–1.94), and private coverage was unchanged (21% and 19%; 0.96; .89–1.03). In New York sites, both RWHAP/Uncomp (20% and 19%) and Medicaid (50% and 50%) coverage were unchanged, while private coverage decreased (13% and 12%; ARRR, 0.86; 95% CI, .80–.92). In nonexpansion state sites, RWHAP/Uncomp (57% and 52%) and Medicaid (18% and 18%) coverage were unchanged, while private coverage increased (4% and 7%; ARRR, 1.79; 95% CI, 1.62–1.99).
Conclusions.In expansion state sites, half of PLWH relying on RWHAP/Uncomp coverage shifted to Medicaid, while in New York and nonexpansion state sites, reliance on RWHAP/Uncomp remained constant. In the first half of 2014, the ACA did not eliminate the need for RWHAP safety net provider visit coverage.
(See the HIV/AIDS Major Article by McManus et al on pages 396–403, and the Editorial Commentary by Eaton and Mugavero on pages 404–6.)
For persons living with human immunodeficiency virus (HIV) (PLWH), being engaged in care with consistent visits to an HIV medical provider promotes medication adherence, prevents HIV complications, and decreases the risk of HIV transmission [1–7]. However, routine ambulatory HIV care is expensive, typically exceeding $15 000 annually [8, 9]. In the United States, low-income, uninsured adult PLWH who do not qualify for Medicaid rely on safety nets such as the federally funded Ryan White HIV/AIDS Program (RWHAP) to cover outpatient HIV provider visits. Before 2014, a quarter of adult PLWH who were engaged in care, 100 000 persons annually, were supported entirely by the RWHAP or by charitable local programs for outpatient HIV provider visits or received uncompensated care (collectively, RWHAP/Uncomp) [10–12].
Before 2014, Medicaid covered 40% of adult PLWH engaged in care [10–12]. Qualifying for Medicaid required demonstrating (and annually redemonstrating) medical need. Most childless adult PLWH who qualified for Medicaid did so because their infection had progressed to the point of AIDS . In addition to Medicaid, another 10%–20% of nonelderly PLWH were covered by Medicare (sometimes dually covered with Medicaid), again with a requirement for an AIDS diagnosis or other disabling condition [5, 10, 11]. Only 25% of PLWH had private insurance [11, 12].
In January 2014, a total of 26 states expanded Medicaid by eliminating the medical need requirement and increasing the income threshold under provisions of the Patient Protection and Affordable Care Act (ACA) . A few states, such as New York, had eliminated the medical need requirement before implementation of the ACA [10, 11, 15, 16]. All states also eliminated private insurance exclusions for preexisting conditions and began providing access to subsidized, private qualified health plan (QHP) insurance.
It is uncertain what fraction of PLWH relying solely on RWHAP/Uncomp to cover HIV provider visits acquired Medicaid or a private insurance plan after implementation of the ACA. Two-thirds of persons with RWHAP/Uncomp coverage have income ≤138% of the federal poverty limit, the financial threshold for Medicaid eligibility in expansion states . The remainder have incomes at or below the 400% limit needed to receive subsidies for QHP premiums. We examined patterns of coverage for outpatient HIV provider visits from January 2011 to June 2014 in 10 HIV Research Network (HIVRN) sites, stratified by whether and when the state of location expanded Medicaid.
The HIVRN is a consortium of 12 adult and 6 pediatric HIV care sites . Each site prospectively collects demographic, utilization, and clinical information from administrative records and structured chart reviews and submits the deidentified data to a coordinating center for combination into a uniform database. Ten adult care sites located in 6 states (California [2 sites], Oregon, Texas, Florida, Maryland, and New York [4 sites]) had healthcare coverage data for HIV provider (physician, nurse practitioner, and/or physician assistant) visits for January 2011 to June 2014. One New York site lacked 2011 and 2012 data, and 1 California site lacked 2011 data; these sites were included for the years for which they had data. Institutional review boards at each site and at the data coordinating center at Johns Hopkins University approved the study.
All patients ≥18 years-old who had ≥1 person-year (PY) of active care engagement from January 2011 to June 2014 were included. Active engagement was defined as having ≥1 HIV provider visit and 1 CD4 cell count result within the calendar year. Because the standard of care for many patients is to perform 1 CD4 cell count annually , a visit but not a CD4 result was required for inclusion in January–June 2014. We excluded 1124 PYs (1.4% of initial PYs) because all HIV provider visits within the year had unknown coverage type; this eliminated 1035 patients (3.5%) from the analysis.
The healthcare coverage for each PY was defined as the primary form of coverage recorded at registration for the last HIV provider visit of the year, or, in 2014, the last visit before 1 July. If the primary form of coverage for the last visit of the year was unknown (1557 PYs; 2% of PYs in the final analytic data set), then the coverage from the most recent prior HIV provider visit with known coverage was counted. Coverage was classified as private, Medicare (including dually enrolled Medicare/Medicaid), Medicaid, or RWHAP/Uncomp.
RWHAP support for HIV provider visits consisted of 3 general forms: grant money used by clinics (occurring at all sites), New York AIDS Drug Assistance Program Plus health coverage (at all New York sites), and “insurance continuation” programs in which the RWHAP paid private insurance premiums on behalf of low-income PLWH (at all New York sites, Oregon, and Maryland). Because this analysis focused on the source of payment rather than how coverage was administered, insurance continuation was coded as RWHAP/Uncomp. We determined the relative contribution of each of these forms of RWHAP support to the RWHAP/Uncomp category, except in the case of New York sites because these sites did not consistently distinguish between forms of RWHAP support in compiling their data. Thus, for New York sites, we described RWHAP support as “RWHAP undifferentiated.” Furthermore, whereas the Oregon and Maryland sites were able to use case management records to distinguish insurance continuation from traditional (employer- and/or beneficiary-purchased) private insurance, the New York sites could not do so. Thus, some RWHAP-funded coverage at New York sites was misclassified as private insurance.
An alternative to RWHAP support existed at both California sites in the form of county-administered/Centers for Medicare and Medicaid Services (CMS) funded demonstration projects established in 2010 to assess the future impact of Medicaid expansion on low-income, uninsured county residents [19, 20]. We categorized this coverage as RWHAP/Uncomp because these PLWH would otherwise have had no option except RWHAP/Uncomp. Uncompensated provider visits were those assessed by sites as not receiving payment after accounting for any RWHAP support for provider salaries.
Sites were stratified according to the Medicaid expansion status of the state of location. “Expansion sites” (4 total) were in 3 states (California, Oregon, and Maryland) that expanded Medicaid in 2014. “Previously expanded sites” (4 total) were in New York, which expanded Medicaid by eliminating the medical need requirement in 2001 for persons with incomes up to 100% of the federal poverty limit . (New York did technically expand Medicaid eligibility in 2014 by increasing this threshold up to 138% federal poverty limit.) “Nonexpansion sites” (2 total) were in Texas and Florida, states that did not expand Medicaid in 2014.
The primary predictor variable was calendar time, classified by calendar year. Other variables included age (assessed on 1 July of each calendar year), sex, race/ethnicity, and HIV risk factors. These variables were categorized as shown in Table Table11.
Analyses were carried out on the PY level. The primary objective was to evaluate changes in the pattern of HIV provider visit coverage before and after 1 January 2014. Based on previous work  showing stability in coverage patterns 2006–2012, we suspected coverage patterns 2011–2013 would be generally stable. We confirmed this stability graphically (Figure (Figure1)1) and then combined the years 2011–2013 into a single category for quantitative analyses. Our data encompassed only the first 6 months of 2014. For analytic purposes, we counted this 6-month period as a full PY while acknowledging that subsequent coverage changes could have happened.
We used multinomial logistic regression to assess the likelihood of RWHAP/Uncomp, Medicaid, and private coverage in the first 6 months of 2014, compared with 2011–2013, relative to Medicare coverage. We used Medicare as the base comparison category because the ACA made no changes to Medicare eligibility and was therefore expected to have the least potential impact on this category. A preliminary model revealed significant interaction between state expansion status and calendar time. Subsequent analyses were therefore stratified by expansion status. To control for potential confounding, we included demographics, HIV transmission risk factors, and an indicator variable for individual HIVRN site in multivariate models. Because the same person could contribute multiple years of coverage information, all models used robust standard errors clustered on persons. Analyses were performed using Stata 12.1 software (StataCorp )  with an α value of .05.
In total, 28 374 patients (7851 expansion, 10 183 previously expanded, and 10 340 nonexpansion) contributed 76 838 PYs of observation time (Table (Table1).1). The median age in 2011 in the expansion sites (46 years; interquartile range, 38–53 years) was similar to that in the previously expanded sites (46 years; 37–53 years), but higher than in the nonexpansion sites (42 years; 32–50 years). Subjects in the expansion sites were more likely to be male and white and to have male-to-male sex or injection drug use as the HIV risk factor than patients in previously expanded or nonexpansion sites.
In the 4 expansion sites, RWHAP/Uncomp coverage decreased from 28% of PYs in 2011–2013 to 13% in 2014, whereas Medicaid coverage increased from 23% to 38% (Figure (Figure1).1). Private coverage (decrease from 21% to 19%) and Medicare (increase from 28% to 30%) were relatively stable. Relative to the change in Medicare and after adjustment for age, sex, race, HIV risk factors, and clinic site (Table (Table2),2), RWHAP/Uncomp decreased by 56% (adjusted relative risk ratio [ARRR] for 2014 vs 2011–2013, 0.44; 95% confidence interval [CI], .40–.48) and Medicaid increased by 82% (1.82; 1.70–1.94). Relative to the change in Medicare coverage, private coverage was not significantly changed in 2014 (ARRR, 0.96; 95% CI, .89–1.03). The pattern of a decrease in RWHAP/Uncomp coverage accompanied by an increase in Medicaid coverage was seen uniformly in each of the individual expansion sites, with ARRRs reaching significance in all cases except for the Medicaid increase in the smallest site (Figure (Figure22).
In the 4 previously expanded sites, RWHAP/Uncomp coverage was generally stable in unadjusted (20% in 2011–2013 and 19% in 2014) and in adjusted (ARRR, 1.00; 95% CI, .94–1.06) analyses (Figure (Figure11 and Table Table2).2). Medicaid was the leading source of coverage (50%) in 2011–2013 and its rates were stable in 2014 in both unadjusted and adjusted analyses. Private coverage rates were generally stable in unadjusted analysis (decrease from 13% to 12%), but in adjusted analysis private coverage decreased (ARRR, 0.86; 95% CI, .80–.92) relative to the change in Medicare. Among the individual previously expanded sites, there was no consistent pattern of change in any type of coverage in 2014 relative to Medicare, and no individual site had any significant change in RWHAP/Uncomp coverage (Figure (Figure22).
In the 2 nonexpansion sites, RWHAP/Uncomp coverage accounted for 57% of coverage in 2011–2013 and 52% in 2014 and was unchanged between time periods in adjusted analysis. Medicaid accounted for 18% of coverage in both time periods and was also unchanged in adjusted analysis. Private insurance accounted for a small portion of coverage in all years but with a large relative increase from 4% in 2011–2013% to 7% in 2014 (ARRR, 1.79; 95% CI, 1.62–1.99). Although both nonexpansion sites had significant increases in private insurance in 2014, they both had ≤10% private insurance coverage in all years (Figure (Figure22).
Regardless of state Medicaid expansion status, Hispanic ethnicity was associated with RWHAP/Uncomp relative to Medicare (Table (Table2).2). Older age was associated with having Medicare compared with all other forms of coverage, and the HIV risk factors of heterosexual sex and injection drug use were both associated with Medicaid and inversely associated with private coverage.
Across all years and all sites, the 26 279 PYs of RWHAP/Uncomp time was 86% RWHAP funded, 9% uncompensated care, 3% funded through CMS demonstration projects, and 2% funded by a county-based charity program. This distribution differed between the expansion status groups (Figure (Figure3).3). Previously expanded sites had 99% RWHAP support with only 1% uncompensated care. Nonexpansion sites had the highest proportion of uncompensated care, 12%. The CMS demonstration projects existed only at the 2 California sites but overall made up 18% of the RWHAP/Uncomp PYs among all 4 expansion sites.
In the first half of 2014, the proportion of RWHAP/Uncomp patients decreased by half in selected HIV clinics in 3 states with Medicaid expansion and changed little or none in clinics in other states. In Medicaid expansion sites, the decline in RWHAP/Uncomp coverage was offset by an increase in Medicaid coverage, while levels of private insurance and Medicare coverage remained stable. In nonexpansion sites, RWHAP/Uncomp status accounted for over 50% of coverage in all years, and appeared to change little, if any, in 2014. The nonexpansion sites witnessed a small absolute increase in private coverage in 2014. These findings are relevant to planning future RWHAP support for PLWH and provide an example of the early impact of the ACA among low-income persons with a serious chronic condition.
The changes in coverage we observed are generally consistent with a National Alliance of State and Territorial AIDS Directors report that described the transition of roughly 12 000 previously uncovered RWHAP clients into Medicaid during January–May 2014 and another 13 000 into QHPs, with 57% of the QHP transitions occurring in nonexpansion states . That report did not examine Medicare or non-QHP private insurance.
Despite expanded Medicaid, 13% of patients in expansion sites and 19% of patients in sites in a previously expanded state remained on RWHAP/Uncomp coverage in the first half of 2014. There may be several reasons for this persistence. First, in the case of the expansion states, additional transition from RWHAP/Uncomp to Medicaid might have occurred in the latter half of 2014. Unlike annual enrollment in QHPs, which was required by 31 March 2014, Medicaid had open enrollment throughout 2014. State and RWHAP policies required RWHAP beneficiaries to apply for Medicaid when their medication coverage (AIDS Drug Assistance Plan) annual reenrollment was due in 2014 [23–26]. In many cases, this would have been after June. RWHAP/Uncomp patients who did not perceive a need for new benefits available through Medicaid may have felt little incentive to apply until their medication coverage expiration date. This rationale, however, does not explain the multiyear stability of RWHAP/Uncomp coverage in the previously expanded state, New York. One factor probably contributing to the persistence of that category in both expansion states and New York is that undocumented immigrants (and some legal immigrants in the United States <5 years) are ineligible for Medicaid or QHPs but are covered by the RWHAP. We do not collect data on immigration status; however, the association of Hispanic ethnicity with RWHAP/Uncomp status supports this supposition. Other contributors may include not knowing how to apply, employment transitions, and changes in income affecting Medicaid eligibility .
In nonexpansion sites, the increase in private insurance from 4% to 7% may signify the effects of a modest amount of enrollment in QHPs. This is speculative, however, because we cannot distinguish QHPs from traditional private plans. In expansion sites, the lack of an increase in private coverage may signify that relatively few patients (theoretically those with incomes >138% but ≤400% of the federal poverty limit) transitioned from RWHAP/Uncomp into a QHP. Again, we cannot be certain without specifically tracking QHPs.
A major implication from this study is the continuing need for RWHAP coverage and other safety nets for provider visits. This need is clearest in nonexpansion states. It is intriguing that our nonexpansion sites not only had little or no transition away from RWHAP/Uncomp but that RWHAP/Uncomp covered half of all provider visits in these sites. Further studies should characterize this need and any shift to QHPs over time. The finding that 12% of the RWHAP/Uncomp category in nonexpansion sites consists of uncompensated care is also intriguing. Clinics and health systems may unpredictably eliminate uncompensated care, which could leave PLWH suddenly without care. In expansion and previously expanded sites, the RWHAP may remain the only option for coverage for immigrants and may be needed at least temporarily for Medicaid or QHP-eligible persons who have so far failed to enroll.
In considering the ongoing need for RWHAP support for provider visits after the ACA, the full extent of RWHAP-funded ambulatory medical services must also be examined. The RWHAP provides extensive “wrap-around” services including extra nursing staff, provider support for longer visit time, adherence counseling, and case management. By improving engagement in care and medication adherence, wrap-around services contribute not only to the health of PLWH but also to HIV prevention through treatment [6, 7, 28–32]. These services are generally not covered by Medicaid, Medicare, or private insurance and are available to all persons who qualify based on income threshold. In 2013, ≥55% of persons with Medicaid coverage for provider visits accessed RWHAP-funded wrap-around medical services (Health Resources and Services Administration [HRSA], HIV/AIDS Bureau, unpublished data). Expanding Medicaid eligibility does not reduce the need for RWHAP funding for wrap-around services.
This study has several limitations, as previously noted. Our sample is limited to 10 large HIV clinics in 6 states. Trends in these states may not represent trends in other states that would be classified in the same Medicaid expansion category, and clinics may not represent trends within a state, especially in states with only a single clinic. However, our sample does include sites from the 4 states (New York, California, Florida, and Texas) with the highest numbers of PLWH, together accounting for almost half of PLWH nationally . Our available data do not extend past 30 June 2014, which may have led to an underestimation of transition away from RWHAP/Uncomp coverage. Finally, we may have missed some transition into or out of RWHAP/Uncomp coverage in the New York sites because of our inability to reliably distinguish RWHAP-funded insurance continuation from traditional (employer and/or beneficiary-purchased) private insurance. Because private insurance was observed to be stable in the New York sites in 2014, we reason there was likely to have been relatively little transition into or out of insurance continuation.
In conclusion, our study provides an early view of change in the pattern of coverage for HIV provider visits after implementation of the ACA. In Medicaid expansion sites, about half of patients relying on RWHAP/Uncomp coverage shifted to Medicaid. Based on the example of the previously expanded sites, expansion sites may see little additional shift from RWHAP/Uncomp to Medicaid. In the 2 nonexpansion sites, more than half of all patients relied on RWHAP/Uncomp coverage for provider visit coverage before 2014, and this did not change in 2014, despite a small absolute increase in private coverage. Our findings suggest the ongoing need for extensive support for HIV provider visit coverage from the RWHAP despite the ACA.
HIVRN members:Participating sites: Alameda County Medical Center, Oakland, California (Howard Edelstein, MD); Children's Hospital of Philadelphia, Pennsylvania (Richard Rutstein, MD); Drexel University, Philadelphia (Jeffrey Jacobson, MD, and Sara Allen, CRNP); Fenway Health, Boston, Massachusetts (Stephen Boswell, MD); Johns Hopkins University, Baltimore, Maryland (K. A. G., R. D. M., and A. L. A.); Montefiore Medical Group (Robert Beil, MD) and Montefiore Medical Center (Uriel Felsen, MD), Bronx, New York; Oregon Health and Science University, Portland (P. T. K.); Parkland Health and Hospital System, Dallas, Texas (A. E. N. and Muhammad Akbar, MD); St Jude's Children's Research Hospital and University of Tennessee, Memphis (Aditya Gaur, MD); St Luke's Roosevelt Hospital Center, New York, New York (Judith Aberg, MD, and Antonio Urbina, MD); Tampa General Health Care, Florida (C. S.); Trillium Health, Rochester, New York (Roberto Corales, DO); and University of California, San Diego (W. C. M.). Sponsoring agencies: Agency for Healthcare Research and Quality (AHRQ) (Fred Hellinger, PhD, J. A. F., and Irene Fraser, PhD) and Health Resources and Services Administration (HRSA) (Robert Mills, PhD, and Faye Malitz, MS), Rockville, Maryland. Data coordinating center: Johns Hopkins University (Richard Moore, MD, J. K., K. A. G., Cindy Voss, MA, and Nikki Balding, MS).
Disclaimer.The views expressed in this article are those of the authors. No official endorsement by the Department of Health and Human Services, the HRSA, or the AHRQ is intended or should be inferred. The funding agencies did not have a role in the design and conduct of the study; the collection, management, analysis, and interpretation of the data; and preparation, review, or approval of the manuscript. Although J. A. F. is an employee of the AHRQ and L. W. C. and H. H. are employees of the HRSA, which administers the Ryan White human immunodeficiency virus/AIDS Program, the views expressed in this article are those of the authors. No official endorsement by the Department of Health and Human Services, the HRSA, or the AHRQ is intended or should be inferred.
Financial support.This work was supported by the AHRQ (contract HHSA290201100007C), the HRSA (contract HHSH250201200008C), the National Institutes of Health (grants K23 AI084854, K23-MH097647, U01 DA036945, P30 AI094189, U01 HD068070, and 2UM1 AI068632-09), and the Clinical Investigation and Biostatistics Core of the UC San Diego Center for AIDS Research (AI036214).
Potential conflicts of interest.S. A. B. has been a consultant for Bristol-Myers Squibb. B. R. Y. has been a consultant and received research support from Gilead. A. L. A. has provided expert testimony unrelated to this work for various attorneys. A. E. N. has received grant support from Gilead. K. A. G. has been a consultant for Bristol-Myers Squibb and Tibotec and an expert witness for the US government. All other authors report no potential conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.