Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Aug 2.
Published in final edited form as: J Health Care Poor Underserved. 2017;28(1):406–429. doi: 10.1353/hpu.2017.0031

Antiretroviral therapy: Racial disparities among publicly insured Californians with HIV

RAPHAEL J LANDOVITZ 1, KATHERINE A DESMOND 2, ARLEEN A LEIBOWITZ 3
PMCID: PMC8328396  NIHMSID: NIHMS1722799  PMID: 28239010

Abstract

Only 43% of Americans with HIV are virally suppressed; the rate is lower for African Americans, even among insured populations. This study uses 2010 Medicare and Medicaid data for HIV-positive Californians to examine how antiretroviral treatment (ART) relates to patient and provider characteristics. Logistic regressions isolated the effect of race/ethnicity on receipt of ART. Over 90% of the full sample received any ART. Nearly 80% of ART users received a recommended combination for at least half the year; half had a recommended combination for 90% of the year. Lacking evaluation and management visits, or seeing only providers with low HIV patient volume lowered the odds of receiving ART. Controlling for other factors, African Americans remained less likely to receive ART at all, or to be covered for 90% of the year with a recommended regimen. The observed racial treatment differentials may lead to important health disparities.

Keywords: HIV, antiretroviral therapy, public insurance, disparities, California

Antiretroviral therapy (ART) has revolutionized medical treatment for people living with HIV/AIDS (PLWHA) by suppressing HIV replication, which is not only optimal for individual health and survival, but also reduces the transmissibility of HIV to sexual partners.1 Despite these advantages and its wide availability, ART was prescribed to only 440,000 of the 1.2 million Americans (37%) living with HIV in 2011,2 in large part because 14% of PLWHA were undiagnosed and only 56% of those diagnosed were engaged in care.2 Overall, only 43% of all diagnosed PLWHA were virally suppressed, after accounting for individuals not retained in care or not prescribed ART.2

The “Gardner cascade” describes the attrition at each progressive step between HIV infection and achievement of durable virologic suppression. African Americans experience lower rates of diagnosis and treatment at every stage in the Gardner cascade.3 A higher percentage of HIV-positive African Americans are undiagnosed compared with Whites (15% vs. 12%), fewer are engaged in care (55.2% vs. 58.5%) and fewer are prescribed ART (50.4% vs. 54.9%).2 These differences combine to create greater differentials in viral suppression than in ART prescription alone. Only 38.9% of African Americans are virally suppressed, compared with 47.3% of Whites.2 The reasons for this differential are not well understood.

Differences in insurance coverage contribute to treatment disparities,4 but disparities in HIV treatment, response, and survival exist even among insured populations and research participants who are provided HIV care and treatment without cost.5,6 Initiation of ART varies with the HIV-specific patient volume and experience of providers,7,8,9 yet the role of providers in contributing to racial/ethnic disparity in viral suppression is unknown. Patient characteristics may also contribute to disparities in viral suppression. Women and younger PLWHA are less likely than their counterparts to obtain ART,10 and women are less likely to receive a recommended ART regimen and to adhere to it.11,12 How much differences in the gender and age composition of African American vs. White populations of PLWHA contribute to the race disparity in viral suppression is also unknown.

This paper analyzes 2010 Medicare and Medicaid claims data from California to determine the extent to which lesser use of antiretroviral medication by African Americans than Whites relates to demographics (younger age, higher proportion of women) or to selected characteristics of providers who treat them. To unpack the racial disparities in viral suppression observed in other studies, this paper relates the achievement of specific intermediate treatment objectives (obtaining ART, using a recommended regimen, and having drug coverage for 90% of the year) to patient and provider characteristics in a publicly insured population.

Methods

Data.

We applied an algorithm that required two outpatient diagnoses (or one inpatient diagnosis) to identify HIV-positive people from Medicare and Medicaid insurance claims data provided by the Centers for Medicare and Medicaid Services (CMS).13 The resulting analysis file of adult Medicare and Medicaid beneficiaries with verifiable HIV included only fee-for-service (FFS) enrollees with full coverage because available data for Medicare managed care enrollees lack diagnosis codes needed to confirm HIV status. The analysis included 2010 full-year enrollees, and required full-year enrollment in a Part D drug plan for Medicare beneficiaries. To exclude people newly diagnosed during 2010, for whom the full-year assessment of medication coverage would be inappropriate, we restricted our analyses to individuals who had an HIV diagnosis code recorded in the claims data prior to 2010 (looking back from 2007 to 2009). Medicare (including dually enrolled) and Medicaid beneficiaries were analyzed separately.

Outcome variables.

For each beneficiary, we created a matrix showing possession of any ARV medications on each day in 2009 and 2010. ARVs were identified by National Drug Code (NDC) in prescription drug claims. Days possessed were identified by fill date and days’ supply. The two-year array allowed prescriptions filled in 2009 to extend into 2010.

The outcome “Any ARV” was defined as having at least one day in 2010 with possession of at least one ARV: nucleoside/nucleotide reverse transcriptase inhibitors (NRTIs), non-nucleoside reverse transcriptase inhibitors (NNRTIs), protease inhibitors (PIs), integrase inhibitors (INSTIs), or fusion inhibitors (FIs).

The outcome “Recommended Combination” was defined as having at least 180 days in 2010 with a recommended ART regimen, defined as two or more NRTIs and at least one NNRTI, PI or INSTI. This outcome applied to those possessing any ARVs during the year. Recommended regimens were defined according to Department of Health and Human Services Antiretroviral Guidelines at the time of visits, interpreted as reasonable and appropriate for either initial or non-initial HIV antiretroviral combination therapy.14

The outcome “90% Coverage” applied to those with any ARVs in 2010 possessing sufficient drug supply of a recommended combination to cover at least 330 days.

Individual characteristics.

The Centers for Medicare and Medicaid Services provided data on age category (< 30, 30–39, 40–49, and 50 years or older), race/ethnicity (non-Hispanic White, African American, Hispanic/Latino, Asian/Pacific Islander, and other), and gender. Contextual variables included median income in ZIP code of beneficiaries’ residence15 and prevalence of living HIV/AIDS cases per 1,000 population in the beneficiaries’ county of residence.16,17 Diagnoses of adjustment or anxiety disorders, mood disorders, schizophrenia or other psychotic disorders, or other mental health disorders18,19 were identified using the Agency for Healthcare Research and Quality (AHRQ) Mental Health and Substance Abuse Clinical Classifications Software (CCS-MHSA).20 Indicators for having 0,1,2,3, or 4 or more Charlson comorbidities were also defined by examining ICD-9 diagnoses.21 The Medicare analyses included an indicator for enrollment in Medicare only vs. dual Medicare-Medicaid enrollment. Medicaid analyses included an indicator for having any outpatient medical clinic visits at a Federally Qualified Health Center or Rural Health Clinic (FQHC/RHC), to account for any differences in outpatient procedure coding.

Provider characteristics.

HIV-related outpatient evaluation and management (E&M) claims were defined by CPT-4 code (or medical visit code for Medicaid FQHC/RHC records), type and place of service, and having an ICD-9 diagnosis code of symptomatic (042) or asymptomatic (V08) HIV. Beneficiaries without HIV-related E&M claims were coded “No E&M visits.” Each provider’s caseload of publicly insured patients with HIV was assessed by counting unique Medicare and Medicaid enrollees seen for HIV E&M visits. To demonstrate a patient’s known access to an experienced HIV provider, we assigned to each beneficiary the case load category of the highest volume provider with whom the patient had an E&M visit related to HIV.

Statistical analysis.

Antiretroviral therapy rates were calculated for each of the three outcomes: any ARV, possessing a recommended regimen for 180 or more days, and a recommended regimen for 330 or more days. Chi-square tests were performed to determine whether outcomes varied significantly by category. We estimated logistic regressions for the Medicare and Medicaid samples to relate each of the three outcomes to a set of demographic, geographic, and provider variables in order to assess the net effect of each. Wald chi-square tests on the logistic regression coefficients indicate whether significant differences exist between specific categories and a reference group, holding constant other factors. We estimated odds ratios, with 95% confidence intervals, and created an appendix with adjusted rates or predictive margins by predictor categories, using the method of recycled predictions.

Results

The Medicare sample included 9,566 PLWHA; the Medicaid sample included 5,883 PLWHA (Table 1). Column A shows the percent of the sample in each predictor group. Columns B-D show the percent of people in each predictor group who achieved each outcome.

Table 1.

Observed Percentages of California PLWHA Receiving ARV Treatment, 2010

A B C D
Group N (%) in group N (% of group) with any ARV use Chi-square p-value N (% of ARV users in group) with 180+ days of recommended ARV use Chi-square p-value N (% of ARV users in group) with 330+ days of recommended ARV use Chi-square p-value
A. Medicare enrollees
Total 9,566 (100.0%) 9,141 (95.6%) 7,193 (78.7%) 4,963 (54.3%)
Age <30 75 (0.8%) 67 (89.3%) .001 53 (79.1%) .67 24 (35.8%) <.001
Age 30–39 612 (6.4%) 578 (94.4%) 443 (76.6%) 253 (43.8%)
Age 40–49 3,459 (36.2%) 3,335 (96.4%) 2,630 (78.9%) 1,711 (51.3%)
Age 50+ 5,420 (56.7%) 5,161 (95.2%) 4,067 (78.8%) 2,975 (57.6%)
White 5,466 (57.1%) 5,245 (96.0%) <.001 4,107 (78.3%) <.001 2,871 (54.7%) <.001
African American 1,809 (18.9%) 1,684 (93.1%) 1,288 (76.5%) 806 (47.9%)
Hispanic 1,887 (19.7%) 1,823 (96.6%) 1,476 (81.0%) 1,062 (58.3%)
Asian/PI 256 (2.7%) 244 (95.3%) 210 (86.1%) 148 (60.7%)
Other 148 (1.6%) 145 (98.0%) 112 (77.2%) 76 (52.4%)
Male 8,563 (89.5%) 8,222 (96.0%) <.001 6,469 (78.7%) .94 4,505 (54.8%) .004
Female 1,003 (10.5%) 919 (91.6%) 724 (78.8%) 458 (49.8%)
No adjustment/anxiety 9,013 (94.2%) 8,622 (95.7%) .045 6,769 (78.5%) .085 4,664 (54.1%) .12
Any adjustment/anxiety diagnosis 553 (5.8%) 519 (93.9%) 424 (81.7%) 299 (57.6%)
No mood disorders 7,952 (83.1%) 7,618 (95.8%) .011 5,984 (78.6%) .47 4,157 (54.6%) .24
Any mood disorders diagnosed 1,614 (16.9%) 1,523 (94.4%) 1,209 (79.4%) 806 (52.9%)
No schizophrenia 9,159 (95.8%) 8,779 (95.9%) <.001 6,909 (78.7%) .91 4,800 (54.7%) <.001
Any schizophrenia diagnosis 407 (4.3%) 362 (88.9%) 284 (78.5%) 163 (45.0%)
No other mental health diagnosis 9,265 (96.9%) 8,868 (95.7%) <.001 6,976 (78.7%) .74 4,818 (54.3%) .69
Any other mental health diagnosis 301 (3.2%) 273 (90.7%) 217 (79.5%) 145 (53.1%)
No comorbidities 5,777 (60.4%) 5,565 (96.3%) <.001 4,436 (79.7%) <.001 3,046 (54.7%) .004
1 comorbidity 2,484 (26.0%) 2,394 (96.4%) 1,889 (78.9%) 1,323 (55.3%)
2 comorbidities 811 (8.5%) 758 (93.5%) 577 (76.1%) 396 (52.2%)
3 comorbidities 310 (3.2%) 279 (90.0%) 200 (71.7%) 139 (49.8%)
4+ comorbidities 184 (1.9%) 145 (78.8%) 91 (62.8%) 59 (40.7%)
Enrolled in Medicare only 2,415 (25.2%) 2,287 (94.7%) .018 1,740 (76.1%) <.001 1,233 (53.9%) .67
Dual Medicare/Medicaid enrollment 7,151 (74.8%) 6,854 (95.9%) 5,453 (79.6%) 3,730 (54.4%)
No E&M visits 862 (9.0%) 731 (84.8%) <.001 587 (80.3%) .067 412 (56.4%) .007
E&M provider <5 583 (6.1%) 533 (91.4%) 405 (76.0%) 256 (48.0%)
E&M provider 5–29 1,426 (14.9%) 1,376 (96.5%) 1,061 (77.1%) 728 (52.9%)
E&M provider 30–49 712 (7.4%) 689 (96.8%) 528 (76.6%) 359 (52.1%)
E&M provider >=50 5,983 (62.5%) 5,812 (97.1%) 4,612 (79.4%) 3,208 (55.2%)
B. Medicaid enrollees
Total 5,883 (100.0%) 5,472 (93.0%) 4,331 (79.2%) 2,790 (51.0%)
Age <30 299 (5.1%) 269 (90.0%) <.001 182 (67.7%) <.001 111 (41.3%) <.001
Age 30–39 691 (11.8%) 629 (91.0%) 494 (78.5%) 269 (42.8%)
Age 40–49 2,268 (38.6%) 2,144 (94.5%) 1,692 (78.9%) 1,045 (48.7%)
Age 50+ 2,625 (44.6%) 2,430 (92.6%) 1,963 (80.8%) 1,365 (56.2%)
White 2,037 (34.6%) 1,915 (94.0%) <.001 1,509 (78.8%) .012 991 (51.8%) <.001
African American 2,026 (34.4%) 1,845 (91.1%) 1,427 (77.3%) 855 (46.3%)
Hispanic 1,157 (19.7%) 1,098 (94.9%) 908 (82.7%) 614 (55.9%)
Asian/Pacific Islander 178 (3.0%) 168 (94.4%) 137 (81.6%) 101 (60.1%)
Other 485 (8.2%) 446 (92.0%) 350 (78.5%) 229 (51.4%)
Male 4,251 (72.3%) 3,976 (93.5%) .012 3,155 (79.4%) .55 2,050 (51.6%) .17
Female 1,632 (27.7%) 1,496 (91.7%) 1,176 (78.6%) 740 (49.5%)
No adjustment/anxiety 5,526 (93.9%) 5,149 (93.2%) .052 4,075 (79.1%) .96 2,615 (50.8%) .24
Any adjustment/anxiety diagnosis 357 (6.1%) 323 (90.5%) 256 (79.3%) 175 (54.2%)
No mood disorders 4,846 (82.4%) 4,514 (93.2%) .38 3,565 (79.0%) .50 2,300 (51.0%) .91
Any mood disorders diagnosed 1,037 (17.6%) 958 (92.4%) 766 (80.0%) 490 (51.2%)
No schizophrenia 5,414 (92.0%) 5,058 (93.4%) <.001 4,019 (79.5%) .049 2,599 (51.4%) .040
Any schizophrenia diagnosis 469 (8.0%) 414 (88.3%) 312 (75.4%) 191 (46.1%)
No other mental health diagnosis 5,612 (95.4%) 5,230 (93.2%) .014 4,141 (79.2%) .80 2,669 (51.0%) .75
Any other mental health diagnosis 271 (4.6%) 242 (89.3%) 190 (78.5%) 121 (50.0%)
No comorbidities 3,805 (64.7%) 3,564 (93.7%) .024 2,851 (80.0%) <.001 1,875 (52.6%) <.001
1 comorbidity 1,473 (25.0%) 1,363 (92.5%) 1,089 (79.9%) 681 (50.0%)
2 comorbidities 400 (6.8%) 360 (90.0%) 263 (73.1%) 160 (44.4%)
3 comorbidities 146 (2.5%) 131 (89.7%) 92 (70.2%) 52 (39.7%)
4+ comorbidities 59 (1.0%) 54 (91.5%) 36 (66.7%) 22 (40.7%)
No FHQC medical visits 2,964 (50.4%) 2,804 (94.6%) <.001 2,197 (78.4%) .14 1,427 (50.9%) .89
Any FQHC medical visits 2,919 (49.6%) 2,668 (91.4%) 2,134 (80.0%) 1,363 (51.1%)
No E&M visits 482 (8.2%) 421 (87.3%) <.001 317 (75.3%) <.001 192 (45.6%) .003
E&M provider <5 218 (3.7%) 194 (89.0%) 142 (73.2%) 88 (45.4%)
E&M provider 5–29 727 (12.4%) 685 (94.2%) 508 (74.2%) 322 (47.0%)
E&M provider 30–49 431 (7.3%) 409 (94.9%) 330 (80.7%) 222 (54.3%)
E&M provider >=50 4,025 (68.4%) 3,763 (93.5%) 3,034 (80.6%) 1,966 (52.3%)
Open in a new tab

Abbreviations: PLWHA, people living with HIV/AIDS; ARV, antiretroviral medication; Dual, dually enrolled in both Medicare and Medicaid, receiving drugs through Medicare Part D; E&M, HIV-related evaluation and management visit; FQHC/RHC, Federally Qualified Health Center or Rural Health Clinic

Analysis of 2010 Medicare and Medicaid claims data, HIV-positive Californians enrolled full year in full coverage fee-for-service programs, Medicare beneficiaries enrolled in Part D. Beneficiaries were required to have an HIV diagnosis recorded in at least one claim prior to 2010.

Antiretroviral therapy use was high in this sample of PLWHA in care: 95.6% of the Medicare sample and 93.0% of the Medicaid sample possessed an ARV for at least one day in 2010 (column B). Fewer than 80% in either sample with a prescription for ART had at least half a year covered with a recommended regimen (column C). Only 54.3% of Medicare recipients receiving any ART filled prescriptions that would cover 90% or more of the year with a recommended ART regimen and just over half (51.0%) of the Medicaid enrollees receiving ART had 90% coverage (column D).

A smaller percentage of African American Medicare enrollees than Whites filled an ARV prescription (93.1% vs. 96.0%); had a recommended regimen, given that they had an ARV prescription (76.5% vs. 78.3%); or had recommended ART coverage for 330 days (47.9% vs. 54.7%). Similar differences were found in the Medicaid sample: 91.1% of African Americans vs. 94.0% of Whites filled any ARV prescription. Of these, 77.3% of African Americans had an approved regimen (vs. 78.8% for Whites) and 46.3% vs. 51.8% for Whites had 330 days of medication coverage. Female PLWHA were less likely to receive ART. Women represented a higher proportion of African American PLWHA than of White PLWHA in both the Medicare (20% vs. 7%) and Medicaid (37% vs. 22%) samples.

Adjustment or anxiety disorders were associated with a lower rate of obtaining ART among both Medicare (93.9% vs. 95.7%) and Medicaid (90.5% vs. 93.2%) recipients. Mood disorders had a small impact on having any-ARV in Medicare (94.4% vs. 95.8% without mood disorders), but not among Medicaid patients. A diagnosis of schizophrenia was associated with lower rates of any-ART and of 90% coverage among both Medicare and Medicaid recipients. Having more comorbid conditions was generally negatively associated with all three ART measures in both samples.

With some exceptions, these unadjusted ARV use rates differed significantly by predictor category. We estimated multivariate equations to isolate the effect of each predictor on ART use, controlling for other demographic and care characteristics.

Any ART Prescription: Who, where, and by whom?

Figures 1–3 present odds ratios for each explanatory factor for each of the three outcomes. The individual characteristics that affected the probability of filling a prescription for ART, shown in Figure 1, were similar for the two programs. People living with HIV/AIDS in the two age categories 40 or older had greater odds of having ART than those who were under 30 (ORs=2.98 with p=.008 and 2.62 with p=.018 in Medicare, ORs=2.26 with p<.001 and 1.76 with p=.009 in Medicaid). Enrollees with a diagnosis of schizophrenia were significantly less likely than others to have an ARV prescription (OR=0.43, p<.001 for Medicare and OR=0.67, p=.018 for Medicaid). The other mental health disorders did not have a significant impact on ART use. Dual Medicare enrollees were more likely to receive ART thatn individuals with Medicare coverage alone (OR=1.74, p<.001). There was strong to moderate evidence that women were less likely to receive any ART in both the Medicare (OR=0.53, p<.001) and Medicaid (OR=0.80, p=.053) samples. Having two or more comorbidities had a strong dampening effect on filling an ART prescription among Medicare recipients (ORs= .19 to .60, p<=.002), while the impact among Medicaid recipients was less pronounced: having two comorbidities had a significant impact on ART use (OR=0.61, p=.009), but having one or three or more was not significant.

Figure 1,

Figure 1,

Figure 1,

Open in a new tab

Results of logistic regressions predicting any ARV use, California PLWHA, 2010

Figure 3,

Figure 3,

Figure 3,

Open in a new tab

Results of logistic regressions predicting 330+ days of recommended regimen, California PLWHA using any ARVs, 2010

Notes for Figures 1–3:

Abbreviations: PLWHA, people living with HIV/AIDS; ARV, antiretroviral medication; Dual, dually enrolled in Medicare and Medicaid, receiving drugs through Medicare Part D; Dx, diagnosis; MH, mental health; HH, household; E&M, HIV-related evaluation and management visit; FQHC/RHC, Federally Qualified Health Center or Rural Health Center

Analysis of 2010 Medicare and Medicaid claims data, HIV-positive Californians enrolled full year in full coverage fee-for-service programs, Medicare beneficiaries enrolled in Part D. Data markers indicate odds ratios with 95% confidence intervals, estimated using logistic regression. Numeric values for the odds ratios and confidence interval are also provided to the right of the markers. Odds ratios represent the odds that the outcome will occur given the particular characteristic, compared to the odds given the reference group. P-values following the confidence intervals are from Chi-square tests of whether the odds ratios are significantly different from one

Even after controlling for these factors, African Americans had significantly lower odds of obtaining ART than Whites (OR=0.63, p<.001 for Medicare; OR=0.70, p=.004 for Medicaid). Latinos, Asian Americans, and other race/ethnicities did not differ significantly from Whites in ART usage. Median income in ZIP code of residence did not influence ART availability for either insurance group. HIV/AIDS prevalence in the county was statistically significant for both Medicare and Medicaid, but odds ratios for cases per 1000 population were not meaningfully different from one (0.97).

Access to higher-volume providers also predicted ART possession. Both Medicare and Medicaid enrollees without HIV-related E&M visits were much less likely to be on ART (OR=0.17 for Medicare, 0.35 for Medicaid, both p<.001). Patients seeing only providers who treated fewer than five HIV patients in our sample were less likely to have a claim for ART than those who saw high-volume providers (OR=0.38, p<.001 for Medicare; OR=0.54, p=.007 for Medicaid). Medicaid enrollees with any visit to a FQHC/RHC were also less likely to receive ART (OR=0.58, p<.001).

Medically recommended ART regimen.

Figure 2 shows that, among Medicare enrollees who received any ART, Asians/Pacific Islanders were significantly more likely than others to receive a recommended ART combination for at least half a year (OR=1.73, p=.004) as were Hispanics, to a lesser extent (OR=1.17, p=.029). There was no evidence that African American Medicare beneficiaries differed significantly from Whites. Other race/ethnicity, gender, and mental health diagnoses were insignificant in both samples. Having two or more comorbidities was significantly associated with a lower likelihood of a having recommended regimen in both groups (ORs = .40 to .79, p<=.01 in Medicare; ORs=.47 to .64 in Medicaid, p<=.012). Medicaid enrollees on ART who were 30 or older were significantly more likely to receive a recommended regimen than patients under age 30 (ORs=1.72 to 2.17; all p<=.001). The only racial/ethnic group in the Medicaid sample that differed significantly from Whites was Hispanics, who were more likely to have half a year of a recommended combination (OR=1.33, p=.004).

Figure 2,

Figure 2,

Figure 2,

Open in a new tab

Results of logistic regressions predicting 180+ days of recommended regimen, California PLWHA using any ARVs, 2010

Medicaid patients seeing providers with larger HIV caseloads were more likely to receive a recommended ART regimen. Those who saw providers with fewer than five HIV patients (OR=0.68, p=.021) or five-29 HIV patients (OR=0.70, p<.001) were less likely to receive recommended regimens than individuals whose providers had larger HIV caseloads. Medicaid patients lacking an E&M visit were also less likely to receive a recommended drug combination (OR=0.73, p=.012), and FQHC/RHC users were as likely as others to receive a recommended regimen. Among Medicare patients, dual enrollees were more likely to have a recommended regimen than individuals with only Medicare coverage (OR=1.27, p<.001), but seeing higher-volume providers did not have a statistically significant impact.

90% ART coverage.

Both provider and individual variables related to having 90% coverage with a recommended ART regimen, as shown in Figure 3. Both Medicaid and Medicare ART users aged 40 and older were more likely to have 90% coverage with a recommended regimen (ORs=1.95 and 1.43 for 40–49 year olds, both p<=.01; ORs=2.63 and 2.08 for those 50 and older, both p<.001), while those aged 30–39 did not differ from those under 30. Enrollees with a diagnosis of schizophrenia were less likely than others to have 90% coverage in Medicare (OR=0.72, p=.004), but there was not a significant difference among Medicaid recipients. There was strong to moderate evidence from both programs that those with adjustment or anxiety disorders were more likely to have full coverage if they filled any prescription for ART than patients without these diagnoses (ORs=1.26, p=.017 in Medicare and p=.060 in Medicaid). Having two or more comorbidities decreased the likelihood of 90% coverage significantly or nearly significantly in both groups (ORs = .52 to .96 with p<.001 to .065 in Medicare; ORs = .53 to .86 in Medicaid with p<.01 to .035 in Medicaid). Women enrolled in Medicare were significantly less likely than men with any ART prescription to have full 330-day coverage (OR=0.83, p=.011), but this was not the case in Medicaid. Among Medicare enrollees, those with dual Medicare/Medicaid coverage were more likely to have 90% coverage (OR=1.17, p=.003). Medicare enrollees living in higher income ZIP codes were slightly more likely to have full coverage (OR=1.03, p=.013), and both Medicare and Medicaid enrollees living in higher prevalence counties were slightly less likely to have full coverage (ORs=0.99 and .98, p<=.001), but these odds ratios indicate that differences of $10,000 in median income or an additional case per 1,000 population had very small effect.

Medicaid enrollees without documented E&M visits were less likely to achieve full ART coverage (OR=0.72, p=.002). Patients seeing providers with larger HIV caseloads were more likely to have a recommended combination of ART for 330 days. Beneficiaries seeing providers with fewer than five HIV patients (OR=0.75, p=.002 in Medicare, OR=0.75, p=.056 in Medicaid), or providers with five-29 patients (OR=0.90, p=.087 in Medicare and OR=0.81, p=.015 in Medicaid) had strong to moderate evidence that they experience lower odds of 90% ART coverage than individuals whose providers had larger caseloads. Users of FQHCs/RHCs did not differ from others in having 90% coverage.

Even after controlling for all these factors, in both programs African Americans had significantly lower odds of having 90% ART coverage compared with Whites (OR=0.83, p=.001 in Medicare; OR=0.81, p=.002 in Medicaid). Latinos (OR=1.2, p<=.01 in both groups) and Asian Americans (OR=1.37, p=.020 in Medicare; OR=1.50, p=.015 in Medicaid) had significantly greater odds than Whites of having 90% ART coverage.

Discussion

Suppression of viral replication to undetectable levels plays a key role in achieving all three principal goals of the National HIV/AIDS Strategy for the United States.22 Viral suppression improves both HIV-related and other comorbidity and mortality among those infected with HIV. In addition, the HPTN 0521 and the Prevention of Mother-to-Child Transmission literature23–27 have definitively identified the impact of virologic suppression in reducing horizontal and vertical transmission of HIV. The clinical and public health benefits of virologic suppression apply equally to all racial and ethnic groups.

As mentioned earlier, the Gardner cascade describes the incremental “loss” at each of the critical steps in identification and treatment that lead to viral suppression: diagnosis, linkage to and retention in care, ART prescription with a recommended combination, and sufficient adherence to achieve virologic suppression.5 These additive “losses” mean that only 43% of PLWHA who know their diagnosis achieve viral suppression, and these rates are lower for African Americans, women and younger PLWHA.2 A detailed understanding of contributors to each step-off in the Gardner cascade clarifies what types of interventions are appropriate; in this paper we have addressed possible losses in ART prescription and adherence.

Our results show that in 2010 ARV use was widespread, with over 90% of Medicare and Medicaid enrollees filling prescriptions for ART. As previously found, African Americans were less likely than others to obtain ARVs28,29 as were women,11,12 younger patients,10 and patients with a diagnosis of schizophrenia.19 Despite the prevalence of ART, more than one-fifth of PLWHA with any ART did not possess a recommended drug regimen. We found that the impact of demographic factors on the appropriateness of the regimen received was attenuated among PLWHA who possessed any ARVs, while provider experience remained a strong predictor (especially for Medicaid enrollees). Patients who did not have evidence of visits with high volume HIV providers were most likely to have non-recommended regimens.30,31

Access to high-volume providers related to each of the three outcomes, and its impact was independent of associations with measured demographic characteristics. Some studies have attributed the lower rates of obtaining a prescription for ART among minority patients to beliefs by providers that such patients would be less adherent to the therapy.32–34 Our results provide some support for this perception. Even after controlling for differences in personal and provider characteristics, African Americans and women with Medicare were less likely than Whites and men who began ART to have obtained sufficient medication to supply 90% of the days in the year.

Although the literature suggests that serious mental health problems, particularly depression, can be associated with non-adherence,19,35we found different mental health diagnoses affected adherence to ART differently. Although a review of the literature found no significant relationship between anxiety disorder and ART adherence,36 our analyses found that among patients using ART, those with anxiety disorders proved more adherent than those without a mental health diagnosis. In contrast, Medicare patients with schizophrenia were less adherent. These findings suggest that it is important for physicians to differentiate among types of mental health diagnoses in predicting which patients are more likely to adhere to therapy.37,38

The racial treatment differentials we observed can lead to important health disparities.39 Combining the likelihood that African American patients will receive and fill an ARV prescription (93% vs. 96% for Whites) with the reduced chance of their having full ART coverage for 90% of the days if they have any ART (49% vs. 54% for Whites), leads to an estimate that overall only 45% of African Americans have sufficient supplies of a recommended ART regimen for 90% of the year, compared with 52% of Whites.

We found that female and younger PLWHA were less likely to have full ART coverage over the year. Although women and younger people represent a greater proportion of African Americans with HIV than Whites with HIV, age and gender do not account for the racial disparities we observe. Being African American was independently related to lower ART adherence, even after controlling for the age and gender distribution of the population.

Statistically significant differences between White and African American beneficiaries were not observed in obtaining a recommended ART regimen. The combination of ARVs prescribed is largely under the control of the physician. The significance of provider experience and the lack of significance of race or gender in predicting receipt of a recommended regimen reflect the provider’s importance. This suggests that disparities in the uptake of ART and in its consistent use over the year may be the result of patient preferences or other barriers to ART utilization that remain in an insured population. These preferences and barriers present important challenges that will have to be overcome if “treatment as prevention” is going to realize its full potential in reducing HIV incidence.

The generalizability of our findings is limited by the exclusion of PLWHA in managed care or with private insurance only. However, Medicare and Medicaid account for over half of PLWHA who are receiving care,40,41 minimizing the impact of this limitation. Secondly, we applied a strict algorithm to identify HIV-positive people from claims,13 which may have resulted in a sample receiving better-than-average care. Third, the provider caseload measures relate only to publicly insured (Medicare and Medicaid) patients. These programs financed HIV patients with more advanced disease, who are the prime candidates for antiretroviral therapy,42 which may also bias our estimates in the direction of over-estimating ARV use in the general HIV population.

Each step of the Gardner cascade represents an opportunity for individual and structural interventions to promote adherence to a recommended ART regimen.43,44 The elusive goal of achieving virologic suppression for all PLWHA is important for both individual and public health. Individuals maintain immunologic integrity, improved longevity and quality of life, and prevent the development of resistant virus;45–49 populations benefit from reduced onward secondary transmission and reduced incident HIV cases.

Current Department of Health and Human Services HIV treatment guidelines recommend initiating ART for all PLWHA, regardless of CD4 count.50 Given the lower rates of engagement in care and in ART prescription and adherence observed among African Americans, health disparities are likely to increase unless providers target increasing ART prescription rates and supporting ART use among minority groups. Physicians can play an important role by prescribing recommended ART regimens and by supporting ART use, particularly for African Americans, women, and youth with HIV.51,52 Absent these provider efforts to support ART use, we are likely to see that disparities between advantaged and disadvantaged PLWHA widen, as was the case following the initial introduction of HAART.4 Optimization at each stage of the Gardner cascade will require a nuanced understanding of the contributors to each successive level of drop-off, as well as interventions to address the shared responsibilities of their origins. These challenges become more vital to understand in detail as larger populations become eligible for ART treatment, with rapidly evolving guidelines in the face of recent clinical trial data.53 Disparities in receipt and/or administration of recommended ART regimens among those who are disabled or ill are likely to be amplified as we aim to treat those who have been heretofore asymptomatic.

Contributor Information

RAPHAEL J. LANDOVITZ, Division of Infectious Diseases, UCLA David Geffen School of Medicine, and UCLA Center for Clinical AIDS Research and Education..

KATHERINE A. DESMOND, Department of Public Policy, UCLA Luskin School of Public Affairs..

ARLEEN A. LEIBOWITZ, Department of Public Policy, UCLA Luskin School of Public Affairs..

References

  • 1.Cohen MS, Chen YQ, McCauley M, Gamble T, Hosseinipour MC, Kumarasamy N, et al. Prevention of HIV-1 infection with early antiretroviral therapy. N Engl J Med. 2011;365(6):493–505. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Centers for Disease Control and Prevention (CDC). Monitoring selected national HIV prevention and care objectives by using HIV surveillance data--United States and 6 dependent areas--2012. HIV Surveillance Supplemental Report 2014; 19(No.3). Accessed December 20, 2015 at http://www.cdc.gov/hiv/library/reports/surveillance/, Published November 2014. [Google Scholar]
  • 3.Gardner EM, McLees MP, Steiner JF, del Rio C, Burman WJ, “The Spectrum of Engagement in HIV Care and its Relevance to Test-and-Treat Strategies for Prevention of HIV Infection”. Clin Infect Dis, 2011; 52(6): 793–800. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Millett G Structural Factors and the HIV Continuum of Care. Presentation at the UCLA Housing and HIV/AIDS Summit. October 31, 2014. [Google Scholar]
  • 5.Giordano TP, Bartsch G, Zhang Y, Tedaldi E, Absalon J, Mannheimer S, et al. Disparities in Outcomes for African American and Latino Subjects in the Flexible Initial Retrovirus Suppressive Therapies (FIRST) Trial. AIDS Patient Care STDs. 2010. May;24(5):287–95. doi: 10.1089/apc.2009.0332. PMID:20438378 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Ribaudo HJ, Smith KY, Robbins GK, Flexner C, Haubrich R, Chen Y et al. Racial differences in response to antiretroviral therapy for HIV infection: an AIDS clinical trials group (ACTG) study analysis. Clin Infect Dis 2013). Clin Infect Dis. 2013. December;57(11):1607–17. doi: 10.1093/cid/cit595. Epub 2013 Sep 17. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Landon BE, Wilson IB, Cohn SE, Fichtenbaum CJ, Wong MD, Wenger NS, et al. Physician specialization and antiretroviral therapy for HIV. J Gen Intern Med. 2003;18(4):233–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Landon BE, Wilson IB, McInnes K, Landrum MB, Hirschhorn LR, Marsden PV, et al. Physician specialization and the quality of care for human immunodeficiency virus infection. Arch Intern Med. 2005;165(10):1133. [DOI] [PubMed] [Google Scholar]
  • 9.Rackal JM, Tynan AM, Handford CD, Rzeznikiewiz D, Agha A, Glazier R. Provider training and experience for people living with HIV/AIDS. Cochrane Database of Systematic Reviews 2011, Issue 6. Art. No.: CD003938. DOI: 10.1002/14651858.CD003938.pub2. Accessed December 20, 2015 at: http://summaries.cochrane.org/CD003938/HIV_training-experience-and-volumes-on-outcomes-for-people-living-with-hivaids-a-systematic-review [DOI] [PubMed] [Google Scholar]
  • 10.Gebo KA, Fleishman JA, Conviser R, Reilly ED, Korthuis PT, Moore RD, et al. Racial and gender disparities in receipt of highly active antiretroviral therapy persist in a multistate sample of HIV patients in 2001. J Acquir Immune Defic Syndr. 2005; 38(1):96–103. [DOI] [PubMed] [Google Scholar]
  • 11.Nicastri E, Leone S, Angeletti C, Palmisano L, Sarmati L, Chiesi A., et al. Sex issues in HIV-1-infected people during highly active antiretroviral therapy: a systematic review. J. Antimicrob Chemother 2007; 60(4):724–32. [DOI] [PubMed] [Google Scholar]
  • 12.Carter A, Min JE, Chau W, Lima VD, Kestler M, Pick N, et al. Gender inequities in quality of care among HIV-positive individuals initiating antiretroviral treatment in British Columbia, Canada (2000–2010). PLoS One. 2014. March 18; 9(3):e92334. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Leibowitz AA, Desmond K. Identifying a Sample of HIV-Positive Beneficiaries from Medicaid Claims Data and Estimating Their Treatment Costs. Am J Public Health. March 2015; 105(3):567–574. NIHMSID #646470. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Panel on Antiretroviral Guidelines for Adults and Adolescents. Guidelines for the use of antiretroviral agents in HIV-1-infected adults and adolescents. Department of Health and Human Services. December 1, 2009; 1–161. Available at http://www.aidsinfo.nih.gov/ContentFiles/AdultandAdolescentGL.pdf. Accessed April 5, 2016. [Google Scholar]
  • 15.U.S. Census Bureau, 2011. American Community Survey 5-Year Estimates, Table S1903, accessed December 20, 2015 using American FactFinder, http://factfinder2.census.gov.2011 values deflated to 2010 using BLS CPI adjustment factor of 0.97.
  • 16.California Department of Public Health, HIV/AIDS Surveillance in California, Table 1, Quarterly Report of HIV/AIDS Cases by County in California, Cumulative as of June 30, 2010. Accessed December 20, 2015 from http://www.cdph.ca.gov/programs/aids/Documents/SSQtr2Jun2010.pdf [Google Scholar]
  • 17.State of California, Department of Finance, E-2. California County Population Estimates and Components of Change by Year – July 1, 2010–2014, Table 4. Accessed December 20, 2015 from http://www.dof.ca.gov/research/demographic/reports/estimates/e-2/view.php. (source of county population). [Google Scholar]
  • 18.Walkup J, Akincigil A, Bilder S, Rosato NS, Crystal S. Psychiatgric diagnosis and antiretroviral adherence among adolescent Medicaid beneficiaries diagnosed with human immunodeficiency virus/acquired immunodeficienty syndrome. J. Nerv Ment Dis May 2009; 187(5):354–61. [DOI] [PubMed] [Google Scholar]
  • 19.Huynh AK, Kinsler JJ, Cunningham WE, Sayles JN.The role of mental health in mediating the relationship between social support and optimal ART adherence. AIDS Care. 2013; 25(9):1179–84. doi: 10.1080/09540121. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.U.S. Agency for Healthcare Research and Quality, Healthcare Cost and Utilization Project. Clinical Classifications Software (CCS) for ICD-9-CM. Accessed December 20, 2015 from http://www.hcup-us.ahrq.gov/toolssoftware/ccs/ccs.jsp. [Google Scholar]
  • 21.Quan H, Sundararajan V, Halfon P, et al. Coding algorithms for defining Comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care. November 2005; 43(11): 1130–9. [DOI] [PubMed] [Google Scholar]
  • 22.The White House Office of National AIDS Policy. National HIV/AIDS Strategy for the United States. July 2010. Accessed December 20, 2015 at: http://www.aids.gov/federal-resources/policies/national-hiv-aids-strategy/nhas.pdf).
  • 23.Cooper ER, Charurat M, Mofenson L, et al. Combination antiretroviral strategies for the treatment of pregnant HIV-1-infected women and prevention of perinatal HIV-1 transmission. J Acquir Immune Defic Syndr 2002; 29:484–94. [DOI] [PubMed] [Google Scholar]
  • 24.Drake AL, Wagner A, Richardson B, John-Stewart G. Incident HIV during pregnancy and postpartum and risk of mother-to-child HIV transmission: a systematic review and meta-analysis. PLoS Med 2014; 11:e1001608. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Roongpisuthipong A, Siriwasin W, Simonds RJ, et al. HIV seroconversion during pregnancy and risk for mother-to-infant transmission. J Acquir Immune Defic Syndr 2001; 26:348–51. [DOI] [PubMed] [Google Scholar]
  • 26.Shapiro RL, Hughes MD, Ogwu A, et al. Antiretroviral regimens in pregnancy and breast-feeding in Botswana. N Engl J Med 2010; 362:2282–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Palombi L, Marazzi MC, Voetberg A, Magid NA. Treatment acceleration program and the experience of the DREAM program in prevention of mother-to-child transmission of HIV. AIDS 2007; 21 Suppl 4:S65–71. [DOI] [PubMed] [Google Scholar]
  • 28.Rosenberg ES, Millett GA, Sullivan DR, Curran JW. Understanding the HIV disparities between Black and White s who have sex with men in the USA using the HIV care continuum modelling study. Lancet HIV; December 2014; 1(3):e112–e118. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Maulsby C, Millet G, Lindsey K, Kelley R, Johnson K, Montoya D, Holtgrave D. HIV Among Black Men Who Have Sex with Men (MSM) in the United States: A Review of the Literature. AIDS Behav. 2014;18:10–25. [DOI] [PubMed] [Google Scholar]
  • 30.Kitahata MM, Van Rompaey SE, Shields AW. Physician experience in the care of HIV-infected people is associated with earlier adoption of new antiretroviral therapy. J Acquir Immune Defic Syndr. 2000;24(2):106–14. [DOI] [PubMed] [Google Scholar]
  • 31.Wilson IB, Landon BE, Hirschhorn LR, McInnes K, Ding L, Marsden PV, et al. Quality of HIV care provided by nurse practitioners, physician assistants, and physicians. Ann Intern Med. 2005;143(10):729–36. [DOI] [PubMed] [Google Scholar]
  • 32.Calabrese SK, Earnshaw VA, Underhill K, Hansen NB, Dovidio JF. The Impact of Patient Race on Clinical Decisions Related to Prescribing HIV Pre-Exposure Prophylaxis (PrEP): Assumptions About Sexual Risk Compensation and Implications for Access. AIDS Behav. 2014;18(2):226–240. doi: 10.1007/s10461-013-0675-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Stone VE. Physician contributions to disparities in HIV/AIDS care: the role of provider perceptions regarding adherence. Curr HIV/AIDS Rep. 2005. November; 2(4):189–93. [DOI] [PubMed] [Google Scholar]
  • 34.Bogart LM, Catz SL, Kelly JA, Benotsch EG. Factors influencing physicians’ judgments of adherence and treatment decisions for patients with HIV disease. Med Decis Making. Jan-Feb 2001; 21(1):28–36. [DOI] [PubMed] [Google Scholar]
  • 35.Gonzalez JS, Batchelder AW, Psaros C, Safren SA. Depression and HIV/AIDS Treatment Nonadherence: A Review and Meta-analysis. J Acquir Immune Defic Syndr. October 1, 2011; 58(2):181–187. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Springer SA, Dushaij A, Azar MM. The impact of DSM-!V mental disorders on adherence to combination antiretroviral therapy among adult persons living with HIV/AIDS: a systematic review. AIDS Behav. 2012. November; 16(8):2119–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Bangsberg DR, Hecht FM, Clague H, Charlebois ED, Ciccarone D, Chesney M, Moss A. Provider Assessment of Adherence to HIV Antiretroviral Therapy. J AcquirImmune Defic Syndr. 2001, 26: 435–42. [DOI] [PubMed] [Google Scholar]
  • 38.Paterson DL, Swindells S, Mohr J, Brester M, Vergis EN, Squier C, et al. Adherence to Protease Inhibitor Therapy and Outcomes in Patients with HIV Infection. Ann Intern Med. 2000; 133(1):21–30. doi: 10.7326/0003-4819-133-1-200007040-00004. [DOI] [PubMed] [Google Scholar]
  • 39.Siddiqi AA, Hu X, Hall HI. Mortality Among Blacks or African Americans with HIV Infection--United States, 2008–2012. MMWR Morb Mortal Wkly Rep 2015;64(4):81–90 [PMC free article] [PubMed] [Google Scholar]
  • 40.Institute of Medicine. 2012. Monitoring HIV Care in the United States: Indicators and Data Systems. Washington, DC; The National Academies Press; [PubMed] [Google Scholar]
  • 41.Quinn K and Skarbinski J. Health Insurance Coverage and Type Predict Durable Viral Suppression among HIV+ Adults in Care: US Medical Monitoring Project 2009, CROI; 2013, Abstract #1028, Atlanta, GA. [Google Scholar]
  • 42.Goldman DP, Leibowitz AA, Joyce GF, Fleishman JA, Bozzette SA, Duan N, Shapiro MF. Insurance Status of HIV-Infected Adults in the Post-HAART Era: Evidence from the United States. Appl Health Econ Health Policy. 2003; 2(2):85–91. [PubMed] [Google Scholar]
  • 43.Bradley H, Hall HI, Wolitski RJ, Van Handel MM, Stone AE, LaFlam M, et al. Vital Signs: HIV Diagnosis, Care and treatment Among People Living with HIV. United States, 2011. MMWR Morb Mortal Wkly Rep November. 25, 2014;63:1–6. [PMC free article] [PubMed] [Google Scholar]
  • 44.Hall HI, Frazier EL, Rhodes P, Holtgrave DR, Furlow-Parmley C, Tang T, et al. Differences in human immunodeficiency virus care and treatment among subpopulations in the United States. JAMA Intern Med. 2013;173(14):1337–44. [DOI] [PubMed] [Google Scholar]
  • 45.Kitahata MM, Gange SJ, Abraham AG, et al. Effect of Early versus Deferred Antiretroviral Therapy for HIV on Survival. N Engl J Med. 2009;360(18):1815–1826. doi: 10.1056/NEJMoa0807252. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Eriksson LE, Bratt GA, Sandström E, Nordström G. The two-year impact of first generation protease inhibitor based antiretroviral therapy (PI-ART) on health-related quality of life. Health Qual Life Outcomes. 2005. May 4;3:32. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Burgoyne RW, Tan DH. Prolongation and quality of life for HIV-infected adults treated with highly active antiretroviral therapy (HAART): a balancing act. J Antimicrob Chemother. 2008. March;61(3):469–73. doi: 10.1093/jac/dkm499. Epub 2008 Jan 3. [DOI] [PubMed] [Google Scholar]
  • 48.St Wright, Carr A, Woolley I, et al. CD4 cell responses to combination antiretroviral therapy in patients starting therapy at high CD4 cell counts. J Acquir Immune Defic Syndr (1999). 2011;58(1):72–79. doi: 10.1097/QAI.0b013e318225ba62. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.HIV-CAUSAL Collaboration, Cain LE, Logan R, Robins JM, Sterne JA, Sabin C, Bansi L, et al. When to initiate combined antiretroviral therapy to reduce mortality and AIDS-defining illness in HIV-infected people in developed countries: an observational study. Ann Intern Med. 2011. April 19;154(8):509–15. doi: 10.7326/0003-4819-154-8-201104190-00001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Panel on Antiretroviral Guidelines for Adults and Adolescents. Guidelines for the use of antiretroviral agents in HIV-1-infected adults and adolescents. Department of Health and Human Services. Accessed December 20, 2015 at http://aidsinfo.nih.gov/contentfiles/lvguidelines/AdultandAdolescentGL.pdf. [Google Scholar]
  • 51.Keller SC, Yehia BR, Momplaisir FO, Eberhart MG, Share A, Brady KA. Assessing the Overall Quality of Health Care in People Living with HIV in an Urban Environment. AIDS Patient Care STDS. 2014;28(4):198–205. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.Sullivan PS, Denniston M, Mokotoff E, Buskin S, Broyles S, McNaghten AD. Quality of care for HIV infection provided by Ryan White Program-supported versus Non-Ryan White Program-supported facilities. PLoS One. 2008;3(9):e3250. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.National Institute of Allergy and Infectious Disease. Starting Antiretroviral Treatment Early Improves Outcomes for HIV-Infected Individuals. NIH News. May 27, 2015. Accessed December 20, 2015 at http://www.niaid.nih.gov/news/newsreleases/2015/Pages/START.aspx. [Google Scholar]

RESOURCES