Despite a population-level decline in HIV incidence in recent years, HIV incidence increased in certain subgroups and geographic regions in the United States.1 The federal Ending the HIV Epidemic (EHE) initiative includes expansion of access to comprehensive HIV prevention approaches, particularly HIV pre-exposure prophylaxis (PrEP), for all individuals who could benefit2 in geographic areas with high HIV incidence.3 Emtricitabine/tenofovir disoproxil fumarate (FTC/TDF) and emtricitabine/tenofovir alafenamide (FTC/TAF) are federally approved medications for PrEP and can prevent HIV infection in at-risk individuals. Nevertheless, implementation of PrEP has been slow4 and inequitable,5–7 limiting the impact of PrEP on HIV incidence at a population level.8
Unmet need for PrEP is particularly high in Black and Hispanic individuals,6 who are at the highest risk of HIV acquisition. Safety-net health systems provided outpatient care to more than 80 million individuals in 2018,9 including ethnically and racially diverse communities that comprise high-priority groups for PrEP implementation. These settings are also a primary source of care for socioeconomically disadvantaged individuals, who have a high risk of HIV infection.10–12 Thus, the number of individuals who could benefit from PrEP in safety-net settings may be higher than in private practice settings, making safety-net health systems particularly important for PrEP implementation. Nevertheless, safety-net health systems may have limited resources to prioritize PrEP implementation. Little evidence is currently available about the frequency of PrEP prescribing in safety-net health systems, particularly in EHE priority regions. This evidence could inform resource allocation for PrEP implementation efforts and identify disparities in PrEP prescribing for future intervention development. We evaluated PrEP prescribing among potential PrEP candidates in a safety-net health system that serves a county designated as an EHE priority area.
JPS Health Network (JPS) is an urban safety-net health system that serves Tarrant County, Texas, which has more than 2 million residents.13 Safety-net health systems provide care for socioeconomically disadvantaged populations including individuals who have low income, are uninsured, or are underinsured. JPS is considered a core safety-net health system because of a legal mandate to provide care regardless of a person's ability to pay.14 In addition, the network is a member of America's Essential Hospitals,15 which is an organization of more than 300 hospitals and health systems across the United States with a mission to serve vulnerable populations. JPS includes a 578-bed academic teaching hospital, a Ryan White–funded HIV Clinic, and more than 40 satellite community health clinics, which include accredited Primary Care Medical Homes (PCMHs). A PCMH is a primary care model that uses an integrated team-based approach to provide patient-centered care.16–18 Tarrant County is an HIV hotspot with increasing HIV incidence19 and a priority area for targeted EHE efforts.3
We used data from electronic health records (EHRs) for patients receiving outpatient primary care within the network. Eligible individuals were aged 18 years or older; had ≥2 visits to a PCMH within the network between January 1, 2015, and December 31, 2019; and had indications for PrEP based on Centers for Disease Control and Prevention (CDC) guidelines20 (see Table S1, Supplemental Digital Content 1, https://links.lww.com/QAI/B697). PrEP indications were defined as a documented behavioral or sexual risk factor of HIV acquisition or a positive laboratory result for bacterial sexually transmitted infections (STIs) at the index visit date. STIs included chlamydia (men only), gonorrhea, or syphilis. The index clinic visit was defined as the first clinic visit at which the individual had a documented indication for PrEP. We excluded individuals who had a previous diagnosis of HIV infection. Individuals who received antiretroviral medications for postexposure prophylaxis, but not PrEP, were only included in the denominator. This study was approved by the North Texas Institutional Review Board.
Our primary outcome of interest was PrEP prescribing, defined as a new prescription for PrEP for potential PrEP candidates from a JPS PCMH at any time after eligibility criteria were met. We reviewed the charts of patients who received a prescription order for an approved formulation for PrEP. Both drug names (FTC/TDF and FTC/TAF) and brand names (Truvada and Descovy) were used to ascertain PrEP prescribing documented in the medication list and provider notes. To capture PrEP indications that were behavioral or sexual risk factors documented in the EHRs, we used the following ICD-10 codes that indicate exposure to STIs, substance use disorders, injection drug use, or “high-risk sexual behavior”: A53, Z20.6, Z20.2, Z72.51, Z72.52, Z72.53, F11, F14, F15 F19.20, W46.1, and W46.0 (see Table S1, Supplemental Digital Content 1, https://links.lww.com/QAI/B697). We extracted sociodemographic data, including age at the index visit (18–24, 25–34, 35–44, 45–54, 55–64, or 65 years or older), race and ethnicity (Hispanic, non-Hispanic Black, non-Hispanic White, and non-Hispanic other), and sex (men or women). Health insurance coverage was defined as the primary payer for the eligible visit and categorized as commercial, public (Medicare or Medicaid), uninsured, or hospital-based managed care plan (offered to eligible individuals who are uninsured).
We estimated cumulative incidence and 95% confidence limits (CLs) of PrEP prescribing based on failure probability (ie, 1—Kaplan–Meier estimate). Person-time was computed as time from the index date (date of encounter associated with first PrEP indication) to the prescription order date, administrative censoring at 365 days, or end of study, whichever occurred first. Individuals with person-time <365 days were censored at the date of their last encounter. We designated 1 day of person-time contribution to the cohort for patients who received a prescription on the same day as the index date. This approach assumes no competing risks (eg, death) for accurate estimation of cumulative incidence and is generally accurate when the frequency of deaths is <10%. Given the 1-year follow-up period, we would not expect to violate this assumption. We originally intended to assess potential disparities by age, sex, and race/ethnicity, but the low outcome frequency precluded analyses to estimate associations.
For a common measure to compare with previous studies, we estimated PrEP prevalence by emulating a repeated cross-sectional survey using EHR. In brief, we applied the same eligibility criteria as used for the primary analysis. For eligible individuals with a primary care encounter within each calendar year, we applied a historical lookback period for each individual to identify whether PrEP was prescribed. Data from each calendar year were pooled to estimate prevalence across all years. We accounted for individuals being eligible for multiple calendar years by using clustered variance to estimate 95% CL for prevalence.
We identified 2957 individuals without a previous HIV diagnosis who had least 2 encounters at a JPS PCMH from 2015 through 2019 and had a documented indication for PrEP. Table 1 summarizes characteristics of the study population, which was predominantly young (58% aged younger than 45 years), women (56%), racial or ethnic minorities (67%), and uninsured or provided care through a hospital-based managed care plan for uninsured individuals (60%).
TABLE 1. -
Characteristics of 2957 Individuals Who Engaged in Primary Care and had Documented Indications for HIV Pre-exposure Prophylaxis
(PrEP) at an Urban Safety-Net Health System, 2015–2019
| ≥ 65
| Non-Hispanic White
| Non-Hispanic Black
| Commercial insurance
| Hospital-based managed care plan
|Indication for PrEP
| Sexually transmitted infection
| Drug use disorders*
| Diagnosis code for “high-risk sexual behavior”†
*Includes diagnosis for drug dependence, methamphetamine, and cocaine use.
†Includes diagnosis for “high-risk sexual behavior” (n = 312); unprotected sex (n = 130); and contact with or exposure to sexually transmitted infection (n = 1278).
We identified 41 individuals who were prescribed PrEP. Among individuals prescribed PrEP, most were aged 25–45 years (61%), men (68%), and racial or ethnic minorities (62%), and nearly half were receiving charity care (42%). In addition, half (50%) of the individuals prescribed PrEP had “high-risk sexual behavior” as a PrEP indication, with men accounting for most of the (74%) individuals with “high-risk sexual behavior” as an indication for PrEP (see Table S2, Supplemental Digital Content 1, https://links.lww.com/QAI/B697). Figure 1 illustrates the cumulative incidence of PrEP prescribing among individuals with PrEP indications. The cumulative incidence of receiving a PrEP prescription was 0.76% (95% CL: 0.50% to 1.2%) within 90 days and 1.3% (95% CL: 0.91% to 1.7%) within 365 days of the first documented PrEP indication. The prevalence of PrEP prescribing in our population over the same period was 1.1% (0.70% to 1.6%) (see Table S3, Supplemental Digital Content 1, https://links.lww.com/QAI/B697).
Our results suggest extremely low frequency of PrEP prescribing among potential PrEP candidates engaged in primary care at an urban safety-net health system. Scaling up PrEP as part of the EHE initiative will require investments in strategies to increase prescribing in safety-net settings, which serve populations who are at disproportionately high risk of HIV infection and have been underserved by PrEP delivery efforts to date.
Several limitations should be considered when interpreting our findings. PrEP indications were based on behavioral or clinical indications documented in EHR data. Sexual behaviors are often underreported and may be sensitive to recall and social desirability biases.21–23 In addition, the ascertainment of behavioral and clinical indications for PrEP is highly dependent on the patient–provider relationship.24 Misclassification of behavioral or clinical indications would manifest as selection bias in our study because these individuals would have been excluded. Misclassification may also have influenced our findings. For example, misclassification of PrEP prescribing is a consideration because we were unable to determine whether patients received a PrEP prescription at another institution, which would result in underestimated cumulative incidence of PrEP prescribing. Our outcome of interest was PrEP prescribing, which aligned with our goal of understanding the provider and health system perspective but does not address PrEP uptake (ie, actual use), which would provide more insight from a patient perspective. Finally, our study focused on people engaged in primary care, and our results may not be generalizable to all people eligible for PrEP-provided care through our health care system (eg, people who only received care in the emergency department).
Our estimates were lower than the few previous studies conducted in urban safety-net settings, which identified frequencies of PrEP prescribing between 13% and 96%.25–29 Several sources of variation may explain the lower estimate in our study. Foremost, most previous studies in safety-net settings25–27 included data from community health centers that were early adopters of PrEP and specialize in care for sexual and gender minorities. In addition, sexual and gender minorities were the most patients in previous studies,25–29 with several studies including primarily non-Hispanic White individuals, for whom PrEP uptake has been the highest. In contrast, although we were unable to identify gender identity or sexual orientation in our study population, our health care system serves an urban population of primarily racial and ethnic minorities, and a minority of our study sample was non-Hispanic White. Finally, we estimated cumulative incidence, whereas most previous studies estimated prevalence. Our study, thus, provides insight into the frequency of new PrEP prescriptions among potential PrEP candidates, whereas studies that reported prevalence provide insight into PrEP prescribing that could be current or past. Nevertheless, for comparison, we estimated the prevalence of PrEP prescribing in our population. Our prevalence estimate was also low (1.1%, 95% CL: 0.70%–1.6%).
The low frequency of PrEP prescribing in our setting may be related to multiple challenges. Barriers to accessing expensive medications such as PrEP may be exacerbated in our study population because Texas is one of the 12 states, primarily in the southern United States, which did not expand Medicaid as part of the Affordable Care Act.30
Providers could be hesitant to prescribe PrEP because of concerns about cost,31 considering that 60% of our study population was uninsured or provided care through a hospital-based managed care plan for uninsured individuals. Patient assistance programs to enhance PrEP access may inadequately address costs because low health literacy creates challenges in navigating the complex application process.32 In addition, low provider knowledge of PrEP is a barrier to PrEP prescribing in safety-net settings.33 Finally, safety-net systems operate with limited financial and staffing resources, which may constrain the ability to include PrEP as an essential medication on formularies,34 hire PrEP navigators, or implement new innovations in care delivery.35
In summary, our results suggest very low cumulative incidence of PrEP prescribing among potential PrEP candidates in an urban safety-net health system. PrEP expansion in health care systems will require investment in strategies to overcome patient-level, provider-level, and system-level barriers to PrEP use in high-priority populations. For example, evidence is emerging that negative attitudes, concerns about cost, and lack of PrEP awareness, knowledge, and skills are provider-level barriers to prescribing PrEP.31,33 Greater understanding of barriers that may be unique to safety-net settings and development of interventions to mitigate these barriers could help improve PrEP implementation in high-priority populations.
1. Centers for Disease Control and Prevention
. CDC Data Confirm: Progress in HIV Prevention
Has Stalled. Available at: https://www.cdc.gov/media/releases/2019/0227-hiv-prevention-stalled.html
. Accessed March 20, 2020.
2. Giroir BP. The time is now to end the HIV
epidemic. Am J Public Health. 2020;110:22–24.
3. Fauci AS, Redfield RR, Sigounas G, et al. Ending the HIV
epidemic: a plan for the United States. JAMA. 2019;321:844–845.
4. Zhang C, McMahon J, Fiscella K, et al. HIV pre-exposure prophylaxis implementation
cascade among health care professionals in the United States: implications from a systematic Review and meta-analysis. AIDS Patient Care STDS. 2019;33:507–527.
5. Sullivan PS, Giler RM, Mouhanna F, et al. Trends in the use of oral emtricitabine/tenofovir disoproxil fumarate for pre-exposure prophylaxis
infection, United States, 2012-2017. Ann Epidemiol. 2018;28:833–840.
6. Huang YA, Zhu W, Smith DK, et al. HIV
preexposure prophylaxis, by race and ethnicity - United States, 2014-2016. MMWR Morb Mortal Wkly Rep. 2018;67:1147–1150.
7. Marcus JL, Hurley LB, Hare CB, et al. Disparities in uptake of HIV
preexposure prophylaxis in a large integrated health care system. Am J Public Health. 2016;106:e2–3.
8. Marcus JL, Paltiel AD, Walensky RP. Has pre-exposure prophylaxis
made a difference at a population level? Jury is still out. Clin Infect Dis. 2020;71:3152–3153.
9. Clark D, Roberson B, Ramiah K. Essential Data: Our Hospitals, Our Patients. Essential Hospitals Institute; 2021. Available at: https://essentialhospitals.org/wp-content/uploads/2021/05/EssentialData2021_Web.pdf
10. Kogan SM, Cho J, Barnum S, et al. Pathways to HIV
-related behavior among heterosexual, rural Black men: a person-centered analysis. Arch Sex Behav. 2017;46:913–924.
11. Pellowski JA, Kalichman SC, Matthews KA, et al. A pandemic of the poor: social disadvantage and the U.S. HIV
epidemic. Am Psychol. 2013;68:197–209.
12. Medicine Io. America's Health Care Safety Net
: Intact But Endangered. The National Academies Press; 2000:302.
13. Bureau USC. United States Census Bureau QuickFacts Tarrant County, Texas. Vol. 2020; 2020. Available at: https://www.census.gov/quickfacts/tarrantcountytexas
. Accessed March 10, 2020.
14. Altman S, Lewin ME. America's Health Care Safety Net
: Intact but Endangered. National Academies Press; 2000.
15. America's Essential Hospitals. Members of America's Essential Hospitals. Available at: https://essentialhospitals.org/about-americas-essential-hospitals/listing-of-americas-essential-hospitals-members/
. Accessed June 12, 2021.
16. Defining the PCMH. Agency for Healthcare Research and Quality. Available at: https://pcmh.ahrq.gov/page/defining-pcmh
. Accessed June 6, 2021.
17. Jackson G, Powers B, Chatterjee R, et al. The patient-centered medical home. Ann Intern Med. 2013;158:169; undefined.
18. Rosenthal TC. The medical home: growing evidence to support a new approach to primary care. J Am Board Fam Med.. 2008;21:427–440.
19. Texas Department of State Health Services. Texas HIV
Surveillance Report, 2018 Annual Report. Texas Department of State Health Services. 2019. Available at: https://www.dshs.texas.gov/hivstd/reports/
20. Centers for Disease Control and Prevention
. US Public Health Service: Preexposure Prophylaxis for the Prevention
Infection in the United States—2017 Update: A Clinical Practice Guideline. Available at: https://www.cdc.gov/hiv/pdf/risk/prep/cdc-hiv-prep-guidelines-2017.pdf
. Accessed June 16, 2021.
21. Catania JA, Gibson DR, Chitwood DD, et al. Methodological problems in AIDS behavioral research: influences on measurement error and participation bias in studies of sexual behavior. Psychol Bull. 1990;108:339–362.
22. Haberer JE, Bangsberg DR, Baeten JM, et al. Defining success with HIV pre-exposure prophylaxis
: a prevention
-effective adherence paradigm. AIDS. 2015;29:1277–1285.
23. Underhill K, Morrow KM, Colleran CM, et al. Access to healthcare, HIV
/STI testing, and preferred pre-exposure prophylaxis
providers among men who have sex with men and men who engage in street-based sex work in the US. PLoS One. 2014;9:e112425.
24. Maloney KM, Krakower DS, Ziobro D, et al. Culturally competent sexual healthcare as a prerequisite for obtaining preexposure prophylaxis: findings from a qualitative study. LGBT Health. 2017;4:310–314.
25. Hojilla JC, Vlahov D, Crouch PC, et al. HIV pre-exposure prophylaxis
(PrEP) uptake and retention among men who have sex with men in a community-based sexual health clinic. AIDS Behav. 2018;22:1096–1099.
26. Marcus JL, Levine K, Grasso C, et al. HIV
preexposure prophylaxis as a gateway to primary care. Am J Public Health. 2018;108:1418–1420.
27. Shover CL, Javanbakht M, Shoptaw S, et al. HIV
preexposure prophylaxis initiation at a large community clinic: differences between eligibility, awareness, and uptake. Am J Public Health. 2018;108:1408–1417.
28. Caponi M, Burgess C, Leatherwood A, et al. Demographic characteristics associated with the use of HIV pre-exposure prophylaxis
(PrEP) in an urban, community health center. Prev Med Rep. 2019;15:100889.
29. Pyra M, Rusie LK, Baker KK, et al. Correlations of HIV
preexposure prophylaxis indications and uptake, chicago, Illinois. Am J Public Health. 2020;110:370–377.
30. Kaiser Family Foundation. Status of State Medicaid Expansion Decisions. Interactive Map; 2020.
31. Pleuhs B, Quinn KG, Walsh JL, et al. Health care provider barriers to HIV pre-exposure prophylaxis
in the United States: a systematic Review. AIDS Patient Care STDS. 2020;34:111–123.
32. Marcus JL, Hurley LB, Dentoni-Lasofsky D, et al. Barriers to preexposure prophylaxis use among individuals with recently acquired HIV
infection in Northern California. AIDS Care. 2019;31:536–544.
33. Agovi AM, Anikpo I, Cvitanovich MJ, et al. Knowledge needs for implementing HIV pre-exposure prophylaxis
among primary care providers in a safety-net health system. Prev Med Rep. 2020;20:101266.
34. Adams D, Wilson AL. Drug selection for safety-net-provider formularies. Am J Health Syst Pharm. 2002;59:1675–1678.
35. Lyles CR, Handley MA, Ackerman SL, et al. Innovative implementation
studies conducted in US safety net
health care settings: a systematic Review. Am J Med Qual. 2019;34:293–306.