Paper Session 7: HIV/AIDS

Background
People who inject drugs (PWID) remain at substantial risk of HIV acquisition, and pre-exposure prophylaxis (PrEP) is critical for HIV prevention. PrEP’s efficacy was demonstrated a decade ago using daily oral tenofovir disoproxil fumarate (TDF) with emtricitabine (FTC). In 2019, co-formulated tenofovir FTC and alafenamide fumarate (TAF) became the second approved daily oral PrEP regimen. However, FTC/TAF is only approved for the prevention of sexually transmitted HIV, excluding individuals at risk of receptive vaginal sex, including PWID. This study explored FTC/TAF feasibility and acceptability for daily oral PrEP among PWID.

Methods
This single-arm, observational, open-label study enrolled 100 PWID from New Haven, Connecticut, to receive FTC/TAF for daily oral HIV prevention. Inclusion criteria were: age ≥18, HIV negative, injection drug use (in the past 6 months), and meeting DSM-V criteria for opioid dependence. Participants meeting clinical criteria received a 90-day supply of FTC/TAF from a community-based syringe services program (SSP). Behavioral and biomedical data were collected at baseline, 3, and for 6 months. Descriptive statistics were used to estimate feasibility, acceptability, side-effects, adherence, and persistence on FTC/TAF.

Results
Participants were mostly male (63%) and non-Hispanic White (52%), with a mean age of 44.4 (SD=9.9). Participants reported injecting drug behavior at least once per day (37%) and engagement in condomless sex (78%) in the past 6 months. Prior use of daily oral PrEP was reported by 21% of participants (TDF/FTC: 14%; FTC/TAF: 3%; unknown: 4%). Although all participants were prescribed FTC/TAF, only 60.0% picked up the drug. Of those, 70% picked up once, 26.7% twice, and only 3.3% at all follow-up visits. Self-reported adherence was high (>90%) across all time points but discordant with urine-based quantification of FTC. Acceptability was high (range: 8-32): 3-months [24.9 (±3.1)]; 6-months [24.4 (±3.4)]. The most frequently reported side effects were tiredness (18.4%) and nausea (17.2%). There were no HIV seroconversions.

Conclusion
FTC/TAF PrEP was positively received among opioid-dependent PWID. Implementation through SSP services was feasible and acceptable, suggesting its viability as an HIV prevention tool for this vulnerable group. Poor adherence, as indicated in earlier studies of PWID using TDF/FTC PrEP, emphasizes the need for enhanced adherence counseling specifically tailored for PWID.

Presenter - Kamal Gautam, MPH,
University of Connecticut

Co-Author - Kiran Paudel, BPH, BPH,
University of Connecticut

Co-Author - Jeffrey A Wickersham, PhD,
Yale School of Medicine

Co-Author - Giselle Bellia,
Yale School of Public Health

Co-Author - Frederick L Altice,
Yale School of Medicine

Co-Author - Luzan JadKarim, MPH, MPH,
Yale School of Medicine

Co-Author - Zeyan Liew,
Yale School of Public Health

Co-Author - Antoine Khati, MD,
University of Connecticut

Co-Author - Michael Copenhaver,
UConn

Co-Author - Roman Shrestha, PhD, MPH, PhD, MPH,
University of Connecticut

Background: Machine learning algorithms (MLAs) are used in multiple sectors, but their application to HIV implementation has been limited. CCMP, a community-based Health Home serving one of the largest and most vulnerable populations of people with HIV (PWH) in New York City, has developed and validated an MLA that reliably predicts which PWH in its system will visit the Emergency Department (ED) in the next two weeks. This pilot study leveraged the MLA as an implementation strategy to guide just-in-time delivery of an established intervention (comprehensive care management; CCM) demonstrated to reduce ED visits and improve clinical outcomes for PWH. We hypothesized that targeting CCM immediately to patients flagged by the MLA would increase the intervention’s efficiency and effectiveness.

Methods: Care managers (CMs) at five agencies (PWH/agency range 209-807; total patients = 2425) were trained to enact the MLA-based implementation strategy for a 6-week pilot (Feb-Mar 2024). CMs received biweekly lists of patients flagged by the MLA, and were incentivized to apply standardized protocols for rapid engagement and intervention.

Results: Application of the MLA implementation strategy increased the number of PWH who received timely CCM, and timely receipt of CCM resulted in significantly fewer ED visits among PWH flagged by the MLA. During the MLA trial, an average of 76% of flagged PWH received CCM (per agency range 68%-85%), compared to a historical control of 53% (44% increase, p < .001). Only 16% of MLA-flagged clients who received CCM went to the ED, compared to 38% of flagged patients who did not receive just-in-time CCM (58% reduction) and 30% of historical controls (47% reduction), both ps <.01. Among PWH not flagged by the MLA, provision of CCM was associated with only a 1% reduction in ED visits (6% visited the ED, p =.09), suggesting that just-in-time CCM intervention matters most for PWH with imminent ED visit risk.

Discussion: To our knowledge, the CCMP MLA is the first validated, real-time algorithm to predict prospective ED visits in a community sample. These data provide evidence that an MLA-driven implementation strategy can enhance the effectiveness of an established, evidence-based intervention by optimizing timely and targeted provision to the patients who need it most. Future implementation research should explore the potential for MLA and other big-data approaches to enhance care for highest-need PWH and other patient populations.

Presenter - Sarit Golub, PhD, MPH,
Hunter College and Graduate Center of the City University of New York

Author - Nathan Ito-Prine, MHA,
Community Care Management Partners

Author - Tavin Weeda, MSDS,
Community Care Management Partners

Author - Carly Wolfer, MA,
Graduate Center of the City University of New York

Author - Melissa Fazzari, PhD,
Albert Einstein College of Medicine

Background: The United States has recently witnessed a dramatic increase in anti-LGBTQ+ legislation. These discriminatory laws are a form of structural stigma, and there is preliminary evidence that they may hinder HIV prevention efforts; however, the mechanisms of this pathway are not understood. We therefore examined whether individual-level internalized stigma mediated the relationship between state-level anti-LGBTQ+ policy climate and individual-level PrEP use in a national sample of sexual and gender minoritized (SGM) youth.

Methods: Keeping it LITE-1 was a prospective, national cohort of 3,330 SGM youth and young adults (ages 13-34) at risk of HIV from 2017-2022. Surveys collected geolocation, internalized HIV-related and sexual identity-related stigma (high/low), and self-reported PrEP use (yes/no). Geolocation was linked with state-level LGBTQ+ policy data from the Movement Advancement Project (MAP). LGBTQ+ laws were characterized as less or more discriminatory based on MAP score distributions. To assess causal mediation, we first estimated the average treatment effect (ATE) of the state-level LGBTQ+ policy environment on individual-level PrEP use, and then the controlled direct effect (CDE) if all participants had both lower internalized HIV- and sexual identity-related stigma. Results were stratified by race and ethnicity.

Results: If all participants lived in states with less discriminatory anti-LGBTQ+ policies, we would expect the prevalence of PrEP use to increase by 14%, compared to if all participants lived in states with more discriminatory policies (ATE: 0.14, 95% CI: 0.09,0.18). Setting all participants to have low internalized HIV- and sexual identity-related stigma did not substantively change this effect (CDE: 0.13, 95% CI: 0.09,0.18). The protective effect of a less discriminatory state policy climate on PrEP use was amplified among Black participants (ATE: 0.23, 95% CI: 0.05,0.38), but did not differ among other ethno-racial groups (Latine: 0.16, 95% CI: 0.06,0.25; non-Hispanic White: 0.11, 95% CI: 0.04,0.17; another marginalized race: 0.14, 95% CI: 0.03,0.24).

Conclusion: Living in states with less discriminatory anti-LGBTQ+ policies increased PrEP use among SGM youth, but internalized stigma did not mediate this effect. Protective LGBTQ+ policies may increase PrEP use and thereby decrease HIV incidence. However, intervening on individual-level stigma will not block the pathway between this structural-level stigma and PrEP use.

Author - Nicole Kelly, PhD, MPH,
University of California, San Diego

Co-Author - Shabbar Ranapurwala, PhD,
University of North Carolina at Chapel Hill

Co-Author - Brian Pence, PhD,
University of North Carolina at Chapel Hill

Co-Author - Lisa Hightow-Weidman, MD, MPH, MD, MPH,
Florida State University

Co-Author - Jaime Slaughter-Acey, PhD,
University of North Carolina at Chapel Hill

Co-Author - Audrey French, MD,
University of Illinois at Chicago

Co-Author - Sybil Hosek, PhD,
University of Illinois at Chicago

Co-Author - Audrey Pettifor, PhD,
University of North Carolina at Chapel Hill