There are 40,000 new HIV infections annually in the U.S., with 70% among men who have sex with men (MSM). Antiretroviral preexposure prophylaxis (PrEP) is up to 99% effective in preventing HIV infection, but uptake has been limited in groups with the highest HIV incidence, such as young, Black, and Hispanic MSM. Scale-up of PrEP in high-risk groups is now a national priority, with the National HIV/AIDS Strategy targeting a 25% reduction in HIV diagnoses by 2020. Thus, NIH has funded the development of interventions to increase PrEP uptake in high-risk groups. However, early data suggest these interventions may yield only modest (?30%) increases in PrEP uptake. The effects of modest increases in PrEP uptake on HIV incidence are unknown. There is also widespread concern that PrEP may decrease condom use, thereby driving rates of bacterial sexually transmitted infections (STIs) among MSM. Estimates of the population-level effects of increasing PrEP uptake on HIV and STIs are needed to 1) determine if realistic increases in PrEP uptake will be sufficient to reduce HIV incidence to target levels, and 2) inform allocation of resources toward efforts to minimize increases in STIs. Ideally, estimates of the effects of increases in PrEP uptake on HIV and STIs would come from randomized trials, but conducting trials with large enough sample sizes in each arm to detect effects on HIV and STIs would be cost-prohibitive. Therefore, estimates must come from observational data from PrEP delivery settings, using analytic methods that can control for time-varying confounders of the effects of PrEP on HIV and STIs, such as rectal STI testing, that are, themselves, affected by past PrEP use. Inverse probability (IP) weighting can handle such confounding, but popular IP weighted methods target effects of deterministic interventions (either 0% or 100% uptake) and cannot accommodate more realistic increases in PrEP uptake (e.g., ?30%). Interventions that result in some but less than 100% PrEP uptake are random dynamic interventions. Using data collected in clinical practice, we will refine novel IP weighted methods designed for realistic, random dynamic interventions to estimate the effects of a range of increases in PrEP uptake in high-risk MSM, relative to observed levels, on incidence of HIV infection (Aim 1) and gonorrhea, chlamydia, or syphilis (Aim 2) in the overall population of MSM. We will use data from patients receiving care during 2012-2019 at 3 sites in Massachusetts: Boston Medical Center, a safety net hospital with the highest HIV diagnosis rate in New England; Atrius Health, an ambulatory practice group caring for 10% of the state?s residents; and Fenway Health, an LGBT-focused community health clinic. The expected outcome of this study is an improved understanding of the increases in PrEP uptake needed to reach or even exceed the target 25% reduction in HIV diagnoses, and the impact of these increases on STIs. By providing benchmarks for success for interventions to increase PrEP uptake, our results will inform decisions about which interventions to scale up for maximal impact, as well as decisions about allocation of resources toward efforts to control STIs.
Efforts are underway to scale up preexposure prophylaxis (PrEP) to prevent HIV in men who have sex with men, but it is unknown how much PrEP uptake will need to increase to substantially reduce the number of new HIV infections. We will use longitudinal data collected from three real-world PrEP delivery settings to estimate the increases in PrEP uptake needed to reduce new HIV infections to target levels among men who have sex with men. Because PrEP may have the unintended consequence of reducing condom use, thereby increasing rates of other sexually transmitted infections, we will also estimate effects of increasing PrEP uptake on gonorrhea, chlamydia, and syphilis in this population.
