Abstract
Advancing HIV behavioral science and its integration with recent biomedical prevention strategies-treatment as prevention and pre-exposure prophylaxis (PrEP)-is complicated because traditional definitions of sexual risk (condomless anal sex [CAS] or CAS with a serodiscordant partner) may not suffice given the success of biomedical prevention in the absence of condoms. To better understand how to define and understand risk in the era of biomedical HIV prevention, we assessed correlates of four different definitions of high-risk sexual behavior among a sample of 401 adult HIV-positive MSM (57% African American) residing in the New York City area: (1) a traditional model of risk defined as CAS with an HIV-negative or unknown serostatus partner; (2) a definition that considered participant viral load (VL); (3) a definition that considered partner PrEP use; and (4) a definition that considered participant HIV VL and partner PrEP use. Among participants reporting anal sex (n = 281) with a man in the past 3 months, 42% were considered high risk under Definition 1, 6% under Definition 2, 25% under Definition 3, and 2% under Definition 4. Correlates of risk varied across definitions. For example, while education was the strongest predictor of high-risk sexual behavior in Definitions 1 and 2, alcohol use and enacted stigma were the strongest predictors in the model that only accounted for partner PrEP use (Definition 3). These findings have implications for whom to reach for intervention efforts, the difficulty in meeting recruitment goals, and the types of interventions that should be used.