This article examines the effects of a multi-level, community-based HIV prevention intervention for crack users residing in low-income neighborhoods in San Salvador, El Salvador conducted between August 2011 and June 2016. The intervention consisted of three components introduced sequentially: (1) rapid HIV testing in community settings; (2) a social network HIV testing intervention; and (3) small group interventions with crack users who were members of the same social network. The intervention was evaluated with an interrupted time series design in which we used respondent-driven sampling to conduct 7 cross-sectional surveys with crack users along a 3-4 month period for each assessment (total n = 1597). Results revealed a significant increase in exposure to the intervention over time with 50% of the participants reporting exposure to one or more of the three components. Getting an HIV test at the community site was associated with reductions in total times each individual had sex without a condom (p < 0.05) compared to those who had been exposed to no intervention components. Being referred by another crack user through the Social Network HIV intervention was also associated with reductions in total numbers of condomless sex (p < 0.05) The cumulative effect of being exposed to more than one intervention component was associated with reductions in total number of times individuals had condomless sex (p < 0.05). In spite of the high level of intervention reach and that self-reported exposure to intervention components was associated with lower sexual risk, reductions in sexual risk over time were not observed in the full sample, indicating that the penetration of HIV prevention components was not sufficient to produce population level change.
Keywords: Crack; El Salvador; HIV; People who use non-injection drugs.