Background Continued increases in HIV among men who have sex with men (MSM) underscore the need for intervention strategies that target those at highest risk of infection.
Methods January–December 2010, we collected information on demographic and behavioural risks in the past 12 months from MSM seeking anonymous HIV testing. We examined risks for HIV infection and calculated population attributable fractions (PAFs) to identify relative contributions of exposures to overall infection.
Results Overall, 81 (2.7%) newly diagnosed infections were identified among 3045 men. Men were median age 31, 64% white, 9% black, 17% Hispanic, and 10% other race/ethnicity. 14% had been diagnosed with an STI in the past year. Among clients for whom behavioural data were available (98%), black race, STI history, receptive anal intercourse (RAI), not always using condoms for RAI, ≥3 partners for RAI, methamphetamine use, sex with an HIV positive partner, and sex with a partner of unknown serostatus were associated with an increased odds of HIV infection in univariate analysis. The univariate association between methamphetamine use and HIV infection was partially mediated by sexual risk behaviour. In multivariable logistic regression, black race (OR, 1.7; 95% CI 1.2 to 2.4), STI history (OR, 2.0; 95% CI 1.1 to 3.6), not always using condoms for RAI (OR, 2.6; 95% CI 1.5 to 4.6), and RAI with ≥3 partners (OR, 2.2; 95% CI 1.3 to 3.9) were significantly associated with HIV infection. Adjusted PAFs were 13.7% (95% CI −1.4 to −26.6) for STI history; 36.7% (95% CI 12.2 to 54.4) for not always using condoms for RAI; and 28.8% (95% CI 5.5 to 46.3) for ≥3 RAI partners. The total combined PAF for these factors adjusted for race/ethnicity was 58.8% (95% CI 28.5 to 72.0). While 81% of HIV-infected men reported at least one risk factor and 11% reported all three, overall, 51% of men screened had at least one of these factors: STI history (14%); not always using condoms for RAI (32%); and ≥3 RAI partners (27%).
Conclusions STI history, inconsistent condom use, and ≥ 3 sex partners for RAI accounted for 59% of new HIV infections, but were present in half those tested. While we identified behaviours for intervention content, we did not identify sub-groups to target. Interventions that address condom use efficacy and reducing numbers of partners for RAI, including the effect of substance use on sexual decision making, should be considered for men reporting these risks.