Background: Effective interventions are needed to prevent acquisition of HIV infection in men who have sex with men. To date, no behavioural interventions specifically for this risk group have been tested with HIV infection as the primary outcome.
Methods: This multisite two-group randomised controlled phase IIb trial tested the efficacy of a behavioural intervention in preventing HIV infection among 4295 men who have sex with men. The experimental intervention consisted of ten one-on-one counselling sessions followed by maintenance sessions every 3 months. The standard condition was twice-yearly Project RESPECT individual counselling. Twice-yearly follow-up visits included testing for HIV antibody and assessment of behavioural outcomes.
Findings: The rate of acquisition of HIV infection was 18.2% (95% CI -4.7 to 36.0) lower in the intervention group than the standard group. Adjustment for baseline covariates attenuated the intervention effect to 15.7% (-8.4 to 34.4). The effect was more favourable in the first 12-18 months of follow-up. The occurrence of unprotected receptive anal intercourse with HIV-positive and unknown-status partners was 20.5% (10.9 to 29.0) lower in the intervention than in the standard group.
Interpretation: The results from the primary analyses allow us to rule out that the experimental intervention is associated with a 35% lower rate of HIV acquisition than in the standard group. The overall estimate of a difference of 18.2%, more favourable estimates of effect in the first 12-18 months, and similar effects on risk behaviours suggest that prevention of HIV infection among men who have sex with men by a behavioural intervention is feasible. Further work should be done to develop more effective interventions.