Impact of male sex worker peer education on condom use in Mombasa, Kenya

Abstract

Background: Over 700 male sex workers are estimated to be actively seeking male clients in Mombasa, Kenya. A 2006 baseline survey of 425 male sex workers found poor knowledge of anal HIV transmission and low consistent condom use with male clients. Methods: After the baseline survey, interventions were implemented including: (a) the opening of a drop-in centre where male sex workers could receive HIV information and testing, (b) the training of 40 peer educators trained in prevention and basic counseling skills, and (c) the distribution of condoms and water-based lubricants. After 14 months, a follow-up behavioural survey of 442 male sex workers was conducted. Pre- and post-intervention measurements were compared, and multivariate logistic regression was used to determine intervention impact on consistent condom use with male clients in the past 30 days. Results: Male sex workers reported improvements in correct knowledge of anal HIV transmission (65% to 73%, P< 0.01) and consistent condom use with male clients in the past 30 days (36% to 50%, P< 0.01). In the follow-up survey, 33% of male sex workers reported exposure to peer educators in the Mombasa area. In multivariate analysis, increasing exposure to peer educators (range: 0 to 5+ yearly contacts) was a key predictor of consistent condom use (OR: 1.14, 95%CI=1.03-1.27, P=0.01), as was having ever been counseled or tested for HIV (OR: 1.70, 95%CI=1.11-2.60, P=0.01). Visiting a drop-in centre failed to be a predictor of consistent condom use (OR: 1.07, 95%CI=0.68-1.69, P=0.77) Conclusions: Exposure to peer education increased consistent condom use among male sex workers in Mombasa. Further strategies to increase peer educator coverage and effectiveness are recommended, including: (a) utilization of peers as outreach HIV counselors, (b) involvement in mobile testing, and (c) increasing referrals to friendly community testing centres. Drop-in centre attendance was not associated with condom use in the multivariate model.