Long-chain peer referral of men who have sex with men: a novel approach to establish and maintain a cohort to measure HIV incidence, Nanjing, China.

Background: HIV prevalence is high among men who have sex with men (MSM) in China. Longitudinal studies are needed to measure HIV incidence, determine causes of acquisition, and test prevention interventions. We described 2 novel methods to enroll and maintain a cohort of MSM, focusing on their abilities to establish a diverse sample, improve retention, and their impact on HIV incidence.

Methods: Employing methods based on respondent-driving sampling, we constructed 2 parallel cohorts measuring HIV incidence over 2 years through 6-month follow-up visits. An initial cohort was constructed using long-chain peer referral from the community; a second phase comprised open cohort peer referral recruitment from MSM completing follow-up.

Results: Three-hundred ninety-seven HIV-negative MSM were enrolled in the initial cohort; 460 were recruited in the open cohort phase. Across recruitment waves, the composition of the cohort was stable. Among initial participants, retention was 72.0%, 68.8%, 49.9%, and 44.8%. Retention was lower in the open cohort. MSM retained in both phases were less risky than those lost to follow-up. HIV incidence was 3.36 per 100 person-years and did not differ by recruitment method.

Conclusions: Our approach efficiently recruited MSM into longitudinal studies with modest improvement in sample diversity from initial recruits. We perceive multiple means to improve diversity and follow-up capitalizing on the network bonds between recruits and through secondary incentives for assistance with referral and retention. Meanwhile, HIV incidence is high among MSM in Nanjing, and interventions need to be developed and tested in longitudinal randomized controlled trials.