Background:  Through 2007, Thailand had an estimated 580,000 persons living with HIV/AIDS and 21,000 AIDS-related deaths. Clients of female sex workers may be at high risk for HIV infection because of risky sexual behaviors and can be a potential transmission source to the general population. A system to monitor HIV prevalence in this population needs to be developed; however, clients are difficult to access.

Methods:  We piloted a novel variation of respondent-driven sampling among female sex workers to recruit their clients in Bangkok. Respondent-driven sampling, a probability-based sampling method, uses a coupon system for peers to recruit their peers. Initial female sex workers, identified through community organizations, recruited peers who referred 1 of their clients and 3 other female sex workers. After providing informed consent, clients completed a standardized questionnaire using a handheld PC, provided an oral sample for HIV testing, and were asked to return for results after 2 weeks. Since there are no standardized data analysis techniques for surveys that utilize respondent-driven sampling to recruit sexual partners of participants, 2 weighted analysis methods were developed based on female sex workers’self-reported social network sizes and were calculated using Respondent-Driven Sampling Analytic Tool (RDSAT). All analyses incorporated RDSAT-generated weights for the dependent variable and were adjusted for survey design.

Results:  Among 549 female sex worker participants, 188 (34%) recruited 1 client. The median age was 38 years (interquartile range, 30 to 50 years). HIV prevalence ranged from 17 to 23% depending on data analysis method, but did not vary independently and significantly by demographics, age at first sex, sexually transmitted disease symptoms, number of partners, and sex with males. Among HIV⁺ clients, 26% were married, 69% used condoms with their regular partners, and 66% with female sex workers. There were 59 serodiscordant client-female sex worker couples and 81% reported condom use during the last sexual encounter.

Conclusions:  High HIV prevalence observed in clients and inconsistent condom use among serodiscordant couples can promote transmission of HIV infection to the general population. Though prevention programs for female sex workers are ongoing in Bangkok, these findings highlight the importance of initiating surveillance and targeted programs for clients of female sex workers and demonstrates a recruitment method for hard-to-reach populations.