Abstract
INTRODUCTION: In South Africa, the rate of HIV in the sex worker (SW) population is exceedingly high, but critical gaps exist in our understanding of SWs and the factors that make them vulnerable to HIV. This study aimed to estimate HIV prevalence among female sex workers (FSWs) in Soweto, South Africa, and to describe their sexual behavior and other factors associated with HIV infection. METHODS: A cross-sectional, respondent-driven sampling (RDS) recruitment methodology was used to enroll 508 FSWs based in Soweto. Data were collected using a survey instrument, followed by two HIV rapid tests. Raw and RDS adjusted data were analyzed using a chi-squared test of association and multivariate logistic regression to show factors associated with HIV infection. FINDINGS: HIV prevalence among FSWs was 53.6% (95% CI 47.5-59.9). FSWs were almost exclusively based in taverns (85.6%) and hostels (52.0%). Less than a quarter (24.4%) were under 25 years of age. Non-partner violence was reported by 55.5%, 59.6% of whom were HIV-infected. Advancing age, incomplete secondary schooling, migrancy and multiple clients increased the likelihood of HIV acquisition: >30 years of age was associated with a 4.9 times (95% CI 2.6-9.3) increased likelihood of HIV; incomplete secondary schooling almost tripled the likelihood (AOR 2.8, 95% CI 1.6-5.0); being born outside of the Gauteng province increased the likelihood of HIV 2.3 times (95% CI 1.3-4.0); and having more than five clients per day almost doubled the likelihood (AOR 1.9, 95% CI 1.1-3.2). CONCLUSION: Our findings highlight the extreme vulnerability of FSWs to HIV. Advancing age, limited education and multiple clients were risk factors associated with HIV, strongly driven by a combination of structural, biological and behavioral determinants. Evidence suggests that interventions need to be carefully tailored to the varying profiles of SW populations across South Africa. Soweto could be considered a microcosm of South Africa in terms of the epidemic of violence and HIV experienced by the SW population, which is influenced by factors often beyond an individual level of control. While describing a hitherto largely undocumented population of FSWs, our findings confirm the urgent need to scale up innovative HIV prevention and treatment programs for this population.