Background: HIV prevalence is 15 times higher among female sex workers (FSWs) than in the overall Brazilian female population. In this article, we investigated factors associated to HIV infection among FSW, recruited by Respondent-driven sampling (RDS), in 10 Brazilian cities in 2008-2009.
Method: The total sampling was 2523 FSWs. The variables considered in the study included sociodemographic characteristics and those related to profession, knowledge, and source of information on sexually transmitted infection/AIDS, access to condoms, sexual behavior, health care and sexually transmitted infection signs, discrimination and violence, and the use of alcohol and drugs. Sampling weights were proportional to the inverse of network size and for the generation of global results for the 10 cities, data were calibrated by the population size. Multivariate logistic regression models were used, taking into account the dependency structure of observations.
Results: The longest period of prostitution (odds ratio [OR], 1.040; 95% confidence interval [CI], 1.013-1.067), the lowest price charged for sex (OR, 0.713; 95% CI, 0.522-0.970), the association with other sexually transmitted infections such as syphilis (OR, 2.186; 95% CI, 1.064-4.488), and the fact of would waive the use of condom on client's request (OR, 3.735; 95% CI, 1.449-9.661) were the most important associated risk factors to HIV infection among FSWs. Purchase of condoms (OR, 0.503; 95% CI, 0.304-0.833) was a relevant preventive factor.
Conclusions: The use of RDS proved to be appropriate for the recruitment of FSWs as long as the statistical analysis takes into account the dependency structure of observations. The HIV infection predictors here depicted can support public health policies focused on this population group in Brazil.