Accessing a diverse sample of injection drug users in San Francisco through respondent-driven sampling

Injection drug users (IDU) are the second most affected population by HIV in San Francisco and the United Stated after men who have sex with men (MSM). Behavioral surveillance data that include the diversity of the population at risk are necessary to develop effective programs for IDU.

We conducted a cross-sectional behavioral survey of IDU using respondent-driven sampling (RDS) in San Francisco. The present analysis focuses the performance of the sampling method in reaching the diversity of the population as a pre-requisite for representative data.

Over 32 weeks, 571 eligible IDU were recruited, of whom 477 (83.5%) with complete records were included in analysis.

The age range was 18-70 years, with 36% age 50 years or older. The majority (56%) were homeless. Male, MSM, African-Americans and Non-Hispanic Whites comprised 71%, 28%, 36% and 35% of IDU, respectively. Twenty-two percent had "ever shared needles in the past 12 months," and 57% reported that they had "shared drugs" in the past 12 months. Peer referral chains were able to cross-recruit IDU by diverse demographic characteristics, drug use related behaviors, program access and use, and other factors relevant to reaching and conducting prevention research on this population.

RDS appears to be an effective sampling tool that reaches diverse populations of IDU, including many who may be missed by drug treatment and HIV prevention services in San Francisco and potentially in other urban areas.