Abstract / Summary
BackgroundRespondent-driven sampling (RDS) is a widely used peer-recruitment strategy to study hard-to-reach populations including people who use drugs. South Africa continues to experience increasing illicit substance use alongside high burdens of tuberculosis (TB) and HIV. Little is known about the factors that influence RDS recruitment success in these settings, especially in rural areas.
MethodsA total of 750 people who smoke drugs (PWSD) were recruited through RDS in a rural town in the Western Cape, South Africa to estimate TB and HIV prevalence. Participant characteristics were compared by recruitment success, defined as distributing at least one study coupon that resulted in peer enrollment, to identify factors associated with successful recruitment.
Recruitment homophily was assessed to quantify within-group recruitment, and the stabilization of selected participant characteristics across recruitment waves were visualized to assess sample diversity over time.
ResultsRecruitment success was more common among unemployed participants (p=0.032), those without an incarceration history (p=0.053), current cannabis users (p=0.086), and participants with moderate-to-severe methamphetamine (p=0.019) or methaqualone (p=0.014) use. High within-group recruitment was observed by neighborhood of residence and racial population group. The proportion of males, individuals with incarceration history, and participants with TB or HIV stabilized across recruitment waves, regardless of seed characteristics.
ConclusionsRecruitment dynamics in this setting were strongly influenced by social and geographic factors, including neighborhood, population group, incarceration history, and substance use type and severity. Identifying characteristics associated with recruitment success may inform more efficient peer-recruitment strategies, such as targeted coupon distribution. These findings support the use of RDS as an effective approach for engaging PWSD in research.