Understanding recruitment: outcomes associated with alternate methods for seed selection in respondent driven...

Area of Science:

• Epidemiology
• Social Sciences
• Biostatistics

Background:

• Respondent driven sampling (RDS) is a method for sampling hidden populations, aiming for unbiased estimates.
• Concerns exist regarding the generalizability of RDS estimates to the broader population.
• This study investigates how seed selection methods influence recruitment and RDS measures.

Purpose of the Study:

• To compare the influence of two distinct seed selection methods on Respondent driven sampling (RDS) outcomes.
• To assess the impact of seed selection on recruitment patterns and population estimates.
• To examine differences in sociodemographic and HIV-related variables between recruitment groups.

Main Methods:

• Two seed groups were established: one with study staff-selected seeds and another with self-presenting seeds.
• Recruitment chains were initiated from each seed group, and RDS estimates were compared.
• Sociodemographic variables, risk behaviors, homophily, and HIV prevalence were analyzed across groups.

Main Results:

• Significant differences were observed in sociodemographic and risk behaviors across the three analytic groups (staff-selected seeds, self-presenting seeds, and their recruits).
• Homophily values varied between seed groups, indicating differential recruitment patterns.
• RDS estimates of population proportions and HIV prevalence differed between recruitment arms, with some non-overlapping confidence intervals.

Conclusions:

• Seed selection significantly influences subgroup access within hidden populations during Respondent driven sampling (RDS).
• Employing multiple seed selection strategies may enhance access to diverse subgroups.
• Further research is needed to optimize RDS methodology for obtaining accurate and representative population estimates.