Dissertation/Thesis Abstract

Evaluation of Respondent Driven Sampling (RDS) to recruit illicit drug users in New York City (2006–2009)
by Rudolph, Abby, Ph.D., The Johns Hopkins University, 2010, 159; 3463580
Abstract (Summary)

This research aims to 1) test the Respondent-Driven Sampling (RDS) assumptions in a drug-using sample and to compare concurrently recruited RDS and targeted street outreach (TSO) samples on demographics, drug use behaviors, and HIV status, 2) investigate the impact of respondent-driven recruitment compared with TSO on health-seeking behaviors and network composition over 6 months, and 3) identify individual-, study-, and neighborhood-level characteristics associated with successful peer recruitment.

This analysis utilized baseline and 6-month follow-up data from Social Ties Associated with Risk of Transition (START), a longitudinal study aiming to identify risk factors for transition into injection drug use. Newly initiated injection drug users (IDUs) and heavy non-injection drug users (NIDUS) (heroin, crack, cocaine), aged 18–40 were recruited using RDS (N=436) and TSO (N=217) in New York City (2006–2009). NIDUs completed surveys every 6 months for 18 months and IDUs were evaluated cross-sectionally. All surveys ascertained demographics and information about his/her drug use and network members.

In the respondent-driven sample, equilibrium was reached for all variables considered, however, several RDS assumptions were not met: 1) reciprocal recruitment ties, 2) random recruitment, 3) accurate self-reported degree weight, 4) independence of seeds and peer recruits, and 5) one peer-recruit per respondent. RDS- and TSO-recruited samples were significantly different with respect to demographics and drug use behaviors at baseline, but there were no significant differences with respect to changes in health-seeking behaviors or network composition over 6 months.

Additionally, our findings demonstrate that neighborhood- and study-level variables are salient predictors of RDS peer recruitment success. While receiving additional study coupons and attending group-facilitated trainings on RDS peer recruitment (RDST) both increased peer recruitment efforts, only RDST attendance improved the probability of peer recruit eligibility. By increasing peer recruitment, RDST attendance facilitated the extension of recruitment chains and the recruitment of a more diverse study population.

These findings 1) provide evidence supporting a need for more formative research prior to implementing RDS studies, particularly those conducted among extremely hidden and marginalized populations and 2) emphasize a need for RDS to be validated with empirical data in a variety of different settings.

Indexing (document details)
Advisor: Celentano, David
School: The Johns Hopkins University
School Location: United States -- Maryland
Source: DAI-B 72/09, Dissertation Abstracts International
Subjects: Behavioral psychology, Public health, Epidemiology
Keywords: Drug use, HIV, Hidden populations, Human immunodeficiency virus, Peer-driven intervention, Respondent-driven sampling
Publication Number: 3463580
ISBN: 9781124764658