Bridging Disciplinary Boundaries (January 11 - 14, 2007)

Research involving HIV and AIDS requires particular attention to subject recruitment and sampling. This attention is in part necessary to assure representation and diversity in samples and increased likelihood of effective translation to “real world” practice. In HIV and AIDS research with Black and African American Men Who Have Sex with Men (BAAMSM), the issue of sampling is relevant for reaching diverse segments of an often “hidden” population.

Respondent-Driven Sampling is form of chain-referral sampling designed for use with hard-to-reach populations that may not be accessed through more traditional probability-based methods and venue-based approaches (Heckathorn, 1997). This strategy engages a core set of participants, “seeds”, to recruit others from their social networks. Participants recruited by seeds are then asked to recruit others in their network. Each subsequent recruitment “wave” results in new participants. With sufficient recruitment waves, a sample independent of influences by the recruiter is achieved.

During a 9-month data collection period BAAMSM were recruited to participate in an epidemiologic study examining socio-cultural determinants of HIV-infection, AIDS and associated risks among Black and African American Men Who Have Sex with Men. Participants completed a study orientation and informed consent process before taking a comprehensive audio computer-assisted self-interview and HIV anti-body test. Through recruitment in their own social networks, seeds, and subsequent waves of participants, recruited more than 570 BAAMSM into the study.

We present data on the samples' demographics, network size and homophily. Homophily, the measure of a person's tendency to affiliate with others of similar circumstance was used to assess the degree to which men of similar circumstances (i.e., HIV antibody status) recruited others of similar statuses and by extension to estimate the extent to which “diversity” was achieved using this strategy. In addition to the data presentation we review procedures that were modified and/or developed to improve sampling outcomes.

Using RDS, 578 men and 36 seeds enrolled in the study. The mean age for this sample was 40.8 years (std = 9.6), ranging from 18 – 60 years. Sexual identification of respondents included: 36- heterosexual; 197-bisexual; 340 gay or homosexual; and 41 “other” or refused to respond. Men entering the study who did not know their HIV-status numbered 58, while those who identified as HIV negative numbered 186 and positives numbered 323. Homophily for the Positive responders was .578; for the Negative responders .473 and for the Did Not Know Group it was .049.

RDS was effectively used to bring a relatively large number of BAAMSM into this research study. The sampling method required staff to repeatedly review and modify the recruitment protocol. RDS clearly has potential for use in research efforts; however, much is yet to be learned about specific applications methods to effectively recruit diverse samples of BAAMSM in HIV & AIDS prevention and intervention research. In this presentation we present lessons learned and thoughts about future recruitment of hard to reach groups using RDS.

References:

Heckathorn, D.D. (1997). Respondent-Driven Sampling: A New Approach to the Study of Hidden Populations. Social Problems(44) 2, 174-199.