Using respondent-driven sampling for behavioural surveillance: Response to Scott

Abstract

s r s t e v p f s In a recent article, Dr. Scott raised ethical questions bout respondent-driven sampling, or RDS (Scott, 2008). r. Scott used data he collected in Chicago, Illinois in 2005 t the time that the National HIV Behavioural Surveillance ystem amongst injecting drug users (NHBS-IDU), sponored by the Centers for Disease Control and Prevention CDC), was being conducted. He asserted that RDS invites, f not promotes, violation of federal guidelines governing he protection of human research subjects and that the RDS esign generally, and the NHBS implementation specifially, resulted in data of questionable quality. CDC places he highest priority on human subjects protection and the thics of surveillance and research. We therefore comment n Dr. Scott’s article and then describe CDC plans for the ext NHBS-IDU cycle. Because NHBS uses a standard proocol in all sites (Lansky et al., 2007), we describe operations enerally, with details about the Chicago experience where elevant. Numerous controls for the protection of participants n research and surveillance activities are fundamental spects of the RDS method (Semaan, Santibanez, Garfein, eckathorn, & DesJarlais, 2008) and many were specifically sed in NHBS-IDU. The controls described below were put n place to assure the ethical conduct of NHBS-IDU. We isagree with Dr. Scott that “RDS assigns the lion’s share f responsibility for ethical conduct. . . to the human subjects