Beth Israel Medical Center (BIMC), in collaboration with the Centers for Disease Control (CDC) and the New York State Department of Health (NYSDOH), used respondent-driven sampling (RDS) in a study of HIV seroprevalence among drug users in New York City in 2004. We report here on operational issues with RDS including recruitment, coupon distribution, storefront operations, police and community relations, and the overall lessons we learned. Project staff recruited eight seeds from a syringe exchange in Lower Manhattan to serve as the initial study participants. Upon completion of the interview that lasted approximately 1 h and a blood draw, each seed was given three coupons to recruit three drug users into the study. Each of the subsequent eligible participants was also given three coupons to recruit three of their drug-using acquaintances. Eligible participants had to have: injected, smoked or snorted an illicit drug in the last 6 months (other than marijuana), aged 18 or older, adequate English language knowledge to permit informed consent and complete questionnaire. From April to July 2004, 618 drug users were interviewed, including 263 (43%) current injectors, 119 (19%) former injectors, and 236 (38%) never injectors. Four hundred sixty nine (76%) participants were men, 147 (24%) were women, and two (<1%) were transgender. By race/ethnicity, 285 (46%) were black, 218 (35%) Hispanic, 88 (14%) white, 23 (4%) mixed/not specified, and four (<1%) native American. Interviews were initially done on a drop-in basis but this system changed to appointments 1 month into the study due to the large volume of subjects coming in for interviews. Data collection was originally proposed to last for 1 year with a target recruitment of 500 drug users. Utilizing RDS, we were able to recruit and interview 118 more drug users than originally proposed in one quarter of the time. RDS was efficient with respect to time and economics (we did not have to hire an outreach worker) and effective in recruiting a diverse sample of drug users.