Respondent-driven sampling to assess characteristics and estimate the number of injection drug users in Bangkok, Thailand

Abstract

Background Since early in Thailand's HIV epidemic, HIV seroprevalence among injection drug users (IDUs) in Bangkok has been around 40%. As Thailand moves to strengthen HIV prevention and care programs for Bangkok IDUs, information on current patterns of drug use and an estimate of the size and composition of the IDU population are essential. Methods We used respondent-driven sampling (RDS) to recruit Bangkok IDUs who reported injecting in the past 6 months. IDUs were interviewed with a standardized questionnaire. Logistic regression was used to compare IDUs currently or previously in treatment with those never treated. RDS software was used to estimate IDU population size based on the proportion in treatment. Results Of 963 IDUs recruited, 828 (86%) were men. One hundred and twelve IDUs (12%) reported never having attended a drug treatment clinic. Young age, unemployment, injection of single drug, and never having been HIV tested were significantly associated with never-in-treatment status. The estimated proportion of IDUs in treatment was 0.55 (95% confidence interval, 0.52–0.60). Dividing the known number of IDUs in treatment (1981 IDUs attending Bangkok drug treatment clinics during October 2003 through March 2004) by this proportion, we estimated the number of IDUs in Bangkok during the period of RDS to be 3595 (95% confidence interval, 3296–3810). Conclusions Data obtained through RDS, although subject to limitations from co-existing government drug control policies and possible under-recruitment of out-of-treatment IDUs, will be useful in informing policy, strengthening prevention approaches, and improving methods to monitor the HIV epidemic among Thai IDUs.