Detoxification center-based sampling missed a subgroup of higher risk drug users, a case from Guangdong, China.

BACKGROUND: Injection drug use remains among the most important HIV transmission risk in China. Representativeness of drug users sampled from detoxification centers is questionable. A respondent driven sampling survey was conducted to compare the results with those from the detoxification center in the same city. METHODS: In 2008, two independent surveys were conducted in Dongguan, China, one for community-based drug users using respondent driven sampling and the other for drug users in a compulsory detoxification center as routine sentinel surveillance. Demographic and behavioral information were collected using the same structured questionnaire. Intravenous blood samples were collected to measure antibodies to HIV-1, and syphilis. RESULTS: Compared to those 400 drug users recruited from the detoxification center, the 303 community-based drug users had higher HIV prevalence (14.7% versus 4.0%, P = 0.04), lower syphilis prevalence (4.7% versus 10.8%, P = 0.07), higher proportion of injection drug use (83.9% versus 60.2%, P = 0.01) and syringe sharing (47.8% versus 36.3%, P = 0.10), more likely to be separated (12.4% versus 3.8%, P = 0.01) and being migrants from Guangxi province (31.4% versus 18.0%, P = 0.09), more engaging in commercial sex (64.4% versus 52.5%, P = 0.04). HIV prevalence and rate of syringe sharing were consistently higher among drug users from Guangxi. CONCLUSIONS: Detoxification center-based sampling missed a subgroup with higher HIV prevalence and higher rate of injection drug use. While detoxification center-based sampled can be used to monitor the trend of HIV prevalence and risk behaviors over time, periodic community-based sampling is still necessary to avoid possible systematic error in detoxification center-based samples