Disassortative mixing patterns of drug-using and sex networks on HIV risk behaviour among young drug users in Yunnan, China

Abstract

ObjectivesThe dominant mode of HIV transmission in China has changed from injection drug use to sexual contact. The objectives of this study were to describe the disassortative and assortative mixing patterns of drug-using and sex networks among young drug users in China. Study designCross-sectional study. MethodsRespondent-driven sampling (RDS) was used to recruit young drug users in an egocentric network study in Yunnan, China. Egos were categorized as having disassortative mixing network patterns if they reported both sex and drug-using networks. Egos who only had a sex network (no drug-using network), or only a drug-using network (no sex network) were categorized as having assortative mixing network patterns. Multiple logistic regression was performed to analyze the relationships between disassortative patterns with risky sexual behaviour and drug-using practices. ResultsA total of 426 participants were recruited into the study. Two hundred forty-two egos reported disassortative mixing patterns and 139 egos had assortative patterns. The RDS-adjusted proportion of having a disassortative pattern was 53.2%. Participants with disassortative patterns were more likely to engage in HIV risk behaviour compared to those with assortative patterns. Specifically, drug users with disassortative patterns reported more multiple sex partners (31.4% vs 19.6%), concurrent partnerships (52.1% vs 39.0%), non-regular sex partners (12.0% vs 4.3%), and sex partners who were IDUs (24.9% vs 12.5%). Consistent condom use with regular or non-regular partners was low (between 18.9% and 47.2%) regardless of the mixing pattern. However, parenteral risk for HIV transmission was relatively low in both groups. ConclusionsThe transition of the HIV epidemic in China from injection drug use to sexual contact may be attributed to disassortative mixing in drug-use and sexual networks. HIV programs should consider disassortative mixing patterns when designing new behavioural interventions.