Network Structure and Proxy Network Measures of HIV, Drug and Incarceration Risks for Active Drug Users

Recent work in the area of drug and HIV risk reduction has demonstrated that ethnographic, ego-centered, and full network analysis of drug using groups fills an important information gap for understanding the impact of both peers and culture on the risky conditions and practices which lead to HIV infection. This paper explores the relationships between individual risk taking behavior and overall network structure measures, specifically measures of density, transitivity, and centralization. These measures are compared with individual risk measures including frequency of drug use, use of common IV drugs, number of IDU sex partners, percent of use of condoms, lifetime number of STDs and time spent in jail. The results indicate there are significant associations between centralization and other network measures and the level of risky behavior of individuals in a network. Taking this finding one step further, the second part of the paper explores the relationships between three proxy measures of network structure and risky behavior. These three proxies are network type, single or multiple network membership, and project recruitment order. The public health implications of these network proxy measures are discussed.