HIV research among internal migrants in China has not fully explored the contexts and perceptions of “risk”. In 2011, urban markets in Liuzhou, China were mapped, and sixty male vendors, age 22 to 56, were selected for in-depth interviews on migration, social and family life, and perceptions and practices of sexual risk behavior. Participants were evenly divided among higher income shop and small stall vendors. All men were sexually active. Only the shop vendors reported non-marital sexual partners, including concurrent partners (n=15), commercial partners (n=10), and other sexual relationships (n=11). Shop vendors engaged in networking activities that facilitated commercial and non-commercial high-risk sex. Perceptions of HIV risk from commercial sex led some men to doubt the protective ability of condoms and rely on local (unproven) self-protection techniques. Networking activities played a role in high-risk sex and shaping migrants' risk perceptions and health practices. The networks created through these processes could also be used to facilitate health promotion activities.

BackgroundSignificant domestic and global research has focused on HIV risk among China’s large internal migrant population. Much of this work takes an individual behavior approach while ignoring the critical role social networks play in shaping HIV risk.MethodsBased on past studies among migrant men in China of yingchou activities (activities that build and reinforce social networks such as eating, drinking alcohol and patronizing commercial sex), we constructed ego-centric networks for a sample of 385 male migrants recruited from multiple worksites in Beijing. We used a nested-model approach to examine the contribution of social network characteristics to HIV risk at both the variable and model levels.ResultsAs compared to an individual-level model, addition of social network variables significantly improved the fit of the models. Commercial sex norms and condom use norms of core yingchou networks were significantly associated with egos’ commercial sex and condom use respectively. The size of yingchou network was associated with egos’ commercial sex. The network models became more sensitive after network norm measures took into account the intimacy of network ties and allowed for egos’ uncertainty when reporting their alters’ sexual behaviors.ConclusionResults suggest the importance of social network factors and core network members in HIV transmission and risk-reduction interventions for male migrants. Future studies could explore other important social networks among male migrants, consider the intimacy of network ties and egos’ uncertainty about alters’ situations in constructing network norms, and refine the measurement of network size and density.