Background. Differences in sexual network structure may help explain racial-ethnic sexually transmitted infection (STI) disparities that are not accounted for by individual-level sexual risk factors. Because of the difficulty of collecting network data, much is unknown about how sexual networks affect STI risk empirically, and what factors shape sexual networks.

Objectives. To better understand how sexual network structure is associated with infection with gonorrhea and chlamydia in urban African American adolescents, and to determine how neighborhood socioeconomic environment is associated with sexual network structure in this population.

Methods. Population-based sexual network data were generated by combining a household survey of African American urban adolescents (14-19 year old) with snowball sampling of their social and sexual networks. Sexual network structure was identified by interviewing partners and partners' partners at two points in time (twelve months apart). We examined associations between network structure and (1) infection with gonorrhea or chlamydia, (2) neighborhood socioeconomic environment and (3) individual-level social correlates of STIs. Logistic regression models were used to estimate multivariate associations, and generalized estimating equations or cluster-based robust standard errors were used to adjust for correlated outcomes.

Results. A total of 814 unique sexually active individuals participated over the course of the two study waves. Participants' sexual network components were small on average and not densely connected. Sexual network position, a reflection of individuals' number of partners and their partners' number of partners, was strongly associated with STI risk. Using individual-level data to measure sexual network position, as opposed to network-level data, led to underestimates of associations of network position with STI risk. Higher neighborhood poverty was associated with increased likelihood of being in sexual network positions that conferred higher STI risk. Furthermore, drug use, drug dealing and gang involvement were associated with an increased likelihood of being linked to multiple partners but not with having a partner with multiple partners.

Conclusions. Risk of STI acquisition can not be conceptualized solely as a function of individual behaviors; it is also a function of the network structure in which individuals are embedded and the social environment that shapes network structure.