Unprotected Sex Among Homeless Youth: Understanding Multiple Levels of Influence Using Multilevel Dyadic Analyses

Background: HIV disproportionately affects homeless youth. Condoms are considered an effective mechanism for preventing HIV transmission. Homeless youth, however, report very low rates of condom-use; 40-70% of youth across various studies indicated engaging in unprotected sexual intercourse. Social network norms have been consistently linked to condom-use among homeless youth. Norms are defined as perceived beliefs about what behaviors are considered acceptable or common within a given group. One important characteristic of social norms is that they are not one-dimensional by nature. More specifically, different kinds of referent groups (such as family, peers, etc.) communicate, encourage, and discourage varying types of norms surrounding both risk and protective behaviors. For example, family norms might encourage condom-use; peer norms, conversely, might discourage use. Most studies regarding condom-use norms among homeless youth have tended to focus on peer-norms, excluding other pertinent influences on condom-use behaviors among this population. Accordingly, the purpose of this study was to understand how these different types of relationships contributed to norms about condom-use and whether such norms were associated with self-reported condom-use among homeless youth.

Method: Egocentric network data were collected from homeless youth accessing services at two drop-in centers in Los Angeles, CA (N=260). Assessment of social norms included 1) perceived condom-use by street-peers (descriptive-norm), 2) encouragement to engage in condom-use from social network members (injunctive-norm) and 3) communication regarding condom-use with social network members (communication-norm). The communicative and injunctive norms were separated by relationship roles (i.e. family, street and home-based peers, sex partner, serious partner, staff etc.). Multilevel dyadic analyses were conducted to understand associations between source-specific communicative and injunctive norms with descriptive peer norms about condom-use. The second multivariate model assessed associations between all types of norms (descriptive, injunctive, and communicative) and condom-use among the sample.

Results: Descriptive analyses indicated that youth communicated to and received encouragement about using condoms from a multitude of sources (such as street and home-based peers, relatives, family, partners, etc.). Bivariate dyadic analyses demonstrated that street peers talking about (OR=3.94) and encouraging condom-use (OR=11.36), talking to a relative (OR=8.80) or a home-based peer (OR=2.94) about condoms were significantly associated with having more protective perceptions of peer norms about condom-use. Multilevel dyadic analyses indicated that talking to street peers (OR=2.31) and receiving encouragement from street peers (OR=11.23) about condom-use were significantly associated with having more protective perceptions of peer norms about condom-use. In the second multivariate model assessing associations between condom-use norms and self-reported condom-use, perception of peer condom-use (descriptive norm) (OR= 6.04) was the only significant factor associated with using a condom during last sexual encounter.

Conclusion: These results suggest that communicating and receiving encouragement from social network members (especially street peers) reinforce condom-use norms within networks, which, in turn, influenced self-reported condom-use among our participants. In light of such findings, HIV interventions intended to promote safer-sex norms may benefit from tapping into these naturally occurring and existing patterns of communication to reduce HIV vulnerability among this at-risk population.