Social Network-Based HIV Interventions for Homeless Youth: Are They Feasible?

Background:HIV prevention is a growing problem for runaway and homeless youth (RHY) populations. HIV/AIDS risk for RHY has been tied to their social networks, identified primarily as other homeless peers who are engaging in HIV risk-taking behaviors. Concerned with the growing rates of HIV infections, many researchers have developed and tested social-network based HIV interventions, recommending that social network-based interventions maybe a promising method of HIV intervention for homeless youth populations. Although data on RHY social networks exist, few studies have examined how RHY transiency may impact the implementation and efficacy of social-network based interventions for homeless youth. This paper examines how RHY network structures’ change overtime and how changes in network structures may impact the feasibility of implementing peer-led social network-based HIV interventions targeting homeless youth.  

Methods:A convenience sample of homeless youth (N=1081) was collected at two drop-in centers in Los Angeles, CA. Three waves of data were collected from October 2011 to December 2012. Participant completed computerized self-administered questionnaire and an interviewer-led social network interview. The social network interview asked participants (egos) to list the individuals (alters) they have interacted within the last month, and to provide demographic and descriptive information for each alter. To build sociometric networks from the egocentric network data collected, alters were matched based on name, nickname, age, gender, race/ethnicity and tattoos. Network analysis proceeded in two stages: (1) Network-level descriptive statistics were calculated and (2) Separable Temporal Exponential-family Random Graph modeling (STERGM) was conducted.  

Results:The number of RHY varied from wave to wave for each site. Site 1: There were 237 RHY during wave 1 (W1), 263 at wave 2 (W2), and 312 at wave 3 (W3). For site 2, there were 138 RHY at W1, 149 at W2, and 131 at W3. A majority of RHY in W2 (65%) and W3 (74%) were new to the networks, with Site 2 making up a majority of the new network members (W2: 82%, W3: 86%). Additionally, the STERGM models determined that there was a significant amount of dissolution of ties in the RHY, and changes in network density and centrality overtime.

Conclusion: The present analyses indicated that RHY social networks are highly unstable. These results suggest that implementing a face-to-face social network-based HIV intervention for RHY maybe a challenge, due to their high transiency. However, it may be feasible to utilize social media-based peer-networks as a mechanism for HIV interventions. Further research should examine, RHY social media-based peer-networks and the possibility of utilizing social media-based peer-networks for HIV interventions.