PURPOSE: We examined the cross-sectional and longitudinal association between social network characteristics and street youths' shelter use, a determinant of health outcomes for homeless youth. METHODS: We analyzed interview data from 138 street youth recruited through venue-based sampling in San Francisco, to assess the cross-sectional relationship between shelter use in youths' social networks and youths' reported shelter use. We also assessed the relationship between baseline network shelter use and shelter use at 6-month follow-up. RESULTS: Low proportions of street youth reported shelter use at baseline (38%) and follow-up (29.6%). Twenty-nine (26.9%) youth were in networks with shelter users at baseline, compared with 17 youth (15.7%) at follow-up. In cross-sectional analysis, youth in networks with shelter users had 5-fold increased odds of reporting shelter use (OR: 5.86, p = .006). A 1-person increase in the number of network shelter users was associated with 2-fold increased odds of youths' shelter use (OR: 2.16, p = .02). In longitudinal analysis, youth in networks with shelter users at baseline had nearly 5-fold increased odds of shelter use at follow-up (OR: 4.95, p = .01). A 1-person increase in the number of network shelter users at baseline was associated with 3-fold increased odds of shelter use at follow-up (OR: 3.15, p = .004). CONCLUSION: Shelter users seem to cluster together. Shelter use by extended network members was associated with increased odds of youths' own shelter use. Understanding how network behaviors influence street youths' health-related behaviors, such as shelter use, could inform network-based interventions encouraging service uptake among marginalized and hard-to-reach street youth populations.
PURPOSE: We examined the cross-sectional and longitudinal association between social network characteristics and street youths' shelter use, a determinant of health outcomes for homeless youth. METHODS: We analyzed interview data from 138 street youth recruited through venue-based sampling in San Francisco, to assess the cross-sectional relationship between shelter use in youths' social networks and youths' reported shelter use. We also assessed the relationship between baseline network shelter use and shelter use at 6-month follow-up. RESULTS: Low proportions of street youth reported shelter use at baseline (38%) and follow-up (29.6%). Twenty-nine (26.9%) youth were in networks with shelter users at baseline, compared with 17 youth (15.7%) at follow-up. In cross-sectional analysis, youth in networks with shelter users had 5-fold increased odds of reporting shelter use (OR: 5.86, p = .006). A 1-person increase in the number of network shelter users was associated with 2-fold increased odds of youths' shelter use (OR: 2.16, p = .02). In longitudinal analysis, youth in networks with shelter users at baseline had nearly 5-fold increased odds of shelter use at follow-up (OR: 4.95, p = .01). A 1-person increase in the number of network shelter users at baseline was associated with 3-fold increased odds of shelter use at follow-up (OR: 3.15, p = .004). CONCLUSION: Shelter users seem to cluster together. Shelter use by extended network members was associated with increased odds of youths' own shelter use. Understanding how network behaviors influence street youths' health-related behaviors, such as shelter use, could inform network-based interventions encouraging service uptake among marginalized and hard-to-reach street youth populations.