The Role of Housing Model in the Social Integration of Formerly Homeless Adults in Supportive Housing

Background and Purpose: Social integration has been identified as a goal and an indicator of programmatic success in supportive housing programs, yet the social integration of formerly homeless persons has also been an ongoing challenge for providers and an area in need of further research. The model of housing residents reside in (i.e., congregate, scatter-site, single-site) may be associated with a greater or lesser likelihood of social integration; some housing models are more likely to host onsite supportive services or have formerly homeless adults clustered together, while others are typically dispersed from other formerly homeless adults with services located outside of residents’ apartments. This study examined different housing models and their associations with sub-constructs of social integration, including interactional (i.e., community participation) and social network dimensions (i.e., size of network and diversity-- different types of persons within the network and kinds of social support).

Methods: Findings come from 3-month follow-up interviews conducted as part of a larger, longitudinal research study of formerly homeless adults living in permanent supportive housing (n=405). Participants took part in interviews 3 months following their permanent supportive housing move-in date. Interviews assessed a variety of topics, including housing location, satisfaction with housing, social network outcomes (i.e., network size, social support, types of network connections), and community integration. Multinomial logistic regression was used to determine the associations between housing model and multiple levels of social integration outcomes in controlled models (e.g., demographics, physical limitations, psychiatric limitations).

Results: Residents in scatter-site housing models were more likely to engage in moderate community participation and more likely to report high emotionally-supportive networks, compared to residents in single-site housing. Of note, no housing models were significantly associated with social network size or diversity in univariable models. Women in the study were less likely to have diverse networks, compared to men. Several substance use measures were associated with increased social integration outcomes, including network size and network diversity.

Conclusions and Implications: Findings indicate that housing models that typically employ a more autonomous approach and are situated within the greater community (e.g. scatter-site) may improve some opportunities for social integration, while other social integration outcomes may be less dependent on a given model of housing. Several trending covariates raised concerns for social integration outcomes; specifically, substance use may increase social integration within risky networks and social spheres.  Results also point to the ongoing need to address social integration for a population that experiences high rates of isolation and that may need assistance with developing and maintaining prosocial relationships.