Abstract: Predictors of Housing (In)Stability Among Adults Experiencing Homelessness (Society for Social Work and Research 25th Annual Conference - Social Work Science for Social Change)

All live presentations are in Eastern time zone.

Predictors of Housing (In)Stability Among Adults Experiencing Homelessness

Wednesday, January 20, 2021
* noted as presenting author
Laura Petry, MSW, PhD Student, University of Southern California, Los Angeles, CA
Hsun-Ta Hsu, PhD, Assistant Professor, University of Missouri-Columbia, Columbia, MO
Eric Rice, PhD, Associate Professor, University of Southern California, Los Angeles, CA
Background and Purpose: On any given night, nearly 370,00 single adults experience homelessness in the U.S. Over the last decade, communities have implemented coordinated assessment in order to ensure equitable access to limited housing resources and to prioritize resources based on levels of vulnerability. The Vulnerability Index-Service Prioritization Decision Assistance Tool (VI-SPDAT) has been adopted to triage individuals to various housing and support services based on their vulnerability assessment scores. However, evidence for the effective use of the tool in identifying vulnerability and helping persons successfully exit homelessness has been limited to single communities. The current study uses a national dataset of VI-SPDAT data for single adults to examine associations among vulnerability scores, placement types, demographics (e.g. race/ethnicity, sexual orientation), and two outcomes: housing stability for 365 days and returns to homelessness crisis services.

Methods: Homeless Management Information System (HMIS) data from 16 communities across the U.S. collected between March 2015 and April 2018 (n=12,814) from single adults over the age of 24 experiencing homelessness was used and included the VI-SPDAT and subsequent housing placements. Housing stability was defined as maintaining initial housing placement for 365 days or longer. Return to the homeless services system was defined as being rehoused by the system following an initial housing placement. Housing placement types included permanent supportive housing (n=8,839), rapid re-housing (n=3,086), living with family (n=219), and self-resolve (n=670). Logistic regression was used to explore associations among vulnerability scores, housing placement type, demographics (e.g., race/ethnicicy, LGBTQ+) and housing stability or returns to the homeless services.

Results: Increases in vulnerability scores were associated with decreased odds of maintaining housing (OR=.751, p<.001) and increased odds of returning to the homeless services system (OR = 1.34, p < .001). Relative to placement in permanent supportive housing, initial exits from homelessness into rapid-rehousing (OR=.409, p<.001), living with family (OR=.560, p<.001), or self-resolve (OR =.342, p<.001) were associated with decreased odds of achieving housing stability. Relative to identifying as white, identifying as Latinx was associated with increased odds of maintaining housing (OR=1.33, p=.032) and identifying as Native American was associated with decreased odds of maintaining housing (OR=.687, p=.038). Identifying as Latinx was also associated with decreased odds of returning to the homeless services system (OR=.489, p<.001). Identifying as LGBTQ+ was not significantly associated with either housing outcome.

Conclusions and Implications: Findings suggest that individuals with higher levels of vulnerability, as measured by the VI-SPDAT, are at increased odds of failing to maintain initial placements and returning to the homeless services system. Further, despite the recommendation of lower-intensity housing interventions or diversion from the homeless services system to personal housing resources for lower-scoring individuals, these placements have lowered odds of success relative to permanent housing options. Lastly, housing outcomes vary across racial and ethnic groups, emphasizing opportunities for homeless services systems to adapt more culturally competent practices in addressing potential disparities.