Homelessness, Racial Equity, and the Vulnerability Index-Service Prioritization Decision Assistance Tool (VI-SPDAT)

Background and Purpose: Over the last 5 years, 81 communities in the U.S. have initiated campaigns to end chronic homelessness. Few cities have met the goal, but efforts have shaped key parts of the homeless service delivery system including prioritization for housing using standard measures. This study examined the implementation of such an effort in Charlotte, North Carolina, specifically including use of the Vulnerability Index – Service Prioritization Decision Assistance Tool (VI-SPDAT) to guide housing decisions. Broad research questions for the study included 1) Who was served by the effort? And 2) How did the project structure impact implementation and outcomes?

Methods: The study used a mixed-methods design to address the research questions. The research team collected data from the Homeless Management Information System (HMIS) on 1660 individuals who were on a By-Name List of chronically homeless individuals from 2015-2018. Univariate and bivariate statistics were used to describe characteristics, including their VI-SPDAT scores. Differences among groups were determined by t-tests or Chi-Square analyses. In addition, stakeholders were purposively sampled for individual (n=33) and focus group interviews (n=103) to examine perspectives on the impact of the project structure. Interviews were analyzed using a constant-comparison technique.

Results: The majority of individuals on the By-Name List during the study period were Men (75%, n=1250), Black and Indigenous People/Persons of Color (BIPOC; 73%, n=1205), and between the ages of 51-64 (46%, n=757). Among those who had VI-SPDAT scores in HMIS (n=1563), 25% (n=390) scored 12-16 on the VI-SPDAT, 24% (n=376) scored 10-11, and 51% (n=797) scored under 10 on the VI-SPDAT. Scores above 10 were prioritized for permanent supportive housing, which includes a permanent housing subsidy and extensive supportive services. Scores under 10 were routed toward short-term subsidies with limited supports or were not prioritized for housing. The distribution of scores was similar between men and women and between older and younger individuals. However, the distribution of scores differed between racial groups. The majority of BIPOC individuals scored under 10 (55%), while most White individuals scored 10 or over (60%). BIPOC individuals scored an average of 1.1 points lower on the VI-SPDAT than White individuals (M=8.9, SD=2.94; M=10.0, SD=3.03; p<.0001) suggesting, on average, a lower priority score for housing. Following scoring patterns, the percentage of BIPOC individuals who were housed in PSH (49%, n=275) was less than White individuals housed in PSH (58%, n=120; X²=5.34, p<.05). Stakeholder interviews suggested provider skepticism of VI-SPDAT scores and a lived experience concern about the weight of the scores in housing decisions. Interviews also documented the push to develop a review process to challenge instrument scoring, without which, housing outcomes may have been more disproportionate.

Conclusions and Implications: These findings suggest that despite significant housing successes through efforts to end chronic homelessness, the widespread adoption of the VI-SPDAT may further exacerbate racial inequities in housing. Similar to a recent study of three communities in the Pacific Northwest, this study suggests the need for a prioritization tool that is sensitive to vulnerabilities that may vary by race and ethnicity.