Examining the Use and Measurement Validity of the VI-Spadat As a Housing Prioritization Tool for Chronically Homeless Adults

Purpose: The Vulnerability Index - Service Prioritization Decision Assistance Tool (VI-SPDAT) is used nationally and internationally to assist in the prioritization of homeless individuals for housing and services. While evidence-informed and widely used, the VI-SPDAT lacks a robust evidence base establishing its measurement validity. While the literature and numerous experts, including individuals with lived homelessness experience, were consulted in the development and refinement of the VI-SPDAT, the instrument remains evidence-informed rather than a validated tool with an extensive evidence-base. Questions among stakeholders remain about the construct measured by the VI-SPDAT and its predictive value. To this end, this study examined 1) how community stakeholders perceive the instrument and 2) how the VI-SPDAT scores correlates with subjective measures of health and mental health among a sample of chronically homeless individuals.

Method: This mixed methods study is part of a larger process, outcomes, and service utilization evaluation project examining a multi-sector collaboration to end chronic homelessness in a large urban community. Semi-structured individual and focus group interviews were conducted with 115 stakeholders including community leaders, agency administration, and front line service providers. Thematic analysis was conducted in Atlas.ti to group data and identify themes, disparities, and relationships within and between participant responses. In addition, data from eleven validated measures from 197 chronically homeless study participants were analyzed to examine correlations with VI-SPDAT scores. A series of correlation matrices were developed in SAS to measure 1) if there was a relationship between the validated measures we collected and VISPDAT scores and 2) the strength of the relationship between these pairs of indicators.

Results: Four keys themes emerged from the focus group and interview data concerning the VI-SPDAT including concerns about instrument validity, variations in administration, mental illness and scoring, and service utilization and scoring. Correlation analyses revealed the VI-SPDAT scores had a very weak correlation with participant responses on the Modified Colorado Symptom Index (r=0.227); PTSD Checklist for DSM-5, Civilian version (r=0.225); and, the Short Form Health- Physical Health Composite (r=-.0555) and Mental Health Composite (r=-0.175). In relation to addiction, the VI-SPDAT scores had very weak correlations with the Alcohol Composite Score (r=0.158) and the Drug Composite Score (0.132).

Conclusions: This study suggests that the VI-SPDAT does not measure the same constructs as validated measures of mental health and substance use, constructs typically associated with vulnerability among homeless individuals. This confirms concerns voiced by front line service providers that the instrument doesn’t capture the true vulnerability of clients, particularly the vulnerability of persons with mental health and substance use disorders. These findings have implications for how persons experiencing homelessness are prioritized for housing resources and which measures are utilized to ensure physical and mental health vulnerability is captured. Future phases of this project will examine the relationship of the VI-SPDAT with objective measures of health, mental health, and service utilization, as well as instrument reliability and factorial validity.