This paper explains the methods used to develop the Silicon Valley Triage Tool for identifying homeless individuals in jails, hospitals and clinics who have continuing crises in their lives that create very high public costs. The model is very robust and accurate, taking advantage of advanced prediction methodologies and a unique and exceptionally valuable database created by Santa Clara County, home to Silicon Valley, linking service and cost records across county departments for the entire population of residents who experienced homelessness over a six-year period – a total of 104,206 individuals. The tool accurately identifies individuals experiencing homelessness whose acute needs create the greatest public costs.
The common failing of many previous efforts to prioritize interventions for the highest need homeless individuals has been their targeting inaccuracy, which has diminished program effectiveness. Many previous screening models have not been sensitive or accurate enough to yield high hit rates without missing a large number of high-risk persons who would benefit from the program while producing cost savings. However, recent technological advances in the fields of predictive analytics and data mining together with the availability of digital integrated administrative data sets with rich service utilization fields allow significant improvement in prediction ability over earlier approaches and models.
In the case of high-cost individuals who are frequent users of public services, as well as other homeless groups, the right solution is often cheaper than the problem. More predictive analytic screening tools are needed to target interventions for:
- Unemployed but employable adults
- Justice-system-involved individuals
- Children who experience trauma
- Children with mental illness
- Transition age foster and probation youth
- Adolescents with high-risk behaviors
- Displaced households
Related Articles
Suffering in Silicon Valley — and how homeless advocates are trying to end it
By Phillip Bane, Smart Cities Council (May 16, 2017)
