Prioritizing Homeless Assistance Using Predictive Algorithms: An Evidence-Based Approach

Halil Toros
Daniel Flaming
Economic Roundtable

In this article, we present a predictive model for identifying homeless persons likely to have high future costs for public services. We developed the model by linking administrative records from 2007 through 2012 for 7 Santa Clara County, California agencies and identifying 38 demographic, clinical, and service utilization variables with the greatest predictive value. We modeled records for 57,259 individuals from 2007 to 2009, and the algorithm was validated using 2010 and 2011 records to predict high-cost status in 2012. A business case scenario shows that two-thirds of the top 1,000 high-cost users predicted by the model are true positives, with estimated posthousing cost reductions of more than $19,000 per person in 2011. The model performed very well in giving low scores to homeless persons with one-time cost spikes, achieving the desired result of excluding cases with single-year rather than ongoing high costs.