In many municipalities, the most vulnerable individuals struggle with multi-dimensional problems, such as homelessness, mental illness, substance abuse, and chronic health conditions. As a result, they frequently interact with multiple public systems, including jails, hospitals, and mental health facilities. Cross-system initiatives that identify these individuals and provide them with comprehensive treatment would also reduce their service needs. However, data traditionally siloed within respective agencies frequently slows such coordinated efforts across entities.
The White House Data-Driven Justice Initiative, a partnership with cities, counties, and states across the country, addressed these obstacles as one of its key goals. We worked with Johnson County, Kansas (and plan to partner with additional jurisdictions in the future) to build a predictive model that combines data from multiple systems and identifies individuals most in need of coordinated assistance. We also analyzed behavioral patterns to understand the factors that best predict interactions with these systems, so that governments can deliver higher quality, targeted services that reduce repeated usage and the associated human and financial costs.