Case Study: IBM achieves deeper insight into SPMI housing shifts with AnyLogic simulation

An Agent-Based Explanation for SPMI Living Situation Changes

IBM Global Research and Otsuka Pharmaceuticals used AnyLogic simulation software to better understand long-term changes in the housing situations of severely and persistently mentally ill (SPMI) patients, including rising homelessness and incarceration. They needed a way to explain the drivers behind these shifts and test possible policy responses.

AnyLogic helped them build an agent-based model with 1,000 patient agents cycling through seven housing types, including prisons and long-term hospitals, while also reflecting declining hospital capacity. The model reproduced historical trends and showed that increasing time-to-crisis by 45 days in 2000 would reduce the projected prison share by 2030, highlighting that expanding long-term hospital capacity or improving patient stability could meaningfully improve outcomes.

IBM