A sizable group of individuals in the United States cycle in and out of jails, prisons, mental health hospitals, homeless shelters, and other expensive public institutions over time. This little-studied population represents significant unmet need and the inadequacy of services for complex consumers. The current study examined a sample ( N = 161) of chronically homeless frequent utilizers of jail and mental health systems in Chicago, Illinois. Cluster analysis was used to differentiate the sample into four reliable subgroups based on measures for prior homelessness, jail incarcerations, mental health hospitalizations, poor current mood, and limitations due to physical health. Logistic regression revealed that clusters differed significantly on reincarceration at 6 months postrelease. Implications for programming and policy for each cluster are discussed, including suggestions for targeting services to distinguishing characteristics for each subgroup. These findings argue for the importance of coordinating efforts across services systems to better identify and serve shared clients.
Trees and discrete subgroups of Lie groups over local fields
