Individuals with substance use disorders (SUDs). including co-occurring disorders, are among the highest-risk populations for medical and psychiatric rehospitalizations, and are often underdiagnosed at initial hospitalization. This study examined predictors for these individuals at baseline hospitalization and subsequent rehospitalizations. Three groups were compared from a sample of individuals admitted to inpatient psychiatry (1982 to 1987) with at least one rehospitalization within a 16-year period. Multivariate logistical regressions were used to determine associations with predictor variables. The data showed that individuals' diagnosed with a SUD after baseline hospitalization were more likely to have more medical hospitalizations and to be diagnosed with schizophrenia compared to those who were diagnosed with a SUD, including co-occurring disorders, at baseline. The results of this study indicate the importance of substance use screening to enhance service resources and treatment outcomes for medically and psychiatrically complex populations.