The authors performed a chart review of 174 patients age 50 and over (mean age, 69 years) to test a model of social and clinical characteristics that predict engagement in outpatient treatment. “Engaged” patients were defined as those who had three or more clinic visits. Logistic regression analysis, using a 14-variable model, was highly significant (P = 0.0003) and able to correctly classify 70% of cases. There were four predictors of clinical engagement: being a Medicaid recipient (P < 0.01), shorter length of current symptoms (P < 0.05), younger age (P = 0.06), and having a diagnosis of schizophrenia (P = 0.06); and these variables accounted for all the explained variance. These findings indicated that predictors of engagement do not invariably follow the patterns reported for younger patients. The authors performed a chart review of 174 patients age 50 and over (mean age, 69 years) to test a model of social and clinical characteristics that predict engagement in outpatient treatment. “Engaged” patients were defined as those who had three or more clinic visits. Logistic regression analysis, using a 14-variable model, was highly significant (P = 0.0003) and able to correctly classify 70% of cases. There were four predictors of clinical engagement: being a Medicaid recipient (P < 0.01), shorter length of current symptoms (P < 0.05), younger age (P = 0.06), and having a diagnosis of schizophrenia (P = 0.06); and these variables accounted for all the explained variance. These findings indicated that predictors of engagement do not invariably follow the patterns reported for younger patients.