BACKGROUND: Inpatient psychiatry is a critical service in a community-based care system for persons with serious mental illness (SMI). Currently, there are few generally accepted or widely used outcomes to assess the effectiveness of inpatient treatment. METHOD: Following a Donabedian Model of Health Care Quality, we utilized eight scales from the RAI Mental Health assessment to derive a clinician-scored outcome measure consisting of 4 domains (Psychosis, Depression, Impairment, and Aggression). We combined subscales measuring these domains into a Composite Measure. We used this measure to assess the entire population (N=719) of our large specialized mental health hospital at the beginning (T1) and end (T2) of three months in the hospital (or admission to discharge in shorter stays). We evaluated the content validity of the measure by comparing items and scales with a list of putative contributors to hospital admission (symptoms and complications). To evaluate concurrent validity, we compared mean scores among hospital units with varying lengths of stay and clinical complexity (acute versus chronic versus complex chronic). We used ROC analysis to evaluate the CIIMHS's ability to predict discharge from the hospital. To evaluate construct validity, we examined the measure's responsiveness to changes among patients after treatment in the hospital. RESULTS: We found strong evidence for all four kinds of validity. CONCLUSIONS: The composite measure represents a valid measure of inpatient mental health status and will serve as a valuable measure of the quality of care for inpatient psychiatry.