Architecture of clinical information and prediction of service utilization and cost.

The state of the art in predicting service utilization and cost in general medical care and psychiatry is reviewed. Diagnosis-related groups (DRGs) and measures of illness severity are specifically examined. Building on this experience and searching for other promising informational elements, key components of clinical information schemas are considered. The interface between patient and treatment domains is examined by using multiaxial classifications connected with the upcoming Tenth Revision of the International Classification of Diseases and the European International Classification of Mental Health Care. Within this framework, univariate and multivariate approaches to the prediction of service utilization and cost are outlined.