| Abstract: | The continued growth of outpatient psychiatric care has been accompanied by a large number of research studies concerning the determinants of psychiatric utilization. One of the major limitations of these efforts has been the inability to go beyond distributional data on the use of services. This article describes a methodology for generating episodes of psychiatric care given a data set with a small amount of routinely collected data present in many medical information systems. Both demographic and medical characteristics are significantly associated with health services resource use as defined by the number of visits in an episode. A model predicting recurrent episodes of care is also described. The general utility of this approach and the substantive implications of the specific results are discussed. |
