Mining hospital data to learn SDA clinical algorithms

Abstract

The practice of evidence-based medicine requires the integration of individual clinical expertise with the best available external clinical evidence from systematic research and the patient's unique values and circumstances. This paper addresses the problem of making explicit the knowledge on individual clinical expertise which is implicit in the hospital databases as data about the daily treatment of patients. The EHRcom data model is used to represent the procedural data of the hospital to which a machine learning process is applied in order to obtain a SDA*clinical algorithm that represents the course of actions followed by the clinical treatments in that hospital. The methodology is tested with data on COPD patients in a Spanish hospital. © 2008 Springer-Verlag Berlin Heidelberg.