Heterogeneous Patient Information Acquisition and Normalization for Shareable, Executable Clinical Decision Support System

The goal of shareable, executable clinical decision support system is both worthwhile and challenging. One of the largest hurdles is that of using the heterogeneous databases available in the general clinical application environment. In this paper, based on the virtual medical record, a solution of patient information acquisition and normalization for clinical decision support system has been raised. We defined a patient information model that supply the distinctions necessary for decision support based on HL7 RIM. To overcome mismatches between the data content and form of a patient information model and those of a local clinical information system, we proposed a highly configurable database mediator method to translate data extracted from locally deployed medical record architectures to a form consistent with the patient-data model. Our solution opening up the possibility of a standardized methodology for embedding guideline-based decision support systems in alternative clinical information systems (CISs) and enables reuse of the decision-support systems. Experimental results on 200 cases in a hospital show that this method proves to be efficient, configurable and flexible, and it enables reuse of the decision-support systems.