Refining disease databases for clinical decision support systems

Clinical decision support systems (CDSSs) require disease knowledge bases that need enormous manual work by costly physicians to develop. To overcome this obstacle, we constructed a disease knowledge base by integrating and processing existing disease databases, Online Mendelian Inheritance in Man (OMIM) and Orphanet. In the present study, the refinement process of the integrated dataset, which was required for CDSS applications, is discussed. Only 43% of the OMIM entries and 50% of the Orphanet entries, in the unified knowledge base, are considered appropriate for CDSS use. The remaining entries include congenital anomalies, genetic variants, and other inappropriate data. The resulting list and additional data of rare diseases suitable for clinical decision support will be made public to facilitate further researches on CDSSs with rare disease coverage.