Clinical pathology data provides a basis for medical decision-making. Uncertainty is introduced into medical-decision-making by irreducible Biologic variability in patient data, ambiguities in communication of data, and the cognition of the decision-making process. Controlled medical terminologies (CMT) and medical decision support systems (DSS) reduce the uncertainty contributed by the latter two sources. Medical DSS are typically comprised of: an interface for data input and output, a knowledge base (KB) for storage of facts, and an inference engine for deriving new information. CMT employ knowledge representation formats similar to those used in DSS KBs. This study will evaluate the semantic network used by the Systematized Nomenclature of Medicine (SNOMED) for use as a KB in a Bayesian DSS to classify rodent liver disorders using clinical chemistry data. The primary aims of this proposal are to 1) assess SNOMED's ability to represent liver disorder concepts, 2) develop a formal methodology for adapting the semantic network of SNOMED for use in a Bayesian network and 3) evaluate the functionality of the resulting DSS using retrospective clinical chemistry data from drug toxicity studies. If SNOMED can functional as a KB to meet these aims, other domain-specific medical DSS can be created using the methodologies developed in this study.
| Dunker, A Keith (2013) Another disordered chameleon: the Micro-Exon Gene 14 protein from Schistosomiasis. Biophys J 104:2326-8 |
| Zimmerman, Kurt L; Wilcke, Jeff R; Robertson, John L et al. (2005) SNOMED representation of explanatory knowledge in veterinary clinical pathology. Vet Clin Pathol 34:7-16 |
