An algebraic approach to detect logical inconsistencies in medical appropriateness criteria

Miguel García-Remesal,Victor Maojo,Luis Laita,Eugenio Roanes-Lozano,José Crespo

DOI: https://doi.org/10.1109/IEMBS.2007.4353499

Abstract:In this paper, we present a computerized approach to detect inconsistencies in medical knowledge bases. The method has been applied to a set of medical appropriateness criteria developed for the review of coronary artery disease management. One of the main problems associated to medical appropriateness criteria is to detect logical inconsistencies in the criteria set, a process often manually carried out by health services specialists. In our approach, appropriateness criteria are automatically translated to rules containing propositional variables, using three-valued Łukasiewicz's logic augmented with modal operators to manage uncertainty. The method assigns a polynomial to each of the rules, integrity constraints, and facts from the rule-based set. This rule set is then checked for inconsistencies. The problem of determining if a formula is a tautological consequence of a set of formulae is reduced by our method into an ideal membership problem in computer algebra. Finally, the set of medical appropriateness criteria is represented in a flowchart format that can be disseminated and remotely accessed over Internet, and can be prospectively used for patient care and management. The method reported in this paper can be applied to other knowledge bases represented by means of IF-THEN rules.