Semantic-Based Conformance Checking of Computer Interpretable Medical Guidelines

Medical recommendations are generally expressed in natural language and therefore their ambiguity can lead to miss interpretations and medical errors. In this paper we propose an approach to 1) disambiguate medical recommendations by specifying them in a declarative language and 2) check the conformance of Computer Interpretable Guidelines (CIGs) with respect to the declarative specification of the medical recommendations. Our approach is based on semantic process mining techniques. Furthermore, we explain two ways to further exploit the information provided by the semantic conformance checker. To increase the accuracy of the model checker we suggest medical scenarios that were not considered for the enactment of the CIG and could show a violation of the medical constraints. Moreover, we discover scenarios which were not covered by the CIG but were considered for the medical recommendations.