A hybrid approach for the verification of integrity constraints in clinical practice guidelines

Abstract

In the last decade, clinical practice guidelines are increasingly implemented in decision support systems able to promote their better integration into the clinical workflow. Despite the attempts involved to detect malformed, incomplete, or even inconsistent implementations of computerized guidelines, none of these solutions is concerned with directly embedding the theoretic semantics of a formal language as the basis of a guideline formalism in order to easily and directly support its verification. In such a direction, this paper proposes a formal framework which has been seamlessly embedded into a standards-based verifiable guideline model, named GLM-CDS (GuideLine Model for Clinical Decision Support). Such a framework hybridizes the theoretic semantics of ontology and rule languages to codify clinical knowledge in the form of a process-like model and, contextually, specify a set of integrity constraints to help to detect violations, errors and/or missing information. Its strong point relies on the capability of automatically verifying guidelines and, thus, supporting developers without the necessary technical background to construct them in a well-formed form. As a proof of concept, an actual guideline for Advanced Breast Cancer has been used to highlight some malformed implementations violating integrity constraints defined in GLM-CDS. © 2013 Springer-Verlag.