Explainable Decision Support Using Task Network Models in Notation3: Computerizing Lipid Management Clinical Guidelines as Interactive Task Networks

William Van Woensel, Samina Abidi, Karthik K. Tennankore, George Worthen, Syed Sibte Raza Abidi

Abstract: Knowledge-driven Clinical Decision Support (CDS) involves the computerization of paper-based clinical guidelines to issue evidence-based recommendations at points-of-care. The computerization of such guidelines in terms of a Task Network Model (TNM) conveniently models them as intuitive workflow models, which can be executed against patient health profiles. We present the GLEAN model that encodes an extensible Finite State Machine (FSM) executional semantics for modular TNM. Extensibility is provided in terms of a high-level formalism for defining execution semantics of custom TNM constructs. GLEAN is implemented using the Notation3 Semantic Web language, which provides powerful features for decisional criteria and queries, and offers integration with the HL7 FHIR standard. We explain CIG workflows as visual, intuitive workflow diagrams that are guided by a concrete patient profile at runtime. As a use case, we computerized guidelines on lipid management for Chronic Kidney Disease (CKD), a challenging problem for many Primary Care Providers (PCPs). To educate PCP on lipid management for CKD, we leverage GLEAN’s easy modularization of CIG and CIG explanations as visual runtime workflows.