MS.14. Modeling and processing of clinical practice guidelines

Clinical practice guidelines (CPGs) are knowledge-based tools for disease-specific patient management that encapsulate best practices in identifying relevant patient data, drawing diagnostic conclusions from these data and prescribing the most appropriate therapies. While it is generally agreed that the use of CPGs at the point of care have a positive impact on patient’s outcomes, their practical adoption is still limited. One of the main obstacles is multi-morbidity (presence of several diseases in the same patient), when application of many CPGs may result in undesirable interactions and deteriorate the quality of provided care. This calls for methods and tools to identify and mitigate possible conflicts between CPGs and combining multiple CPGs for multi-morbidities – development of such methods constitutes one of the “grand challenges” for clinical decision support in the coming years.

A CPG can be seen as a specialized form of a workflow and there have been attempts to employ well established solutions (e.g. BPMN) to model and process clinical guidelines. However, specific characteristic of CPGs calls for dedicated methods that are better suited to capture and deal with this characteristic. Thus, the aim of this research topic is to: (1) propose a formal representation of CPGs and domain knowledge that facilitates identification and mitigation of possible conflicts, (2) develop algorithms for identifying and mitigating conflicts between multiple concurrently applied CPGs, (3) develop a decision support system for planning and simulating therapies that implements these algorithms and (4) verify the developed system in several real-life scenarios.

Main Advisor at Poznan University of Technology (PUT)
Co-advisor at Universitat Politècnica de Catalunya (UPC)