We developed a framework for identifying and mitigating adverse interactions in multi-morbid patients managed according to multiple clinical practice guidelines (CPGs). The framework relies on first-order logic (FOL) to represent CPGs and secondary medical knowledge and FOL theorem proving to establish valid patient management scenarios. It handles many complexities of CPGs (e.g. time-based interactions) and also considers patient preferences. One limitation is its inability to capture hierarchical dependencies between concepts at different levels of granularity. This limitation results in a very detailed specification of secondary knowledge. In this work we address this shortcoming by expanding the FOL-based knowledge representation to handle hierarchical representations of drug classes.
Learning Objective 1: After participating in this session, the learner should be better able to understand the issues present when representing clinical knowledge for clinical decision support, and be able to discuss the use of logical inference in the context of a mitigation framework used to concurrently apply multiple clinical practice guidelines to a multi-morbid patient encounter.
Martin Michalowski (Presenter)
University of Minnesota
Szymon Wilk, Poznan University of Technology
Wojtek Michalowski, University of Ottawa
Marc Carrier, Ottawa Hospital Research Institute