A CIG Integration Framework to Provide Decision Support for Comorbid Conditions Using Transaction-Based Semantics and Temporal Planning

Managing comorbid conditions, i.e., patients with multiple medical conditions, is quite challenging for Clinical Decision Support Systems (CDSS) based on computerized Clinical Practice Guidelines (CPG). In case of comorbidity, CDSS will need to recommend treatments from multiple different CPG, which may adversely interact (e.g., drug-disease interactions), or introduce inefficiencies. A-priori, static integration of computerized comorbid CPG is insufficient for clinical practice. In this paper, we present a solution for dynamic integration of CPG in response to evolving health profiles. Using Description and Transaction Logics, we define a set of CIG integration semantics for encoding integration decisions that cope with comorbidity issues at execution-time. These dynamic, transaction-based semantics are well-suited to roll back prior decisions when no longer safe or efficient; or, inversely, apply new decisions when relevant. Moreover, comorbid CIG integration should consider temporal properties of CIG tasks—at execution-time, these properties will be influenced by a range of temporal constraints. Given all temporal constraints, optimal task schedules will be calculated that will determine the feasibility of CIG integration decisions.