Medical treatment process modeling based on process mining and treatment patterns

Abstract

Clinical practice guidelines (CPGs) are statements relating to evidence-based and economically reasonable medical treatment processes (MTPs) for certain clinical circumstances. The executable MTPs in healthcare information systems can assist the clinical processes. In our previous work, several treatment patterns and their modeling proposals were proposed to reduce the effort spent in modeling MTPs. However, given a CPG document, all the process elements are modeled manually. Besides, business process mining can extract MTP models automatically from execution logs. However, the existing process mining algorithms focus on the control-flow of structured processes. This paper proposes an integrated framework for modeling executable MTPs based on process mining and treatment patterns, taking both the efficiency of mining and the quality of modeling into account. In this framework, an execution log processing approach is presented to identify the subsequences and decision points conforming to the treatment patterns and represent them with abstract activities. Experiments on a synthetic log of the non-secondary hypertension MTP and empirical findings demonstrate the effectiveness of our approach. The results show that the process mining in our approach framework can automatically generate more accurate MTP models, and the subprocess models based on treatment patterns make the models easy to understand.