Acute coronary syndrome (ACS) is considered a life-threatening disease that contributes to high morbidity and mortality worldwide. Approximately 1.5 million hospital discharges involve patients with ACS. Hence, early mortality predictions for ACS patients can benefit both patients and hospitals. Existing approaches only build predictive models based on data flow while ignoring sequential relations (i.e., control-flow) between treatments, which leads to information loss. Moreover, they focus on post hoc analysis, in which patients have already been cleared or discharged from hospitals. This study investigates the application of process mining techniques to mortality predictions. We propose a process pattern extraction approach to discover common treatment pathways and a predictive process monitoring framework to early predict mortality for ACS patients currently under treatment based on data flow and discovered process features. We achieve 97% and 86% accuracy for short- and long-term mortality predictions for ACS inpatients utilizing a real-life EHR dataset.
CHEN, QIFAN; Lu, Yang; Tam, Charmaine; and Poon, Simon, "Predictive Process Monitoring for Early Predictions of Short- and Long-Term Mortality for Patients with Acute Coronary Syndrome" (2022). PACIS 2022 Proceedings. 22.
When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.