Machine Learning For Diagnosis Of Myocardial Infarction Using Cardiac Troponin Concentrations

This article presents the development of a clinical decision support system, CoDE-ACS, using machine learning with single or serial high-sensitivity cardiac troponin measurements to inform the probability of acute myocardial infarction. The model was trained to identify patients with an adjudicated diagnosis of acute myocardial infarction and was tested in a stepped-wedged cluster-randomized, controlled trial. The results showed that CoDE-ACS was superior to pathways that use fixed cardiac troponin thresholds or risk scores and performed consistently across different health care systems and patient subgroups. The proposed care pathway could reduce time spent in emergency departments, prevent unnecessary hospital admissions and improve the early treatment of myocardial infarction.