Using Machine Learning Technologies to Develop a Smart System for Predicting Nstemi Patients Re-Admission Within 90 Days

Cardiovascular disease (CVD) is one of the greatest threats to human health, of which coronary heart disease (CHD) is the most common and accounting for about 13% of the total deaths globally each year. Acute coronary syndrome (ACS) is the most urgent and serious one among CHD. According to whether persistent ST-segment elevation on Electrocardiogram (ECG/EKG) and molecular biomarkers are present or not, there are three major types of ACS, including ST elevation myocardial infarction (STEMI), non-ST elevation myocardial infarction (NSTEMI) and unstable angina. The NSTEMI patients are very heterogeneous. Their conditions may be spontaneous remission, or worsen and develop into STEMI or even death. How to correctly classify the patients and provide proper care afterward is an important issue. However, though several types of risk scoring were applied to it, a perfect solution remains unavailable. For example, previous research shows that for patients admitted to the emergency departments of 39 hospitals in Taiwan, NSTEMI patients (10.1%) have higher mortality rate than STEMI patients (6.1%) who were actually more severe.This project aims to establish a novel method for precision risk evaluation for NSTEMI (PREN) to accurately assess the risk of NSTEMI patients in the emergency department and to correctly distinguish between high- and low-risk patients for providing them appropriate treatment recommendations. The project is expected to be implemented in three phases over three years. The first year will be focused on developing algorithms and comparing them with other existing scoring systems. In the second year, the study cohort will be expanded to further verify the algorithms and preparations as well as evaluations for introducing the method into the hospital information systems (HIS) will be performed. In the third year, clinical trial of the method will be conducted in hospitals. The new method is expected to be an effective tool to improve the medical quality of NSTEMI patients at high risk and to reduce the medical costs of patients at low risk.