Investigating Application of Change Point Analysis in Monitoring Health Condition of Acute Coronary Syndrome Patients

Abstract

This study investigated the application of Change Point Analysis (CPA) in monitoring the health condition of Acute Coronary Syndrome (ACS) patients. In order to risk stratification and reduce the damage of the heart tissue, monitoring Troponin level of ACS patients and detecting subtle and potentially important changes as early as possible is critical during the diagnosis and treatment process. However, since humankind has fairly limited cognitive processing capacity, medical experts have been challenged to manually analyze each patient’s data, detect subtle changes and make a decision about next treatment steps as early as possible. For this reason, this paper proposes a Clinical Decision Support System (CDSS) by developing an automated way of monitoring the health condition of ACS patients using change point analysis. A time series data about Troponin test results of 461 ACS patients were collected from Almazov National Medical Research Centre. CPA was employed as a combination of Cumulative Sum (CUSUM) charts and bootstrapping to detect changes. As a result, single and multiple changes in the level of Troponin of ACS patients were detected with 95% confidence level.