An on-line computerized ECG program for the prediction of acute ischemic heart disease

Abstract

A logistic regression-based predictive instrument (PI) was developed to compute an emergency room (ER) patient's likelihood of acute cardiac ischemia (ACI) on the basis of clinical presentation and ECG (electrocardiography). The purpose of this study was to incorporate the PI into a computerized ECG so that a patient's probability of ACI could be provided automatically in real time with the ECG. The authors modified the original PI's variables to use the capabilities of a computerized ECG and programmed the PI's formula and variables into a Hewlett-Packard ECG using Electrocardiographic Language. Between the ECG and an electrocardiographer there was 85% testing agreement on the exact classification and measurement of ECG waveforms and agreement within 2% for the calculated probability of ACI. This application has considerable promise as a real-time decision aid for ER triage.