Evaluation of AI-enhanced tele-ECG response time and diagnosis in acute chest pain patients

Comprehensive Summary

Accorsi et al. evaluated the impact of artificial intelligence (AI) on the response times and diagnostic accuracy of Tele-ECG services for patients with acute chest pain. At a single Brazilian Telemedicine center, 22,159 ECG tracings of patients with suspected ischemia at external emergency departments (ED) were analyzed. Seventeen cardiologists reviewed the ECGs, receiving assistance from a convolutional neural network (CNN) by automatically measuring ECG segments, detecting abnormalities, and alerting to potential ST-segment elevation myocardial infarction (STEMI). The primary endpoint was the turnaround time for AI-assisted ECG report generation and characterization of diagnoses at EDs not specialized in cardiology. Approximately 12% of tracings were uninterpretable due to technical issues including artifacts or lead reversal. The median response time for an ECG report was 75 seconds, and in STEMI cases, the median report time was 375 seconds (6.25 minutes). Most common diagnoses included diffuse ventricular repolarization changes (21.92%), sinus tachycardia (9.32%), and complete branch block (4.56%). STEMI was diagnosed in 202 cases (0.9%).

Outcomes and Implications

The study demonstrates that AI-enhanced Tele-ECG workflows resulted in quick response times, likely explained by the software’s ability to detect abnormalities early. These findings have particular implications for resource-limited EDs, where AI can support non-specialized staff in rapid clinical decision making for high-risk patients. In addition, the study’s low prevalence of STEMI cases shows that many low-risk patients access clinical care through EDs. The study is limited by its retrospective design, the exclusion of clinical history and longitudinal outcomes, and the inability to analyze 12% of tracings due to technical issues. Ultimately, AI enhances efficiency through early detection and prioritization of high-risk cases, but oversight by certified cardiologists remains necessary for diagnosis validation and final clinical decision making.