Artificial Intelligence-based Remote Electrocardiogram MonitoringImproves the Diagnosis of Arrhythmias and Reduces Morbidity Caused by Fatal Arrhythmias - Abstract
Background: Remote ECG monitoring can provide valuable information for diagnosis in patients with cardiovascular diseases (CVD), but manually studying large amounts of
ECG data can be tedious and time-consuming. The aim of this study was to evaluate the values of AI-based ECG monitoring and data analysis in clinical practice.
Methods: Between January 2018 and October 2019, a total of 16,408 patients evaluated for suspected CVD at Hefei Hi-Tech Cardiovascular Hospital and its associated
community hospitals/clinics underwent i-Holter, a heart remote mobile monitoring device from Yocaly Information Science & Technology Co., Ltd, Lepu Medical, China. The i-Holter
monitor is a remote, AI-based, real-time heart monitoring system with built-in pre-warning functions. The ECG data obtained from remote monitoring was screened and analyzed
by this AI-based ECG analysis system and the final diagnosis was confirmed by ECG specialists. The number of pre-warning alerts, classification of arrhythmias, subsequent patient
management, and outcomes were also analyzed.
Results: Of the 16,408 patients, 68.3% were detected to have different arrhythmias. The most common is supraventricular arrhythmia (40.7%), followed by ventricular
arrhythmia (30.2%), ST-T segmental alternation (20.5%), cardiac pause (3.7%), atrial-ventricular blockage (2.5%), and branch blockage (2.2%). A total of 3,351 patients were
alerted due to reaching critical ECG values and 8,874 phone calls were taken for the alerts. Among the alerts, 37.1% were supraventricular tachycardia, 36.4% were ST-T
alternation, following by atrial flutter/atrial fibrillation (7.4%), ventricular tachycardia (6.4%), sinus tachycardia (6.1%), III° A-V block (3.3%), cardiac pause (1.98%), and ventricular
flutter/fibrillation (0.03%). 1,229 patients (36.7%) with alerts needed promote management and obtained good outcomes.
Conclusion: AI-based remote monitoring can manage high volume ECG data, provide prompt diagnosis with high accuracy, initiate alert timely if critical values are reached.
Patients in fatal situations can be managed in a timely manner, improving patient prognosis and leading to reduced mortality.