Role of Artificial Intelligence in Reducing Data Deluge From Cardiac Implantable Electronic Devices

BACKGROUND: Remote monitoring (RM) is the standard of care for patients with cardiac implantable electronic devices (CIEDs). Data deluge from RM of CIEDs is one of the major challenges being faced by cardiac device clinics. OBJECTIVES: The purpose of this study was to highlight the role of artificial intelligence (AI) in reducing data deluge from CIEDs and to elucidate its performance in augmenting the traditional device clinic workforce. METHODS: Deidentified data from RM of CIEDs from 78 device clinics across the United States were included. Transmissions processed by AI and device technicians (DTs) were compared. Around 15% of all transmissions processed were examined at random by 2 cardiac electrophysiologists who were blinded to the source of the interpretation. RESULTS: Over 8 months, 690,673 transmissions were generated from RM of 74,217 devices. AI processed 28.3% of the transmissions, whereas DT processed 71.7% (n = 495,054). AI forwarded 22.8% (n = 44,550) while DT forwarded 38.5% (n = 190,492) of the total processed transmissions to the device clinics. As such, only 34% (n = 235,042) of the total transmissions were forwarded to the device clinics. Considering electrophysiologist interpretation/adjudication as the gold standard, AI was 93.8% accurate with 99.1% sensitivity, 93% specificity, and an F1 score of 81%. CONCLUSIONS: This study presents a real-world analysis of incorporating AI in device clinic workflow to assist in interpretation and classification of RM data. AI interpretation of transmissions was found to have an accuracy similar to that of DT with superior sensitivity but lesser specificity.