Safe automatic one-lead electrocardiogram analysis in screening for atrial fibrillation

Area of Science:

• Cardiology
• Medical Informatics
• Public Health

Background:

• Atrial fibrillation (AF) screening via electrocardiogram (ECG) can prevent AF-related stroke.
• Intermittent ECG recordings offer a potential method for mass AF screening.

Purpose of the Study:

• To validate an automated AF screening algorithm's performance against manual ECG analysis.
• To assess the algorithm's efficiency in identifying individuals at risk of AF-related morbidity.

Main Methods:

• The STROKESTOP study involved mass screening for AF using intermittent, 30-second, one-lead ECG recordings.
• A computerized algorithm analyzed 80,149 ECG recordings from 3,209 individuals, comparing results to manual interpretation by experts.

Main Results:

• The algorithm correctly identified normal ECGs (sinus rhythm/minor rhythm disturbances) in 87.1% of recordings.
• The algorithm demonstrated a 99.9% negative predictive value and 97.8% sensitivity for detecting AF.
• Manual interpretation workload was reduced from 288 to 35 ECGs per pathological finding.

Conclusions:

• Automated ECG screening using a computerized algorithm is safe and effective for identifying normal ECGs.
• The algorithm significantly reduces the need for manual ECG evaluation, improving screening efficiency.
• The algorithm achieves 100% sensitivity on an individual basis for AF detection.