Artificial Intelligence Enabled ECGs for Atrial Fibrillation Identification and Enhanced Oral Anticoagulant Adoption A Pragmatic Randomized Clinical Trial summary

In this PACUPod episode, we review a pragmatic, cluster-randomized trial evaluating AI-assisted interpretation of standard 12-lead ECGs to identify atrial fibrillation and prompt guideline-directed oral anticoagulant therapy in hospitalized patients managed by non-cardiologists across two Taiwanese hospitals. The AI sends alerts with actionable recommendations to consider AF diagnosis and initiate NOACs. Primary outcomes include NOAC prescription within 90 days post-discharge, new AF diagnoses, echocardiography orders, and cardiologist referrals; secondary outcomes cover ischemic stroke, cardiovascular death, and all-cause mortality. Results show a significant increase in NOAC initiation (23.3% vs 12.0%; HR 1.85) and new AF diagnoses (HR 1.40) in the AI-alert group, with no significant differences in imaging or referrals or short-term clinical outcomes. The episode discusses the implications for real-world hospital workflows, the potential to close the AF treatment gap, and the need for longer follow-up, while situating findings within prior AI AF detection research and emphasizing multidisciplinary collaboration to optimize stroke prevention through timely anticoagulation.