Point of care ai driven er2ep referrals for non-valvular atrial fibrillation patients in the emergency department

Emergency Room to Electrophysiology (ER2EP) Atrial Fibrillation (AF) referrals may result in shorter times to ablation and optimal medical therapy. An effective way to facilitate ER2EP referrals can be the use of point of care digital health solutions like an Artificial Intelligence/Machine Learning (AI/ML) clinical decision support tool. To determine if a cloud based Artificial Intelligence/Machine Learning (AI/ML) clinical decision support tool can aid in guideline-driven care and procedures. This RCT enrolled patients > 18 years old who presented to the ER with a diagnosis of Atrial Fibrillation. Subjects were excluded if they were receiving anticoagulation (AC) for an indication other than AF, hemodynamically unstable, or lacked capacity to provide informed consent. Patients were randomized 1:1 to receive either standard of care or cloud based AI/ML ER2EP clinical support tool (Lucia AF mobile app) to detect and treat AF. Follow up phone calls up to 30 days evaluated for ER2EP referral compliance, adverse events, and readmission. Overall, 40 patients were enrolled, 43% female, 57% male, 70% Hispanic, 30% Non-Hispanic. Twenty-two patients cared for by physicians receiving ER2EP output (55%) were compared to controls. ER2EP patients had more primary AF patients (22.3% vs 54.6%, P=0.025), although both cohorts received similar rates of anticoagulation (66.7% vs 72.7%, p=.68). Anticoagulation, scoring of CHA2DS2-VASc, EP referrals, LAAC device placement, and RF ablation rates were higher in the ER2EP cohort (95% CI = -55.1, -11.6, P = 0.003). One hundred percent of ER2EP recommendations were followed. ER2EP have a greater rate of AF diagnosis, with higher rates of AC, scoring of the CHA2DS2-VASc, EP referrals, LAAC device placement, and RF ablation within 30 days.