Cluster Randomized Trial Examining the Impact of Automated Best Practice Alert on Rates of Implantable Defibrillator Therapy

Background Implantable cardioverter-defibrillators (ICDs) are indicated in patients with left ventricular ejection fraction ≤35%, but many eligible patients do not receive this therapy. In this cluster randomized trial, we investigated the impact of a best practice alert (BPA) through the electronic health records on the rates of electrophysiology referrals, ICD implantations, and all-cause mortality in severe cardiomyopathy patients. Methods and Results Providers in the Heart and Vascular Institute (n=106) and in General Internal Medicine (n=89) were randomized to receive or not receive a BPA recommending consideration for ICD implantation. Patients belonging to the BPA and no BPA groups of providers were followed to the end points of electrophysiology referrals, ICD implantations, and all-cause mortality. Between 2013 and 2015, patients with reduced left ventricular ejection fraction were managed by 93 providers in the BPA (n=997 patients) and 102 providers in the no BPA (n=909 patients) groups. Patients in the 2 groups had comparable baseline characteristics. After a median follow-up of 36 months, 638 (33%) patients were referred to electrophysiology, 536 (27%) received an ICD, and 445 (23%) died. Patients in the BPA group were more likely to be referred to electrophysiology (hazard ratio=1.23; P=0.026), to receive ICD therapy (hazard ratio=1.35; P=0.006), and exhibited a trend towards slightly lower mortality (hazard ratio=0.85; P=0.091). Conclusions Delivering a BPA through the electronic health record recommending to providers consideration of ICD implantation when the left ventricular ejection fraction is ≤35% improves the rates of electrophysiology referrals and ICD therapy in patients with severe left ventricular dysfunction.