Abstract 15634: Primary Prevention Defibrillator Use in a Safety-Net Setting

Abstract

Introduction: Disparities in use of implantable cardioverter-defibrillators (ICD) for primary prevention have been documented; however, whether ICDs may be underutilized within a safety-net healthcare system has not been studied. Methods: Using an electronic health record cohort of all patients with heart failure (HF) who received care within the public safety-net system in San Francisco from 2001-2018 linked to device implantation records from two health systems, we identified all patients with ICD implantation (adjudicated with manual chart review) and primary prevention ICD-eligible with no indication for secondary prevention ICD, EF<35% on 2 echocardiograms at least 90 days apart, HF hospitalization (to increase specificity for symptomatic HF), and prescribed beta blocker and ACEi/ARB. For exploratory survival analysis, eligibility and follow up started at the second echocardiogram. Results: Out of 14,230 patients with HF, at least 384 patients were potentially eligible for primary prevention ICD, but only 9 (2.3%) had an ICD implanted (2 CRT-D, 7 ICD). Of these 6 were nonischemic, and 3 had revascularized ischemic cardiomyopathy. Those who received ICD were younger (47 years vs 55 years; p=0.05), with lower rates of substance use (Table). Among 4 patients who died after ICD, median survival after device placement was 2.8 years (range 0.73 to 7.3); among 5 still alive, median follow up is 6.2 years (range 5.5 to 7.0). In an exploratory analysis, survival did not greatly differ by ICD placement (logrank p=0.53; age-adjusted HR 1.55; 95%CI 0.57-4.24). Conclusion: Among individuals receiving care at a safety-net system in San Francisco, even acknowledging that not all barriers to ICD placement are measured, very few guideline-eligible patients received primary prevention ICD. Additional studies are needed to identify patient-level and systems-level barriers to referral and implantation specific to a diverse, publicly insured patient population.