Abstract 241: Registry vs. Real World: Revascularization Treatment and In-Hospital Mortality Among STEMI Patients

Abstract

Objective: Hospital-based cardiovascular disease registries provide important clinical data on treatments and outcomes. However, the voluntary paid-membership nature of these registries makes them vulnerable to selection bias. In this study, recent ST-elevation myocardial infarction (STEMI) patient treatment rates were compared between registry and non-registry hospitals. Methods: The published list of Florida hospitals in the Acute Coronary Treatment and Intervention Outcomes Network (ACTION) Registry during 2013-2014 was obtained from the ACTION website. All-payer hospital discharge data with 100% coverage were analyzed. STEMI patient characteristics, treatment, and outcomes were compared for registry vs. non-registry hospitals for one year, using detailed ICD-9-CM diagnosis and procedure codes. Treatments identified were primary percutaneous coronary intervention (PCI), coronary artery bypass graft surgery (CABG), and infusion/injection of fibrinolytic drugs (LYTICS). Results: A minority of acute-care hospitals participated in the ACTION registry in Florida during 2013-2014, admitting 1/3 of STEMI patients. Statewide, there were 10,592 STEMI patients hospitalized. In the ACTION hospitals, treatment rates were: same-day PCI = 70.5%, CABG = 8.7%, PCI on a later day = 6.2%, PCI with timing unknown = 2.5%, LYTICS alone = 0.1%. Patients who received no treatment comprised 12.1% of patients in registry hospitals vs. 15.6% in non-registry hospitals. Registry no treatment rates were lower for all race-gender groups, with the greatest difference in Black women (15.7% vs. 21.8% no treatment). Treatment rates varied by age (see Figure). After Medicare eligibility age (65+), treatment differences narrowed. Unadjusted in-hospital mortality rates were greater in non-registry hospitals for patients who received same-day PCI (5.3% vs. 4.1%) and PCI later (5.5% vs. 2.5%), but were lower for CABG patients (5.3% vs. 8.0%). Conclusions: Patients treated at non-ACTION hospitals were somewhat less likely to receive any revascularization therapy, and had higher overall unadjusted rates of mortality. Registry participating institutions may be a selected “best practices” group of hospitals, from which national generalizations should be derived with caution.