Do patients treated at academic hospitals have better longitudinal outcomes after admission for non-ST-elevation myocardial infarction?
Am Heart J. 2014 May;167(5):762-9
Authors: O'Brien E, Subherwal S, Roe MT, Holmes DN, Thomas L, Alexander KP, Wang TY, Peterson ED
BACKGROUND: Prior studies have found that academic hospitals provide more consistent use of guideline-recommended therapies in patients with non-ST-segment myocardial infarction (NSTEMI) compared with nonacademic centers, yet it is unclear whether these care differences translate into longer-term outcome differences.
METHODS: Using data from the CRUSADE Registry linked to Center for Medicare & Medicaid Services claims, we compared 30-day and 1-year all-cause mortality among 12,194 older patients with NSTEMI (age ≥65 years) treated at 103 academic centers and 28,335 patients treated at 302 nonacademic centers from February 2003 to December 2006. Outcomes were first adjusted for clinical characteristics, followed by adjustment for hospital performance, on 13 acute and discharge guideline-recommended therapies using a shared frailty model (an extension of the Cox proportional hazard model).
RESULTS: Compared with older patients with NSTEMI treated at nonacademic hospitals, those treated at academic hospitals had greater and more consistent use of evidence-based acute and discharge therapies, were more likely to receive in-hospital revascularization (61.1% vs 54.2%; P < .0001), and had modestly lower risk-adjusted 30-day mortality after adjustment for patient-level clinical characteristics (8.9% vs 10.2%, adjusted hazard ratio [HR] 0.89, 95% CI 0.80-0.99). These differences were attenuated (HR 0.94, 95% CI 0.83-1.02) after further adjustment for hospital delivery of evidence-based treatments, yet did not persist out to 1 year (unadjusted HR 0.92, 95% CI 0.84-1.01, P = .089).
CONCLUSIONS: Patients with NSTEMI treated at academic centers are more likely to receive guideline-recommended therapies and had modestly better 30-day outcomes. Nevertheless, these differences do not persist out to 1 year.
PMID: 24766988 [PubMed - in process]