Quality of Care at Hospitals Identified as Outliers in Publicly Reported Mortality Statistics for Percutaneous Coronary Intervention

BACKGROUND: Public reporting of percutaneous coronary intervention (PCI) outcomes may create disincentives for physicians to provide care for critically ill patients, particularly at institutions with worse clinical outcomes. We thus sought to evaluate the procedural management and in-hospital outcomes of patients treated for acute myocardial infarction before and after a hospital had been publicly identified as a negative outlier. METHODS: Using state reports, we identified hospitals that were recognized as negative PCI outliers in 2 states (Massachusetts and New York) from 2002 to 2012. State hospitalization files were used to identify all patients with an acute myocardial infarction within these states. Procedural management and in-hospital outcomes were compared among patients treated at outlier hospitals before and after public report of outlier status. Patients at nonoutlier institutions were used to control for temporal trends. RESULTS: Among 86 hospitals, 31 were reported as outliers for excess mortality. Outlier facilities were larger, treating more patients with acute myocardial infarction and performing more PCIs than nonoutlier hospitals (P< 0.05 for each). Among 507 672 patients with acute myocardial infarction hospitalized at these institutions, 108 428 (21%) were treated at an outlier hospital after public report. The likelihood of PCI at outlier (relative risk [RR], 1.13; 95% confidence interval [CI], 1.12-1.15) and nonoutlier institutions (RR, 1.13; 95% CI, 1.11-1.14) increased in a similar fashion (interaction P= 0.50) after public report of outlier status. The likelihood of in-hospital mortality decreased at outlier institutions (RR, 0.83; 95% CI, 0.81-0.85) after public report, and to a lesser degree at nonoutlier institutions (RR, 0.90; 95% CI, 0.87-0.92; interaction P< 0.001). Among patients that underwent PCI, in-hospital mortality decreased at outlier institutions after public recognition of outlier status in comparison with prior (RR, 0.72; 9% CI, 0.66-0.79), a decline that exceeded the reduction at nonoutlier institutions (RR, 0.87; 95% CI, 0.80-0.96; interaction P< 0.001). CONCLUSIONS: Large hospitals with higher clinical volume are more likely to be designated as negative outliers. The rates of percutaneous revascularization increased similarly at outlier and nonoutlier institutions after report of outlier status. After outlier designation, in-hospital mortality declined at outlier institutions to a greater extent than was observed at nonoutlier facilities.