Differences in Hospital Risk-standardized Mortality Rates for Acute Myocardial Infarction When Assessed Using Transferred and Nontransferred Patients

Abstract

Background: One in 5 patients with acute myocardial infarction (AMI) are transferred between hospitals. However, current hospital performance measures based on AMI mortality exclude these patients from the evaluation of referral hospitals.

Objective: To determine the relationship between risk-standardized mortality for transferred and nontransferred patients at referral hospitals.

Research design: This is a retrospective cohort study.

Subjects: Fee-for-service Medicare claims from 2011 for patients hospitalized with a primary diagnosis of AMI, at hospitals admitting at least 15 patients in transfer.

Measures: Hospital-specific risk-standardized 30-day mortality rates (RSMRs) for 2 groups of patients: those admitted through transfer from another hospital, and those natively admitted without a preceding or subsequent interhospital transfer.

Results: There were 304 hospitals admitting at least 15 patients in transfer. These hospitals cared for 77,711 natively admitted patients (median, 254; interquartile range, 162-321), and 11,829 patients admitted in transfer (median, 26; interquartile range, 19-46). Risk-standardized mortality rates were higher for natively admitted patients than for those admitted in transfer (mean, 11.5%±1.2% vs. 7.2%±1.1%). There was weak correlation between hospital performance as assessed by RSMR for patients natively admitted versus those admitted in transfer (Pearson r=0.24, P<0.001); when performance was arrayed by quartile, 102 hospitals (33.6%) differed at least 2 quartiles of performance across the 2 patient groups.

Conclusions: For Medicare patients with AMI, hospital-specific RSMRs for natively admitted patients are only weakly associated with RSMRs for patients transferred in from another hospital. Current AMI performance metrics may fail to provide guidance about hospital quality for transferred patients.

Figure 1

Cohort Selection Flow Diagram

Figure 2. Comparison of Hospital Performance

Hospitals that perform well for natively admitted patients do not necessarily perform well for patients admitted in transfer. Each data point represents an individual hospital, with risk-standardized 30-day mortality for natively admitted patients (X-axis) vs. risk-standardized 30-day mortality for patients admitted in transfer (Y-axis). Hospitals ranked in the top quartile for one group and the bottom quartile for another group are shaded in black. Hospitals that moved up or down two performance quartiles are shaded in grey.