In fiscal year (FY) 2013, the Centers for Medicare and Medicaid Services (CMS) began penalizing US hospitals for excess 30-day readmissions in Medicare patients with heart failure (HF), pneumonia, and acute myocardial infarction (AMI). Since then, 30-day risk standardized readmission rates (RSRR) have decreased for these conditions. Because readmissions and deaths are competing events, we investigated how 30-day risk standardized mortality rates (RSMR) have changed for HF and AMI over this time period. Methods:
Publically reported hospital data was obtained for FY 2013 and 2016: RSRR, RSMR, and excess readmission ratios (ERR) for HF and AMI, calculated as 3-year averages. Pneumonia was excluded because the risk model changed to include aspiration and sepsis. For FY 2013 and 2016, RSMRs at hospitals that received penalties for excess readmissions (ERR > 1) were compared to those that received no readmission penalties (ERR < 1). Next, the difference in RSRRs and RSMRs for both conditions between 2013 and 2016 were calculated for individual hospitals. For HF and AMI, ANOVA was used to compare the mean RSMR change among hospitals with increased (>1%), stable (−1 to 1%), and decreased (>−1%) RSRRs. Results:
We identified 3467 and 2039 US hospitals with RSMRs and RSRRs for HF and AMI respectively. ERRs for HF and AMI were available for 2819 and 1815 hospitals. For both 2013 and 2016, mean RSMRs for HF were greater (by 0.4% and 0.5% respectively) at institutions that received no readmission penalties compared to those that were penalized (P
< .001). Similar findings were not present for AMI. Between 2013 and 2016, 73% of hospitals experienced a reduction in RSRRs for HF. However, RSMRs increased in 58% of them. Conversely, 67% of hospitals improved in both RSMR and RSRR for AMI (Figure A, B
). The mean increase in RSMRs for HF was 0.3% whereas RSRRs decreased by an average of 1% (P
< .001). For AMI, RSMRs and RSRRs decreased by an average of 1% and 1.4% respectively (P
< .001). RSMRs for HF increased irrespective of changes in RSRRs. For AMI, RSMR decreased for all three readmission categories (Figure C
While 30-day readmission rates have improved for HF, 30-day HF mortality rates have increased at more than half of US hospitals since the advent of CMS readmission penalties. Furthermore, 30-day HF mortality remains higher in hospitals that do not receive CMS readmission penalties, which have historically been driven by HF readmissions. In contrast to AMI, successful recent efforts to prevent “excess” 30-day HF readmissions have not been accompanied by lower HF mortality.