Abstract 371: Proportion of Readmissions to Non-Index Hospitals is Associated with All-Cause Risk-Adjusted Readmission Rate for Medicare Patients with AMI and CHF

Abstract

Background: The Affordable Care Act introduced the Hospital Readmissions Reduction Program (HRRP), implementing financial penalties for hospitals with excess risk-standardized readmission rates (RSRRs) for specific diagnoses including congestive heart failure (CHF) and acute myocardial infarction (AMI). Hospitals may have more ability to reduce readmission rates for readmissions at their own hospitals as opposed to other hospitals. We hypothesize that hospitals with higher proportions of patients re-admitted to a different (“non index”) hospital also have higher all-cause 30-day RSRRs. Methods: We used a national sample of Medicare data to identify hospitalizations for CHF and AMI and subsequent all-cause 30-day readmissions using data from 5,291 hospitals for the period 1999 to 2010. For each hospital we calculated the proportion of readmissions at a different (“non-index”) hospital and the all-cause 30-day risk-adjusted readmission rate (RSRR). We then fitted a regression model to estimate the relationship between the proportion of admissions to the non-index hospital and all-cause RSRRs across hospitals. Results: We identified 2,863,696 discharges for AMI and 6,981,895 discharged for CHF for the period 1999 to 2010. On average across hospitals, the percentage of readmissions to non-index hospitals was 28.0% (SD 25.4%) for AMI and 22.7% (SD 17.8%) for CHF. Hospitals with higher RSRRs had higher percentages of their readmissions occurring at a different hospital for both AMI and CHF (p < 0.001 for both conditions). Due to this association, hospitals in the highest decile for 30-day RSRR had 29.1% and 20.7% of readmissions occur at non-index hospitals for AMI and CHF, respectively, while the corresponding proportions for hospitals in the lowest decile were 19.1% and 18.4%. Conclusions: Compared to high performing hospitals, hospitals with the worst 30-day readmission rates for AMI and CHF also have the highest percentage of those readmissions occurring at hospitals other than their own. The relationship was particularly strong for AMI. These results suggest that the proportion of readmissions to other hospitals might be a structural disadvantage associated with higher readmission rates.