Relationship between hospital mortality and readmission rates after transcatheter aortic valve replacement

There is substantial hospital-level variation in 30-day risk-standardized mortality rate (RSMR) and risk-standardized readmission rate (RSRR) after transcatheter aortic valve replacement (TAVR). However, the relationship between hospital RSMRs and RSRRs has not been well characterized.

We analysed data on 141 905 Medicare fee-for-service beneficiaries who underwent TAVR across 512 hospitals between 1 October 2015 and 31 December 2020. The primary and secondary outcomes of interest were 30-day all-cause mortality and 30-day all-cause readmissions, respectively. Hierarchical logistic regression models with random hospital-level intercepts were used to estimate RSMRs and RSRRs for each hospital. We used Pearson correlation coefficient (r) and restricted cubic spline regression to determine the relationship between RSMR and RSRR in the overall cohort and within subgroups based on hospital characteristics. The median [interquartile range (IQR)] hospital-level 30-day RSMR was 2.2% (2.1-2.4%), ranging from 1.3 to 3.5%. Similarly, the median (IQR) hospital-level 30-day RSRR was 13.2% (12.7-13.8%), ranging from 10.6 to 16.8%. In the overall cohort, there was weak correlation between 30-day RSMR and RSRR after TAVR (r = 0.25, 95% CI 0.17-0.33, P < 0.001). Subgroup analyses by hospital characteristics demonstrated the weakest correlation between RSMR and RSRR for non-JCAHO accredited hospitals (r = 0.07), hospitals in the Midwest (r = 0.12), and West (r = 0.14), and hospitals with low TAVR volume (r = 0.15).

Risk-standardized mortality and readmission rates after TAVR are weakly correlated, suggesting that hospital practices and processes of care influencing mortality are likely different from those influencing readmissions after TAVR, thereby necessitating measurement of both outcomes and developing specific interventions to decrease mortality and readmissions.