Evaluation of hospital readmission rates as a quality metric in adult cardiac surgery

ObjectiveTo assess the reliability of 30-day non-elective readmissions as a quality metric for adult cardiac surgery.BackgroundUnplanned readmissions is a quality metric for adult cardiac surgery. However, its reliability in benchmarking hospitals remains under-explored.MethodsAdults undergoing elective isolated coronary artery bypass grafting (CABG), surgical aortic valve replacement/repair (SAVR) or mitral valve replacement/repair (MVR) were tabulated from 2019 Nationwide Readmissions Database. Multi-level regressions were developed to model the likelihood of 30-day unplanned readmissions and major adverse events (MAE). Random intercepts were estimated, and associations between hospital-specific risk-adjusted rates of readmissions and were assessed using the Pearson correlation coefficient (r).ResultsOf an estimated 86 024 patients meeting study criteria across 298 hospitals, 62.6% underwent CABG, 22.5% SAVR and 14.9% MVR. Unadjusted readmission rates following CABG, SAVR and MVR were 8.4%, 9.3% and 11.8%, respectively. Unadjusted MAE rates following CABG, SAVR and MVR were 35.1%, 32.3% and 37.0%, respectively. Following adjustment, interhospital differences accounted for 4.1% of explained variance in readmissions for CABG, 7.6% for SAVR and 10.0% for MVR. There was no association between readmission rates for CABG and SAVR (r=0.10, p=0.09) or SAVR and MVR (r=0.09, p=0.1). A weak association was noted between readmission rates for CABG and MVR (r=0.20, p