Measuring readmissions after surgery: do different methods tell the same story?

• We found that hospitals’ risk-adjusted hospital-wide readmission rates and potentially preventable readmission rates after surgery were highly correlated (r = .85, P < .0001).
• The overall agreement between these 2 methods in categorizing hospital performance was 82%.
• The disagreement was not significantly associated with hospital’s surgical volume (P = .475), surgical complexity (P = .543), or average length of stay (P = .355).

BackgroundReadmission is widely used as a quality metric to assess hospital performance. However, different methods to calculate readmissions may produce various results, leading to differences in classification with respect to hospital performance. This study compared 2 commonly used approaches to measure surgical readmissions: the 30-day all-cause hospital-wide readmissions (HWRs) and the potentially preventable readmissions (PPRs).MethodsWe examined the correlation between hospitals' risk-adjusted HWR and PPR rates and whether there was agreement in categorizing hospital performance between these measures among 111 hospitals with inpatient surgical programs in the Veterans Health Administration.ResultsWe found that hospitals' HWR and PPR rates were highly correlated (r = .85, P < .0001). The overall agreement between these 2 methods in categorizing hospital performance was 82% for all surgeries, 82% for colectomy, 84% for coronary bypass, and 87% for hip/knee replacement, respectively.ConclusionsDespite differences in methodologies, the HWR and the PPR measures provided relatively consistent perceptions of hospitals' performance on surgical readmissions.