Quality of surgical care and readmission in elderly glioblastoma patients.

Abstract

Thirty-day readmissions post medical or surgical discharge have been analyzed extensively. Studies have shown that complex interactions of multiple factors are responsible for these hospitalizations. A retrospective analysis was conducted using the Surveillance, Epidemiology and End Results (SEER) Medicare database of newly diagnosed elderly glioblastoma multiforme (GBM) patients who underwent surgical resection between 1991 and 2007. Hospitals were classified into high- or low-readmission rate cohorts using a risk-adjusted methodology. Bivariate comparisons of outcomes were conducted. Multivariate analysis evaluated differences in quality of care according to hospital readmission rates. A total of 1,273 patients underwent surgery in 338 hospitals; 523 patients were treated in 228 high-readmission hospitals and 750 in 110 low-readmission hospitals. Patient characteristics for high-versus low-readmission hospitals were compared. In a confounder-adjusted model, patients treated in high- versus low-readmission hospitals had similar outcomes. The hazard of mortality for patients treated at high- compared to low-readmission hospitals was 1.06 (95% CI, 0.095%-1.19%). While overall complications were comparable between high- and low-readmission hospitals (16.3% vs 14.3%; P = .33), more postoperative pulmonary embolism/deep vein thrombosis complications were documented in patients treated at high-readmission hospitals (7.5% vs 4.1%; P = .01). Adverse events and levels of resection achieved during surgery were comparable at high- and low-readmission hospitals. For patients undergoing GBM resection, quality of care provided by hospitals with the highest adjusted readmission rates was similar to the care delivered by hospitals with the lowest rates. These findings provide evidence against the preconceived notion that 30-day readmissions can be used as a metric for quality of surgical and postsurgical care.