Inefficiency of Care in the Hub-and-Spoke System

Abstract

INTRODUCTION: Complex surgical care is often centralized at one high volume hospital within a system, commonly referred to as the hub. We hypothesized that the potential performance advantages of hub hospitals seen for complex operations would not extend to common operations. METHODS: Using the Healthcare Cost and Utilization Project’s State Inpatient Databases, adult surgical patients who underwent one of 10 general surgery operations in 13 geographically diverse states (2016 to 2018) were identified. System status was assigned using the American Hospital Association survey. Hub designation was assigned to the hospital performing the greatest number of each specific operation. Multivariable logistic regression was used to evaluate the risk-adjusted odds of death or serious morbidity (DSM) and prolonged length of stay (LOS) at hub hospitals relative to other system hospitals (spokes), stratified by operation complexity. RESULTS: A total of 128,248 patients across 44 hub-and-spoke systems, including 133 hospitals were included, and 77.9% of complex and 61.6% of common operations occurred at a hub. Overall, 22.2% of patients experienced prolonged LOS and 12.8% experienced DSM. For complex operations, risk-adjusted odds of DSM were significantly lower in hubs. For common operations, risk-adjusted odds of DSM were similar between hubs and spokes while the odds of prolonged length of stay were greater at hubs (Table). CONCLUSION: The DSM benefit of hub hospitals for complex operations did not extend to common operations. Instead, patients undergoing common operations at hub hospitals had similar rate of DSM with higher rate of prolonged LOS. This finding shows a clear opportunity for improved system efficiency. Table. - Risk-Adjusted Outcomes of Common and Complex Operations at Hub Hospitals (Reference = Spokes) Variable Death or serious morbidity for common operations Prolonged length of stay for common operations Death or serious morbidity for complex operations Prolonged length of stay for complex operations Observations (n) 118,764 118,762 9,484 9,483 Hub hospitals odds ratio 1.01 1.08* 0.84* 1.06 Confidence interval (0.97, 1.05) (1.05, 1.12) (0.72, 0.97) (0.93, 1.20) *Significance at the p < 0.05 level.Model covariates include race, age, gender, income quartile, insurance type, state, admission status, and operation type. Common operations include cholecystectomy, inguinal hernia repair, mastectomy, colon resection, and small bowel resection. Complex operations include total gastrectomy, Whipple, hepatic lobectomy, esophagectomy, and rectal resection.