Defining Operative Thresholds for High-Volume Pancreaticoduodenectomy Centers to Improve Surgical Outcomes

Abstract

Background: Performance of pancreaticoduodenectomy at high-volume (HV) hospitals has been shown to reduce operative mortality. However, what threshold of case volume meaningfully impacts surgical outcomes is unknown. We determined a clinically meaningful cut-point for defining HV pancreaticoduodenectomy centers and assessed the impact of hospital volume on surgical outcomes. Methods: We identified adult NCDB patients undergoing pancreaticoduodenectomy between 2004-2015. A multivariable model with restricted cubic splines was built to predict 5-year overall survival according to average yearly case volume, adjusting for demographic/clinicopathologic factors. This data was then analyzed using a change-point procedure which identified two regression lines and a distinct segmentation at 24.6 cases/year. Hospitals were subsequently divided into HV (>/=25 cases/year) and low-volume (LV, <25 cases/year) groups. Chi-square tests were used to analyze group differences. Logistic regression was utilized to assess the impact of hospital volume on surgical outcomes. Results: Our cohort consisted of 32,333 patients (13.2% treated at HV centers). Treatment at a HV center was associated with increased likelihood of R0 resection and adequate lymph node evaluation. HV centers also had significantly decreased 30- and 90-day postoperative mortality (OR 0.34, 95% CI 0.25-0.46 and OR 0.48, 95% CI 0.40-0.58; respectively) after adjusting for age, sex, race, comorbidities, histology, and stage. Conclusion: Treatment at a HV pancreaticoduodenectomy center is associated with significantly improved surgical outcomes. Our study identified 25 cases/year as a clinically meaningful distinction between HV and LV centers which should be considered when evaluating regionalization of pancreatic cancer care to improve patient outcomes.