A four-step method to centralize pancreatic surgery, accounting for volume, performance and access to care

Background: Adequate criteria for pancreatic surgery centralization are debated. This retrospective study aimed to define a reproducible method for complex care centralization, accounting for hospital performance and access to care. Methods: The method consisted in: 1. Analysis of overall outcome and mortality-related factors. 2. Assessment of volume and adjusted mortality of each hospital. 3. Definition of different centralization models. 4. Final adjustments to guarantee access to care, evaluating travel times and waiting lists. This method was tested on Lombardy, the most populous Italian region (about 10 million inhabitants, 24 000 km2). Results: According to Ministry of Health data, 79 hospitals performed 3037 resections in 2014–2016. Mean overall mortality was 5.0%, increasing from 2.3%, of seven high-volume facilities (>30 resections/year) to 10.7% of 56 low-volume facilities (<10 resections/year). Five centralization models were tested (range: 7–23 hospitals): the best performing model included seven high-volume facilities, providing both low mortality (<2%), and easy access to care, namely reasonable travel time (≤60 min for >90% of the population), and limited impact on waiting list (1.1 extra-resection/hospital/week). Conclusion: The four-step method appears as a flexible tool to centralize pancreatic surgery, allowing regulatory institutions to estimate the effect of different models.