The effect of centralization on short term outcomes of pancreatoduodenectomy in a universal health care system

Abstract Background Centralization of pancreatic resections is advocated due to a volume-outcome association. Pancreatic surgery is in Norway currently performed only in five teaching hospitals. The aim was to describe the short-term outcomes after pancreatoduodenectomy (PD) within the current organizational model and to assess for regional disparities. Methods All patients who underwent PD in Norway between 2012 and 2016 were identified. Mortality (90 days) and relaparotomy (30 days) were assessed for predictors including demographic data and multi-visceral or vascular resection. Aggregated length-of-stay and national and regional incidences of the procedure were also analysed. Results A total of 930 patients underwent PD during the study period. In-hospital mortality occurred in 20 patients (2%) and 34 patients (4%) died within 90 days. Male gender, age, multi-visceral resection and relaparotomy were independent predictors of 90-day mortality. Some 131 patients (14%) had a relaparotomy, with male gender and multi-visceral resection as independent predictors. There was no difference between regions in procedure incidence or 90-day mortality. There was a disparity within the regions in the use of vascular resection (p = 0.021). Conclusion The short-term outcomes after PD in Norway are acceptable and the 90-day mortality rate is low. The outcomes may reflect centralization of pancreatic surgery.