Improving outcome for patients with pancreatic cancer through centralization

Background: High-volume institutions are associated with improved clinical outcomes for pancreatic cancer. This study investigated the impact of centralizing pancreatic cancer surgery in the south of the Netherlands. Methods: All patients diagnosed in the Eindhoven Cancer Registry area in 1995-2000 (precentralization) and 2005-2008 (implementation of centralization agreements) with primary cancer of the pancreatic head, extrahepatic bile ducts, ampulla of Vater or duodenum were included. Resection rates, in-hospital mortality, 2-year survival and changes in treatment patterns were analysed. Multivariable regression analyses were used to identify independent risk factors for death. Results: Some 2129 patients were identified. Resection rates increased from 19.0 to 30.0 per cent (P < 0.001). The number of hospitals performing resections decreased from eight to three, and the annual number of resections per hospital increased from two to 16. The in-hospital mortality rate dropped from 24.4 to 3.6 per cent (P < 0.001) and was zero in 2008. The 2-year survival rate after surgery increased from 38.1 to 49.4 per cent (P = 0.001), and the rate irrespective of treatment increased from 10.3 to 16.0 per cent (P < 0.001). There was no improvement in 2-year survival in non-operated patients. After adjustment for relevant patient and tumour factors, those undergoing surgery more recently had a lower risk of death (hazard ratio 0.70, 95 per cent confidence interval 0.51 to 0.97). Changes in surgical patterns seemed largely to explain the improvements. Conclusion: High-quality care can be achieved in regional hospitals through collaboration. Centralization should no longer be regarded as a threat by general hospitals but as a chance to improve outcomes in pancreatic cancer.