Background: To study the effect of hospital volume and surgeon volume on postoperative hospital mortality, morbidity and long-term survival after resection of the head of the pancreas in a nationwide study (case record study), taking into a consideration risk factors found important in series based on experience in one hospital. Method: The case record investigation of 374 patients identified from the National Hospital Discharge Database as having undergone resection of the head of the pancreas between 1990 and 1994 in Finland. Results: The records of 350 patients were obtained for analysis. Operations were performed in 33 hospitals by 98 surgeons (average 2.1/year/hospital and 0.7/year/surgeon). Hospital mortality was 36/350 (10%), increasing from 4 and 7 to 13% with decreasing hospital volume from > 10 and 5-10 to < 5 respectively ( P < 0.05) and increasing from 3 and 10 to 14% with decreasing surgeon volume from > 3 and 1-3 to < 1, respectively ( P < 0.05). Most deaths were caused by surgical or technical complications (31/36 = 86%). Besides hospital mortality, postoperative complications, re-operations and hospital stay were also affected by surgeon volume. In the univariate analysis, also the age of the patient had an effect on the hospital mortality, and preoperative biliary stenting on the uncomplicated recovery, but in the multivariate analysis hospital mortality was independently affected by age (OR 0.94, P = 0.004) and surgeon volume (OR 1.3, P = 0.04), re-operations by surgeon volume (OR 1.10, P = 0.05) and hospital volume (OR 1.03, P = 0.05), postoperative complications by using the preoperative stent (OR 0.45, P = 0.02). Long-term survival was dependent on the histology of the specimen and by uncomplicated recovery, but not by hospital volume or surgeon volume. Conclusion: To decrease postoperative morbidity, mortality and hospital stay, pancreatic head surgery needs to be concentrated to only a few hospitals and to a few surgeons.