King's College London

Objective: We assessed the association between population resection rates, hospital procedure volume and death rates in pancreatic cancer patients in England. Design: Patients diagnosed with pancreatic cancer were identified from a linked cancer registration and Hospital Episode Statistics dataset. Cox regression analyses were used to assess all-cause mortality according to resection quintile and hospital volume, adjusting for sex, age, deprivation and comorbidity. Results: There were 31,973 pancreatic cancer patients studied, 2580 had surgery. Increasing resection rates were associated with lower mortality among all patients (χ²(1df) = 176.18, p_trend < 0.001), with an unadjusted hazard ratio (HR) of 0.78 95%CI [0.75 to 0.81] in the highest versus the lowest resection quintile. Adjustment changed the estimate slightly (HR 0.82, 95%CI [0.79 to 0.85], (χ²(1df) = 99.44, p_trend < 0.001)). Among patients that underwent surgery, higher procedure volume was associated with lower mortality (HR = 0.88 95%CI [0.75-1.03] in hospitals carrying out 30+ versus <15 operations a year, shared frailty model, χ²(1df) = 1.82, p_trend = 0.177). Conclusion: Higher population resection rates were associated with lower mortality. The association with hospital procedure volume was less clear possibly due to small number of patients who underwent surgery. Nevertheless these results suggest survival is higher in hospitals that carry out a greater number of operations a year, particularly those doing 30+ operations, supporting the benefit of centralising perioperative expertise in specialist centres. Ensuring people are increasingly diagnosed when they are suitable candidates for surgery, and have access to these specialist centres may lead to an increase in the proportion of patients that undergo surgical resection which could plausibly increase survival of pancreatic cancer patients.