Determinants of Outcomes Following Resection for Pancreatic Cancer—a Population-Based Study

Mary A Waterhouse, Elizabeth A Burmeister, Dianne L O'Connell, Emma L Ballard, Susan J Jordan, Neil D Merrett, David Goldstein, David Wyld, Monika Janda, Vanessa L Beesley, Madeleine E Payne, Helen M Gooden, Rachel E Neale

Patient and health system determinants of outcomes following pancreatic cancer resection, particularly the relative importance of hospital and surgeon volume, are unclear. Our objective was to identify patient, tumour and health service factors related to mortality and survival amongst a cohort of patients who underwent completed resection for pancreatic cancer. Eligible patients were diagnosed with pancreatic adenocarcinoma between July 2009 and June 2011 and had a completed resection performed in Queensland or New South Wales, Australia, with either tumour-free (R0) or microscopically involved margins (R1) (n = 270). Associations were examined using logistic regression (for binary outcomes) and Cox proportional hazards or stratified Cox models (for time-to-event outcomes). Patients treated by surgeons who performed <4 resections/year were more likely to die from a surgical complication (versus ≥4 resections/year, P = 0.04), had higher 1-year mortality (P = 0.03), and worse overall survival up to 1.5 years after surgery (adjusted hazard ratio 1.58, 95 % confidence interval 1.07-2.34). Amongst patients who had ≥1 complication within 30 days of surgery, those aged ≥70 years had higher 1-year mortality compared to patients aged <60 years. Adjuvant chemotherapy treatment improved recurrence-free survival (P = 0.01). There were no significant associations between hospital volume and mortality or survival. Systems should be implemented to ensure that surgeons are completing a sufficient number of resections to optimize patient outcomes. These findings may be particularly relevant for countries with a relatively small and geographically dispersed population.