Extending Quality Improvement for Pancreatoduodenectomy Within the High-Volume Setting: The Experience Factor.

ABSTRACT

OBJECTIVE: To analyze the association of a surgeon's experience with postoperative outcomes of pancreatoduodenectomies (PDs) when stratified by Fistula Risk Score (FRS). SUMMARY BACKGROUND DATA Centralization is now well-established for pancreatic surgery. Nevertheless, the benefits of individual surgeon's experience in high-volume settings remains undefined. METHODS: Pancreatoduodenectomies performed by 82 surgeons across 18 international, specialty institutions (median140 PD/year) were analyzed. Surgeon cumulative PD volume was linked with postoperative outcomes through multivariable models, adjusted for patient/operative characteristics and the FRS. Then, surgeon experience was also stratified by the ten, previously defined, most clinically impactful scenarios for clinically-relevant pancreatic fistula (CR-POPF) development. RESULTS: Of 8,189 PDs, 18.7% suffered severe complications (Accordion≥3), 4.8% were reoperated upon and 2.2% expired. Although the most experienced surgeons (top-quartile; >525 career PDs) more often operated on riskier cases, their experience was significantly associated with declines in CR-POPF (P<0.001), severe complications (P=0.008), reoperations (P<0.001), and length of stay (LOS) (P<0.001) - accentuated even more in the most impactful FRS scenarios (2,830 patients). Risk-adjusted models indicate male gender, increasing age, ASA class and FRS, but not surgeon experience, as being associated with severe complications, failure-to-rescue and mortality. Instead, upper-echelon experience demonstrates significant reductions in CR-POPF (OR 0.66), reoperations (OR 0.64) and LOS (OR 0.65) in moderate-to-high fistula risk circumstances (FRS≥3, 68% of cases). CONCLUSIONS: At specialty institutions, major morbidity, mortality and failure-to-rescue are primarily associated with baseline patient characteristics, while cumulative surgical experience impacts pancreatic fistula occurrence and its attendant effects for most, higher-risk pancreatoduodenectomies. These data also suggest an extended proficiency curve exists for this operation.