Background: Recent support for centralization of complex operations, such as pancreaticoduodenectomy (PD), is based on surgeon-specific volume-outcome relationships. This study examined whether volume of anatomically related operations (operative mix), besides PD, is also independently associated with postoperative outcomes after PD.
Methods: The study queried the Nationwide Inpatient Sample (2004-2009) for surgeons performing PD. Operative mix (OM) was defined as the year-specific number of other pancreatic, hepatic, biliary, and gastric operations performed by individual surgeons. Regression models included surgeon and hospital PD volume, adjusted for other hospital- and patient-specific factors.
Results: Among 1747 surgeons, 88.3% had low PD volume (≤ 5 cases/year), 8.9% had moderate PD volume (6-16 cases/year), and 2.8% had high PD volume (≥ 17 cases/year). Low-PD-volume surgeons with high OM (≥ 21 cases/year) (4.4%), moderate-PD-volume surgeons with high OM (3.4%), and high-PD-volume surgeons with high OM (2.7%) each had lower mortality than low-PD-volume surgeons with low OM (9.3%; all p ≤ 0.02). The frequency of prolonged hospitalization among low-PD/high-OM surgeons (45.3%) was lower than among low-PD/low-OM surgeons (61.6%; p < 0.001). Increasing OM volume was associated with decreased inpatient mortality, shorter hospital stay, and lower likelihood of any postoperative complication, using unadjusted regression (all p < 0.001). Adjusted regression results indicated that increasing OM volume is a significant predictor of decreased odds of a prolonged hospital stay (odds ratio [OR] 0.81; 95% confidence interval [CI] 0.73-0.90; p < 0.001).
Conclusions: Surgeon PD volume is an important predictor of outcomes after PD. However, surgeon OM volume identifies a subset of lower-PD-volume surgeons with more favorable outcomes.