Assessing surgical quality: comparison of general and procedure-specific morbidity estimation models for the risk adjustment of pancreaticoduodenectomy outcomes.

Introduction: The use of outcomes to evaluate surgical quality implies the need for detailed risk adjustment. The physiological and operative severity score for the enumeration of mortality and morbidity (POSSUM) is a generally applicable risk adjustment model suitable for pancreatic surgery. A pancreaticoduodenectomy (PD)-specific intraoperative pancreatic risk assessment (IPRA) estimates the risk of postoperative pancreatic fistula (POPF) and associated morbidity based on factors that are not incorporated into POSSUM.

Objective: The aim of the study was to compare the risk estimations of POSSUM and IPRA in patients undergoing PD.

Methods: An observational single-center cohort study was conducted including 195 patients undergoing PD in 2008-2010. POSSUM and IPRA data were recorded prospectively. Incidence and severity of postoperative morbidity was recorded according to established definitions. The cohort was grouped by POSSUM and IPRA risk groups. The estimated and observed outcomes and morbidity profiles of POSSUM and IPRA were scrutinized.

Results: POSSUM-estimated risk (62 %) corresponded with observed total morbidity (65 %). Severe morbidity was 17 % and in-hospital-mortality 3.1 %. Individual and grouped POSSUM risk estimates did not reveal associations with incidence (p = 0.637) or severity (p = 0.321) of total morbidity or POPF. The IPRA model identified patients with high POPF risk (p < 0.001), but was even associated with incidence (p < 0.001) and severity (p < 0.001) of total morbidity.

Conclusion: The risk factors defined by a PD-specific model were significantly stronger predictive indicators for the incidence and severity of postoperative morbidity than the factors incorporated in POSSUM. If available, reliable procedure-specific risk factors should be utilized in the risk adjustment of surgical outcomes. For pancreatic surgery, generally applicable tools such as POSSUM still have to prove their relevance.