Pancreatic resections are not only safest but also most cost-effective when performed in a high-volume centre: A...

It is not known whether the treatment costs of pancreatic surgery can be reduced by centralisation. The aim of this study was to analyse the impact of hospital volume on the short-term prognosis and costs in a nationwide study.

The National registry was searched for patients undergoing pancreatoduodenectomy (PD) in Finland between 2012 and 2014. Patient data was recorded up to ninety days postoperatively and Charlson comorbidity index (CCI) calculated. Complications were classified according to Clavien-Dindo. A CCI was calculated for each patient. The hospitals were categorized by yearly resection rate: high (≥20, HVC), medium (6-19, MVC) and low (≤5, LVC). Costs were calculated according to the 2012 billing list.

The study population comprised 466 patients. Demographics were similar in the HVC, MVC and LVC groups. Mortality was lower in the HVCs than in MVCs and LVCs at 30 days (0.8% vs. 8.8-12.9%; p < 0.01) and at 90 days (1.9% vs. 10.5-16.1%; p < 0.01). Hospital volume and CCI were significant factors for mortality in multivariate analysis. Median costs among all patients were lower in the HVC group than in the MVC/LVC groups (p = 0.019), among Clavien-Dindo class III (0.020), among patients over 75 years (p < 0.001) and among patients who survived over five days (p = 0.015).

Thirty- and 90-day mortality is 10 times lower when the patient is operated on in an HVC. The study shows that the median overall costs of surgical treatment are 82-88% of the median costs in lower volume centres.