Annals of surgery | 2019

Specificity of Procedure Volume and its Association With Postoperative Mortality in Digestive Cancer Surgery: A Nationwide Study of 225,752 Patients.

 
 
 
 
 
 
 
 

Abstract


OBJECTIVES\nWe aimed to examine whether the improved outcome of a digestive cancer procedure in high-volume hospitals is specific or correlates with that of other digestive cancer procedures, and determine if the discriminant cut-off of hospital volume may influence postoperative mortality (POM) regardless of the procedure.\n\n\nBACKGROUND\nPerforming complex surgeries in tertiary centers is associated with improved outcome. However, the association between POM and hospital volume of nonspecific procedures is unknown.\n\n\nMETHODS\nPatients who underwent colectomy, proctectomy, esophagectomy, gastrectomy, pancreatectomy, and hepatectomy for cancer between 2012 and 2017 were identified in the French nationwide database. Chi-square automatic interaction detector was used to identify the cut-off values of the annual caseload affecting the 90-day POM. A common threshold was estimated by minimization of chi-square distance taking into account the specific mortality of each procedure.\n\n\nRESULTS\nOverall, 225,752 patients were identified. Hospitals were categorized according to the procedure volume (colectomy: ≥80\u200acases/yr, proctectomy: ≥35/yr, esophagectomy: ≥41/yr, gastrectomy: ≥16/yr, pancreatectomy: ≥26/yr, and hepatectomy: ≥76/yr). The overall 90-day POM was 5.1% and varied significantly with volume. The benefits of high volume were transferable across procedures. High-volume hospitals for colorectal cancer surgery significantly influenced the risk of death after hepatectomy (P < 0.001) and pancreatectomy (P < 0.001). The common threshold for all procedures that influenced POM was 199\u200acases/yr (odds ratio 1.29, P < 0.001).\n\n\nCONCLUSION\nIn digestive cancer surgery, the volume-POM relationship of one procedure was associated with the volume of other procedures. Thus, tertiary hospitals should be defined according to the common threshold of different procedures.

Volume None
Pages None
DOI 10.1097/SLA.0000000000003532
Language English
Journal Annals of surgery

Full Text