Risk adjusted benchmarking of abdominoperineal excision for rectal adenocarcinoma in the context of the Belgian PROCARE improvement project.

OBJECTIVE The abdominoperineal excision (APE) rate, a quality of care indicator in rectal cancer surgery, has been criticised if not adjusted for confounding factors. This study evaluates variability in APE rate between centres participating in PROCARE, a Belgian improvement initiative, before and after risk adjustment. It also explores the effect of merging the Hartmann resections (HR) rate with that of APE on benchmarking. DESIGN Data of 3197 patients who underwent elective radical resection for invasive rectal adenocarcinoma up to 15 cm were registered between January 2006 and March 2011 by 59 centres, each with at least 10 patients in the registry. Variability of APE or merged APE/HR rates between centres was analysed before and after adjustment for gender, age, ASA score (3 or more), tumour level (rectal third), depth of tumour invasion (cT4) and preoperative incontinence. RESULTS The overall APE rate was 21.1% (95% CI 19.7 to 22.5%). Significant variation of the APE rate was observed before and after risk adjustment (p<0.0001). For cancers in the lower rectal third, the overall APE rate increased to 45.8% (95% CI 43.1 to 48.5%). Also, variation between centres increased. Risk adjustment influenced the identification of outliers. HR was performed in only 2.6% of patients. However, merging of risk adjusted APE and HR rates identified other centres with outlying definitive colostomy rates than APE rate alone. CONCLUSION Significant variation of the APE rate was observed. Adjustment for confounding factors as well as merging HR with APE rates were found to be important for the assessment of performances.