Time to rethink surgeon-specific outcome data for colorectal surgeons in England: Cross-sectional data of 73,842 resections for colorectal cancer

Abstract

Since 2013, individual surgeon and NHS Trusts outcomes following elective colorectal cancer surgery have been in the public domain in England. The 90-day operative mortality rates following colorectal resections are available for the public to view online.

Aim: The aim of this study is to evaluate the quality of the published data. It also aims to find whether this data will show the expected pattern of inverse correlation between case volume and the postoperative 90-day mortality rate.

Methodology: The postoperative 90-day mortality data was taken from the Association of Coloproctology of Great Britain and Ireland (ACPGBI) website. Surgeons and Trusts were categorized according to case volume. Completeness of data at Trust and surgeon levels was analysed. All statistical analyses were performed in Statistical Package for Social Science (SPSS, version 23, IBM, Armonk, NY).

Results: 788 colorectal surgeons performed 73,842 resections for colorectal cancer in 143 hospitals over a 5-year period (1st April 2010 and 31st March 2015). The mean national 90-day mortality after colorectal resections was 2.6%. No significant effect was identified when mortality rates were correlated with the surgeon or Trust volume. There was a missing data of 3874 patients in the individual surgeon level analysis when compared to the number of procedures included in Trust analysis (73,842 vs 69,968 cases). About one-third of hospitals (n = 43) had a case ascertainment of less than 90%. Out of the 788 surgeons, there were only two outliers whose mortality rates were outside the "funnel limit".

Conculsion: The expected relationship between case volume and mortality rates could not be established. The completeness of data and low numbers of procedures per surgeon are major concerns. Additional outcome metrics should be utilized to assess performance quality. Failure to Rescue approach should be explored and utilized. It is crucial to have more rigorous and streamlined methods for data collection and case ascertainment to present to the public reliable, complete and relevant data.