How good are surgeons at identifying appendicitis? Results from a multi-centre cohort study

Background: Convincing arguments for either removing or leaving in-situ a macroscopically normal appendix have been made, but rely on surgeons' accurate intra-operative assessment of the appendix. This study aimed to determine the inter-rater reliability between surgeons and pathologists from a large, multicentre cohort of patients undergoing appendicectomy. Materials and methods: The Multicentre Appendicectomy Audit recruited consecutive patients undergoing emergency appendicectomy during April and May 2012 from 95 centres. The primary endpoint was agreement between surgeon and pathologist and secondary endpoints were predictors of this disagreement. Results: The final study included 3138 patients with a documented pathological specimen. When surgeons assessed an appendix as normal (n=496), histopathological assessment revealed pathology in a substantial proportion (n=138, 27.8%). Where surgeons assessed the appendix as being inflamed (n=2642), subsequent pathological assessment revealed a normal appendix in 254 (9.6%). There was overall disagreement in 392 cases (12.5%), leading to only moderate reliability (Kappa 0.571). The grade of surgeon had no significant impact on disagreement following clinically normal appendicectomy. Females were at the highest risk of false positives and false negatives and pre-operative computed tomography was associated with increased false positives. Conclusions: This multi-centre study suggests that surgeons' judgements of the intra-operative macroscopic appearance of the appendix is inaccurate and does not improve with seniority and therefore supports removal at the time of surgery.