Objective Our aim was to determine aetiology of post-colonoscopy colorectal cancers (PCCRCs) identified from population-based data through local root cause analysis at a high-volume mixed secondary and tertiary referral centre.

Design/method A subset of national cancer registration data, collected by the National Cancer Registration and Analysis Service, was used to determine PCCRCs diagnosed between 2005 and 2013 at our centre.

Root cause analysis was performed for each identified PCCRC, using World Endoscopy Organisation recommendations, to validate it and assess most plausible explanation. We also assessed whether patient, clinician and/or service factors were primarily responsible.

Results Of 107 ‘PCCRC’ cases provided from the national dataset, 20 were excluded (16 missing data, 4 duplicates). 87 ‘PCCRC’ cases were included of which 58 were true PCCRCs and 29 false PCCRCs.

False PCCRCs comprised 17 detected cancers (cancer diagnosed within 6 months of negative colonoscopy) and 12 cases did not meet PCCRC criteria. Inflammatory bowel disease was the most common risk factor (18/58) and the most common site was rectum (19/58). The most common explanation was ‘possible missed lesion, prior examination negative but inadequate’ (23/58) and clinician factors were primarily responsible for PCCRC occurrence in most cases (37/58).

Conclusion Our single-centre study shows, after local analysis, there was misclassification of PCCRCs identified from a population-based registry. The degree of such error will vary between registries. Most PCCRCs occurred in cases of sub-optimal examination as indicated by poor photodocumentation. Effective mechanisms to feedback root cause analyses are critical for quality improvement.