P36 Validating the post-colonoscopy colorectal cancer (PCCRC) rate as a metric of endoscopy quality

Abstract

Introduction Endoscopists PCCRC rates have been shown to inversely correlate with their adenoma detection rates. The World Endoscopy Organisation (WEO) has recently published methodology for comparing unadjusted PCCRC rates between different organisations. These whole system rates may not reflect endoscopists’ performance. This study aimed to produce a validated PCCRC rate for individual endoscopists. Methods All cases of Colorectal cancer (CRC) diagnosed between 2010 and 2018 at our Trust were ascertained from Somerset Cancer Database using ICD10 codes C18–20. From the Endoscopy reporting system all colonoscopies performed in the same years were identified. By SQL queries within a MS Access database the following were determined for the 6 years 2010- December 2015. • unadjusted PCCRC cases i) True +ve CRC diagnoses by colonoscopy ii) Cases with 2 colonoscopies within 6/12 of diagnosis. Cases reviews took place for group i) and iii). The following variations in the WEO method were used to produce an Endoscopy-related PCCRC rate. Exclusions Genetic syndromes, IBD and follow-up recurrent EMR cases. Delays in management not due to failure of endoscopic assessment Errors in recorded timings for date of CRC diagnosis. Inaccurately coded cases Results From 2010–2015, 21267 colonoscopies were performed. From 2010–2018 1916 CRC cases diagnosed, 1246 (65%) were diagnosed by colonoscopy. 39 unadjusted PCCRC cases were identified. After case review 18 cases were excluded by the criteria above, a (5), b (1), c (9), d (3). However, in reviewing cases with 2 colonoscopies within 6/12 of diagnosis 3 additional misses were found making a total of 24 endoscopy related PCCRC. The unadjusted PCCRC was 5.1% vs Endoscopy-related 3.1% The funnel plot of Endoscopy related PCCRC rates (figure 1) indicated 3 outlying endoscopists Conclusions The limitation of this study was only locally diagnosed CRC were ascertained. Around 5% of our colonoscopies were on patients from a neighbouring trust and additionally patient migration is possible. However, the Audit released by Bowel Cancer Intelligence for 2011–2013 recorded our unadjusted PCCRC at a similar 3.8%. The case reviews identified 12/39 cases that were clearly not PCCRC (category c & d). This is a substantial barrier to the non-case reviewed PCCRC methodology if used to critique endoscopist’s practice. 46% of unadjusted PCCRC cases were not related to endoscopic recognition. 3 Endoscopists (2 locums) were identified with statistically higher PCCRC rates. Approximately 1 in 1000 colonoscopies may miss a lesion leading to a PCCRC. For individual endoscopist with >200 colonoscopies in this period, PCCRC cases per colonoscopy varied from 3/419 to 0/3176.