Article Text

PWE-012 What is the best key performance indicator (kpi) for polyp detection at colonoscopy?
  1. A Fraser,
  2. T Rose,
  3. MercyAscot endoscopists
  1. Medicine, University of Auckland, Auckland, New Zealand


Introduction Adenoma (TA) detection has been a ”gold standard” KPI. Guidelines suggest that finding >20% in a screening population aged >50 years is acceptable. This target of TA detection is now outdated as post-colonoscopy colorectal cancer (PCCRC) rates are unacceptable at this level of TA detection. Audit of individual endoscopists is often performed using polypectomy rates as a surrogate of TA detection and often in a mixed (non-screening) population. There is uncertainty about the need for collecting data on the number of TA removed and whether this data should be analysed per total number of procedures or using only procedures were polyps were found as the denominator. This latter option tests the idea that some colonoscopists have a “one and done” policy. There is no data on the utility of recording the number of sessile serrated polyps (SSP).

Method Data was obtained from a continuous audit process using histology over 5 years. The number of polyps per procedure was collected for TA and SSP for 3 years. Procedures were divided into “first procedure” for evaluation of new symptoms and “planned follow-up of polyps”. A typical practice will have endoscopists with different case-mix – in particular age, number of follow-up procedures compared to new patients and gender. One simple approach is to filter the adenoma results according to these parameters.

Results 12555 procedures were analysed. KPI data for the 15 endoscopists was ranked by polyp detection rate (highest rank at 89% and lowest rank at 34%). The ranking was then re-adjusted according the other potential KPIs. The ranking order showed modest differences for% of procedures with TA and the mean number of TA per total procedures but was very different for the mean number of TA per procedure with polyps found. The rankings were completely different for% of procedures with SSP and mean number of SSP per total procedures. The same endoscopists were ranked for% of TA detected and compared with the subgroup of TA “if first procedure” and age >50. There were significant differences suggesting casemix affects audit data for TA (especially using a mixed population rather than a screening population). There was good correlation between% of procedures with TA and the mean number of TA per procedure (r=0.72, p=0.001) but there was no correlation of% of procedures with TA and mean number of TA in procedures were there was a polypectomy. There was an good correlation between% of procedures with SSP detected and mean number of SSP per procedure (r=0.56, p=0.015).

Conclusion The only true test of any KPI will be showing a correlation with PCCRC. This has been shown for% adenomas but it is unknown if other KPIs will be more (or less) predictive. The rankings are different for each KPI and it is this comparison with the performance of peers that is often viewed with most interest.

Disclosure of Interest None Declared

  • colonoscopy quality
  • polypectomy

Statistics from

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.