Impact of endoscopist annual procedure numbers and withdrawal time on colonoscopy quality: analysis of data from the UK national endoscopy database (NED)

Abstract

Introduction This nationwide study aimed to assess UK colonoscopy quality using key performance indicators (KPIs), including comparison with previous national audit (2011).[1]

The impact of endoscopist annual procedure numbers and median withdrawal time for completed negative procedures upon selected KPIs was also assessed.

Aims & Methods Data on all colonoscopies performed between 1st March 2019 and 29th February 2020 was extracted from the National Endoscopy Database (NED). Three KPIs - caecal intubation rate, patient discomfort, polyp detection rate - were calculated (with 95% confidence intervals): Associations between endoscopists’ annual procedure numbers and median withdrawal times and each KPI was assessed using two-level mixed binomial logistic regression models with endoscopist as random effect with other variables (patient age, patient sex, procedure indication) included as fixed effects to correct for case mix. Results were presented as odds ratios (OR) with 95% confidence intervals.

Results 592,764 colonoscopies were analysed, uploaded from 400 sites by 3,872 endoscopists. 62.2% were carried out to investigate patient symptoms, 9.9% for bowel cancer screening, 21.4% for other screening/surveillance (polyps, post-CRC, IBD, etc), and 6.6% performed for planned therapy or due to abnormal imaging.

29.6% were performed with no sedation, 69.1% using conscious sedation, and 1.4% under deep sedation or general anaesthetic.

Unadjusted caecal intubation rate was 93.5% (95% CI 93.4-93.6), polyp detection rate was 37.3% (37.2-37.4), and 4.8% (4.7-4.8) of colonoscopies were recorded with moderate or severe patient discomfort; all have improved since the previous national audit in 2011 (p<0.05 for all). Median withdrawal time for a completed negative colonoscopy was 8 minutes (IQR 7-10) and rectal retroflexion was performed in 90.1% of colonoscopies.

Compared to endoscopists performing under 100 colonoscopies annually, endoscopists with higher annual procedure numbers had a lower likelihood of incomplete procedures and procedures with moderate/severe discomfort, and higher likelihood of polyp detection (Table 1).

Endoscopists with longer withdrawal times for negative procedures were more likely to identify polyps (6-8mins vs 1-5 mins, OR=1.41 (1.25-1.57); 9-11mins, OR=1.72 (1.69-2.18); >11mins, OR=1.92 (1.69-2.18)).