Search in the Abstract Database

Abstracts Search 2017

* = Presenting author

P450 Reliability assessment of endoscopic scoring tools using central video review of colonoscopies in paediatric patients with ulcerative colitis: data from the Canadian Children IBD Network

Carman N.*1, Huynh H.2, Walsh C.M.1, Ricciuto A.1, Mouzaki M.1, Crowley E.1, Church P.C.1, Walters T.D.1

1The Hospital for Sick Children, University of Toronto, Gastroenterology Department, Toronto, Canada 2University of Alberta, Edmonton, Canada

Background

Reliable and consistent endoscopic assessment of mucosal disease severity is important in the evaluation of patients with Ulcerative Colitis (UC). Of the commonly used assessment tools, the Ulcerative Colitis Endoscopic Index of Severity (UCEIS) is likely more responsive to change than the Mayo Endoscopic Score (Mayo-ES). Neither have been formally evaluated in paediatric patients. Using videos of colonoscopies performed in patients from the Canadian Children IBD Network, we undertook to assess inter-rater reliability (IRR) for the UCEIS and Mayo-ES amongst Pediatric IBD physicians familiar with the tools, as well as non-IBD pediatric gastroenterologists.

Methods

Video recordings of ileo-colonoscopies of paediatric patients with UC undergoing endoscopic assessment at Network sites were utilised for the analysis. 8 physicians (4 IBD experts) reviewed the videos, blinded to clinical information, collecting data encompassing the UCEIS and Mayo-ES. A global assessment of endoscopic lesion severity (GELS) was also recorded using a visual analogue scale. IRR was measured using Intraclass correlation coefficients (ICCs). Correlation between scoring tools was measured using Spearman's test of correlation (r).

Results

There was a broad range of endoscopic severity within the endoscopic assessments (median UCEIS 6 (range: 3 to 8). The IRR for both Mayo-ES and UCEIS was excellent amongst IBD physicians. However, whilst the IRR for Mayo-ES was very good amongst non-IBD gastroenterologists, for UCEIS it was only moderate (see table). Amongst IBD physicians, there was good correlation between the UCEIS score and Mayo-ES (r=0.75, p<0.001), as well as between each score and the GELS (Mayo ES: r=0.78, p<0.001; UCEIS: r=0.72, p<0.001). Within the 3 items of the UCEIS, the most common sources of disagreement between readers were estimation of the degree of bleeding by all physicians, and evaluation of erosions/ulcers by non-IBD gastroenterologists (see table).

Table 1. Inter-rater reliability using ICC of UCEIS and Mayo Endoscopic Score for IBD physicians and non-IBD gastroenterologists

IBD gastroenterologistsNon-IBD gastroenterologists
Median UCEIS score (min–max)6 (3–8)7 (3–8)
ICC – Total UCEIS score (95% CI)0.87 (0.74–0.94)0.55 (−0.25–0.84)
p<0.001p=0.01
ICC – Vascular pattern (95% CI)0.80 (0.60–0.91)0.81 (0.52–0.93)
p<0.001p=<0.001
ICC – Bleeding (95% CI)0.50 (0.11–0.77)0.51 (−0.89–0.83)
p=0.01p=0.03
ICC – Erosions and ulcers (95% CI)0.88 (0.76–0.95)0.15 (−0.35–0.55)
p<0.001p=0.20
ICC – Mayo endoscopic score (95% CI)0.88 (0.77–0.94)0.72 (0.27–0.89)
p<0.001p=0.01

Conclusion

Centralised video review of colonoscopy is feasible for assessing endoscopic severity in paediatric UC. Assessment of the scoring tools (UCEIS and Mayo-ES) using video recordings showed excellent IRR in the hands of IBD physicians familiar with the tools. For non-IBD gastroenterologists, IRR of the Mayo-ES was good, however reliability with the UCEIS score was poorer, perhaps reflective of unfamiliarity with the tool. Repeat assessments following training in the application of the tool are planned.