P213. Influence of clinical history on scoring of the Ulcerative Colitis Endoscopic Index of Severity (UCEIS)
S.P. Travis1, D. Schnell2, P. Krzeski3, J.-F. Colombel4, B.G. Feagan5, W. Reinisch6, B.E. Sands7, B.R. Yacyshyn8, C.A. Bernhardt9, J.-Y. Mary10, W.J. Sandborn11, 1Translational Gastroenterology Unit, John Radcliffe Hospital, Oxford, United Kingdom, 2Middletown, United States, 3Warner Chilcott, Weybridge, United Kingdom, 4Hôpital Claude Huriez, Centre Hospitalier Universitaire de Lille, Lille, France, 5University of Western Ontario, Robarts Clinical Trials, Robarts Research Institute, Ontario, Canada, 6Universitatsklinik Innere Medizin III, Abteilung Gastroenterologie und Hepatologie, Medical University of Vienna, Vienna, Austria, 7Mount Sinai Hospital, New York, United States, 8University of Cincinnati, Division of Digestive Diseases, Cincinnati, Ohio, United States, 9Bernhardt Regulatory Consulting, Cincinnati, Ohio, United States, 10INSERM U717 Biostatistics and Clinical Epidemiology, Universite Paris Diderot, Paris, France, 11Division of Gastroenterology, University of California San Diego, La Jolla, California, United States
The UCEIS is a 3-item (vascular pattern, bleeding and ulceration) validated tool for assessing endoscopic severity of UC. The effect of symptom knowledge and patients' history on disease scoring with the UCEIS was assessed.
20 investigators, successfully trained on the UCEIS by scoring 4 or, if necessary, 7 reference videos, each performed 28 evaluations, including 2 repeats of non-control videos, from a set of 44 randomly assigned videos (stratified by Mayo Clinic scores 0–11), including 4 common controls. Investigators were randomised to have knowledge of symptoms and history (unblinded group; n = 10) or not (blinded group; n = 10). 2 control videos with Mayo score = 1 were assigned symptom information somewhat more severe than reported and 2 with Mayo score = 11 were attributed milder symptoms. UCEIS scores (range 0 [normal] to 8 [most severe]) and overall severity, scored on a 100-point visual analogue scale (VAS), were compared between groups using nonparametric methods and permutation-based p-values. The coefficient of variation (CV%; StdDev/Mean*100%) was also calculated for both groups for each non-control video (n = 44). Inter-investigator variability was summarised using kappa (κ) statistics.
Median UCEIS scores ranged from 0–8, whilst mean VAS scores ranged from 1 (normal subject) to 96 (most severe patient). Correlation between UCEIS and VAS severity evaluations (0.91 vs. 0.93) and median CV% (25.3% vs. 25.6%) were similar between blinded and unblinded groups. There were no significant differences in UCEIS scoring between groups for the 4 control videos and for VAS scoring in 3/4. There was a significantly lower VAS score (p = 0.03) for the unblinded (median = 80.5) than the blinded group (median = 88) for the other control video (severe disease with milder symptoms). Of the 44 videos, there was 1 for which the UCEIS score was significantly different (p = 0.048) and 2 others for which VAS scores were significantly different (p = 0.01, 0.02) between groups. The groups were broadly consistent for individual descriptors in the UCEIS (κ control videos, unblinded: 0.69;0.52;0.62 vs. blinded: 0.65;0.58;0.62).
UCEIS scoring was largely unaffected by knowledge of patient information, although there were some differences in the individual components. A high level of inter-observer agreement for the UCEIS in scoring an overall assessment of endoscopic activity in UC was confirmed in this independent cohort.