Search CDDW 2008 Abstracts

HOME

Return to Table of Contents

341

RELIABILITY STUDY TO ASSESS A MEASURE OF SMALL BOWEL CLEANLINESS IN VIDEO CAPSULE ENDOSCOPY

R Chan, L Hookey, J Louw, W Paterson
Department of Gastroenterology, Queen’s University, Kingston, Ontario

The optimal bowel cleansing regimen for VCE has yet to be determined. Prior studies have been heterogeneous in many areas including the instrument measure used to assess bowel cleanliness. Prior to conducting a further larger RCT, there is a need to validate the tool chosen to assess the quality of small bowel preparation. The objective of this study was to determine the interobserver reliability of two measures of bowel cleanliness, one being a 5 point Likert scale (scored 0-4) and the other a proportion of time during which the small bowel is obscured.
Twenty VCE studies were randomly retrieved from our database (Dec/05-Apr/07). Baseline characteristics were diverse reflecting different indications for the procedure and bowel preparation regimen. Reviewers were blinded to these characteristics including method of bowel preparation. Three gastroenterologists with local VCE experience independently rated the same 20 studies with the two measures described. Additional measures included a score for the presence and degree of bubbles (bubble scale) and the recorded time to initial appearance of turbid fluid.
Five of the twenty VCE studies have been reviewed thus far:
CaseColyteScoreAScoreBScoreC%A%B%C
14L3332.295.454.01
23L21233.8633.8017.15
33L01127.6943.7331.69
42L01253.9368.7848.97
54L3124.8644.4569.69

We feel that the number of reviewed studies is too small to conduct a meaningful statistical analysis currently. Preliminary findings and reviewer opinion suggest some discrepancy that may be stemming from ambiguity inherent within the proposed scales. Thus we have identified an area of improvement that should lead to further refinement of these measuring instruments. We have planned a formal meeting with the reviewers to discuss these aspects further. Once these issues are addressed, we will proceed as originally intended to evaluate the complete composite of retrieved VCE studies, and conduct a formal statistical analysis.

PREVIOUS     NEXT