Interobserver variability in colonoscopy quality assessment: a retrospective standardized multicenter video-based...

Area of Science:

• Gastroenterology
• Medical Imaging
• Quality Improvement

Background:

• Colonoscopy quality is crucial for patient outcomes, relying on bowel preparation and complete examination.
• Validated tools like the Boston Bowel Preparation Scale (BBPS) are used but can be subjective.
• Assessing interobserver variability in expert colonoscopy evaluation is essential.

Purpose of the Study:

• To evaluate interobserver variability among expert endoscopists using a standardized multicenter colonoscopy video dataset.
• To assess the reliability of the Boston Bowel Preparation Scale (BBPS) in expert evaluations.
• To determine variability in recognizing key anatomical landmarks during colonoscopy.

Main Methods:

• Retrospective analysis of 64 colonoscopy videos from two academic centers.
• Eight experienced gastroenterologists independently scored videos in randomly assigned pairs.
• Interobserver agreement was measured using Cohen's kappa and intraclass correlation coefficient (ICC) for BBPS scores, preparation adequacy, and landmark recognition.

Main Results:

• Significant inter-reviewer variability was found in Boston Bowel Preparation Scale (BBPS) scores across different colon segments and the total score.
• While overall bowel preparation adequacy classification showed no significant difference, discordant judgments occurred in 34% of cases.
• Recognition of anatomical landmarks, particularly the appendiceal orifice and colonic flexures, also exhibited variability.

Conclusions:

• Expert-based colonoscopy quality assessment demonstrates clinically relevant variability, even under standardized review conditions.
• The findings underscore the need for developing more objective and reproducible methods for colonoscopy quality control.
• Improving standardization and objectivity in colonoscopy assessment is vital for enhancing patient care and outcomes.