Published online Sep 14, 2017. doi: 10.3748/wjg.v23.i34.6281
Peer-review started: April 17, 2017
First decision: June 7, 2017
Revised: June 9, 2017
Accepted: July 12, 2017
Article in press: July 12, 2017
Published online: September 14, 2017
Processing time: 154 Days and 12.9 Hours
To determine the inter-observer variability for colon polyp morphology and to identify whether education can improve agreement among observers.
For purposes of the tests, we recorded colonoscopy video clips that included scenes visualizing the polyps. A total of 15 endoscopists and 15 nurses participated in the study. Participants watched 60 video clips of the polyp morphology scenes and then estimated polyp morphology (pre-test). After education for 20 min, participants performed a second test in which the order of 60 video clips was changed (post-test). To determine if the effectiveness of education was sustained, four months later, a third, follow-up test was performed with the same participants.
The overall Fleiss’ kappa value of the inter-observer agreement was 0.510 in the pre-test, 0.618 in the post-test, and 0.580 in the follow-up test. The overall diagnostic accuracy of the estimation for polyp morphology in the pre-, post-, and follow-up tests was 0.662, 0.797, and 0.761, respectively. After education, the inter-observer agreement and diagnostic accuracy of all participants improved. However, after four months, the inter-observer agreement and diagnostic accuracy of expert groups were markedly decreased, and those of beginner and nurse groups remained similar to pre-test levels.
The education program used in this study can improve inter-observer agreement and diagnostic accuracy in assessing the morphology of colon polyps; it is especially effective when first learning endoscopy.
Core tip: Paris classification is used worldwide in clinical practice to categorize the morphology of gastrointestinal polyps. However, few studies regarding the inter-observer variability associated with this classification have been reported. In this study, we identified that an education program could be helpful in improving inter-observer agreement and diagnostic accuracy when determining the morphology of colon polyps, and that the educational program was especially effective when first learning endoscopy. We suggest that daily practice of morphology assessment of colon polyps and the proper education of the Paris classification are essential in maintaining the quality of colonoscopic examination.