Procedural performance in gastrointestinal endoscopy: live and simulated.

Abstract

The purpose of the present study was to compare the use of a human and a computer-based technical skills assessment tool in live and simulated gastrointestinal endoscopies performed by consultants and trainees. Validated human-based Likert scales were used individually for generic and specific technical skills for two procedures. Two observers assessed each procedure independently and blindly. Computer-based assessment tools were also used, assessing aspects of the mucosa as well as various procedural times. The two tools were then compared. A total of 210 live and simulated endoscopies (107 esophagogastroduodenoscopy [OGD], 103 sigmoidoscopy) were performed by 18 consultants and 37 trainees. Mean inter-rater reliability using Cronbach alpha was good for the human-based tool (range: p = 0.62-0.75). Construct validity was good for the human-based tool (analysis of variance [ANOVA] range: p = 0.000-0.002) but not for the computer-based tool (ANOVA range: p = 0.263-0.701). This human-based technical skills assessment tool of gastrointestinal endoscopies seems to have face, content, concurrent, and construct validities compared to computer-based assessment parameters. The tool has the possibility of being used in training, self-appraisal, and revalidation.