Procedural performance in gastrointestinal endoscopy: live and simulated.

Area of Science:

• Gastroenterology
• Medical Education
• Surgical Skills Assessment

Background:

• Assessing technical skills in gastrointestinal endoscopy is crucial for patient safety and effective training.
• Current assessment methods include both human-based evaluations and emerging computer-based technologies.

Purpose of the Study:

• To compare the efficacy and validity of a human-based technical skills assessment tool against a computer-based tool for gastrointestinal endoscopies.
• To evaluate these tools in both live and simulated settings for consultants and trainees.

Main Methods:

• Utilized validated human-based Likert scales for generic and specific technical skills in esophagogastroduodenoscopy and sigmoidoscopy.
• Employed computer-based tools to assess mucosal aspects and procedural times.
• Two independent observers blindly assessed each procedure using the human-based tool, followed by comparison with computer-based metrics.

Main Results:

• A total of 210 procedures were assessed by 18 consultants and 37 trainees.
• The human-based tool exhibited good inter-rater reliability (Cronbach alpha: 0.62-0.75) and construct validity (ANOVA: p = 0.000-0.002).
• The computer-based tool showed poor construct validity (ANOVA: p = 0.263-0.701).

Conclusions:

• The human-based technical skills assessment tool possesses face, content, concurrent, and construct validities.
• This tool is a viable option for use in endoscopy training, self-appraisal, and professional revalidation.
• Human-based assessment offers a more reliable method for evaluating technical proficiency in gastrointestinal endoscopy compared to current computer-based alternatives.