Development and evaluation of acceptance scale for artificial intelligence in digestive endoscopy by subjects.

Abstract

Digestive endoscopy is an important diagnostic and therapeutic tool for digestive system diseases. The artificial intelligence (AI)-assisted system in endoscopy (hereinafter referred to as AI in digestive endoscopy) has broad application prospects in the field of digestive endoscopy. The trust and acceptance of endoscopic subjects are the cornerstone of the research, application, and promotion of AI in digestive endoscopy. Currently, the tools for measuring the acceptance of AI in digestive endoscopy by subjects are limited at home and abroad. This study aims to develop a scale for measuring the acceptance of AI in digestive endoscopy by subjects, then to evaluate its reliability and validity. By conducting literature research, an item pool and dimensions were constructed, and a preliminary scale was constructed using Delphi method. Through the first stage of the survey on the subjects, the reliability and validity of the scale were tested, and the revised scale was used for the second stage of survey on the subjects to further verify the structural validity of the scale. The acceptance scale for AI in digestive endoscopy included 11 items in 3 dimensions: accuracy, ethics, benefit and willingness. In the first stage of the survey, 351 valid questionnaires were collected, and the Cronbach's α was 0.864. The correlation coefficient between the total score of the scale and the score of the test item was 0.636, and the Kaiser-Meyer-Olkin (KMO) value in exploratory factor analysis was 0.788. In the second stage of the survey, 335 valid questionnaires were collected, and in confirmatory factor analysis, the χ2/df was 3.774, while the root mean squared error of approximation (RMSEA) was 0.091. Acceptance scale for AI in digestive endoscopy by subjects developed in this study has good reliability and validity.