A self-assessment tool for predicting discomfort and tolerance in Chinese patients undergoing esophagogastroduodenoscopy

Background For patients taking esophagogastroduodenoscopy (EGD), sedation should ideally be used individually based on patients’ comfort and tolerance level. However, currently there is no valid predictive tool. We undertook this study to develop and temporally validate a self-assessment tool for predicting discomfort and tolerance in Chinese patients undergoing EGD. Methods We recruited 1522 patients undergoing routine diagnostic EGD without sedation. We collected candidate predictor variables before endoscopy and evaluated discomfort and tolerance with a 5-point visual analogue scale after the procedure. We developed logistic regression predictive models based on the first 2/3 of participants, and evaluated the calibration and discrimination of the models in the later 1/3 of patients. Results 30.2% and 23.0% participants reported severe discomfort or poor tolerance to EGD respectively. The predictive factors in the model for discomfort included sex, education, expected level of discomfort, and anxiety before endoscopy. The model for tolerance included income, expected level of discomfort, and anxiety before endoscopy. In the validation population, the established models showed a moderate discriminative ability with a c-index of 0.74 for discomfort and 0.78 for tolerance. Hosmer–Lemeshow test suggested the models had fine calibration ability (discomfort: P = 0.37, tolerance: P = 0.41). Conclusions Equations for predicting discomfort and tolerance in Chinese patients undergoing EGD demonstrated moderate discrimination and variable calibration. Further studies are still required to validate these tools in other population. Trial registration Chinese Clinical Trial Registry (ChiCTR1800020236).