Development and validation of a scoring system to predict esophagogastroduodenoscopy necessity

Abstract

AimThis study aimed to develop and validate a scoring system for predicting the necessity of performing esophagogastroduodenoscopy (ESD) to enhance accuracy and reduce misapplications.MethodsBetween February 2021 and April 2022, outpatients attending the gastroenterology department, scheduled for EGD, were recruited. Patients completed the system evaluation by providing clinical symptoms information, relevant medical history, and their final endoscopic findings, considered as the diagnostic gold standard. Patients were randomly divided (2:1) into the training and validation cohort. The optimal algorithm, chosen from five alternatives including a parallel test, was used. Six physicians participated in human‐computer‐controlled validation. Sensitivity and negative likelihood ratios (‐LR) were used as the primary indicators.ResultsA total of 865 patients were enrolled, with 578 in the training and 287 in the validation cohorts. The scoring system comprised 21 variables, including age, thirteen typical clinical symptoms, and seven relevant medical histories. The parallel test was selected as the final algorithm. Positive EGD findings were 54.5% (315/578) in the training cohort and 62.7% (180/287) in the validation cohort. The scoring system demonstrated a sensitivity of 79% in the training cohort and 83.9% in the validation cohort, with corresponding ‐LR values of 0.627 and 0.615. Compared to the physicians, the scoring system exhibited higher sensitivity (84% vs 68.7%, P=0.02) and a lower ‐LR (1.10 vs 2.41, P=0.439).ConclusionsWe developed and validated a scoring system to predict the necessity of EGD using a parallel test algorithm. The results from a single‐center validation demonstrated that this system is a user‐friendly and effective tool for pre‐endoscopic patient evaluation.This article is protected by copyright. All rights reserved.