急性非静脉曲张性上消化道出血患者急诊内镜干预需求的预测模型构建OA

Objective To identify influencing factors for emergency endoscopic intervention in patients with acute non-variceal upper gastrointestinal bleeding(ANVUGIB)and to develop a predictive model.Methods Clinical data from patients with ANVUGIB admitted to the General Hospital of Central Theater Command from January 2022 to April 2025 were retrospectively collected,they were divided into intervention group and non intervention group according to whether endoscopic intervention was performed within 24 hours after admission.Patients admitted from January 2022 to August 2024 were included in the training set,and those admitted from September 2024 to April 2025 were included in the validation set.In the training set,propensity score matching(PSM)was used to balance confounders,and clinical outcomes were compared between the intervention and non-intervention groups.Lasso regression was used to screen predictor variables,and a nomogram was developed using multivariable Logistic regression.The predictive ability and accuracy of the model were assessed using the receiver operating characteristic(ROC)curve,area under the curve(AUC),calibration curve,and Hosmer-Lemeshow test.Decision curve analysis(DCA)and clinical impact curve(CIC)were used to evaluate the clinical utility of the model,which was further validated in the validation set.Results A total of 732 patients were included,with 584 in the training set and 148 in the validation set.After PSM,there were 68 patients in the intervention group and 136 patients in the non-intervention group in the training set.Compared to the non-intervention group,the intervention group showed better clinical outcomes in terms of rebleeding rate and length of hospital stay.Lasso and Logistic regression analyses showed that hematemesis,pulse rate,systolic blood pressure,history of peptic ulcer,and platelet count were factors associated with endoscopic intervention.The AUC of the predictive model was 0.783.The calibration curve was close to the ideal diagonal line,and the Hosmer-Lemeshow test showed good model fit(P=0.153).The AUC for internal cross-validation was 0.775,and the AUC in the validation set was 0.778.Both DCA and CIC indicated that the model had potential clinical application value.Conclusion The Endoscopic Intervention Score is composed of simple and objective predictors that can be readily obtained early after admission and has good predictive performance.