Objective To explore the risk factors associated with rebleeding in elderly patients with gastric ulcer, construct a predictive model and validate it.

Methods A retrospective analysis was conducted on 200 elderly patients with gastric ulcer, and all patient data were collected and analyzed. LASSO regression analysis was used to screen the risk factors for rebleeding in elderly patients with gastric ulcer, and a predictive model for the influencing factors of rebleeding in elderly patients with gastric ulcer was established. Based on the results of the predictive model, a ROC curve was plotted, and the area under the curve (AUC) was calculated.

Results Among the 200 patients, 43 were rebleeding patients and 157 were non-rebleeding patients. Univariate analysis revealed 11 significantly different indicators. After dimensionality reduction using LASSO regression, 8 optimal modeling indicators were identified, and logistic regression analysis was performed on them, ultimately resulting in 5 indicators, Logi(P) = –1.026 + 0.451 × age –0.527 × albumin + 0.302 × presence of Helicobacter pylori infection + 0.375 × long-term NSAIDs use + 0.476 × stage A1 of gastric ulcer. The evaluation showed that the model had a C-index of 0.903, an AUC of 0.899, and a 95%CI of 0.868 to 0.949, indicating good predictive ability.

Conclusions This study established a predictive model for the risk factors of rebleeding in elderly patients with gastric ulcer based on LASSO-logistic, which has good predictive value.