Development and validation of a risk prediction model for new-onset gastrointestinal bleeding in critically ill...

To explore risk factors for new-onset gastrointestinal bleeding (GIB) in critically ill patients and construct a predictive model.

A retrospective study of 241 intensive care unit patients was conducted. Clinical data, laboratory indicators, treatments, and outcomes were collected. Risk factors were analyzed via univariate and multivariate logistic regression. A nomogram was established, and its performance was assessed using receiver operating characteristic curves, concordance index (C-index), calibration plots, Hosmer-Lemeshow test, bootstrap resampling, and decision curve analysis (DCA). Internal and external validation were performed.

Age ≥ 65, shock, sepsis, renal dysfunction, hepatic failure, mechanical ventilation > 48 h, and hemoglobin < 8 g/dL were independent risk factors for new-onset GIB, while albumin < 30 g/L emerged as a predictive factor. The nomogram demonstrated strong discrimination (C-index 0.825 in the training cohort, 0.804 in the validation cohort), good calibration, and favorable clinical utility. DCA confirmed its benefit in guiding clinical decisions.

Multiple comorbidities and treatment-related factors contribute to new-onset GIB in critically ill patients. The developed nomogram provides an effective tool for individualized risk assessment, supporting early intervention and improved outcomes.