Table 1_Screening of risk factors and construction of a prediction model for enteral nutrition intolerance in elderly patients with comorbidities and sarcopenia.docx

A visual nomogram prediction model was developed incorporating these independent predictors. Four statistical methods were applied to fully assess model performance: ROC curves for discrimination evaluation, calibration curves for risk fitting verification, DCA for clinical utility assessment, and SHAP analysis to interpret the predictive weight of each variable.</p>Results<p>Severe dependence in activities of daily living (ADL) and concomitant psychiatric disorders were independently linked to higher ENI risk, whereas red blood cell (RBC) levels produced a protective effect.

The predictive model delivered solid discriminative performance, registering an AUC of 0.77 (95% CI: 0.69–0.85) in the training cohort and 0.70 (95% CI: 0.57–0.84) in the validation cohort. The Hosmer–Lemeshow goodness-of-fit test confirmed adequate model calibration (p > 0.05), and net clinical benefit analysis validated steady clinical value across clinically relevant risk threshold ranges. SHAP analysis indicated that concomitant psychiatric disorders served as the primary variable driving the model’s predictive results.</p>Conclusion<p>The developed nomogram exhibited acceptable discrimination and acceptable calibration upon internal validation, with detectable net clinical benefit.

This visual risk scoring tool enables personalized ENI risk estimation in elderly multimorbid patients with sarcopenia. Additional multicenter external validation work should be finished prior to clinical adoption.</p>