Development and Validation of a Nomogram for Predicting Oral Frailty Risk in Elderly Patients With Ischaemic Stroke

ABSTRACT Aim To develop and validate a risk prediction model for oral frailty in elderly patients with ischaemic stroke. Design A cross‐sectional study. Methods A temporal cohort of 633 elderly isachemic stroke patients from May 2024 to February 2025 was chronologically divided into a training set ( n = 443) and validation set ( n = 190). Participants were classified into oral frailty and non‐oral frailty groups based on the Oral Frailty Index‐8. In the training set, feature selection combined least absolute shrinkage and selection operator regression and random forest recursive feature elimination, followed by Nomogram Construction via Binary Logistic Regression. The model underwent internal validation using bootstrap resampling, and its generalizability was assessed with the validation set. The model was comprehensively evaluated using Receiver Operating Characteristic ( ROC ) curves, the Hosmer‐Lemeshow Test, Calibration Plots, and Decision Curve Analysis ( DCA ). Results In both the training and validation sets, the prevalence of oral frailty among elderly ischaemic stroke patients was 63.2% and 62.1%, respectively. Wearing dentures, tooth brushing frequency, dry mouth symptoms, chewing difficulty, swallowing function, oral health literacy, and oral health status were identified as significant predictors of oral frailty. ROC analysis demonstrated strong discriminative ability of the nomogram. The Hosmer‐Lemeshow Test confirmed model consistency, and the calibration curve indicated excellent and stable calibration performance. DCA revealed that the model provided significant net clinical benefit in clinical practice. This free, interactive dynamic nomogram is accessible at: https://xiaowen.shinyapps.io/dynnomapp/ . Conclusion This study presents a reliable, accessible model to assess oral frailty risk in elderly ischaemic stroke patients, facilitating clinical identification of high‐risk individuals and providing a scientific foundation for oral health interventions. Relevance to Clinical Practice The nomogram helps healthcare professionals identify high‐risk patients, understand risk factors, and improve oral health management. Reporting Method TRIPOD‐AI checklist. Patient or Public Contribution No patient or public contribution.