Development and internal validation of a risk prediction model for oral frailty in hospitalized older adults with chronic diseases

Background This study aimed to investigate the prevalence and influencing factors of oral frailty, a significant geriatric syndrome, among hospitalized older adults with chronic diseases, and to develop a corresponding risk prediction model to facilitate early screening and intervention. Methods From September 2024 to May 2025, we recruited 443 older adult patients with chronic diseases from two tertiary grade A hospitals in Wuhu City, Anhui Province, using a convenience sampling method. Data were collected using a general information questionnaire, the Oral Frailty Index-8 (OFI-8), the FRAIL scale, the Appetite Scale, and the Clinical Physiological Resilience Scale. Binary logistic regression analysis was conducted to identify factors associated with oral frailty. Results The prevalence of oral frailty among hospitalized older adult patients with chronic diseases was 69.3%(307/443). Appetite, employment status, age, Clinical physiological resilience, and frailty were the best predictors of oral frailty in hospitalized older adult patients with chronic diseases. These influencing factors were utilized to construct a nomogram model, which demonstrated excellent consistency and accuracy. The area under the curve (AUC) in the training set was 0.805 (95% confidence interval [CI] = 0.753–0.857). In the validation set, the AUC was 0.900 (95% CI = 0.843–0.957). The Hosmer–Lemeshow test values were p = 0.465 and p = 0.161 (both > 0.05). The calibration curve indicated a significant agreement between the nomogram model and the actual observations. Both the ROC curve and decision curve analysis (DCA) demonstrated that the nomogram possesses outstanding risk predictive performance. Conclusion The prevalence of oral frailty is notably high among hospitalized older adult patients with chronic diseases, with key influencing factors including age, employment status, appetite, and clinical physiological resilience. The developed risk prediction model demonstrates good calibration and discriminative ability, supporting its potential use in early screening and targeted intervention of oral frailty in this population.