Construction and validation of an oral frailty risk prediction model for community-dwelling older adults

OBJECTIVE: The aim of this study was to explore the factors that influence oral frailty (OF) in community - dwelling older adults, and to construct and validate a risk prediction model for oral frailty. METHOD: Through univariate and logistic regression analyses, we identified predictors of oral frailty in community-dwelling older adults. A nomogram was drawn based on the logistic regression results, and the predictive ability of the model was assessed by the area under the receiver operating characteristic curve (ROC), the Hosmer-Lemeshow (H-L) test, and calibration curves. RESULTS: Variables in the model included history of falls within one year, sarcopenia risk, oral health knowledge, oral health beliefs and oral health behaviours (P < 0.05). The model showed good discrimination; the AUC for the training, validation and test groups was 0.895 (95% CI = 0.854–0.937), 0.880 (95% CI = 0.816–0.944) and 0.835 (95% CI = 0.760–0.910), respectively; H-L test results were χ2 = 6.126(P = 0.633), χ2 = 7.475( P = 0.486), χ2 = 5.603(P = 0.692), respectively.The calibration ability of all three groups of models was good. CONCLUSION: The predictive nomogram model constructed in this study exhibits excellent discrimination and calibration capabilities. It can accurately and conveniently screen the elderly population with oral frailty in the community, providing a reference for healthcare professionals to conduct early identification and implement prevention and control measures.