Construction of a frailty risk prediction model based on oral frailty scores for older adults in Hubei Province: a cross-sectional study

This study aimed to investigate the relationship between oral frailty scores and frailty risk among older adults in Hubei Province, as well as to construct and validate a frailty diagnostic prediction model based on oral frailty scores. A total of 433 elderly individuals were recruited, of whom 384 aged 60 years and above from Hubei Province were ultimately included in the cross-sectional study. Frailty status was assessed using the FRAIL scale. Prior to performing multivariate logistic regression to identify independent associated factors, the data were balanced using the SMOTE-Tomek Links technique. A nomogram prediction model was constructed based on key variables. Its predictive performance and calibration were evaluated using Receiver Operating Characteristic (ROC) curve and calibration curves. Multivariate regression analysis identified age ≥ 80 years, limited physical activity duration, and fewer than 20 remaining teeth as independent associated factors for frailty. ROC curve analysis indicated that both the D-E-N-T-A-L score (AUC = 0.812) and the OFI-8 score (AUC = 0.693) were predictive of frailty status, with the former demonstrating stronger discriminative ability in this sample. The nomogram model integrating age, physical activity level, number of remaining teeth, social interaction, frequency of outings, and the D-E-N-T-A-L score achieved an AUC of 0.838. The nomogram model based on D-E-N-T-A-L oral frailty score demonstrated good discriminative ability for concurrent frailty status in a sample of community-dwelling older adults. It may serve as a potential screening aid for early frailty screening and personalized health management, thereby facilitating early intervention for frailty among older adults.