Current situation and risk prediction model of oral frailty among elderly people in rural areas

Objective: To investigate the current situation of oral frailty in elderly people adults and to establish a corresponding risk prediction model. Methods: Using convenience sampling method, 595 rural elderly people were selected as study subjects in Bengbu City, Anhui Province, from July to December 2024, and divided into the modeling group(n=445) and the validation group(n=150), and were surveyed using the general information questionnaire, oral frailty index screen-8, FRAIL scale, eating assessment tool-10, mini-nutritional assessment short form, geriatric depression scale-5, perceived social support scale, and geriatric self-efficacy scale for oral health were used to conduct the survey.Single- and multivariate logistic regression was used to comprehensively screen the influencing factors, and the screened influencing factors were plotted in dynamic line graphs using the R software ′rms′ package, and the predicton performance of the model was evaluated by ROC curves and calibration curves. Results: Multivarinte analysis determined that age ≥80 years (OR=3.804, 95%CI: 1.275-7.459), polypharmacy (OR=7.979, 95%CI: 3.766-16.903), oral health-related self-efficacy (OR=0.460, 95%CI: 0.231-0.917), swallowing dysfunction (OR=2.429, 95%CI: 1.168-5.051), wearing denture (OR=19.713, 95%CI: 9.982-38.927), and having a history of chronic disease (OR=1.928, 95%CI: 1.081-3.441) were the six factors as independent influences.The AUC value was 0.883 for the modeling set and 0.827 for the validation set, the calibration curves for the modeling and validation sets converged to the ideal curve, and the Hosmer-Lemeshow goodness-of-fit results (χ²=6.31, P=0.61) suggested that the prediction model was well fitted. Conclusions: The risk prediction model constructed in this study has good prediction performance, which can provide a theoretical basis for primary healthcare workers to prevent the occurrence of oral frailty at an early stage and to develop corresponding measures.