养老机构老年人低舌压的影响因素分析及决策树模型的构建OA

Objective:To investigate the factors influencing the maximum tongue pressure of elderly residents in nursing homes and to construct a decision tree model.Methods:A convenience sample of 206 elderly residents meeting the inclusion and exclusion criteria was recruited from two nursing homes in Nanchang,Jiangxi province,between April and July 2025.Participants were divided into a normal tongue pressure group(n=104)and a low tongue pressure group(n=102)based on the median maximum tongue pressure(28.13 kPa).Factors influencing reduced maximum tongue pressure were analyzed.IBM SPSS Modeler was used to construct a decision tree model for these factors.Results:The prevalence of low tongue pressure among the elderly residents was 49.5%.Both the logistic regression model and the decision tree model indicated that age,presence of chewing difficulties,EAT-10 scores,and the number of"ta"repetitions per second were associated with maximum tongue pressure.The areas under the receiver operating characteristic(ROC)curves for the two models were 0.829 and 0.821,respectively.DeLong's test showed no statistically significant difference between the two models(Z=0.471,P=0.677).Conclusion:Advanced age,chewing difficulties,higher EAT-10 scores,and fewer than six"ta"repetitions per second are risk factors for low tongue pressure in elderly residents of long-term care facilities.Both the logistic regression model and the decision tree model constructed in this study demonstrate good predictive performance,facilitating the early clinical identification of the risk of reduced tongue pressure and the implementation of personalized interventions.