Influencing factors for dysphagia in the elderly and establishment of a predictive model
10.13406/j.cnki.cyxb.003733
- VernacularTitle:老年吞咽障碍影响因素探究及其预测研究
- Author:
Peng PENG
1
;
Xinrui CHEN
;
Yilin ZHOU
;
Xiaoqin TIAN
;
Yuqin TANG
;
Dan DENG
Author Information
1. 重庆医科大学公共卫生学院卫生统计学教研室,重庆 400016
- Keywords:
dysphagia;
influencing factors;
predictive model;
XGBoost model
- From:
Journal of Chongqing Medical University
2025;50(4):501-510
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To investigate the influencing factors for dysphagia in the elderly,to construct a predictive model for dysphagia,and to provide a theoretical basis for clinical practice.Methods:In this case-control study,the patients with dysphagia who attended Department of Geriatrics in the first affiliated hospital of Chongqing Medical University from March 2016 to June 2023 were enrolled as case group,and the patients without dysphagia who attended the same department during the same period of time were enrolled as con-trol group.The correlation analysis,least absolute shrinkage and selection operator(LASSO)regression,and multivariate logistic re-gression analysis were used to investigate the influencing factors for dysphagia;the 10-fold cross-validation Extreme Gradient Boosting(XGBoost)model was used to predict dysphagia,and the SHapley additive exPlanations(SHAP)method was used for model visualiza-tion.Results:There were 1009 cases in the case group and 2125 cases in the control group.The correlation analysis and LASSO re-gression analysis identified 12 factors for the multivariate logistic re-gression analysis,and the results showed that sarcopenia,increasing age,children or caretakers as caregivers,frail health,poor oral health,poor self-care ability,depression,and cognitive impairment were risk factors for dysphagia(odds ratio[OR]>1,P<0.05),and fe-male sex and participation in community activities were protective factors against dysphagia(OR<1,P<0.05).The XGBoost model had a good predictive efficacy,with an accuracy rate of 0.795,a preci-sion rate of 0.711,a sensitivity of 0.613,a specificity of 0.881,an F1 value of 0.661,and an area under the ROC curve of 0.855.The SHAP plot showed that the top five important characteristics were caregiver,oral score,frail health condition,activities of daily living,and cognitive function.Conclusion:There are various influencing factors for dysphagia in the elderly,and the elderly patients with poor oral health,frailty,dependence on others for daily life,and cognitive impairment should be taken seriously in clinical practice.The XGBoost model has a good performance in predicting dysphagia in the elderly,which can provide a reference for clinical practice.
