急性上呼吸道感染加重住院风险中西医预测模型的LASSO和Logistic回归构建OA

Objective:To construct a prediction model for hospitalization risk of aggravated acute upper respira-tory tract infection based on LASSO and logistic regression,so as to provide reference for early identification of se-vere infection and early prevention.Methods:A total of 19 178 patients with acute upper respiratory tract infec-tion admitted to the emergency department from November 2023 to May 2025 were enrolled and divided into the training set(n=13 424)and validation set(n=5 754)at a ratio of 7:3.LASSO regression was used to screen core predictive variables,and Logistic regression was combined to establish the model.The receiver operating character-istic curve(AUC),Hosmer-Lemeshow test and decision curve analysis were adopted to verify the model efficacy.Results:Multivariate logistic regression analysis showed that age,white blood cell count(WBC),platelet(PLT),C-reactive protein(CRP),general aching symptoms,baloxavir marboxil treatment,in-hospital traditional Chinese med-icine preparations and TCM syndrome types were independent influencing factors for aggravated hospitalization of acute upper respiratory tract infection.The AUC values of the model in the training set and validation set were 0.96 and 0.95 respectively[95％CI(0.95,0.97)(0.93,0.97)].Hosmer-Lemeshow test and DCA indicated that the model had good calibration degree and high net clinical benefit.Conclusion:The integrated traditional Chinese and western medicine prediction model for hospitalization risk of aggravated acute upper respiratory tract infection constructed based on LASSO and Logistic regression presents good discrimination,calibration and clinical practica-bility.The nomogram can quantitatively assess hospitalization risk intuitively,which has important clinical value for early identification,diagnosis and treatment.