Risk prediction model for spontaneous preterm birth in patients with gestational diabetes based on logistic regression and decision tree
WANG Chang
XU Jian
XUE Xiumei
Abstract：Objective:To analyze the risk factors for spontaneous preterm birth(SPB)in patients with gestational diabetes(GDM),and to construct a logistic regression model and a decision tree risk model.Methods:A total of 218 GDM patients who received regular prenatal examinations and gave birth in our hospital from April 2022 to May 2024 were retrospectively selected as the research subjects.They were divided into SPB group and non-SPB group according to whether SPB occurred.Clinical data of the two groups were collected.Univariate and multivariate logistic regression were used to analyze the risk factors for SPB in GDM patients.Logistic regression and decision tree algorithms were used to construct its risk prediction model,and the predictive value of the two models for SPB was compared.Results:A total of 55 cases of SPB occurred in 218 GDM patients,with an incidence of 25.23%.Logistic regression analysis showed that age,pre-pregnancy body mass index(BMI),gestational hypertension,pre-pregnancy menstrual disorders,premature rupture of membranes,anemia and pregnancy infection were risk factors(P＜0.05).A decision tree model was constructed based on the risk factors.The model selected five explanatory variables,including age,pre-pregnancy BMI,pre-pregnancy menstrual disorders,premature rupture of membranes,and pregnancy infection,with a total of 4 layers and 13 nodes.Age was the most important influencing factor for GDM patients with SPB.The area under the curve(AUC)of the decision tree model for predicting GDM patients with SPB was 0.909[95%CI(0.862,0.943)],and the AUC of the Logistic regression model was 0.832[95%CI(0.776,0.879)].The DeLong test results of the two models were Z=2.486,P=0.0129.Conclusion:The decision tree risk prediction model was constructed based on the risk factors of age,pre-pregnancy BMI,gestational hypertension,pre-pregnancy menstrual disorders,premature rupture of membranes,anemia,and infection during pregnancy.The prediction efficiency of the decision tree risk prediction model was significantly higher than that of the Logistic regression model.
Keywords：Logistic regressiondecision tree modelgestational diabetes mellitusspontaneous preterm birthrisk prediction model
Publication Date：2025-08-15
Online Publishing Date：2025-08-25(First online date of this platform, not the publication date of the document)
Pages：6( 2831-2836 )