Construction of decision tree model for subclinical hypothyroidism in pregnancy
Objective:To investigate the risk factors of subclinical hypothyroidism(SCH)during pregnancy and establish a decision tree model.Methods:The clinical data of 261 parturient women were retrospectively analyzed,the risk factors of SCH were screened by Logistic regression analysis,and the decision tree model was constructed by Modeler software.Results:The incidence of SCH was 12.64%.Logistic regression analysis showed that age≥30 years old,BMI＜24kg/m2,first delivery,gestational diabetes mellitus,preeclampsia and anemia were all influential factors for SCH in pregnant women(P＜0.05).Decision tree model shows that anemia is the most important factor affecting maternal SCH during pregnancy.ROC results showed that the AUC value of multi-factor was higher than that of decision tree(Z=2.470,P=0.140).Conclusion:Age,BMI,pregnancy and childbirth history,gestational diabetes,preeclampsia,and anemia are all influencing factors for SCH during pregnancy.The decision tree risk prediction model constructed based on these factors has good risk prediction performance.
spregnancySubclinical hypothyroidismRisk factorsDecision tree
国家科技期刊平台
NSTL
万方数据
