Abstract
AimsTo develop a GDM risk stratification model in Chinese pregnant women using machine learning algorithm, for judgment of the risk of GDM before 16 gestation weeks. MethodsA retrospective study of 17005 pregnant women with 1965 women developed GDM. Maternal clinical routine examination indicators, disease history and other clinical characteristics of pregnant women were obtained before 16 gestation weeks. Maternal clinical parameters were analyzed, selected and divided into 6 groups. The prediction models were constructed using LR (logistic regression) and RF (random forest), and were evaluated using areas under the receiver-operating characteristic curve (AUC). The cut-off value of the predicted probability of GDM was calculated by interquartile range. The performance of models was internal validated. ResultsWe developed a GDM risk stratification prediction model in Chinese pregnant women before 16 gestation weeks, with the AUC 0.746 and 15 parameters included. The model presented reliable ability to predictively stratify GDM risk of population. And the ≥ 7.77% predicted risk cut-off showed a strong ability to rule out GDM in women who predicted negative before 16 gestational weeks. ConclusionsOur study provide a simple and effective screening method for clinical GDM risk stratification in Chinese pregnant women before 16 gestation weeks.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have