 This study performed a meta-analysis on machine learning models used to predict gestational diabetes mellitus in the general population and identified logistic regression as the most commonly employed method, with maternal age, family history of diabetes, BMI, and fasting blood glucose being the for most commonly used features. The study found that ML methods are attractive for predicting GDM, but emphasized the importance of quality assessments and unified diagnostic criteria to expand their use. This article was authored by Zheqing Zhang, Luchin Yang, Wen Tao Han, and others.