Gestational hyperglycemia and EOS

Dataset description This dataset contains the de-identified clinical data used for the development of a prenatal prediction model for neonatal early-onset sepsis (EOS). The dataset was derived from the development cohort of a multicenter mother–infant cohort study conducted in China between 2017 and 2023. It includes maternal clinical characteristics, pregnancy-related information, laboratory measurements, glucose-related indicators, delivery factors, and neonatal EOS outcomes. Maternal variables include demographic characteristics, obstetric history, pregnancy complications, fasting plasma glucose measurements during different gestational stages, oral glucose tolerance test parameters, routine blood parameters, biochemical indicators, and inflammation-related biomarkers. Selected derived metabolic and inflammatory indices were also included for model development. The dataset was used to construct and evaluate prenatal EOS prediction models based on maternal pregnancy information. Multiple statistical and machine-learning approaches were applied, and the final prediction model was developed using variables available before delivery to identify pregnant women at increased risk of neonatal EOS. All data were de-identified before release. No directly identifiable personal information is included in this dataset. The dataset is intended to facilitate reproducibility of the prediction model and support further research on prenatal risk assessment of neonatal early-onset sepsis.