Predicting Intrapartum High Fever in Chinese Parturients Receiving Labor Analgesia

A multicenter retrospective study of 747 parturients from three Chinese hospitals who developed intrapartum fever while receiving labor analgesia. The dataset includes 238 cases (31.9%) of intrapartum high fever and was used to develop an XGBoost model with six predictors. The model achieved an AUC of 0.674 in external validation.

Use Cases

• Training predictive models for intrapartum high fever based on BMI, meconium-stained amniotic fluid, hypertension, anemia, MLR, and PLR.
• Evaluating feature importance for clinical prediction models using SHAP analysis.
• Benchmarking machine learning model performance (AUROC, AUPRC, balanced accuracy, F1-score) in external validation cohorts.
• Studying risk factors associated with adverse outcomes in parturients receiving labor analgesia.

Strengths

• 747 patient records from a multicenter study across three hospitals.
• Model performance metrics reported for training (AUC 0.771), test (AUC 0.716), and external validation (AUC 0.674) sets.
• Six key predictors identified via LASSO regression, with BMI highlighted as the most important factor.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Data may reflect geographic bias inherent to the three Chinese hospitals.

Provenance

Source

Electronic health records (EHR) from three Chinese hospitals.

Collection Method

Multicenter retrospective study.

Freshness

Last updated 2026-05-07 05:24:40; freshness should be verified.

Geography

China (three hospitals).