Machine Learning Model for Predicting Intrapartum High Fever in Chinese Parturients

A multicenter retrospective study from three Chinese hospitals includes 747 parturients who received labor analgesia and developed intrapartum fever. The dataset was used to develop and validate an XGBoost model predicting high fever (≥38.5 °C), which achieved an AUC of 0.674 in external validation. Authored by Bo Liu and last updated on 2026-05-07.

Use Cases

• Training a predictive model for intrapartum high fever based on EHR-derived features like BMI and blood cell ratios.
• Benchmarking machine learning models for maternal health risk assessment using a dataset with external validation.
• Interpreting feature importance in clinical models using SHAP analysis on predictors such as BMI and meconium-stained amniotic fluid.
• Studying the epidemiology of intrapartum fever and its progression to high fever in a Chinese hospital cohort.

Strengths

• Dataset includes 747 patient records with a specific outcome (31.9% developed high fever).
• Model was validated on an independent external cohort from a third hospital.
• Six key predictors were identified via LASSO regression, including BMI, MLR, and PLR.
• Model performance metrics (AUC, AUPRC, balanced accuracy, F1-score) are reported for training, test, and validation sets.

Limitations

• Row count and column-level documentation are unknown, limiting suitability assessment.
• Data is limited to 21.1 KB, suggesting a very small scope, likely containing only model results or summary statistics.
• Geographic scope is limited to three hospitals in China, which may limit generalizability.

Provenance

Source

figshare

Collection Method

Multicenter retrospective study extracting candidate variables from electronic health records.

Freshness

Last updated 2026-05-07 05:24:41; freshness should be verified.

Geography

Three hospitals in China.