Key points are not available for this paper at this time.
Background Intrapartum high fever (≥38.5 °C) is associated with adverse outcomes, but predicting its occurrence in parturients with intrapartum fever remains difficult. We developed a machine learning model to assess this risk. Methods This multicenter retrospective study included parturients who received labor analgesia and developed intrapartum fever (≥38.0 °C) from three Chinese hospitals. The derivation cohort comprised parturients from two hospitals, with parturients from the third hospital serving as an independent external validation cohort. Candidate variables were extracted from electronic health records (EHR). Least absolute shrinkage and selection operator (LASSO) regression was used for feature selection. An extreme gradient boosting (XGBoost) model was developed with hyperparameters optimized via five-fold cross-validation and random search. Model performance was evaluated using area under the receiver operating characteristic curve (AUROC), area under the precision-recall curve (AUPRC), balanced accuracy, and F1-score. The SHapley Additive ExPlanations (SHAP) method was applied to interpret the model. Findings A total of 747 parturients were included in this study, of which 238 cases (31.9%) developed intrapartum high fever. Using LASSO regression, six predictors were retained: body mass index (BMI), meconium-stained amniotic fluid, hypertension, anemia, monocyte/lymphocyte ratio (MLR), and platelet/lymphocyte ratio (PLR). The XGBoost model achieved an area under the curve (AUC) of 0.771 in the training set, 0.716 in the test set, and 0.674 in the external validation set. SHAP analysis indicated that BMI was the most important predictive factor. Conclusion The XGBoost model demonstrated good performance in predicting intrapartum high fever in parturients with intrapartum fever receiving labor analgesia. SHAP analysis further revealed that BMI was the most important predictive factor in the model.
Liu et al. (2026) studied this question.