Machine learning-based prediction of out-of-hospital births in prehospital emergency care.

BACKGROUND

Unplanned out-of-hospital births constitute rare but high-risk obstetric emergencies managed by emergency medical services (EMS). Rapid assessment of labor progression in prehospital settings is challenging due to limited diagnostic resources and time pressure, increasing the risk of adverse maternal and neonatal outcomes. Machine learning (ML) may support early risk stratification using routinely collected prehospital data.

OBJECTIVES

To develop and validate supervised ML models for predicting prehospital birth and to evaluate whether these models reflect clinically intuitive obstetric reasoning.

MATERIAL AND

METHODS

This retrospective observational study analyzed 3,002 EMS-attended labor cases in Poland (August 2021.January 2022). The outcome was birth occurring before hospital arrival. Candidate predictors included maternal characteristics, obstetric history, stage of labor, vital signs, and intrapartum findings. Penalized logistic regression (elastic net), random forest (RF), support vector classifier with radial basis function kernel (SVC-RBF), Gaussian naive Bayes (GNB), and k-nearest neighbors (kNN) models were trained using stratified fivefold cross-validation. Model performance was evaluated using discrimination metrics (area under the receiver operating characteristic curve (ROC-AUC) and precision-recall AUC (PR-AUC)) and calibration metrics (Brier score and logarithmic loss (log loss)). Nested cross-validation was applied to reduce overfitting. Model interpretability was assessed using standardized coefficients, permutation importance, and Shapley Additive Explanations (SHAP) values.

RESULTS

Penalized logistic regression demonstrated robust performance (ROC-AUC: 0.97 }0.01; PR-AUC: 0.81 } 0.04; Brier score: 0.036 }0.015). Random forest and SVC-RBF models achieved comparable discrimination (ROC-AUC up to 0.97), whereas kNN performed less well (ROC-AUC = 0.84). The 2nd stage of labor was the dominant predictor (Ŕ = 1.39), followed by amniotic fluid status (Ŕ = .0.44). Sensitivity analysis excluding the stage of labor reduced model performance but retained moderate discrimination (ROC-AUC . 0.76), indicating that additional clinical variables contributed to prediction.

CONCLUSIONS

Machine learning models demonstrated high internal predictive performance for prehospital birth using routinely available EMS data and reproduced clinically intuitive decision patterns. Such tools may support, but not replace, prehospital obstetric decision-making.