The Application of Machine Learning Models to Predict Stillbirths.
Oguzhan Gunenc1, Sukran Dogru1, Fikriye Karanfil Yaman2
1Konya City Hospital, Konya 42020, Turkey.
Medicina (Kaunas, Lithuania)
|March 27, 2025
Summary
Machine learning accurately predicts stillbirth using maternal and obstetric data. The Random Forest model achieved 96.8% accuracy, aiding early intervention to reduce stillbirth rates.
Area of Science:
- Perinatal Medicine
- Medical Informatics
- Machine Learning in Healthcare
Background:
- Stillbirth remains a significant challenge in obstetric care.
- Predictive models can improve early detection and intervention strategies.
Purpose of the Study:
- To evaluate the predictive value of obstetric clinic data for stillbirth detection.
- To assess the performance of machine learning models in predicting stillbirth.
Main Methods:
- Retrospective study of 951 pregnancies (452 stillbirths, 499 live births).
- Utilized maternal, fetal, and obstetric characteristics from electronic health records.
- Developed and compared four machine learning models: logistic regression, Support Vector Machine, Random Forest, and multilayer perceptron.
Main Results:
- Key risk factors for stillbirth included consanguinity, fetal anomalies, previous stillbirth, maternal thrombosis, oligohydramnios, and placental abruption.
- Previous stillbirth (OR: 7.31) and thrombosis (OR: 14.13) significantly increased stillbirth risk.
- The Random Forest model demonstrated the highest accuracy (96.8%), sensitivity (96.3%), and specificity (97.2%).
Conclusions:
- The Random Forest model effectively predicts stillbirth with high accuracy.
- Integrating maternal, neonatal, and obstetric risk factors into predictive models can aid healthcare providers.
- Improved stillbirth prediction can lead to timely prenatal interventions and reduced stillbirth rates.
