Machine learning models for predicting preeclampsia: a systematic review.

Area of Science:

• Obstetrics and Gynecology
• Medical Informatics
• Artificial Intelligence in Healthcare

Background:

• Preeclampsia poses significant risks to maternal and fetal health.
• Accurate prediction of preeclampsia is crucial for timely intervention and improved outcomes.
• Existing prediction methods have limitations, necessitating exploration of advanced techniques.

Approach:

• This systematic review followed PRISMA guidelines, searching major databases up to February 2023 for studies on machine learning for preeclampsia prediction.
• Included studies were assessed for risk of bias and applicability using PROBAST.
• Four retrospective cohort studies involving nine distinct machine learning models were analyzed.

Key Points:

• Machine learning models utilized maternal characteristics, medical history, medication, obstetrical history, and lab/ultrasound findings.
• Effective models included Elastic Net, Stochastic Gradient Boosting, Extreme Gradient Boosting, and Random Forest.
• Area Under the Curve (AUC) for ML models ranged from 0.860 to 0.973, indicating high predictive performance.

Conclusions:

• Machine learning models demonstrate high prediction performance for preeclampsia risk.
• Early pregnancy information is sufficient for training effective ML models.
• These findings support the integration of ML into routine prenatal care for preeclampsia prediction.