Predictive analysis on the factors associated with birth Outcomes: A machine learning perspective.

Area of Science:

• Medical Informatics
• Machine Learning in Healthcare
• Maternal and Neonatal Health

Background:

• High global stillbirth rates, particularly in Sub-Saharan Africa, hinder progress toward Sustainable Development Goal 3 (SDG3).
• Ghana faces challenges in meeting maternal and neonatal mortality targets, including the World Health Organization's 2030 goal.
• Existing maternal healthcare interventions require enhancement with predictive tools for high-risk pregnancies.

Purpose of the Study:

• To identify critical factors influencing childbirth outcomes (stillbirth vs. live birth).
• To develop and evaluate a machine learning model for predicting high-risk pregnancies.
• To support Ghana and other Sub-Saharan African nations in improving maternal and neonatal healthcare.

Main Methods:

• Comparison of four machine learning classifiers: Extreme Gradient Boosting, Random Forest, Logistic Regression, and Artificial Neural Network.
• Utilized data from a tertiary health facility in Ghana for model training and validation.
• Addressed class imbalance in childbirth outcomes using the Synthetic Minority Over-sampling Technique (SMOTE).

Main Results:

• Fetal heartbeat and gestation age at birth were identified as the most significant predictors of childbirth outcomes.
• Maternal age, number of babies, and delivery method showed no significant association with birth outcomes.
• The Random Forest model demonstrated superior performance with high accuracy (0.98), F1-score (0.99), and AUC (0.90).

Conclusions:

• Machine learning offers a powerful tool for early detection of high-risk pregnancies in clinical settings.
• The study provides crucial insights for improving maternal and neonatal healthcare in Ghana and similar regions.
• Findings can inform policy and research to accelerate progress towards global maternal and neonatal health objectives.