A Meta-Analysis of Machine Learning Models in Pregnancy Care and Applications

Examining the efficacy of machine learning-based box models in pregnancy care, this meta-analysis delves into diverse models—white, grey, and black box variants—to predict childbirth types, delivery modes, and mortality risks. The methodology integrates Random Forest, Logistic Regression, Neural Networks, and others, analyzing data ranging from demographics to genetic information. Results showcase model versatility in the accurate risk assessments for maternal and fetal health. Challenges include interpretability, healthcare system integration, and data quality. In conclusion, the study underscores machine learning's transformative potential in pregnancy care, emphasizing the need for ongoing research to enhance interpretability, address limitations, and optimize integration into routine prenatal care for improved outcomes.