Exploring the Role of Abuse, Substance use and General Health in Pregnancy Risk Prediction using Machine Learning Algorithms

Pregnancy needs continuous health assessment for mothers because it affects maternal wellbeing and baby development. Correct evaluation of pregnancy risks allows doctors to intervene at the right time for proper healthcare distribution. This research examines the performance behavior of today’s most advanced machine learning systems. After extensive analysis of multiple approaches, the model selection process was finalized. Its primary task is to classify pregnancy risk levels into three categories: The system generates risk levels of High and Moderate and Low Risk by evaluating the expectant mother’s vital signs. Approximately 2,49,654 samples were extracted from the 2016-2022 PRAMS dataset that National Center for Chronic Disease Prevention and Health Promotion (CDC) created. Key input parameters begin with maternal abuse followed by wants to get pregnant then proceed to drinking behavior drug intake and smoking patterns alongside oral health and interactions with doctors about diet ingredients exercise depression anxiety occupational status workplace leave and stress and several preconception and pregnancy vaccine sequences from the mother. Historical patient records containing NIH and WHO defined vital signs and risk assessment data will feed the model which uses Multiple Imputation by Chained Equations (MICE) to handle null values when learning complex input-output connections between variables. An advance version of GBM, Extreme Gradient Boosting shows a highest Accuracy of 91% . This outperformed all other individual classifiers evaluated including Decision Tree (88.65%), Random Forest (82%), Support Vector Machine (71.27%), Gradient Boosting Machine (88.22%), Neural Network (90.43%). Through this model healthcare providers maximize resource management by tailoring interventions for pregnant patients. This novel ML system produces accurate pregnancy risk forecasts which can enhance both maternal and fetal healthcare by enabling quick interventions while streamlining personalized care approaches for better worker quality.