Maternal health risk analysis, automated pregnancy risk level identification and status monitoring system based on multivariable clinical data

Background: Ensuring high-quality care and regular monitoring of maternal health are essential during pregnancy to promptly detect potential complications. Unfortunately, the lack of specialist obstetricians and necessary expertise poses challenges in providing frequent follow-up and monitoring for pregnant women in low-resource areas. This research aimed to address these issues by designing and developing a pregnancy risk level and status monitoring system using a machine learning approach. Methods: A system for classifying pregnancy risk levels and monitoring status was developed using data from pregnant women, including age, systolic and diastolic blood pressure, blood sugar levels, body temperature, and heart rate. Six machine learning models were trained to classify risk levels. Additionally, a web-based user interface was developed to enable physicians to identify risk levels and track the trends of each clinical parameter throughout the pregnancy period. Results: Based on our experimental analysis, we found a strong correlation between high pregnancy risk and elevated systolic and diastolic blood pressures, higher blood sugar levels, and older ages compared to other clinical parameters. Out of the six machine learning models trained and validated, the random forest classifier demonstrated superior performance with a 91% classification accuracy and was chosen for risk level classification. Conclusions: Our results highlight the accuracy and effectiveness of the machine learning model in predicting pregnancy risk levels using clinical data. The proposed risk level and status monitoring systems can serve as valuable decision support tools for managing maternal complications, particularly in resource-limited settings.