<![CDATA[Maternal Mortality Ratio (MMR) remains a strategic issue in Indonesia's national health system. Efforts to reduce MMR are often hindered by the slow early detection of high-risk pregnant women at primary healthcare facilities. To bridge this gap, this study developed "Bunda Dini", a web-based Decision Support System (DSS) integrating a predictive machine learning model. The application was built using the framework Laravel and MySQL database management system. The system architecture was designed with a multi-tenant approach, enabling secure data isolation between various clinics or hospitals, as well as structured access permissions for Midwives, Obstetricians, Clinic Admins, and Super Admins (Developers). The system's inference engine is powered by an optimized Extreme Gradient Boosting (XGBoost) algorithm, which has been proven capable of classifying pregnancy risk levels based on six vital parameters with an accuracy rate of 84.24% and a high-risk class Recall reaching 96.36%. Key features of the application include an intelligent anamnesis form, risk level visualization real-time, as well as a comprehensive reporting module supporting patient data recapitulation in both PDF and Excel formats. Test results show that the system interface is highly responsive, easy for medical staff to use, and guarantees the confidentiality of medical records across clinics. The implementation of the "Bunda Dini" web platform has high potential to become a reliable medical triage instrument, accelerating emergency obstetric care responses, and directly contributing to efforts to reduce maternal mortality.]]>
Copyrights © 2026
