Geospatial Causal Inference for Barangay-Level Maternal Care Assessment: A Causal Forest Framework on Antenatal Intervention Outcomes in Davao City

Abstract

Maternal mortality is a critical concern globally, particularly in rural and economically disadvantaged countries. In the Philippines, despite national and local interventions, the situation remains a critical concern due to the country’s reported Maternal Mortality Rate. This study develops a geospatial causal inference framework to assess the effectiveness of maternal care programs in Davao City, particularly in the barangay level, providing insights into local variations in maternal health outcomes. Using the Routine Health Information System dataset covering from 2020 to 2024, the usage of Causal Forest estimates heterogeneous treatment effects of antenatal interventions, including prenatal care, anemia management, skilled birth attendance, and laboratory screenings. Furthermore, Spatial cross-validation and falsification tests are applied to address spatial autocorrelation and validate causal robustness. Additionally, results are visualized through an interactive dashboard, enabling clear identification of high and low-impact areas for targeted interventions. The integration of causal machine learning with geospatial analysis demonstrates the potential to uncover localized intervention effectiveness, optimizing maternal care resource allocation and strategic planning, as well as improving maternal health outcomes in Davao City.