A Burundian Health Intelligence Platform (BHIP): Leveraging technologies to track and improve maternal and neonatal outcomes in a very low-income setting

I demonstrated in my DPhil that the Burundi’s health information system for maternal and newborn health (MNH) remains poor; records are missing or likely inaccurate. Plus, only women who contact health facilities are captured, leaving significant community-level events and outcomes unrecorded. To transform routine health surveillance in Burundi, one of the world’s poorest countries, I propose designing and testing how a Health Data Intelligence System (HDIS), focused on MNH, could be implemented leveraging on proven digital technologies to capture and integrate community health worker and facility data and on evolving large language models (LLM) that, for example, enable obstetric risk stratification and other forms of patient and provider decision support. Operationally, CHWs will record antenatal and postnatal information into a smartphone App (CHW MNH App), while facility personnel will access a web platform to add peri-partum information. These will stream into existing facility information systems; OpenClinic and DHIS2. A pilot implementation articulated in a ‘Programme Theory’ will generate learning for scalability and transferability while producing actionable MNH data useful for building and testing a ‘data to decision-making pipeline’. This research will foster local data governance, create and operationalise a data protection policy, and strengthen South-South and South-North partnerships