Cause of Death Determination Ascertainment (CODA)

Role
Core team member; multimodal fusion evaluation
Problem
Many deaths occur without reliable cause documentation, limiting mortality statistics and public-health planning.
Contribution
Test the performance of multimodal fusion approaches; contribute to discussions of model choice, development, and interpretation and make recommendations; coauthor consortium papers.
Methods & tools
Cause-of-death classification, multimodal fusion, model evaluation and interpretation.
Outputs
A three-year international consortium project launched in 2026; methods and tools are in development.

CODA is a Gates Foundation initiative to improve cause-of-death assignment using artificial intelligence, coordinated through a global consortium led by Vital Strategies. It aims to strengthen death data in settings where reliable cause-of-death information is still limited.

My role centers on evaluating how well multimodal fusion approaches perform. I also take part in the consortium’s discussions of model choice, development, and interpretation, contribute recommendations, and coauthor resulting papers.

Project focus

  • Improve cause-of-death assignment in low-resource settings
  • Develop and test AI-enabled approaches for death data systems
  • Support more reliable mortality evidence for policy and planning