The CODA project aims to provide more accurate cause-of-death data in low-income settings by combining artificial intelligence with household and clinical information. Funded for three years by the Gates Foundation and led by Vital Strategies with an international consortium, CODA addresses large gaps where only eight per cent of deaths have documented causes and many deaths occur outside health facilities. Philip Setel and others describe this lack of data as a major blind spot for health policy.
CODA trains its models on historical records and rigorously validated datasets of deaths confirmed through post-mortem investigation. The tool can run offline and upload records later. In community settings, community health workers carry out post-mortem interviews with family members; the system manages interviews in real time, adapts suggested questions as new details appear, and converts spoken testimony into structured inputs for the cause-of-death algorithm. In clinical settings, physicians can enter patient history, observations and test results. Rather than a single final verdict, CODA provides a level of confidence for each recommendation.
Partners include Northeastern University, the University of Washington, IS Global, RTI International and the CHAMPS Project Office. Limited trials are planned in South Africa and Bangladesh starting in September. Project leaders note that autopsies offer the highest-quality data but are costly and hard to scale. The project will form a scientific advisory committee to address ethical, legal and cultural issues, and experts have emphasised transparency, data protection, government engagement and collaboration with end users.
Difficult words
- consortium — group of organizations working together on a project
- post-mortem — process done after a person's death
- validate — check information or data to make sure it is correctvalidated
- autopsy — medical examination of a body after deathautopsies
- transparency — openness and clear sharing of information or decisions
- confidence — degree of certainty in a result or decision
Tip: hover, focus or tap highlighted words in the article to see quick definitions while you read or listen.
Discussion questions
- What are the advantages and limits of using AI combined with household interviews to determine causes of death in areas with few medical records?
- How should a project like CODA protect personal and family information while collecting sensitive health data?
- What practical or cultural challenges might the project face when running trials in different countries such as South Africa and Bangladesh?
Related articles
Study: Industrial Revolution pollution affected people unevenly in England
A study in Science Advances used bones and historical records to track pollution exposure during the Industrial Revolution. Researchers analysed remains from 94 people in industrial South Shields and rural Barton-upon-Humber and found uneven, gendered exposure.
Hantavirus linked to international cruise ship
WHO reports 11 hantavirus cases, including three deaths, connected to the Dutch-flagged MV Hondius that sailed from Argentina to Cape Verde. Some passengers later travelled through South Africa; health authorities say public risk remains low.
