CARE-Africa - Novel Interventions to Identify and Control Diarrhoeal Diseases in Sub-Saharan Africa
CARE-Africa is developing an AI-based clinical decision-support tool for diarrhoeal diseases in children under five across sub-Saharan Africa. Diarrhoeal disease s remain a major cause of illness and death in the region, with lasting effects on children's growth, development and schooling. Because many different pathogens can cause diarrhoea and outcomes vary widely, clinicians often struggle to identify the cause and judge which children are most at risk. The project's tool combines patient data entered by a clinician with automatically retrieved environmental, climate, socioeconomic and demographic data to suggest a likely diagnosis and treatment, including an estimate of antibiotic resistance risk and an explanation of its reasoning. The tool is designed to run on a tablet, deliver results within five minutes, and work offline where connectivity is limited. The project is coordinated by King's College London, alongside partners in Uganda (Infectious Diseases Institute), Ethiopia (Addis Ababa Health Bureau), South Africa (Jembi Health Systems), Italy (University of Perugia) and Spain (Causal Foundry).
Investigators
Infectious Diseases Institute (Kampala, Uganda)
- Dr Francis Kakooza
Head Department of Global Health Security,
Infectious Diseases Institute - Dr Dathan Byonanebye
Deputy Head Department of Global Health Security,
Infectious Diseases Institute
Addis Ababa Administrative Health Bureau AACAHB (Addis Ababa, Ethiopia)
JEMBI Health Systems (Cape Town, South Africa)
- Dr Christopher Seebregts
Founder and CEO
Jembi Health Systems
University of Perugia UNIPG
- Prof. Nicola Senin
- Prof. Michele Moretti
Causal Foundry
Aims
To reduce the burden of paediatric diarrhoeal disease in sub-Saharan Africa by building an AI tool that helps clinicians identify the likely cause of a child's illness, judge severity and risk, and choose appropriate treatment including guidance on antibiotic resistance risk without needing specialised laboratory testing at the point of care.
Methods
The team is training AI models on multi-scale, multi-modal clinical, environmental, climate, socioeconomic and demographic data, drawn from historical records in Ethiopia and Uganda and newly collected data from sentinel sites. The tool's interface is being co-designed with clinical end-users for use on a tablet device. It will first be piloted in eight healthcare facilities in Ethiopia and Uganda, then rolled out for further evaluation across 40 facilities, generating results for approximately 4,700 children. A parallel modelling strand will produce risk maps and scenario tools to support policymakers, informed partly by community consultations on local risk factors and barriers to care.
Trials Design
Two-stage clinical evaluation: an initial pilot across 8 healthcare facilities in Ethiopia and Uganda, followed by a larger evaluation across 40 facilities involving approximately 4,700 children.
Impact
If validated, the tool could help ensure children receive more appropriate treatment for diarrhoeal diseases, accelerating recovery, reducing unnecessary antibiotic use, easing pressure on antimicrobial resistance, and making better use of limited health system resources. It aims to give clinicians a decision-support tool that draws on a wide range of data sources while requiring minimal manual input, including rapid assessment of likely antibiotic resistance without waiting days for lab-based susceptibility testing.
Principal Investigator
Professor of Artificial Intelligence for Science
Affiliations
Funding
Funding Body: EDCTP3 - European & Developing Countries Clinical Trials Partnership
Amount: €4,808,785.75
Period: May 2026 - April 2029