Every year, diarrhoea kills hundreds of thousands of children under five in sub-Saharan Africa. Not because there are no treatments. But because, far too often, no one knows which treatment to use.

When a child arrives at a clinic with diarrhoea, a nurse or doctor has to make a fast call. Is it a bacterial infection or a viral one? Which bug is causing it? The wrong answer means the wrong medicine. In the worst cases, it means death.

A new €4.8m research project wants to take that guesswork away — and the Infectious Diseases Institute (IDI) is part of it.

What is CARE-AFRICA?

CARE-AFRICA is a research project led by King’s College London. It is funded by the European Union through the Global Health EDCTP3 programme and brings together six partners from Uganda, Ethiopia, South Africa, the UK, Italy and Spain.

The team’s aim is to build an AI tool that can tell a health worker, within five minutes, what is most likely making a child sick — and what the best treatment is. The tool will run on a tablet.

“Ultimately, we want to equip a frontline healthcare worker with a tool that can predict the most likely cause of a child’s diarrhoeal infection from their symptoms. Is it a bacterium or a virus? If it’s a bacterium, which antibiotics work — and which ones should be avoided?” said Professor Tania Dottorini, the project lead and a professor of AI at King’s College London.

Why does this matter?

Diarrhoea is one of the biggest killers of young children in the world. The problem is tricky: it can be caused by many different bugs — bacteria such as Shigella and E. coli, or viruses like norovirus — and each needs a different response.

When health workers cannot quickly identify the cause, they often prescribe broad antibiotics and hope for the best. This is dangerous. Giving the wrong antibiotic does not just fail the child in front of you — it feeds a bigger crisis called antimicrobial resistance (AMR). Over time, overuse of antibiotics makes the drugs stop working. That affects everyone.

The CARE-AFRICA tool is designed to break this cycle. By helping workers choose the right drug for the right bug, it could save lives and slow AMR simultaneously.

How will the tool work?

The AI will be trained on data from children in Uganda and Ethiopia who have had diarrhoeal infections. The data will cover clinical symptoms as well as broader factors such as the local environment, climate, and community health trends.

When a health worker enters a child’s details, the tool will show which bugs are most likely responsible. It will also flag which antibiotics work against them and which to avoid. All of this in five minutes, on a tablet.

Testing it in the real world

The tool will first be tested in eight clinics in Uganda and Ethiopia. Based on those results, the team will improve it before running a larger trial across 40 clinics with up to 4,700 children.

If it works, the tool could speed up diagnosis, cut unnecessary antibiotic use and take pressure off overstretched health systems.

“This project is truly going beyond what we do now. We’re moving AI out of the laboratory and into the real world, in countries where it can make a real difference,” said Professor Dottorini.

What does it mean for Uganda?

Makerere University is one of the six partners in this project. Uganda is one of the two countries where the tool will be built and tested. For IDI, this is what good research looks like — science that starts in communities, is tested in communities, and ends up helping them.

Professor Dottorini, who grew up in Africa, put it simply: “As a mother, to see your child suffering is terrible. I want to make a difference for children, mothers and families in places where access to doctors and healthcare is limited. This project gives us a chance to do that.”

For children across Uganda and Ethiopia, CARE-AFRICA offers something real: faster answers, better treatment, and a better chance of survival.