Sankho Mvundula’s wooden sculpture, its curves carved to represent breath and resilience, stands in Malawi’s Kungoni Center of Culture and Art—a quiet testament to a new hope in child health. At its core is BIOTOPE, an AI tool developed by researchers at University College Dublin that’s proving far more accurate than existing methods at identifying Malawian children with severe pneumonia who urgently need hospital care. Pneumonia still kills nearly 1 million children under five each year worldwide, and in Malawi—where one doctor serves roughly 28,000 people—spotting the sickest kids early is a matter of life and death. BIOTOPE, trained on data from over 2,500 children across primary clinics in the country, doesn’t just offer predictions; it offers precision where it’s needed most.
The algorithm, detailed in a study published in PLOS Medicine, uses a machine-learning technique called a "random forest" model to analyze a constellation of factors: breathing rate, temperature, heart rate, oxygen levels, nutritional status, and even household conditions. Unlike current international guidelines, which can miss critically ill children who don’t display classic warning signs, BIOTOPE significantly outperforms existing risk assessment tools. Crucially, it’s built to integrate seamlessly into Malawi’s existing Integrated Community Health Information System (iCHIS), meaning health workers can use it without extra paperwork or training—a vital consideration in overstretched clinics.
Led by Dr. Joe Gallagher of UCD’s School of Medicine, the BIOTOPE project is a collaboration spanning Mzuzu University, the University of Galway, Queen’s University Belfast, the World Health Organization, the Malawi Ministry of Health, and Luke International Norway. The team didn’t work in isolation: parents, caregivers, and local artists were involved from the start, ensuring the tool reflects community needs. "This is science that starts and ends with the communities it is designed to serve," said Professor Balwani Mbakaya of Mzuzu University.
One of BIOTOPE’s most powerful features is its adaptability. As new patient data accumulates, the algorithm can be retrained, allowing it to evolve with changing disease patterns and health system demands. "Machine learning gives us the ability to create something that improves over time rather than becoming obsolete," said Professor Cathal Seoighe of the University of Galway. For Dr. Chris Watson of Queen’s University Belfast, the project exemplifies global collaboration at its best—"what can be achieved when researchers, clinicians and communities work together across borders."
With pneumonia still claiming too many young lives, BIOTOPE offers more than technological innovation; it offers a practical, scalable lifeline. As it moves from research to real-world use, it carries not just data and code, but the hopes of families and artists alike—breathing new life into the fight for child survival.