Kenya Adopts Machine Learning Tool to Forecast Child Malnutrition Ahead of Droughts

Jameel Observatory-backed model becomes the first forward-looking child nutrition forecast in the country’s early warning system

Over 740,000 children in Kenya between the ages of six and 59 months required treatment for acute malnutrition in the 12-month period ending March 2026, according to national food and nutrition analysis. Kenya’s early warning infrastructure has historically flagged these crises only after conditions deteriorated — relying on data that reflected the present rather than anticipating what was coming.

A machine learning tool developed by Susana Constenla-Villoslada, a doctoral researcher at the University of California, Berkeley, now changes that. The model combines historical acute malnutrition rates with data on weather patterns, conflict, and food prices to generate forecasts at one, three, and six-month intervals. Because malnutrition shifts gradually and its underlying drivers tend to persist, past measurements prove reliably predictive of future risk — a property the model exploits to give officials earlier warning than previous approaches allowed.

Co-developed with Kenya’s National Drought Management Authority (NDMA), the tool has been formally woven into the authority’s drought early warning processes. Outputs feed into meetings of the Kenya Food Security Steering Group, bringing together government departments, U.N. agencies, donors, and NGOs to coordinate response. Deployment was supported by the University of Edinburgh and the Jameel Observatory for Food Security Early Action, an international partnership that includes ILRI, Save the Children, J-PAL, and Community Jameel.

The NDMA can now operate, update, and maintain the tool independently, with the model refreshed monthly. Piloting of anticipatory action trigger mechanisms is planned across three Kenyan counties — Isiolo, Tana River, and Marsabit.

George Richards, director of Community Jameel, described the tool as “harnessing machine learning to forecast when children in Kenya are at risk of acute malnutrition” and turning “complex data into life-saving action.” Community Jameel was founded by Mohammed Abdul Latif Jameel KBE. The work was presented at the One Health Summit in Lyon, convened by French President Emmanuel Macron, alongside the broader Nutrition for Growth Summit commitments that set global targets for data-driven nutrition action.