Overview

• Researchers integrated clinical data from over 17,000 Kenyan health facilities with satellite-derived crop health indicators to create the predictive AI model.
• The model achieves 89% accuracy for one-month forecasts and 86% accuracy for six-month predictions, outperforming traditional methods reliant on historical trends.
• A prototype dashboard visualizing regional malnutrition risks is being integrated into Kenya's health systems to enable targeted responses.
• The tool leverages Kenya’s DHIS2 health data platform, with potential for adaptation in more than 125 countries using the same system.
• The findings, published in PLOS One, mark a critical step in addressing malnutrition, which affects 5% of Kenyan children under five, with rates up to 25% in some regions.