Artificial Intelligence Enhances Prediction and Prevention of Child Malnutrition

A team of interdisciplinary researchers from the University of Southern California (USC), in collaboration with the Microsoft AI for Good Lab, Amref Health Africa, and Kenya's Ministry of Health, has developed an innovative AI model capable of predicting acute child malnutrition in Kenya up to six months in advance. This groundbreaking tool harnesses clinical data from over 17,000 Kenyan health facilities combined with satellite imagery capturing crop health and productivity, delivering a comprehensive approach to early warning systems. The AI achieves an impressive accuracy of 89% when forecasting malnutrition one month ahead and maintains an 86% accuracy over six months, surpassing traditional methods that often rely solely on recent trends. This advanced forecasting capability is particularly valuable in regions experiencing fluctuating malnutrition rates, enabling timely intervention and resource allocation.

The importance of this development is underscored by the fact that approximately 350,000 Kenyan children under five currently suffer from acute malnutrition, which significantly compromises their immune system and increases mortality risk from common illnesses. In certain areas, the rate reaches as high as 25%. Global undernutrition contributes to nearly half of all deaths in children under five, marking it as a critical public health challenge.

Current prediction methods mainly depend on expert judgment and historical data, which often fall short in capturing emerging hotspots or sudden shifts in malnutrition prevalence. The new AI tool leverages Kenya's routine health information system (DHIS2) data and satellite-derived indicators to identify high-risk zones with greater precision. A prototype dashboard visualizes regional malnutrition risks, supporting health authorities and humanitarian agencies in making faster, more targeted responses.

Dr. Bistra Dilkina, co-director of USC's Center for Artificial Intelligence in Society, emphasized the transformative potential of this technology: "By utilizing data-driven AI models, we can better understand complex relationships affecting child nutrition and improve our predictive capabilities."

The study detailing this innovation was published in PLOS One, reflecting extensive collaboration among experts from Microsoft, USC, Amref Health Africa, and the Kenyan Ministry of Health. Girmaw Abebe Tadesse from Microsoft highlighted the significance of the project: "Malnutrition is a major challenge on the African continent, exacerbated by climate change and food insecurity. This AI tool has the potential to make a profound difference."

In Kenya, malnutrition affects around 5% of children under age five—about 350,000 children—leading to weakened immunity and increased risk of death from illnesses such as diarrhea and malaria. The most affected regions experience prevalence rates up to 25%. Dr. Laura Ferguson from USC's Institute on Inequalities in Global Health pointed out that current forecasting methods are inadequate for real-time response, underscoring the importance of integrating health data with satellite imagery for proactive interventions.

A notable outcome of this research is the development of a prototype dashboard that displays regional malnutrition risks, aiming to facilitate faster and more efficient responses. The team is working closely with Kenyan health authorities and Amref Health Africa to embed the AI model into health planning and decision-making processes, striving to establish a sustainable, regularly updated public resource. Ferguson emphasized that solving complex health issues like malnutrition requires a collaborative approach involving public health experts, medical personnel, nonprofits, and engineers, with every partner playing a crucial role.

Since over 125 countries utilize Kenya's DHIS2 system, this AI-based framework has the potential for global application, allowing other nations to adopt similar models for combating child malnutrition. As Dr. Dilkina stated, "If we can implement this in Kenya, it can be adapted elsewhere worldwide—collaboration and commitment are key."

Source: https://medicalxpress.com/news/2025-05-ai-child-malnutrition-efforts.html