An artificial intelligence tool (AI) that predicts acute child malnutrition up to six months earlier could help fight the situation in Kenya and throughout Africa, almost half of the children under the five associated with acute malnutrition.
However, gaps in the data may find it difficult to know where to focus resources in countries such as Kenya.
Five percent of children in Kenya are strongly malnourished, according to Kenya’s demographic survey, a level that is considered to be concerned about public health.
Scientists have come up with a model of mechanical learning that uses clinical health data and satellite images to predict malnutrition trends across the country.
The tool was developed by a team from the University of Southern California (USC), in collaboration with Microsoft’s AI for Good Research Lab, Amref Health Africa and the Kenyan Ministry of Health.
Lead researcher Laura Ferguson, Research Director at the USC Institute for global health inequalities, says the goal is to equip health authorities with timely warnings that support effective prevention and treatment answers.
“The tool is designed to predict malnutrition in all counties in Kenya [and]… Prepare prevention and treatment strategies“Ferguson said Scidev.net.
To make these forecasts, the model pulls data from the government’s health software (DHIS2) and combines it with satellite images to identify where and when malnutrition is likely to occur.
Unlike traditional models that depend solely on historical trends, this AI tool incorporates clinical data from more than 17,000 Kenya health facilities.
It was accurately 89 % for one month forecasts and 86 % accuracy in six months, marking a significant improvement over basic models.
The tool can also incorporate available in common data on agricultural vegetation derived from satellite images to the model to indicate available food sources, Ferguson added.
Encouraging by the results in Kenya, researchers hope that the tool can be adapted for use in about 125 other countries that also use DHIS2- especially in the 80th and medium-income nations where malnutrition remains the main cause of child mortality.
“This model is a game-changer“, said Bistra Dilkina, Associate Professor of Computer Science and co-director of the USC Center for AI in society.
“Using data -based AI models, you can record more complex relationships between multiple variables that work together to help us predict malnutrition more accurately,“He explained.
To maximize the impact of the tool, the cooperation between the sectors is the key, says Samuel Mburu, head of the digital transformation in Amref Health Africa, who also worked on the project. Proposes to align health services with agriculture and disaster management efforts.
“Ongoing investment in digital infrastructure and training are also critical“Mburu said Scidev.net.
Peter ofware, a Kenyan country director for Helen Keller International, a non -profit -based US -based US -based Nutrition and Health Company, agrees that integrating vegetation data with DHIS2 improves prediction accuracy.
“This improves the accuracy of forecasts“said ofware, who did not participate in the research.
“However, DHIS data, which is their main source, have many restrictions on quality – especially for malnutrition.“
Children are usually examined only for malnutrition in facilities where there is treatment, which limits how representative the data is, he added.
