Two countries adopt malaria forecasting system developed at Duke

Two countries along a potentially dangerous Central American corridor for malaria transmission will begin using a forecasting system developed by researchers at the Duke Global Health institute as part of their national efforts to control the disease.

Panama and Honduras will be the first countries to officially adopt the system, which uses real-time climate, health and environmental data to predict when and where malaria outbreaks will occur. The forecasts will help the countries’ health ministries target interventions such as the deployment of rapid diagnostic tests and distribution of medications. 

“The model is trained to predict what variables are most likely to drive malaria cases in a particular area,” says William Pan,  Dr.P.H., a professor of population studies and global environmental health at Duke, who developed the system. “We’re hoping to show that stakeholders can use this information to tailor responses and control future transmission.”

The forecasting system is based on tools Pan and his collaborators have developed over the past decade to assess and predict malaria trends, which have been previously deployed in Peru and other parts of the Amazon. The new model, developed with a grant from the National Aeronautics and Space Administration (NASA), adds several layers, including local malaria case counts and interventions, that refine its ability to make district-level predictions up to nine months into the future. 

“Any time we can be thinking about what’s coming down the pipeline, it’s helpful,” says Justin Lana, Ph.D., an epidemiologist who worked on the project and is with the Clinton Health Access Initiative (CHAI), a nonprofit global health organization that supports programs in more than 35 countries, including malaria control programs in several Central American countries.  “If we can think about what we need to do in the next one, two or three months, we can support the ministries in getting the resources where they need to be and respond pre-emptively.”

Malaria has resurfaced as a significant threat in parts of Central America, particularly in Panama, which saw malaria cases rise from less than 1,000 in 2017 to more than 15,000 in 2024. A majority of those cases occurred in communities in and around the Darien Gap, a heavily forested region along Panama’s border with Colombia that has become a primary route for migrants traveling to the U.S. southern border. 

Countries north of Panama have so far not experienced similar spikes, and both El Salvador and Belize have been declared malaria-free by the World Health Organization since 2021. Nine countries – including Panama and Honduras – have been working to eliminate malaria as part of a regional campaign launched in 2014, which is supported by CHAI and other public and private partners. 

But with the number of migrants traveling through the region now in steep decline, the next few years will be critical to reestablishing Panama’s firewall against malaria spreading further north. “the question now is, how quickly the interventions and programmatic responses will be able to bring those case numbers back down?” says Lana, who was a DGHI doctoral scholar while completing a Ph.D. in environmental health at Duke in 2021. 

It’s likely that Panama’s surge is a result of many factors, including the large number of Panamanians who traveled to the Darien Gap to sell goods and services to migrants, Pan says. His team is working with researchers at Panama’s Instituto de Investigaciones Científicas y Servicios de Alta Tecnología to study whether the increase in cases may be connected to the presence of new variants of the parasite or its mosquito hosts. 

Pan’s forecasting system may be a powerful tool in answering those questions. The models use AI-based learning to analyze more than 3,000 variables that can impact disease spread, including variations in air and ocean temperatures, precipitation patterns and satellite images showing changes in land cover. The forecasts can help health officials tease out where local interventions are having an effect on disease trends, as well as the extent to which outbreaks may be driven by environmental factors beyond their control.

“The ability to bring in climate information to help the Ministries of Health evaluate the impact of our interventions is huge,” says Lana. “Unless you’re able to account for these very complex climate relationships, you’re kind of in the dark about how impactful any particular intervention is.”

But ultimately, Panama’s ability to regain its progress in controlling malaria will depend not on science or data, but sound decisions, says Lana. 

“Fancy dashboards and predictions only drive down cases if the right people are looking at the data and taking the right actions,” he says. “It’s really up to us to ensure the tools Duke’s created are as useful as possible to the health ministries.”