Cellphone data helps researchers target likely spots for malaria control

October 11, 2012

GAINESVILLE, Fla. — Cellphone records could be a valuable tool for controlling and eliminating malaria in sub-Saharan Africa, according to a new study co-authored by a University of Florida researcher.

Mosquitoes that carry malaria have a limited flight range, but that doesn’t stop the disease from traveling long distance. Humans infected with the disease can carry it anywhere a car or plane can reach. That makes eliminating the disease challenging, especially when limited resources for health care and mosquito control are available to cover a large geographic region. In Kenya, a team of researchers has shown how cellphone records can be used to identify which regions should be targeted first to maximize the benefit of malaria control and elimination efforts.

The study appears in the Oct. 12 issue of the journal Science.

“Over 30 countries around the world have stated a national goal of eliminating malaria,” said Andy Tatem, an associate professor of geography at UF’s Emerging Pathogens Institute and co-author of the study. “But it’s difficult to eliminate the disease when new cases are constantly being imported.”

It is also impossible to pin down how a disease is spreading without accurate information about where people live.

“Figuring out where people live sounds trivial, but it’s actually a very difficult thing to do in sub-Saharan Africa,” said Caroline Buckee, assistant professor of epidemiology at Harvard School of Public Health and the senior author of the study.

Tatem was able to provide the needed population maps through his AfriPop Project that uses satellite imagery, census data and land cover maps to create detailed population distribution maps of sub-Saharan Africa. The team then used records provided by a Kenyan cellphone company to identify popular travel routes between population centers. The records gave the team a year’s worth of data on 15 million anonymous cellphone users’ movements around Kenya.

“Researchers have used GPS trackers, surveys and traffic flow on highways to try to understand how people are moving, but that gave us information about a few hundred people at best,” Buckee said. “Using cellphone records gave us billions of data points.”

The next step was to apply the population and movement information to a simple malaria transmission model that predicts risks of infection using probability mathematics.

The result was a new map that shows how malaria is most likely to move between different regions in Kenya and which locations, if targeted for malaria control or elimination, would yield the biggest benefit nationally.

“Malaria control programs have very effective tools today to prevent transmission, but unfortunately, resources for implementing them are quite constrained,” said Justin Cohen, the senior technical adviser from the Clinton Health Access Initiative Malaria Control Team.

“The technique used in this study gives us a way to optimize the impact of our limited resources,” he said.