How Will Disease Spread?
Cell phone study offers glimpse into how people -- and illnesses -- move.
June 5, 2008 — -- In a study designed to track how large-scale disease might spread, researchers have made a not-so-surprising discovery: Most people are slaves to routine and tend to stay close to home.
Over the course of six months, three Northeastern University researchers tracked cell phone users in Europe in search of information on human travel patterns that could be used in disease epidemic prevention and urban planning. Epidemiologists have looked to human mobility studies as a way to predict how an epidemic might spread as a population moves around.
"Most individuals spend most of their time in a very restricted area, while some individuals spend most of their time over a very large area," said Cesar Hidalgo, physics researcher at the university's Center for Complex Network Research and co-author of the study that was published in the journal Nature Wednesday.
With the help of a European cell phone company in an unnamed country, Hidalgo, along with co-authors Marta Gonzalez and Albert-Laszlo Barabasi, followed the cell phone usage patterns of 100,000 anonymous users over six months by monitoring the cell towers that route calls and text messages. Researchers estimated the approximate location of the user according to which cell phone tower handled the call.
Additionally, for one week, the researchers followed a group of 206 cell phone users equipped with programs that allowed their location to be recorded every two hours by their cell phone company. The numbers were disguised with what Hidalgo called "ugly code" – a long series of letters and numbers unique to each individual.
The smaller group was studied "to check [that] the results we were getting were accurate," Hidalgo said.
According to Hidalgo, over the six months, 73 percent of people spent most of their time within the same 10-mile radius. Between 2 and 3 percent of people moved regularly over several hundred miles.
Researchers could also reliably predict which users would be in the high-traveling group or the stay-near-home group after viewing their activity for four or five days.