Whenever he can, Jan Medlock relies on official sources. He gathers data on infectious disease from the World Health Organization, the U.S. Centers for Disease Control and Prevention (CDC) and government health ministries. But last summer, the assistant professor in the College of Veterinary Medicine at Oregon State found himself turning to the fount of family photos and puppy videos: Facebook. In Liberia, the Ministry of Health and Social Welfare was reporting Ebola cases on the global social network.
With colleagues at Yale University, Medlock assembled data from Facebook posts, the 2008 Liberian housing census and the previous Ebola outbreak in the Democratic Republic of the Congo. The researchers scoured medical literature for data on transmission factors such as how many people, on average, are likely to be infected by one sick person. Finally, they calculated the risks of catching Ebola in three different circumstances: communities, hospitals and funerals.
The findings were stark. The study found that a healthy person was about 15,000 times more likely to catch Ebola at a funeral than in the general population, says Medlock.
“To stem Ebola transmission in Liberia,” the scientists wrote in the journal Science, “it is imperative to simultaneously restrict traditional burials, which are effectively serving as super-spreader events.”
Numbers are Medlock’s stock in trade. He uses them to create models — sets of mathematical equations — that inform health-care providers about how to get the most benefit from limited budgets. He has looked at flu, West Nile virus and dengue fever.
Medlock sees a need for household level surveys to increase the accuracy of future disease projections. “Our tools fit the general patterns of outbreaks well,” he says, “but they are quite bad at getting the numbers right.”