Using technology and big data to predict and prevent the next Ebola outbreak

Andrey Rzhetsky, PhD, (top row, right) with attendees at the Technology 4 Ebola conference in Cairo, December 2014

Andrey Rzhetsky, PhD, (top row, right) with attendees at the Technology 4 Ebola conference in Cairo, December 2014

The Ebola epidemic has disappeared from the headlines in the United States, but it’s far from over in West Africa. The World Health Organization reports that the current epidemic has lasted longer than every previous outbreak, and has accounted for five times as many deaths as all previous outbreaks combined. While new cases seem to be tapering off in Liberia, in the first week of January there were 74 new cases identified in Guinea and 250 in Sierra Leone.

As this article in Vox points out, it will take more than sending doctors to these countries and pouring money into infrastructure to slow transmission. It will take cultural changes, traditional door-to-door contact tracing and education in rural areas to finally quell the epidemic.

In December, the Egyptian government and Microsoft organized the Technology 4 Ebola conference in Cairo to take another approach to fighting Ebola, gathering politicians, public health experts, scientists, economists and activists to think of ways to apply technology to deal with the current crisis, and better prepare for the inevitable future outbreaks.

Andrey Rzhetsky, PhD, professor of genetic medicine and human genetics and principal investigator/director of the Conte Center for Computational Neuropsychiatric Genomics at the University of Chicago, was one of those in attendance. While he is quick to point out that he is not an expert on Ebola or other infectious diseases, the type of work he does analyzing large-scale data sets could help build better epidemiological models and guide the public health response before an outbreak turns into a crisis.

“Ideally we would better understand where these outbreaks happen, why they happen, and learn to predict and warn in advance so we could have those teams on the ground before something horrible happens,” he said.

The problem with this kind of predictive computational modeling is that it relies on tons of data inputs, and data collection in places like Sierra Leone, where cell phone networks and infrastructure are spotty, can be difficult.

One of the recommendations drafted by conference participants was to develop custom-built social technologies that health workers could use on the ground in affected areas to spread information about preventing the spread of the disease. A second focused on using mobile technologies, sensors and cloud-based systems to improve contact tracing and collecting epidemiological data. The third was to develop a shared data platform that could be used by researchers working on those predictive models.

The key to making these recommendations work, Rzhetsky said, is giving both health workers on the ground and researchers from afar as much data to work with as possible.

“Basically you’re trying to collect as much evidence as possible and then piece it together in time and space,” Rzhetsky said. “It’s prudent to start with obvious known factors, but if you’re looking only at known factors, you will never find anything new.”

About Matt Wood (468 Articles)
Matt Wood is a senior science writer for the University of Chicago Medicine and editor of the Science Life blog.
%d bloggers like this: