How crowdsourcing and supercomputing are helping Sierra Leone combat its Ebola epidemic

It’s trickier than it should be.
It’s trickier than it should be.
Image: Reuters/Christopher Black/WHO/Handout via Reuters
We may earn a commission from links on this page.

For US social media users, it’s been impossible to escape loony Ebola misconceptions (the virus is airborne) and inane preventative measures (e.g. uninstalling Uber, shunning hipster bowlers). But if an Ebola outbreak really does hit the US, this collection of misconceptions could actually prove useful.

“Behavioral adaptation is critical when dealing with Ebola,” says Madhav Marathe, an expert on epidemiology and computing at Virginia Tech. “The first step is to find out the ‘pulse’ of the society, their perception, their willingness to comply with public health measures.”

That’s much tougher in Sierra Leone, where Ebola is now suspected of killing more than 3,000 people. Illiteracy and spotty communications infrastructure make it difficult to know exactly what Sierra Leoneans understand about Ebola—an information void that is key to the mystery of Ebola’s swift spread—and how to stop it.

A new project that harnesses crowdsourcing and supercomputing may help Sierra Leone’s government close that gap. IBM Research and Sierra Leone’s Open Government Initiative (OGI)—in partnership with Airtel, Cambridge University’s Africa’s Voices radio project, and Kenya’s Echo Mobile—have created a new platform that lets the government and citizens communicate via radio, text and voice messaging, and then deploys sophisticated analytics to find useful trends from the data.

The idea is to create an “Africanized solution,” says Dr. Uyi Stewart, chief scientist at IBM Research-Africa. “We’ve got to understand how technology is being consumed,” he says, in order to customize data collection and, ultimately, help the OGI tailor its public health messages.

A sample of the maps generated through the initiative that allow the government to pinpoint what’s needed where.
A sample of the maps generated through the initiative that allow the government to pinpoint what’s needed where.
Image: IBM Research

Take hand-washing, for instance. Before the initiative launched, the Sierra Leone government had been encouraging people to wash their hand with soap. But by using radio broadcasts to query listeners about specific beliefs, the new initiative discovered rampant confusion not about whether soap would work, but about which kinds of soap were effective. So the government changed its PSAs to say that any soap would work. “We just added one word and it was successful,” says Stewart.

To cite another example, the media has chalked up the presence of unburied corpses to cultural taboos against allowing strangers to bury dead relatives. Though there’s some merit to that theory, “it’s not the whole story,” says Stewart.

The data analytics suggested because that of the stigma attached to Ebola, people were waiting until the authorities tested their relatives’ bodies for Ebola before burying them. The problem was much less culture than infrastructure. ”Since the government only can conduct 68 tests a day, this creates a backlog of people waiting to exonerate their dead relative,” says Stewart. As a result, the Sierra Leone government is requesting more testing kits from the international community.

Since the initiative is only five weeks old, it’s hard to tell how it might apply to public health emergencies in other poor countries, says Virginia Tech’s Marathe, though he notes that computing clearly role to play in public health epidemiology. “The basic project that IBM has started can be used to work with citizens in other developing nations as well as response to other diseases,” he says.