Big data predicted the coronavirus outbreak and where it would spread
Big data predicted the coronavirus outbreak and where it would spread
Three weeks before the Chinese government announced travel restrictions on the coronavirus epicenter of Wuhan, medical data nerds were already detecting signs of an outbreak. Employees at the Canadian startup BlueDot found their computers picked up early warnings from online chatter across China.
“Chinese language newspapers had reported on some sick people who may have been associated with a seafood market in Wuhan,” Isaac Bogoch, a BlueDot consultant, physician and clinical investigator at the University of Toronto, said. “And then public health officials discussing meetings between hospitals.”
BlueDot notified the World Health Organization, which issued an early virus alert.
People like Bogoch feed computers all kinds of data — animal disease information, airline schedules, web articles — find patterns of outbreaks and where they might spread. Even before this coronavirus showed up outside China, Bogoch had co-authored a short paper predicting it might go first to Thailand or Japan.
On Jan. 17, just days later, came this announcement from Nancy Messonnier, director of the National Center for Immunization and Respiratory Diseases at the Centers for Disease Control and Prevention:
“Three cases outside of China have been identified, two in Thailand one in Japan, all travelers from Wuhan.”
Bogoch and his team got it exactly right. How? First, they ran the numbers on historical travel patterns.
“We looked at commercial flight data, just to see where people from Wuhan were going to,” Bogoch said. “In addition to that, we looked at the destinations, you know, what is the capacity of a country to handle an infectious diseases threat?”
We looked at commercial flight data, just to see where people from Wuhan were going.
Isaac Bogoch, physician and clinical investigator, University of Toronto
By his metrics, countries with less developed public health systems have a higher score in terms of epidemic risk.
Employing artificial intelligence as an early warning system for global outbreaks isn’t just for government. Airlines, hotels and cruise lines stand to lose millions of dollars if people stop traveling, and many companies purchase data and analysis from AI firms to assess financial risk.
The AI firm Metabiota assesses a disease — its symptoms, death rate and availability of vaccine — and then surveys people on how much that disease scares them. It found this coronavirus has a high “scariness” index.
“SARS, when that hit China and caused a 30 to 40% loss in revenue, it’s actually a very similar score to what we’re seeing to the current coronavirus,” Nita Madhav, CEO of Metabiota, said.
SARS … it’s actually a very similar score to what we’re seeing in the current coronavirus.
Nita Madhov, CEO, Metabiota
Like BlueDot, Metabiota uses AI to forecast epidemics. Sometimes, the data sends a clean and clear signal.
“The most powerful information is when the unofficial sources [and] the official sources are all pointing towards the same sort of picture about what’s going on,” Madhav said.
But data predictions are only as good as, well, the underlying data. And in this outbreak, some experts worry China is underreporting cases, as doctors are too busy for patients just mildly ill.
“You know, the difference between how big this is going to be if there’s 250 cases versus 500 cases versus 1,000 cases,” Eric Lofgren, computational epidemiologist at Washington State University, said. “Those are very different numbers early in an epidemic. And those will change the forecasts pretty dramatically.”
So there’s risk in these prediction models. They could give countries and businesses false confidence that everything is OK or send a false alarm.
For what it’s worth, a new paper co-authored by Bogoch in Toronto suggests the coronavirus could spread next to Turkey, Egypt or New Zealand.
There’s a lot happening in the world. Through it all, Marketplace is here for you.
You rely on Marketplace to break down the world’s events and tell you how it affects you in a fact-based, approachable way. We rely on your financial support to keep making that possible.
Your donation today powers the independent journalism that you rely on. For just $5/month, you can help sustain Marketplace so we can keep reporting on the things that matter to you.
