Virginia Tech flu forecasting technology to be used by AccuWeather

Researchers create a synthetic population to predict the real impact of the flu

BLACKSBURG, Va. – A forecasting technology created by researchers at Virginia Tech could warn people days and weeks before a flu outbreak hits their area.

The tool will soon be available to the public as researchers are teaming up with AccuWeather to make the flu forecast a part of the weather app. While it's still in the testing phases, by early next year the tool will be accessible to the entire country -- letting people know when the flu is expected to hit their area and how bad the outbreak is expected to be.

Technology behind the flu forecast was originally developed by researchers at the Biocomplexity Institute of Virginia Tech in 2014 to help track the Ebola outbreak in West Africa. 

The disease surveillance tool is already being used by federal, state and local governments to track and predict the impact of epidemics. While researchers often work with the government or government entities, teaming up with a private company like AccuWeather is something new for this project.

"For the first time we can see our work reaching out to the masses through these very popular apps and interfaces," says Srini Venkatramanan, a postdoctoral associate at the Biocomplexity Institute. "It's a way of keeping us on track. When people start looking at it, we have our name on the line and can keep improving our forecasts and get an idea of how much impact it's have on society."

The partnership between the Biocomplexity Institute and AccuWeather is expected to have a big effect on the level of flu knowledge and awareness nationwide. The research will give people more time to prepare once they find a major flu outbreak is headed their way, allowing users to get a flu shot or take other precautions to avoid getting sick.

Unlike weather forecasts, which have become accepted in our everyday lives, forecasts for the flu or other diseases haven't been as accessible. When companies or the government do release a flu forecast, it tends to be pretty broad.

Researchers at Virginia Tech are working to update the AccuWeather flu forecaster on a weekly basis. They'll use confirmed cases reported to the Centers for Disease Control and Prevention, combined with social media posts and Google searches related to the flu.

One of the most complex aspects to this flu forecasting technology is the ability to create synthetic populations.

In simple terms, researchers are able to use computers at the Biocomplexity Institute to set up a population similar to what actually exists in the United States. They will use that population to model real behaviors, predicting the response that each person will have to the outbreak based on where they live, travel and commute.

"Private behavior [modeled in the technology] is people socially distancing themselves from people that are infected," says Venkatramanan. "People that are infected don't go to workplaces or don't go to schools."

The computers model the impact on the spread of the flu that having those people stay home or having a caregiver take care of a sick loved one.

Researchers can also model government and business reactions as well. They are able to look at the real impact a flu vaccine campaign could have or look at what would happen if the CDC rolled out a second, more effective vaccination later in the flu season. Researchers are able to compile all of that information and narrow it down to see the impact the flu outbreak would have on each person created in the model.

The flu forecaster is expected to be available on the app in March.