With the influenza outbreak in the U.S. now officially considered an epidemic by the Centers for Disease Control and Prevention, the jury is still out on whether the performance of Google Flu Trends is up to snuff.

Launched in the U.S. in 2008, Google Flu Trends uses aggregated web search data to estimate flu activity in near real-time. Estimating the start, peak, and duration of each flu season, the company claims that Google search terms are good indicators of flu levels and that their online service is more finely grained geographically and is more immediate—up to 1-2 weeks ahead of traditional methods such as the CDC’s official reports. At the same time, Google Flu Trends emphasizes that it “is not designed to be a replacement for traditional surveillance networks or supplant the need for laboratory-based diagnoses and surveillance."  

However, Google Flu Trends predictions during the H1N1 pandemic in 2009 were off-base and during the 2012/2013 season GFT significantly overestimated the severity of the influenza outbreak compared to the CDC’s reported U.S. flu levels. Google’s estimates in 2012/2013 predicted nearly double the proportion of healthcare visits for influenza-like illnesses reported by the CDC, which bases its estimates on surveillance reports from laboratories across the country.

Ultimately, a company investigation blamed the overly high estimates on the media and its coverage of the severity of the flu season that resulted in an extended period in which users were searching for terms which Google Flu Trends identified as correlated with flu levels. As a result, for the 2013/2014 season, Google launched a retrained model—still using the original method—that more closely approximated CDC data.

More recently, for the 2014/2015 season, Google launched a new Flu Trends model in the U.S. that takes official CDC flu data into account as the flu season progresses. Nonetheless, we’ll have to wait until the end of flu season, when Google evaluates the performance of its models, to get an official evaluation of performance.

A July 2014 study published in the American Journal of Preventive Medicine concluded that Google Flu Trends may be inaccurate, but improved methodologic underpinnings can yield accurate predictions by “revising the inner plumbing” of GFT that can improve the accuracy of forecasts. According to the study, existing shortfalls in the predictions result from the methodologies that GFT employs, not from the data themselves.

Register or login for access to this item and much more

All Health Data Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access