Google flu tracker faces hurdles

Google Flu Trends, a tracking system based on research by Google and the Centers for Disease Control and Prevention, drastically overestimated peak flu levels this season, demonstrating the continued importance of traditional surveillance networks.

Since its creation in 2008, Google Flu Trends' estimates have almost exactly matched the CDC's own surveillance data over time, but the latest U.S. flu season confused the algorithms of the system. It estimated almost double the CDC's national peak of flu for Christmas with some even larger discrepancies at the state level, Nature reports.

Several researchers suggested that widespread media coverage of this year's severe U.S. flu season may have triggered flu-related searches by people who were not sick.

"You need to be constantly adapting these models, they don't work in a vacuum," John Brownstein, an epidemiologist at Harvard Medical School in Boston, said, according to Nature. "You need to recalibrate them every year."

Disease experts said that the overestimate is a reminder that new systems work as complements but not substitutes for traditional epidemiological surveillance networks.

"It is hard to think today that one can provide disease surveillance without existing systems," Alain-Jacques Valleron, the founder of France's Sentinelles monitoring network, said, according to Nature. "The new systems depend too much on old existing ones to be able to live without them."

Despite the setback, crowdsourced and data mining tracking systems are becoming a part of the norm for flu surveillance, even for people in the field, Nature reports.

"I'm in charge of flu surveillance in the United States and I look at Google Flu Trends and Flu Near You (another tracking system) all the time, in addition to looking at U.S.-supported surveillance systems," Lyn Finelli, the head of the CDC's Influenza Surveillance and Outbreak Response Team, said, according to Nature. "I want to see what's happening and if there is something that we are missing, or whether there is a signal represented somewhat differently in one of these other systems that I could learn from."