FLICKR, WOODLEYWONDERWORKSDuring flu season last year, a weather-forecasting-like prediction system for flu infections demonstrated that it could predict US cities’ peak outbreaks with about 63 percent accuracy two to four weeks in advance, and could sometimes accurately forecast rises in flu cases some nine weeks ahead of time, according to a study published this week (December 3) in Nature Communications.
“Having greater advance warning of the timing and intensity of influenza outbreaks could prevent a portion of these influenza infections,” study coauthor Jeffrey Shaman, an assistant professor of environmental health sciences at Columbia University in New York, told LiveScience. For example, public health officials could “determine areas that are in greater need of vaccine supplies, where antiviral drugs should be directed and whether or not school closing is needed in the face of a highly virulent outbreak,” he said.
The system, a previous version of which was tested on flu data in New York City, draws on reports of lab-tested influenza cases from the Centers for Disease Control and Prevention and data on flu-related search queries from Google Flu Trends, along with basic information about how the virus spreads. Beginning in late November last year, Shaman and his colleagues tested the predictor on 108 American cities. They found that the system’s overall accuracy topped 60 percent, and that it worked better for cities with smaller, denser populations. The forecaster also became more accurate later in the season, once it was working with more data. By early spring, it boasted an accuracy of 74 percent.
The researchers told LiveScience that they plan to make the flu-forecasting system freely available to the public online.