28 Nov Researchers develop new method for forecasting flu outbreaks
By Jodie Tillman, Times Staff Writer
Scientists have developed a computer model for predicting flu outbreaks weeks in advance, raising the tantalizing possibility of “flu forecasts” that might one day help guide such decisions as when to ramp up vaccine production or close schools.
The new flu model, described in a study published this week in theProceedings of the National Academy of Sciences, incorporates techniques used in weather prediction to forecast flu outbreaks up to seven weeks in advance.
Researchers at Columbia University and the National Center for Atmospheric Research focused their study on the winters between 2003 and 2009 in New York City.
The formula used data from the Google Flu Trends project, which estimates outbreaks based on the number of flu-related search queries in a given region. The model also incorporated findings from a previous study that found wintertime U.S. flu epidemics tended to occur following very dry weather.
Scientists say the model represents a first step in flu forecasting and would need to be adjusted based on such factors as geography — peak flu season varies from region to region — and the strain of influenza.