Prediction model combines Google Flu Trends, weather forecasting techniques
A model using data from Google Flu Trends and weather forecasting techniques could predict the peak of flu outbreaks in specific areas more than seven weeks in advance, according to researchers from Columbia University and the National Center for Atmospheric Research.
Their work was published Monday at Proceedings of the National Academy of Sciences.
Previously, researchers in Pakistan asserted that while Google Flu Trends serves as a good "baseline indicator" of epidemic trends, it could become an effective early warning system through application of "sophisticated statistical analysis."
The new model takes near-real-time data about flu-related searches from Google and applies weather-forecasting techniques to make predictions based on past flu trends. It can provide "a window into what can happen week to week as flu prevalence rises and falls," Jeffrey Shaman, Ph.D., an assistant professor of Environmental Health Sciences at Columbia University's Mailman School of Public Health, said in an announcement.
They created estimates of flu trends in New York City based on the Google data from 2003 to 2008 and concluded that in the future, the prediction model could make flu forecasts for specific areas as common as those for the weather.
Because the study made predictions for flu outbreaks within various areas of New York City, the researchers could test the degree of certainty for their predictions, just as a weather forecaster can say there's a 70 percent chance of rain, the Los Angeles Times points out.
Knowing about an upcoming outbreak could lead people to perhaps get a flu shot or be more careful around others who are coughing and sneezing. It could help healthcare providers to more accurately stock vaccines and could provide warning to others, such as when schools might need to close.
The researchers expect others to build upon and refine their model. They plan to test it various locales, because, as Shaman said, "there is no guarantee that just because the method works in New York it will work in Miami."
Elsewhere, researchers from the University of Rochester in New York have turned to Twitter to make flu predictions. They analyzed 4.4 million tweets over one month in 2010 that were tagged with GPS location data from more than 630,000 users in New York City to predict flu outbreaks up to eight days beforehand.
Meanwhile, researchers at the University of Pittsburgh put sensors on children to study their interaction patterns in an effort to develop strategies for stopping flu outbreaks.