Researchers from the University of California San Diego have found a way to predict how the flu will spread a week in advance with the same level of accuracy Google Flu Trends is displaying right now.
In a study published in the journal Scientific Reports, researchers combined the use of traditional flu monitoring information from the U.S. Centers for Disease Control and Prevention and big data from Google Flu Trends and analyzed social networks to come up with a means of predicting how the flu season would progress in a week's time.
Michael Davidson, a political science doctoral student and corresponding author for the study, explained that their research resulted into an innovation that involved constructing a network tying together health regions in the United States guided by data from the CDC. By analyzing which regions reported flu cases around the same time last season, the researchers were able to pinpoint which areas had the strongest connections, giving them analytic power that could greatly improve how Google predicts flu trends.
Google Flu Trends is very good at showing in which parts of the country people are looking up information on the flu and its related symptoms, all valuable because of the real-time component, but sometimes appear inflated because it shows flu concerns more widespread than confirmed cases.
By comparing the network that researchers were able to derive from DCD data with projections from Google Flu Trends, predictions should be refined and improved.
By showing that accuracy can be improved when predicting flu patterns are involved, researchers are hopeful that their work will be utilized to benefit the public, and be used by data scientists and epidemiologists for better treatment efforts and target prevention, most especially when epidemics are in the picture.
The study received funding support from the James S. McDonnell Foundation and the Robert Wood Johnson Foundation. Co-authors for the study include: Dotan Haim, also a political science graduate student, and Jennifer Radin, a part of the joint doctoral program between UC San Diego and San Diego State University.
Every year, up to 20 percent of the U.S. population are affected by seasonal influenza, resulting in more than 200,000 hospitalizations. The current season is about halfway through but the current flu vaccine just around 23 percent effective against the prevailing strain. This poses problems as the flu vaccine is the top line of defense people have against the virus, although a vaccine has never been expected to provide 100 percent protection against the flu.
