When FiveThirtyEight's political minesweeper Nate Silver analyzes the data from Google Flu Trends, there are bound to be some notable results.
We're previously written about how Google Flu Trends uses flu-related search data to estimate the number of infections in different geographical areas. Rather than just linking to the regular Google Flu Trends map, though, Silver decided to plug the publicly available data into a spreadsheet and create his own map.
The result shows when each state hit 5,000 on the Google index, a figure Silver says would correspond to a bad seasonal flu in a normal year. Only two states had yet to reach the 5,000 mark: Florida and Utah.
Now, as we've detailed in earlier posts, Google Flu Trends is not the most scientific flu surveillance method out there. But other flu surveillance methods, including the International Society for Disease Surveillance's DiSTRIBuTE program, also suggest that Florida has not been as affected by swine flu as have other states. Why is this?
Silver speculates that Florida's large number of senior citizens may be helping reduce the number of H1N1 infections in the state. (Research suggests people who lived through other H1N1 flus between 1918 and 1957 may be less susceptible to the current H1N1 virus.) Meanwhile, a recent article in the Miami Herald suggested Florida's humid weather may be keeping the virus at bay and that the flu could spread more widely in the winter months. Is that really what's protecting Florida from swine flu? And what will happen, Silver asks, when the 2009-2010 flu season peaks in January?