Wednesday, November 12, 2008

Google Flu

It looks like someone already beat me to the punch by making a blog post related to the CDC/Google collaboration.

This afternoon I was listening to an interesting discussion about Google Flu http://www.google.org/flutrends/ on NPR's Talk of the Nation. Note: To understand what google did click on "How does this work?" from the above link.

There is no arguing that the correlation between the number of people searching for words related to the Flu on Google and the number of people who actually had the flu during the last 5 years is astonishing but what is the practicality of this data?

Currently, the CDC uses a combination of data sources such as phyisician reports and over-the-counter cold medicine sales to track flu outbreaks. The CDC/Google collaboration is attempting to find an earlier way to detect flu outbreaks, which would be useful for preventing the spread of the outbreak. The prospects of the CDC/Google collaboration are exciting but remember that the data analysis was retrospective. Therefore, what happens when the study becomes prospective? Also, could the data be negatively affected by searches related to other types of flu such as Bird Flu? It seems to me that the best way to collect data would be through an electronic medical records system. Why isn't there a universal electronic medical records system?

0 comments: