Alright, google is entering the medical profession. No. But there is a group within google and CDC that are attempting to predict how bad the flue season will be or rathere follow it.

It does make sense. Consider the search terms used. If one examines google trends for certain topics, you find that human behavior can be reasonably predictable.

Take a search term like Jesus Christ, it peaks twice in the year. Easter and Christmas. The super bowl starts climbing for January. Valentines too. So now if people are starting to suffer from the flu.

• Fever (usually high)
• Chills
• Headache
• Tiredness (can be extreme)
• Cough usually dry
• Sore throat
• Runny or stuffy nose usually common in children
• Body/Muscle aches
• Diarrhea, nausea and vomiting (more common among children than adults)

Besides the treatments and vaccinations being sought for. The symptons appearing first, the concerned parent or ill adult may start searching to see what they have before going to the doctor where it will get documented, and eventually reported to the CDC.

So on that note Google search terms could be an indicator of a rise in flu incidents. Odds are you will be. So lets take a look at some of the Google's findings.

You can read more about google's flu trends at their flu site or if you really into the researh part their is a paper on it too. download pdf here or you can go to Nature's web site for the reseach on the article to find a faster way of detecting the oncoming flu assault.

Now using search engines to correlate to flu isn't new.  Maybe new to us, but it has been done in the past, but not as extensive as this one on google's logs. 

CDC's historical data is available from US Influenza Sentinel Provider Surveillance Network (http://www.cdc.gov/flu/weekly).  This was used for the CDC portion of the charts. 

The problem with the CDC data base is it lags about 2 weeks in real time from what is occurring now.  So when the flu season starts to really heat up, they may still reporting everything is OK.  Using search engines as a way of early detection is, more timely.  Here we they are using in their terms "the collective intelligence of millions of users".  It provides a broader reaching and way of monitoring influenza.  It will be used for the 2008-09 flue season.

Now their privacy disclaimer.  None of the queries in the Google data base for this project can be associated  with a particular individual.  For more about this you can read Google's privacy policy.  at (http://www.google.com/privacypolicy.html).

There are other health related searches that can be performed as indicators on the general health of our population.  Search terms related to allergy, and sunburn.  As expected during those seasons when they occur the searches go up.

A list of queries they use are below to establish the Google patterns. 

Table 1

Topics found in search queries which were found to be most correlated with CDC ILI data

Search query topic------------------------------------ Top 45 queries --Next 55 queries

---------------------------------------------------------------n -Weighted-----n-Weighted

• Influenza complication ----------------------------11 ---18.15 ------5 ----3.40
• Cold/flu remedy ------------------------------------8 -----5.05 -------6 ----5.03
• General influenza symptoms --------------------5 -----2.60------- 1 ----0.07
• Term for influenza ---------------------------------4 -----3.74 -------6 ----0.30
• Specific influenza symptom ----------------------4 -----2.54 -------6 ----3.74
• Symptoms of an influenza complication -------4 -----2.21 -------2 ----0.92
• Antibiotic medication ------------------------------3 -----6.23 -------3 ----3.17
• General influenza remedies--------------------- 2 -----0.18 -------1 ----0.32
• Symptoms of a related disease -----------------2 -----1.66------- 2 ----0.77
• Antiviral medication -------------------------------1 -----0.39 -------1 ----0.74
• Related disease -----------------------------------1 -----6.66 -------3 ----3.77
• Unrelated to influenza ----------------------------0 -----0.00 -----19 ----28.37
• Total -----------------------------------------------45 -----49.40 ----55 ----50.60

The top 45 queries were used in our final model; the next 55 queries are presented for comparison purposes. The number of queries in each topic is indicated, as well as queryvolume-