The value of predictions stems not only from the early indications it can provide about the nature and the timing of an event to occur. Satisfying predictions should also present the time in which the event is expected to pass.
As it seems, Google Flu Trends can provide an early-warning system for outbreaks of influenza but can it provide signs of its upcoming decline?
Looking at Google’s raw data provided on Mexico – the search for flu related information peaked at January 18, and since shows a constant decline. Interestingly, while the world was hysterically counting cases of swine flu – Google’s “warning system” indicated a relief.
Can web measures assist in real time decision making?
Can search volume decline be counted as signs for risk reduction of an epidemic?
Should governments rely on such measures to prepare but also withdraw from unnecessary actions as those taken during the 1976 flu?
– February 1976:
“There is evidence there will be a major flu epidemic this coming fall. The indication is that we will see a return of the 1918 flu virus that is the most virulent form of the flu. In 1918 a half million Americans died. The projections are that this virus will kill one million Americans in 1976”
(– F. David Matthews, secretary of health, education, and welfare)
– On Oct. 1, 1976, the immunization program began.
– By Oct. 11, approximately 40 million people had received swine flu immunizations.
– The Great Swine Flu Epidemic of 1976 never took place.
And last – will we be able receive true indications to the flu suspected to return on winter 2009?