Google is questioning users to track the levels of interest in searches for information about particular health conditions. It’s an attempt to flesh out a theory that Internet searches can predict illness trends.
Users who search for keywords relating to illness and medicines may, in randomly selected cases, be asked whether their search was prompted by a medical condition (either their own or someone they know.) The question is designed to differentiate between, for example, somebody searching for ‘ibuprofen’ because they are suffering from pain, and somebody researching the term for a college product.
The hope is that this will produce percentages which can be extrapolated to get a better idea of how many people overall are searching for information about a particular illness of symptom at any one time, allowing Google to track significant patterns.
The questions will only be asked of users in the United States. The firm says no personally identifiable data will be kept, and none of the data gathered will be used for advertising purposes.
The project is an extension of Google Flu Trends, a longstanding project to keep track of the level of searches for flu-related information and details about associated symptoms. The firm has tracked such data since 2003 and says there’s a clear match between its findings and the actual levels of flu cases recorded by the Centers for Disease Control and Protection.
Intriguingly, although the data matches up, there is a one to two week lag between Google’s figures and those of the CDCP. One possible explanation is that users are searching for information about the symptoms before the infection develops to the point that they are diagnosed with flu. If that proves the case, the Google data could potentially be used to predict likely outbreaks, particularly on a regional basis.