An article in today’s New York Time’s announces the release of Google Flu Trends, a tool developed by the philanthropy wing Google.org. Google Flu Trends tracks a specified set of search terms for “ebbs and flows”, broken down geographically by regions and states. Monitored search terms include “flu symptoms” and “muscle aches” (A complete list of monitored terms does not seem to be publicly available, perhaps because of the possibility of manipulation). At Google.org, a new page explains how the Flu Trends system works and provides a compelling comparison of Google’s data fluctuations with data on “influenza-like illness” collected by CDC. CDC’s data is collected from labs, health care providers, death certificates; according to Google.org, last year a prototype version of Flu Trends detected the onset of flu season (“a significant ) two weeks before “published” CDC reports. The “published” there might indicate that CDC had awareness earlier than they released it to the public, but it does seem that Flu Trends is real-time while CDC ILI surveillance is not. In addition, Google.org presents an argument about the publicity of this information: not only do they argue that Flu Trends will help public health departments, they also claim it could be used by individuals to help make decisions about managing their own health.

However, this openness argument is interestingly contested in the Times article by Farzad Mostashari, Director of Epidemiology at New York City and one of the early developers of syndromic surveillance. Mostashari criticizes Google.org precisely for its lack of openness: “If Google provided health officials with details of the inner workings of the system so that it could be validated scientifically, the data could serve as an additional way to detect influenza that was free and might prove valuable, said Mr. Mostashari.” Interestingly, this is precisely the criticism Mostashari and others levelled at Federal syndromic programs like BioSense: without providing access to the analytic algorithms, the parts of the system that determine and define what constitutes an epidemic event, it is difficult for the presented information to be taken up in public health practice.