The power of the Internet as a real-time international information network has long excited researchers looking to track disease outbreaks. The challenge with fast-progressing outbreaks of any disease, such as influenza, is that parameters of where it is spreading and how fast are often not available until well after the outbreak has actually occurred.
For several years, my wife, an Infectious Disease specialist, participated in a flu prediction market. The market was run as a research project using the Iowa Electronic Markets, a real-money futures market. Participants were invited based on their professional role as someone likely to have early information about the progression of the flu. They attempted to predict flu severity by providing participants with financial rewards for guessing correctly.
Google is taking this idea one step further, disintermediating the process by removing the health professional and going direct to individuals nationwide. As health consumers turn overwhelmingly to the Web as a first stop for health information, Google looks for search patterns and extrapolates a likely spread of the disease. By watching for a set of terms related to the flu, Google Flu Trends maps out current flu severity with a timeliness never before possible.
It’s one of several efforts by Google to begin to use the notion of the wisdom of crowds to provide new services (along with other efforts such as Google Searchwiki). In this case, mining search data in an intelligent way creates an entirely unique data product with unique value. I presume that they could do the same thing to track other diseases, consumer trends, terrorist activity…the possibilities are endless!
The challenge here is there is no expectation that search queries represents wisdom being provided to a crowd. Searchers typically assume a high level of anonymity when using Google. It is unclear if that level of trust with the tool is earned.
Such concerns prompted some pushback from privacy advocates recently as reported in iHealthBeat.
Which leads to the question: when does the value of this sort of transparency outweigh the risk to privacy? Would the answer be different if Google was aggregating Google Health data to provide such an analysis? Would the answer be different if Google were tracking Avian influenza rather than the more typical, generally non-lethal varieties?
None of us will ever know with certainty what data is or is not part of the analysis for Google Flu Tracker and who works with that data. What is clear is that Google will continue to face tight scrutiny for each new data product it makes available.