Flu Tracking Highlights Google Search Analysis as a Tool for Medicine
According to the Google search engine, one of the most popular health stories in the past week has been on the use of Google’s own search analysis capabilities.
It seems that over the course of recent flu seasons, searches on Google for “flu symptoms” and related terms mirror actual flu cases reported by the US Centers for Disease Control and Prevention (CDC). In fact, for some regions, the retrospective tracking of Google searches pointed to jumps in flu-related searches before the CDC had finished compiling data and reporting on flu trends for the same regions. This suggests that Google search analysis might serve as a sentinel for the earliest detection of flu outbreaks in some regions.
Population centers generally have their own sentinel laboratories, which are major sources for flu data used by the CDC. “In our locale, with our clinical lab, we get as early a warning as anyone in our country, and we know exactly when it is influenza and not another respiratory infection,” says Lawrence Drew, MD, PhD, director of the clinical virology laboratory at UCSF. “We would not be advantaged by utilizing this tool. It could help people in public health get a handle on flu activity in other parts of the country.”
There are many flu cases and plenty of people searching online using flu-related terms, which might provide a critical mass needed for effective and accurate analysis. Would the same strategy work to identify other disease trends?
What about those drugs that get pulled from the market after Food and Drug Administration approval? Could an analysis of Google searches combining drugs and side effects provide an early heads-up for dangerous side effects that are too rarely reported by patients and physicians?
Without more information, it’s tough to speculate on how useful to medicine the search analysis approach may become. The press office for the peer-reviewed scientific journal Nature confirms that the report has been accepted for publication, but the report has not yet been made available.
Sentinel Influenza Lab
The UCSF clinical virology lab is one of California’s sentinel laboratories. These labs provide data on confirmed influenza and other respiratory viruses. Most of the cases reported by the UCSF lab are detected in UCSF clinic patients or in patients admitted to the hospital. Classifying flu infections as influenza A or influenza B is routine at the sentinel labs. At UCSF’s lab, this is done using a fluorescent antibody test. However, tracking drug resistance and individual flu strains is a task generally left to the CDC.
Analysis of Google search trends, of course, cannot be used to identify flu types. And what about outbreaks of flu-like illnesses that are not the flu?
A few weeks ago, I developed a high fever, chills, sweats, deep muscle aches and headaches — a few weeks after having a flu shot. Incredulous, I nonetheless Google-searched “flu symptoms” and “muscle aches” — and I never saw a doctor.
Multiply this experience among multitudes across the country, and it would seem to illustrate the power of analyzing Google searches to find disease outbreaks, even among people who do not come to the attention of medical care providers.
However, my case may illustrate a different problem. I might not have had influenza at all, Drew suggests. It’s more likely that I had some kind of parainfluenza virus infection or something else entirely.
“The Google approach could lead to a false result when another disease has symptoms similar to influenza,” Drew says.
Seasonal flu tracks a course that ebbs and flows with striking regularity from year to year. It is conceivable that a significant increase in Google searches on particular symptoms at unexpected time points during the yearly flu cycle might signal that some unusual pathogen is on the loose.
What else is on the horizon in the realm of sentinel disease detection? The capabilities for sequencing the genes of pathogens and providing positive identification is expanding exponentially — and becoming much faster and more comprehensive. Maybe before too long, we’ll have better lab tests. Stay tuned.