There’s a scene in an old 30 Rock episode where we see Liz Lemon’s most recent Google searches: “singles yoga” and “scalp pain.” Hilarious, yes, but the moment also illustrated how often we appeal to the Internet with nebulous health concerns like “bleary vision,” “weird snoring,” and yup, “scalp pain.”
This behavior is so ubiquitous, a group of scientists at Microsoft, Stanford, and Columbia University are examining search data as a way to identify drug interactions and side effects that doctors might miss. In a study called “Web-Scale Pharmacovigilance: Listening to Signals from the Crowd, the scientists proved that by analyzing several million search queries, they could predict that two medications (an antidepressant and a cholesterol drug) would cause harmful side effects.
Why does this matter? For one thing, it’s a huge improvement on the existing infrastructure for identifying side effects. According to the New York Times, “the F.D.A. asks physicians to report side effects through a system known as the Adverse Event Reporting System. But its scope is limited by the fact that data is generated only when a physician notices something and reports it.”
And even if a doctor does take notice, it might not change what’s being prescribed for years, which means that potentially millions of patients are being affected by unreported side effects. Multiply that by thousands of drugs being prescribed in millions of different combinations, and you’ve got an impossibly complex network of interactions. Most patients will rarely go so far as to report a side effect, but many will search the Internet for advice on their symptoms, creating a massive reservoir of data that contains clues about these conditions. The Times explains:
The researchers said they were surprised by the strength of the ‘signal’ that they detected in the searches and argued that it would be a valuable tool for the F.D.A. to add to its current system for tracking adverse effects. ‘There is a potential public health benefit in listening to such signals,’ they wrote in the paper, ‘and integrating them with other sources of information.’
The researchers said that they were now thinking about how to add new sources of information, like behavioral data and information from social media sources. The challenge, they noted, was to integrate new sources of data while protecting individual privacy.
‘I think there are tons of drug-drug interactions–that’s the bad news,’ Dr. Altman said. ‘The good news is we also have ways to evaluate the public health impact.
Of course, they aren’t the first to use Internet search data to quantify information about health. At the beginning of the flu season, we wrote about the incredible prescience of Google Flu Trends, which is able to predict when flu outbreaks will hit far more quickly than the CDC, simply by measuring how often people searched for “flu symptoms” and the like. Interestingly, though, media hype can skew the data. When Flu Trends sparked a flurry of blog posts, the map spiked a false positive. And as scientists (as well as the curious public) begin searching for news about particular drugs or interactions, they could–at least theoretically–funk up the data in a similar way.
Check out the full story here.