In case you think that, by avoiding more active engagement in social media, adverse event reporting isn’t going to catch up with you – think again.
The Wall Street Journal reports that, “Scientists are sifting through massive quantities of freely available data scattered across the Internet, aiming to catch potentially deadly problems with prescription drugs more quickly—even ahead of federal regulators.”
Researchers from the University of Virginia and West Virginia University have developed mathematical recipes that computers use to filter billions of pieces of data from patient comments in online chats, websites and news stories to detect serious adverse drug reactions.
They then catalog the complaints and determine when they rise to a level that deserves medical or regulatory attention. Similar techniques have been used to detect bioterrorism threats, study crowd and consumer behavior and map out how infections spread.
The goal is to be able to alert the medical community about possible negative drug reactions much earlier so they can use it their clinical practices, as well as to warn the FDA about potential new adverse reactions.
The challenge for researchers, they say, is to sort through tons of "noise," the reams of information that aren't relevant, accurate or important, and to recognize useful signals. Researchers hope to use continually improving computer algorithms—programs that detect key patterns or relationships between words—to make recognition of important signals more accurate, says Ahmed Abbasi, a professor of information technology at the University of Virginia's McIntire School of Commerce in Charlottesville.
In preliminary data, they found that 80% of the time, their formulas detected potential adverse event patterns three months earlier than the FDA issued warnings, said Dr. Adjeroh, a computer science professor at West Virginia University in Morgantown. In some cases, they were years ahead of the FDA's warnings. The researchers were recently awarded a $130,000 grant from the National Science Foundation Smart Health and Wellbeing program to launch a bigger project in this area.
Such big-data approaches move away from a reliance on voluntary reporting and clinical intuition. "There's always been this struggle between intuition and data," says Dr. Abbasi. "There's a paradigm shift where data-driven decision-making is being increasingly adopted."
Note to industry: Lead, Follow, or Get Out of the Way.
Suggestion to industry: Lead.
