The Food and Drug Administration is launching full bore into new data mining techniques as evidenced by three solicitation documents posted recently.
The agency posted a sources sought document on Tuesday seeking a vendor that could crawl through more than 20 million biomedical journal abstracts and citations housed on a National Library of Medicine database to uncover drugs that are disproportionately associated with “adverse events.”
FDA also solicited for a separate data mining program on Monday within its Office of Clinical Pharmacology. That solicitation is aimed at uncovering “safety signals” associated with FDA-approved drugs or drugs the FDA is considering approving. A safety signal is any event, positive or negative, that sometimes accompanies taking a new drug. The solicitation doesn’t describe what data sources would be mined.
In late March, FDA’s Office of Criminal Investigation began soliciting for a program to monitor the Web for illegal sales of drugs, medical devices, cosmetics and veterinary products as well as counterfeit food and adulterated or misbranded vaccines.
FDA is among many government agencies that have been pressing recently to use the troves of new data available on the Web and related technology to make better predictions and provide better services.