Researchers at Microsoft, Stanford University, and Columbia University found evidence of unreported prescription drug side effects before the U.S. Food and Drug Administration's (FDA's) warning system tagged them by analyzing data from Google, Microsoft, and Yahoo search queries by 6 million Internet users.
The researchers' data-mining methods build on work by a Stanford lab investigating automated "drug-drug" interaction discovery using software to sift through the data found in FDA reports. Microsoft researchers developed software for scanning anonymized data gathered from a software toolbar installed in Web browsers by users who allowed the collection of their search histories. They were able to probe 82 million individual searches for drug, symptom, and condition information.
The researchers determined that the interaction of paroxetine and pravastatin probably led to hyperglycemia, based on the higher likelihood that people searching for both drugs over the year-long study period also would search for hyperglycemia-related terms, versus those who searched for just one of the drugs.
The researchers say the strength of the signal they identified in their searches would make it a valuable enhancement to the FDA's adverse-effect tracking system.
From The New York Times
View Full Article
Abstracts Copyright © 2013 Information Inc., Bethesda, Maryland, USA
No entries found