Data from Web Searches Could Help Identify Drug Side Effects and Drug-Drug Interactions

Researchers at Stanford and Columbia Universities teamed up with Microsoft to show that information from internet searches could potentially be used to identify novel drug-drug interactions and aid surveillance of drug safety.

In 2011, Tatonetti et al utilized adverse drug event reporting system of the FDA to identify novel drug-drug interactions. One potential novel drug-drug interaction discovered using this method was between the commonly prescribed drugs paroxetine and pravastatin. When prescribed together they were associated with hyperglycemia, and this finding was confirmed in subsequent mouse models and when examining patient electronic medical records.

The latest study continues from this previous work. The hypothesis was that users experiencing side effects from a drug will carry out internet searches for terms related to the drugs they are taking and their symptoms, and that this information could be used to identify potential drug side effects and novel drug-drug interactions. Logs of browser searches (including Google, Bing, Yahoo) from consenting web users were mined to identify users who had searched for hyperglycemia-related symptoms and terms. (These logs pre-dated the aforementioned publications regarding paroxetine and pravastatin drug interaction). 82 million drug, symptom and condition queries were analyzed. They found that users searching  for both paroxetine and pravastatin in a 12-month window also searched  for hyperglycemia-related symptoms more often than those who searched  for just one of these drugs. The study shows that logs of web searches could contribute to pharmacovigilance.

Read more:
Web-scale pharmacovigilance: listening to signals from the crowd
Ryen W White, Nicholas P Tatonetti, Nigam H Shah, Russ B Altman, Eric Horvitz
J Am Med Inform Assoc doi:10.1136/amiajnl-2012-001482
 
Learn about the pharmacogenetics of pravastatin and paroxetine