I posted a note nearly six years ago about how the the FDA had created the Sentinel Network to scan the electronic medical records of military personnel and their families to search for unreported drug adverse events (AEs) and assess drug efficacy (see: DOD Will Share Electronic Medical Records with the FDA). Here's a current link on the FDA web site for those interested in viewing the current "deliverables" from the Sentinel Network contracts (see: FDA's Sentinel Initiative - Deliverables from Completed Contracts). I had a hunch even back then that using EHR data might not be productive because these systems have not been designed for medical research (see: The Major Incentive for Mining and Analyzing "Raw Data" in Hospitals). Now comes news about an effort to search for unreported prescription drug effects using the web search queries (see: Unreported Side Effects of Drugs Are Found Using Internet Search Data). Below is an excerpt from the article:
Using data drawn from queries entered into Google, Microsoft and Yahoo search engines, scientists at Microsoft, Stanford and Columbia University have for the first time been able to detect evidence of unreported prescription drug side effects before they were found by the Food and Drug Administration’s warning system. Using automated software tools to examine queries by six million Internet users taken from Web search logs in 2010, the researchers looked for searches relating to an antidepressant, paroxetine, and a cholesterol lowering drug, pravastatin. They were able to find evidence that the combination of the two drugs caused high blood sugar.The study...is based on data-mining techniques similar to those employed by services like Google Flu Trends, which has been used to give early warning of the prevalence of the sickness to the public. 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. The new approach is a refinement of work done by the laboratory of Russ B. Altman, the chairman of the Stanford bioengineering department. The group had explored whether it was possible to automate the process of discovering “drug-drug” interactions by using software to hunt through the data found in F.D.A. reports. The group reported in May 2011 that it was able to detect the interaction between paroxetine and pravastatin in this way. Its research determined that the patient’s risk of developing hyperglycemia was increased compared with taking either drug individually....The scientists were able to explore 82 million individual searches for drug, symptom and condition information.The researchers first identified individual searches for the terms paroxetine and pravastatin, as well as searches for both terms, in 2010. They then computed the likelihood that users in each group would also search for hyperglycemia as well as roughly 80 of its symptoms — words or phrases like “high blood sugar” or “blurry vision.” They determined that people who searched for both drugs during the 12-month period were significantly more likely to search for terms related to hyperglycemia than were those who searched for just one of the drugs.
People often turn to Dr. Google to search about symptoms they may be experiencing (see: Paging Dr. Google! We Are Waiting for a Second Opinion). Some of them may be drug side effects. Why turn to the web for such research rather than hospital EHR databases? First of all, many patients may not want to "bother" their physician with a report of a non-specific symptom like blurry vision, perhaps assuming that it might be quickly dismissed. They may thus go unrecorded in the EHRs. Many such patients may also be taking multiple drugs and their physicians may be trying to cram a lot of decision-making into a short office visit. But when patients such as these get home to their computer, they may be willing to put general questions to Dr. Google such as whether "he" knows about any untoward reactions with paroxetine and pravastatin.
If you ask me, many researchers are going to ignore EHRs for drug research and go directly to Google if they can "explore 82 million individual searches for drug, symptom and condition information." This will be particularly relevant as researchers become more proficient in the use of Big Data (see: Big Data: The Hottest New Thing in Computing Attracts Big Money; More Discussion about Big Data; Relevance for Healthcare). The key aspect of such Google research is understanding how to interpret search engine queries. Here's a graph of the search term gonorrhea from Google Trends with volume peaking in October, 2010. A correlation with relevant news headlines is also provided.
Comments