Mining Social Media for Adverse Events
In February, the FDA released a Request for Proposal asking firms to help the agency monitor social media activity for adverse event reporting, among other content. Many have been critical about the utility of FDA examining user-generated content for adverse events. While such an effort may seem like a “needle in the haystack” approach, we think social media and online activity could lead to some fruitful information.
An “adverse event” is any serious undesirable experience a patient has associated with the use of a medical product. In the time after FDA approves a drug, unexpected safety problems can emerge once a product goes to market and is used widely among potentially millions of people. As a result, a critical aspect of FDA’s responsibilities includes postmarket safety monitoring—analyzing the risks and benefits of products once they go to market.
“The rise of social media on the Internet, including especially user-generated content such as blogs, forums, message boards, wikis and podcasts, has created new opportunities to interface with the public with respect to emerging hazard situations involving FDA-regulated products,” FDA states in their solicitation. “The increasing presence of social media promises new capabilities to monitor the effectiveness of FDA's ongoing risk communication efforts.” Thus, FDA noted that they are “in need of both historical and ‘real-time’ monitoring and analyses of a representative sample of social media web sites.”
FDA is especially interested in early detection of adverse events as well as food-borne illness. “The scope of work includes social media buzz reports, a social media dashboard, and quarterly surveillance reports related to specific product classes.”
New Study – More people report adverse events on Twitter than to FDA
FDA may be headed in the right place. A study entitled Digital Drug Safety Surveillance: Monitoring Pharmaceutical Products in Twitter released earlier this year suggests that patients are more likely to post drug-related adverse events on Twitter than report them to the FDA.
The researchers compared adverse events related to a diverse range of prescription and over-the-counter drugs identified in Twitter posts with corresponding data collected from the FDA Adverse Event Reporting System (FAERS). The study identified 4,401 reports with resemblance to adverse events, which they deemed “Proto-AEs,” among a sample of 61,402 posts collected from 6.9 million tweets sent between November 2012 and May 2013. In comparison, 1,400 adverse events (AEs) were reported to FAERS during a similar timeframe.
Twitter users can direct their conversations to specific medical products or government agencies. The authors note that Twitter hashtags to #Pfizer, #FDA, or on specific issues, such as “#accutaneprobz” show that patients are increasingly using Twitter in order to be heard.
“I don't understand it when people say they're always hungry. I can't even remember the last time I had an appetite #AccutaneProblems”
The authors note: “[a]t the same time that there is significant under-reporting of AEs through official channels, new Internet services have given voice to patients who routinely share information in public forums, including their experiences with medical products,” note the authors.
As anyone who scans Twitter knows, often you have to work through a lot of garbage to land on meaningful information. The authors nonetheless believe Twitter holds very valuable information; the trick is to know how to search for it:
“One of the key aspects of this research was to develop an ontology that allowed translation between social media vernacular and MedDRA [Medical Dictionary for Regulatory Activities is a clinically validated international medical terminology dictionary], yielding strong automated classification performance as noted above. As an illustrative example of this cross-ontology translation process, we consider the tweet “Humira never really worked for me. Orencia was good. Xeljanz was the best but ate a hole in my stomach. #RABlows.” In this example with rheumatoid arthritis medicines, the first product could be reported as ineffective, but there is also a more serious event.”
Others are not so sure
Depending on social media data mining for adverse event reports is a long shot, says Steven Nissen, MD, department chair of cardiovascular medicine and patient advocate with the Cleveland Clinic. Nissen was recently interviewed in HealthLeaders Media.
"This is like listening to rumors," he states. "Someone tells you something, you tell someone else, it gets repeated as fact." Tweets about the negative side effects of a specific drug could actually cause widespread misinformation about it.
Furthermore, due to privacy concerns, Nissen argues that "[d]octors aren't going to be tweeting adverse effects, and the Twitter Generation doesn't take too many drugs yet." Nissen concludes: “If the FDA wants to get better data about adverse reactions, it will be through improving their current reporting systems, not by listening to random chatter on the Internet."
Industry duties to report adverse events
According to their 2001 draft guidance for industry entitled Postmarketing Safety Reporting for Human Drug and Biological Products Including Vaccines, adverse event information submitted via the Internet to an entity with postmarketing reporting obligations should be reported to FDA if there is knowledge of the four basic elements for submission of an individual case safety report, namely:
- 1. An identifiable patient
- 2. An identifiable reporter
- 3. A suspect drug or biological product
- 4. An adverse experience or fatal outcome suspected to be due to the suspect drug or biological product
Because Tweets and Facebook posts arguably do not cover a number of these required categories, companies have sought further instruction from FDA about their responsibilities. FDA’s most recent release of social media guidance makes limited mention of companies' adverse event obligations. They focus instead on how companies must convey risks within 140-characters and what obligations companies have when correcting inaccurate Facebook and twitter posts.
As more people turn to social media for medical information, it will be interesting to follow how both FDA and industry respond. With the breadth of disease-specific chat rooms and online forums, vital risks are certainly being discussed online.
We are also interested in the potential for google searches to reveal important disease related information. Analysis of searches for particular symptoms within, for example, particular geographic areas, would offer useful information concerning the spead of disease.