Social Media Tested by FDA as Potential AE Discovery Tool
For years, the irony has not been lost on us that, while the Food and Drug Administration (FDA) is fairly visible and active itself in social media, it has been slow to provide clear guidelines to pharmaceutical companies. Now the FDA is looking at social media as a tool that goes beyond a communications vehicle and functions as a way to better discover and report adverse events (AEs).
In response to a report that found 86 percent of adverse events (AEs) went unreported by Medicare enrollees, the FDA funded a study to help better understand whether social media could help shrink the gap between actual and reported AEs. The study results were released at the end of April.
With two goals in mind — determining the reliability of AEs discovered through automated tools and comparing the rate of AE mentions found on Twitter to AEs reported through the FDA Adverse Event Reporting System (FAERS) — the study examined 6.9 million English-language Twitter posts published between Nov. 1, 2012, and May 31, 2013. Of these, 61,402 tweets were flagged as relevant to 23 pre-identified prescription and over-the-counter drug products and were then manually analyzed.
Human analysis found that 4,401 of the drug mentions — or 7.2 percent — qualified as potential AEs based on the FDA’s four AE reporting criteria. During that same time period, 1,400 AEs were reported to FAERS using traditional means. Researchers found, however, that the Twitter data collected during the study was “too noisy” to be meaningfully compared with FAER-reported data.
Context is crucial to identifying AEs
Despite reports to the contrary, the potential for discovering an overwhelming number of AEs in the social space has been one factor behind some pharmaceutical companies’ hesitation to engage in social media. For many, social monitoring technologies such as Brandwatch, Sysomos, Radian6, and Visible Intelligence are viewed as fairly straightforward, technology-driven solutions. However, this FDA-funded study emphasizes that technology must be used in conjunction with human resources.
After sorting through more than 60,000 tweets, the researchers came to the conclusion that an automated tool simply cannot accurately identify AEs. Context is critical, and only human analysis can correctly interpret tone, the nature of the problems reported, the true meaning in the event of misspellings, and whether the comment is actually a clinician talking about cases that involve AEs (rather than a patient discussing an AE that occurred).
Combining tech and human touch
Intouch Solutions has offered a combined review-and-analysis approach to social listening for several years, consisting of carefully vetted listening technologies and a team of pharma-specialized analysts. Automated tools are used to pull aggregated mentions and organize the data, but human analysts are still needed to correctly interpret and code the data. Humans provide insights; they create a full picture with meaningful findings for future action.
While AE mentions are just one of the many reasons pharmaceutical companies may choose to limit their social media engagement, the concern is understandable, and the FDA’s study somewhat validates the concern. The risk can be mitigated, however, by applying existing AE operation procedures, as well as ensuring the right combination of listening tools — both technology and human — are in place.
For more information about Intouch Solutions’ social media service offerings, contact your Intouch Solutions representative.