Is social media the answer to finding side effects? After all, there is a large cohort of individuals willing to publicly share intimate details of their lives online, including medical details.
Interest in mining social media for signals of drug safety has existed for years, with advanced tools for analysis being developed. Regulators on both sides of the Atlantic having been trawling through social media, trying to net safety signals.
On social media, you are the product. Your data has been used to target advertising at you, and for political campaigns, amongst others things. Facebook is currently in the midst of an old school advertising campaign using TV ads, billboards, and newspaper ads to try and recover trust.
In the case of drug safety, who would object to their data being used to improve the safety of medicines? The simple answer is no one, but it’s not clear after recent controversies that it really is that simple.
A recent paper has looked at drug signals on internet forums and compared them to drug signals found with the spontaneous reporting system in France. Drug signals from the forum were less informative than those reported to the regulator, and had a different profile. Psychiatric side effects were the most common in the forum, nervous system disorders in the regulator database. Cases in the forums had poor information, reducing their utility in terms of assessing causation or validity.
There are other problems. Are people on Twitter and Facebook that publicly share health data representative of the general public? Demographic analysis would suggest not. That’s before we even tackle personality trait differences, that might influence the nature of side effects and differential reporting rates.
Drug safety signals can be increased by media attention. Reports of muscle pain with statins, may lead to increase reports of muscle pain associated with statins. “Hey, I think I’ve got that!” Reports on social media may be similarly subject to such notoriety bias.
Platforms also atrophy, and sometimes die. Rapid generational shifts can occur. Users leave established social media networks, for new networks. Some of these networks may be less conducive for research for drug safety signals, either being closed networks (Whatsapp) or more visual picture-orientated networks (Such as Instagram or Snapchat). Regulators may end up in a constant game of cat and mouse, as the user base shifts from platform to platform, and tools become non-functional due to technical issues, and tightened third party access to user data. Users may also start to change their relationship to social media, and privacy may, once again, become more fashionable.
Social media is a supplementary source of drug safety signals, which will require agility, care to prevent undermining of public trust, and careful interpretation. Long established pharmacovigilance systems, including spontaneous reporting by suspicious clinicians and patients to regulators will remain important.
You can report side effects at the Yellow Card Site in the UK.
Photo“The Art of Social Media” by MKHMarketing.