A Social Media Question IRBs Must Ask

I don’t envy IRBs. In the absence of clear guidance on social media and patient recruitment, they are in a difficult spot.

Despite the difficulty, the solution is not to hope social media goes away. It won’t. Internet users have developed a taste for dynamic interactive content, and they have no desire to revisit the old Web 1.0 days.

So it’s time to adapt.

With social media, we have a real opportunity to engage and educate patients in ways that we have not in the past. When used properly, I believe social media can help patients be more informed.

Unfortunately, social media can also be used improperly. That’s what concerns IRBs. And it should.

So how can IRBs distinguish between proper and improper use of social media? I’ll be providing a key piece of the answer in the following paragraphs.

A Social Media Issue IRBs Must Consider

In a previous blog post, I summarized a Quorum IRB webinar on the regulatory and IRB considerations of social media. The webinar explained how existing regulations can be applied to social media. If you are new to this issue, check out my summary of the webinar.

Existing regulations provide a good foundation, but they are not enough. As Mitchell Parrish noted in the Quorum webinar, IRBs may ponder additional issues specific to social media. And they should.

One additional issue, more than any other, should be assessed by IRBs. That issue is spam. Wikipedia defines spam as:

…the use of electronic messaging systems to send unsolicited bulk messages indiscriminately.

With electronic mediums like email, spamming can be identified with relative ease. But because social media is an especially fluid and evolving medium, spam is more difficult for a casual observer to spot. As a result, marketers with an unsophisticated understanding of social media often cross a line.

And that’s exactly the sort of scenario I’ve seen play out in patient recruitment. Someone affiliated with a clinical research study, in an attempt to recruit patients, will spam prospective subjects with information about a study.

Patient Recruitment and the Importance of Social Media Context

I assume patient recruitment spamming is the unintentional byproduct of unsophisticated marketing. But it is still very damaging.

For one, this practice is an intrusion on patients. And second, it is bad for the entire clinical research industry. Prospective subjects who are spammed may become distrustful of clinical research. And given the viral nature of social media, they may relay that feeling of distrust to their social network as well.

So I urge IRBs to look at the context of social media posts for patient recruitment, in addition to the content of those posts.

To help you assess that context, I’ll go over a key question to consider during the IRB approval process. But first, let’s look at an example of patient recruitment spamming.

What Does Social Media Spamming Look Like?

As I mentioned previously, the line for social media inappropriateness can be blurry. And you’ll find some disagreement on exactly where that line is. But the following example is very clear cut. Under no circumstances is this appropriate use of social media for patient recruitment.

I’m not going to show a live example. Instead, I set up fake Twitter accounts to mimic a scenario that I’ve witnessed in patient recruitment.

As you can see in the screenshot below, a patient recruiter is conversing with a social media user. The patient recruiter is tweeting to a prospective subject about an Alzheimer’s trial.

Except the patient recruiter is not conversing at all. In reality, the patient recruiter is using a bot to send out hundreds, or possibly thousands, of automated messages about the clinical trial. The screenshot you see above is just one in a very long line of tweets with the same text, all directed at different twitter users.

So how does the bot know who to send tweets to? It searches public tweets for particular keywords. In this case, the bot is searching all tweets for the keyword “Alzheimer’s”. Once the bot finds a tweet with that keyword, it sends a patient recruitment tweet to that user.

From Bad To Worse

This practice is a good way to piss people off, not to mention get you banned from a social media platform. But in some instances, it gets worse.

Because the bot is not human and thus does not understand the context of keywords, it tweets indiscriminately. A twitter user who tweets something like this can trigger a message from the bot:

So basically what happens is the social media user tweeted about his grandfather’s passing. As a result, that user received a solicitation for participation in an Alzheimer’s trial. And the patient recruiter probably has no idea that his automated social media presence is producing offensive content.

This example illustrates the importance of context in social media. The content of the post is alright, but the context is not.

As you can see, this practice is a not good for patients and not good for clinical research, which is why I’m writing this post.

A Key Social Media Question For IRBs

IRBs can’t just ask whether social media content will be used for spamming. The term spam is subject to too much interpretation. Instead, IRBs need an objective concrete way to elicit relevant information during the approval process.

Spamming is often associated with some type of automation, as described in the example we used above. But the use of automation is not necessarily a good predictor of spammy behavior. In fact, thoughtful use of automation in moderation can even enhance the social media experience for prospective subjects. And spamming can still occur without the use of automation.

But most commonly, problems arise when automation is used for interaction with social media users. So the question I suggest IRBs ask is this:

Is software being used to automate interaction with prospective subjects?

The word interaction is key here. Interaction needs to be distinguished from broadcasting. To illustrate the difference, here is an example of me broadcasting a message to my Twitter followers:

And here is an example of me interacting:

See the difference? IRBs need to pay special attention to the use of automation for interaction (although it would not hurt to consider automation of broadcasting as well). If automation is being used to interact with prospective subjects, further scrutiny is definitely warranted.

For IRBs reading this, feel free to contact me if you need clarification related to social or other digital media. I am sympathetic to the challenge of establishing a policy. And I want to ensure that digital media is used in a manner that benefits both patients and the clinical research industry.

Comments? What other issues specific to social media should IRBs be looking at? Put your feedback below.