Unfortunately, the rapid communication enabled by various devices and platforms could undermine clinical trials, allowing research subjects to share notes and possibly manipulate the study, warns Holly Fernandez Lynch, JD, MBE, a medical ethicist at the University of Pennsylvania in Philadelphia.

“The key is thinking about social media issues that might arise, before they actually arise,” she says. “It is easier to prevent than respond.”

Lynch and colleagues warn that research subjects sharing clinical trial information could affect participant recruitment and dropout rates, and alter study methodology like blinding. These problems could be worsened if the information shared is inaccurate or confusing.1

As a result, participants who believe they are in the placebo arm of the study may seek medication being administered to another arm of the study. By the same token, those who think they are in the treatment arm of the study may not report adverse effects that could affect their trial status.

Private Message Groups Trigger Concerns

For example, Lynch and co-authors cited a study in their institution of breast cancer patients who had recovered, but faced a risk of recurring disease. Trial participants were randomized into several arms to receive drugs approved by the FDA for other indications.

At the beginning of the trial some research participants posted information on an already established social media page for breast cancer patients. This public posting of general information was initially viewed as beneficial by researchers.

However, some of the research subjects formed a private group and did not invite the researchers to join. Investigators became concerned that the information-sharing about perceived trial arms and clinical results could undermine the study.

The situation was resolved when investigators discussed their concerns with the participants and stressed the importance of compliance to generate meaningful data that could guide care for the disease.

To address such issues, they designed a framework that could be used to approach this issue, using the acronym ADEPT to outline key components.

“This could be incorporated into templates for IRB submission,” Lynch tells IRB Advisor. “Have investigators thought about the way social media may impact their studies? A social media plan will not be appropriate for every study, but the role of the IRB should be to help investigators know which questions to ask and think in advance about what kind of plan would be appropriate for their protocol.”

The ADEPT framework encourages a structured and systematic approach, including the following steps and key features:

• Assess if social media use could pose risks for the study.

A potential red flag for impending social media issues is research on subjects with rare diseases or established patient advocacy efforts.

• Design collaboratively.

Bringing in research subjects as “partners” in a study diminishes the likelihood of social media activity that could undermine the trial.

• Educate research subjects.

In the informed consent process, emphasize the ethical duty of participants to adhere to the study parameters to maximize the benefits to themselves and future patients.

• Pre-empt problems, offer alternatives.

For example, consider proactive measures like moderated online discussions to quickly address research questions with accurate information.

• Take additional measures if needed.

Take these actions with prudence, enforcing needed measures without eroding a good faith relationship with research participants.

Gaming the System

Social media can aid in trial recruitment and can give subjects a sense of support in a shared community in an online forum, Lynch notes. However, with these benefits come the attendant risks.

“People are worried that they are going to get kicked out of the study or not be found eligible,” Lynch says. “So they may find ways to contravene the eligibility criteria. If you maximize eligibility, you are going to minimize that kind of behavior.”

Indeed, the design of a trial may set up temptations for participants to circumvent the rules if they see some personal advantage.

“We talked to a patient advocate in the context of writing this paper,” Lynch says. “Her perspective was that when patients in clinical trials engage in behavior that could be damaging to the research, it is often because the trial wasn’t designed with their concerns in mind.”

As these concerns are expressed on social media between participants, there could be efforts to game the eligibility requirements or not report adverse events.

“One of the most substantial concerns is that participants might share advice about how to avoid triggering valid and well-tailored exclusion criteria that otherwise would bar some individuals from enrolling in a desirable trial or jeopardize their continued participation,” the authors concluded.

Right to Try Raises Questions

This is a particularly topical issue in light of the Right to Try (RTT) issues being raised by dying patients seeking experimental drugs.

“My advice for both the RTT context and the social media context is that it is better for everyone for the investigators to carefully select their inclusion and exclusion criteria,” Lynch says. “Make sure that eligibility is as strong as practically appropriate, so you are not cherry-picking patients or being too restrictive.”

Bringing patients in as partners through collaboration and education is an important aspect of the program, but it will not solve all problems, she says.

“If the patient population says we all want the experimental intervention, then the response is that would not be good science because we wouldn’t have a control arm,” she says.

“It is the responsibility of the investigator and the researcher to explain why that is not what patients should want. There is a fallacy that whatever is new and experimental will be beneficial. It could actually be worse.”


1. Lynch HF, Largent EA, Joffe S, et al. Protecting Clinical Trial Participants and Study Integrity in the Age of Social Media. Cancer 2018: https://doi.org/10.1002/cncr.31748.