What the SAFA trial found about social media recruitment

Social media recruitment for clinical trials is often discussed as a reach question: can it get more participants, faster, than traditional channels. The SAFA trial, an acne treatment study, measured something more specific and arguably more useful: not just how many participants social media brought in, but who they were, and how that compared to participants recruited through primary care.

The volume was substantial, and the composition was different

Social media accounted for over half of all participants in the trial, 53.9%, already a significant share of total recruitment on its own. The more striking finding sits in who those participants were relative to those recruited through primary care.

Ethnic minority participants were recruited through social media at a rate of 9.0%, compared to just 1.5% through primary care, a sixfold difference. That's not a marginal improvement in diversity metrics. It suggests that whatever barriers were limiting ethnic minority participation through the traditional primary care recruitment pathway in this trial, they weren't operating in the same way, or to the same degree, through social media channels.

Social media recruitment also reached participants further along in their condition than primary care did: 57.5% of social media recruits had severe acne (an Investigator's Global Assessment score of 3 or higher), and 60.2% had been living with the condition for more than five years. Both figures point at social media reaching people who had been dealing with a more serious, more entrenched version of the condition, potentially people who hadn't found their way into a clinical pathway through primary care at all, or had disengaged from one.

Retention held up, at least in the short term

A recruitment channel that brings in different participants is only useful if those participants actually stay in the study. At 12 weeks, social media-recruited participants matched primary care performance almost exactly, both at around 85% retention. That's a meaningful finding in itself: a common assumption about non-traditional recruitment channels is that participants recruited more casually, through an advertisement rather than a clinical referral, might be less committed. This trial's data doesn't support that assumption, at least at the 12-week mark.

The picture shifted by 24 weeks, where all recruitment routes saw substantial attrition, with dropout rates of 25% to 32% across methods. That's worth being precise about: the increased attrition wasn't specific to social media-recruited participants. It affected every recruitment channel roughly equally, suggesting whatever was driving longer-term dropout was about the study itself, or about acne treatment studies generally, rather than something specific to how a participant found their way in.

Trust mattered, and it was buildable

The trial's authors also captured something about participant experience that's easy to overlook in a purely quantitative recruitment analysis: participants found targeted social media advertisements acceptable and convenient, and their trust in the advertisement was strengthened specifically by high-quality graphics and recognisable NHS and university logos.

That's a concrete, actionable detail rather than a vague endorsement of "good marketing." Participants encountering a study through a social media advertisement were evaluating its legitimacy partly through visual cues, professional design and institutional branding, that a study team can directly control. A social media recruitment effort that skips this, treating it as a lower-effort channel that doesn't need the same visual polish as an official study website, may be leaving real trust, and real recruitment, on the table.

What this means for planning recruitment strategy

A few concrete implications follow from these findings:

• Social media recruitment can materially improve diversity, not just volume. The sixfold difference in ethnic minority recruitment here is large enough to be a primary reason to include social media in a recruitment plan, not just a secondary channel for filling numbers.
• Different channels may reach genuinely different populations along the disease spectrum. The gap in severity and disease duration between social media and primary care recruits in this trial suggests these aren't interchangeable routes to the same pool of potential participants.
• Short-term retention concerns about "casually recruited" participants aren't well supported here. Recruitment channel didn't predict 12-week retention in this trial; whatever predicts longer-term attrition, it wasn't which door a participant walked through.
• Institutional branding and visual quality genuinely affect trust in social media recruitment materials. This is a design and production consideration, not an afterthought, if a study wants social media recruitment to work as well as it did here.

The broader point isn't that social media should replace primary care or other traditional referral pathways. It's that these channels appear to reach meaningfully different participants, and a recruitment strategy relying on only one is very likely leaving a specific, describable segment of the eligible population unreached.