Combating Fraudulent Participant Enrollment When Using Electronic Recruitment

Objectives: To describe an unanticipated response to electronic survey recruitment and strategies we implemented to identify and exclude ineligible participants.

Methods: For an NIH-funded trial testing an online continence promotion program, we sought to recruit 500 participants for an exclusively electronic study. Eligible participants had to be assigned female at birth, 50 years or older, able to read and understand English, and have access to an internet-connected device with which to access the continence promotion. Participants were recruited through a study website promoted via online and in-person outreach by community organizations and social media advertisements. The study website provided information about the study and contained a link to an electronic screening survey where eligible participants provided electronic consent and then were invited to complete a baseline questionnaire and create a study account. Once participants created a study account, they were randomized and received their first participation stipend ($25).

In response to a sudden influx of participants overnight and subsequent inquiries about stipends, we suspected and then confirmed fraudulent activity and implemented a participant verification process. Participants who could not confirm their identity were removed from the trial.

Results: Figure 1 presents the recruitment timeline and participants entering the screener daily. Recruitment launched 11/30/23 with 2 weeks of scheduled Facebook advertisements and emails sent to 1,361 community partner organizations. We closed enrollment on 12/22/23 following a surge of participants that entered the study between 12/14 and 12/22, having met our enrollment target.

As we processed the surge of new participants, several flags prompted us to question whether they were truly unique, eligible participants (Table 1). We worked with IRB and REDCap officials to implement a data integrity recovery plan. Almost 70% of enrolled participants did not pass an identity verification survey and were subsequently excluded. We re-opened recruitment with 3 strategies: 1) reCAPTCHA, 2) minimum survey completion time threshold, 3) identity confirma-tion survey with additional personal screening if discrepant. Even with these strategies, fraudu-lent participants continued to enter the screening link long after Facebook advertisements had stopped, and we ultimately created a new study link and sent the invitation only to university em-ployees to complete enrollment.

We ultimately had 1,393 fraudulent participants (373 inconsistent information; 433 survey time; 587 failed personal screen) attempt to enroll, over 400 of whom received a participant stipend. Most (986, 71%) used Gmail or Outlook (368, 26%) emails.

Conclusions: Research recruitment via social media is particularly vulnerable to exploitation by those who seek to earn compensation through participation regardless of eligibility. When recruiting for electronic studies, it is imperative to implement identity and eligibility confirmation checks to protect human subjects and preserve data integrity.