Social Media versus Traditional Clinic-Based Recruitment for a Dyadic STI Prevention Trial: Results from the Sexperience Study

Abstract

Background

The enrollment of youth in clinical trials has generally been achieved through conventional in-person recruitment but is evolving with the surge in the use of social media and presents an alternative resource for research recruitment for sexually transmitted infection (STI) prevention trials.

Objective

To compare the direct costs and performance of social media recruitment versus clinic-based recruitment method for a dyadic behavioral intervention for STI among heterosexual couples.

Methods

In the clinic-based recruitment arm spanning 60 weeks, Patients aged 16–25 years were recruited through an adolescent/young adult clinic. Social media adverts targeting college students within the city were also posted online over 23-weeks, using Facebook ad software. We compared the direct costs and performance of both recruitment methods to assess feasibility.

Results

381 individuals were approached, of which 21 completed the dyadic intervention (11 from social media-based recruitment and 10 from clinic-based recruitment). Clinic-based recruitment accounted for 91.0% of total recruitment cost and 9.9% of the total cost was spent on social media recruitment via Facebook ad.

Conclusion

Recruitment of adolescents and young adults for a dyadic behavioral STI intervention trial using social media is feasible, has lower direct costs, and results in similar outcomes compared to clinic-based recruitment.

Background

Recruitment of research subjects, both young and old have historically been conducted in brick and mortar settings.^1–3 However, engaging youth in research studies through this method present many challenges. For one, many young people consider themselves invincible, and as a result, rarely seek medical services unless extremely necessary or dire. According to a recent survey, 93% of youth do not schedule medical check-ups, and nearly half self-diagnose online.⁴ Youth recruitment for sexual and reproductive health research is even trickier, as the discussion of sex and sexuality is considered private and shame evoking by many young people. The changing trend in technology and social media usage, and the widespread adoption of social media may offer a way out of these challenges. The vast majority of adolescents in the U.S report access to mobile phones and daily use of the internet⁵, making social media an indispensable tool for study recruitment. This has been harnessed by many studies involving youth.^6,7 In this study, we describe the preliminary effectiveness and direct costs of social media-based recruitment compared to clinic-based recruitment for engaging adolescents and young adults (AYA) in a dyadic sexually transmitted infection (STI) trial.

Methods

Sexperience is an exploratory randomized trial designed to examine the feasibility, acceptability, and differences in short-term sexual behavioral outcomes between urban youth who receive individual sexual risk reduction counseling versus those who receive both individualized counseling and a structured dyadic debrief that includes condom negotiation role-plays and problem-solving support by trained research staff counselors. The study was conducted in the Baltimore Maryland Metropolitan area (BMA) which is ranked the city with the highest incidence rate of Neisseria gonorrhea and Chlamydia in the U.S.⁸ Heterosexual AYA aged 16–25 years were recruited by trained staff using two methods: (1) in-clinic recruitment and (2) social media recruitment. For the former, youth were approached and enrolled at an academic pediatric/adolescent medical practice during office visits. Patients were referred for study contact by their providers. Over a period of 60 weeks, Recruiters approached referred patients in person during office and some evening hours or via phone calls. In the latter, social media advertisements were placed on Facebook for 23 weeks to target youth in the BMA. Media advertisements were also disseminated to 5 universities/colleges in the BMA using map location feature on Facebook. Facebook ads were in the form of picture(s), a video, or a combination of these in a drop-down menu called an “Instant Experience.”

Participants are eligible if they live in the BMA, have a prior STI diagnosis, have no report of intimate partner violence with their current sexual partner, and have a current sexual partner within 5 years of their age (partner must be 16–30 years adhering to pre-specified age differentials to coincide with state law and developmental expectations). Participants are ineligible if they have language or cognitive disabilities that impair their ability to consent or participate, do not have a working cell phone, or their sexual partner is unavailable to participate in both visits. This study was approved by the Johns Hopkins Institutional Review Board and the trial is registered at ClinicalTrials.gov (Identifier: NCT03275168).

Direct costs for each strategy were calculated based on the costs for the staff hours to execute study recruitment procedures and the cost of using the Facebook ad software. Facebook ad software metrics were also used for a descriptive analysis on 1) Reach (number of times an ad achieves an outcome), 2) Impression (number of times the ad was shown on screen across all chosen platforms), and 3) Clicks (the number of clicks within the ad itself). The average cost per click metric was used to determine utility for the campaign. To assess feasibility, the total number of recruitments via social media was compared to in-clinic recruitment.

Results

Overall, 381 individuals were approached which yielded 145 eligible participants, of which 21 completed the intervention. Seventy-five percent of approached individuals were females with mean age of 20.9 years (SD=2.3). 243 individuals were approached for recruitment through the clinic. Of these, 125 (51.4%) were eligible and 10 (8.0%) completed the intervention. 138 people were approached through social media. Of these, 20 (14.5%) were eligible and 11 of the eligible persons (55.0%) completed the intervention (Table 1). Persons recruited via social media were older with a mean age of 23.6 years (SD =1.9) compared to those recruited through the clinic (mean age, 20.9 years (SD=2.1).

Table 1.

Demographics of Participants

Demographics	Total	Social Media	Clinic
Overall Age, Years (SD)	20.9 (2.3)	23.6 (2.3)	20.9 (2.0)
Male, N (%)	74 (19.5)	17 (23.0)	57(77.0)
Age, Years (SD)	21.1 (2.2)	24.2 (2.3)	20.5 (2.2)
Female, N (%)	285 (75.0)	94 (33.0)	191 (67.0)
Age, Years (SD)	20.9 (2.2)	23 (2.3)	20.2 (2.2)

Clinic recruitment was conducted for 60 weeks and cost $20,768 (91.3% of the total cost of overall recruitment efforts), reached 243 participants ($85/person reached), and yielded 10 participants ($2077/person enrolled). Social media recruitment was conducted over a period of 23 weeks and cost $3,558 (8.7% of total cost of recruitment) reached 138 participants ($26/person reached) and yielded 11 participants ($323/person enrolled). Total social media recruitment costs were generated from the cost of staff effort and Facebook ad placement, which cost $1,849, representing 1.2% of total recruitment cost. Table 2 presents the cost breakdown for both recruitment methods. Majority of cost savings for recruitment resulted from the limited cost of social media campaigns and reduced staff hours required to manage social media recruitment compared to the clinic approach (15 hours versus 40 hours per week).

Table 2.

Cost of Advertisement by Method of Recruitment of Adolescents and Young Adults in the Sexperience Study

Cost & Analytics	Overall	Social Media N (%)	Clinical Practice N (%)
Overall Time Span of Recruitment (Weeks)		23.0 (100)	60.0 (100)
Recruitment cost $		$1,850.0	0.0
Employee Cost $		$1708.8	20,768.0
Total Cost $ (%)	24,326.0	3,558.8 (8.7)	20,768.0 (91.3)

Discussion

Our findings indicate that social media recruitment is feasible and executed at lower direct costs compared to clinic-based recruitment. These findings are consistent with emerging research comparing conventional to social media recruitment. For example, a comparative study of pregnant women using both social media and clinic-based approaches demonstrated that media-based recruitment was faster and 40% less expensive than clinic-based recruitment.⁷ Participants were highly responsive to research advertisements on social media, probably because it allowed for self-selection based on the interest of the user. Most social media studies report similar trends in research interest among the youth.⁹ This is to be expected given the surge in internet use in the age group, making social media a ready source of young subjects for clinical research studies.⁵ Subjects of media recruitment who met all eligibility criteria were also more likely to participate in the study.

This study must be considered in light of some limitations. This was an exploratory study designed to recruit a specific cohort of dyads with STI histories for intervention that despite the broad outreach efforts yielded a small sample size from which to draw data. The study has limited generalizability as it targeted heterosexual, mostly African American population in the BMA. However, this is an important target population given the experience of health disparities associated with STI and their sequelae in this community.

Additionally, our inability to assess for eligibility through social media prior to screening subjects via e-mails or the initial phone call led to a significant drop-off between the many subjects approached versus those who were eventually eligible and therefore, enrolled. Subsequent study will need to account for this by recruiting a greater proportion of patients to correct for possible drop-offs or may have to incorporate all eligibility criteria into future advertisements. Nonetheless, social media recruitment has greater reach as demonstrated in the paper. Interested participants can self—select to be screened for study eligibility and to engage in the informed consent process on their own terms unrushed by clinical (e.g., illness, need for labs or other services) or personal issues (e.g., transportation, pending work and school return for the day) that may impair their ability to engage that day.

In conclusion, this study demonstrated that use of social media recruitment of AYA for a dyadic sexual health intervention was feasible, able to reach more people and enroll an equal number of participants in less time, and the direct costs of social media recruitment was lower per reached and enrolled participant. These findings further build the case for alternative recruitment strategies given the paucity of research focused on the recruitment of heterosexual dyads and youth at risk for STI into clinical trials.

Supplementary Material

1

Fig 1. Steps for Social Media Recruitment

Fig 1.

Schematic of Social Media and Clinic-Based Recruitment

Implications and Contribution:

Whereas the recruitment of adolescents and young adults into sexually transmitted infections (STI) research studies have largely been clinic-based, the current study shows that the use of social media for recruitment of adolescents and young adult heterosexual couples for a dyadic STI behavioral intervention is both feasible and of low-cost.