Strategies for improving recruitment of pregnant women to clinical research: An evaluation of social media versus traditional offline methods in Vancouver, Canada (Preprint)

BACKGROUND Social media is an effective alternative to offline methods for participant recruitment to research. However, the effectiveness of social media compared with offline strategies among pregnant women is unclear. Further, it is unclear whether recruitment strategy alters demographic characteristics of participants. OBJECTIVE We aimed to estimate recruitment rates from social media and offline methods and to explore the whether participant demographics differed according to recruitment strategy in a clinical nutrition trial that recruited 60 healthy pregnant women in Vancouver, Canada. METHODS Facebook was used to run 9 social media campaigns, 10-18 days each (15-weeks total) and costing $50-$100 CAD ($675 CAD total). Offline methods were used concurrently over 64-weeks. A total of $300 CAD was spent on printing. Demographic characteristics of those recruited via each method was compared using bivariate statistics. Cost, rate of recruitment and conversion rate in each group was calculated. Performance metrics of social media campaigns, including reach, impressions, clicks, inquiries, and enrollments, were recorded. Linear regression was used to explore the association between metrics and dollars spent per campaign. RESULTS In total, n=481 inquiries were received (n=51 [11%] via offline methods; n=430 [89%] via social media). Enrollees (n=60) included n=24 (40%) and n=36 (60%) via offline and social media methods, respectively. Gestational weeks was provided by n=251 women (52%) upon inquiry (mean ± SD gestational weeks was 13.3 ± 4.7 and 13.2 ± 5.6 in the offline and social media groups, respectively, P=.96). There were no statistically significant differences in age (33 ± 3.2 and 33 ± 3.6, P=.67), ethnicity (58% and 56% Caucasian, P=.97), education (88% and 78% had University-level education, P=.64), household income (58% and 47% >$100,000 CAD/year, P=.26), pre-pregnancy BMI (22.2 ± 2.6 and 23.4 ± 2.8, P=.11), or parity (75% and 72% nulliparous, P=.81); results are presented for offline and social media, respectively. Direct cost/enrollee was $13 and $19 in those who were recruited via offline and social media methods, respectively (however, this does not include cost of labour). Rate of recruitment was ~6x faster via social media than offline methods, however, the conversion rate was higher via offline methods than social media (47% versus 8%). Overall, campaign metrics (reach, impressions, clicks, and inquiries) improved over time. Amount spent per campaign (controlling for campaign duration) was significantly associated with improved clicks (P=.01), and inquiries (P=.04), but not enrollments (P=.19). CONCLUSIONS Social media was more efficient and effective for recruitment of pregnant women than offline methods. We gained numerous insights for optimization of social media campaigns (dollars spent, attribution setting, photo testing, automatic optimization) to increase clicks and inquiries, however this does not necessarily increase enrollments, which was more dependent on study specific factors (e.g. time of year, study design, and intervention). CLINICALTRIAL ClinicalTrials.gov (identifier: NCT04022135). Registered on July-14-2019. https://clinicaltrials.gov/ct2/show/NCT04022135