Background: Patients and providers demonstrate significant misconceptions and lack of knowledge regarding emergency contraception (EC), contributing to its underutilization.  There is a need to increase awareness and disperse correct information about EC to women who are at high risk of unintended pregnancy.  Social media has established itself as a leading source of news and health information for a significant proportion of the population, and offers access to a large audience that may not otherwise interface with the medical community.  Public health campaigns utilizing social media are becoming increasingly common, and may be particularly relevant for reproductive health topics that are sensitive or not commonly discussed.
Study Purpose: The overarching study goal was to assess the real-world effectiveness of Facebook advertising as a strategy to reach reproductive-aged people at risk of unintended pregnancy with a brief EC educational intervention.  The primary objective of the study was to compare efficacy in increasing EC knowledge and engagement on social media between a video and a website about EC.  Secondary objectives were to compare subjects’ satisfaction and appeal of the advertisements, assess current knowledge of EC amongst a local study population and a targeted online population, and explore if a brief Facebook advertising campaign improved population-level EC knowledge.
Methods: The study compared two existing interventions used for EC education, a 2-minute video and a website, and included women ages 18-29 years old.  Part 1 was a randomized clinical trial utilizing pre- and post-surveys, looking at participants’ ability to comprehensively identify an EC method correctly (including when to use and where to access it), before and after the intervention.  Part 2 utilized paid Facebook split-test advertising to conduct a $5,000 campaign, in which a Facebook ad featuring an embedded EC video was compared to a Facebook ad with a website link, using social media metrics commonly used to gauge online advertising success.  Surveys assessing EC knowledge amongst the target Facebook population before and after the ad campaign were also administered.
Important Findings:  Results from Part 1 are still being analyzed; information on the impact of each intervention on knowledge gained, satisfaction, likelihood of Facebook engagement and likelihood of EC use will be reported subsequently.  The Facebook ad campaign was run for 2 weeks and completed in April 2019.  The Facebook campaign demonstrated extensive reach – reaching an estimated 177 thousand people – with the average user seeing the ad 2.9 times.  The percentage of times people saw the ad and performed a click on it (click-through rate, reflecting actions such as reactions, expanding the screen, clicking on profile) was 0.36 and 0.32 for the video and weblink ads, respectively;  there were very few reactions, shares or comments on either ad.  The engagement with the video was very brief; of the estimated 199,317 times the video started playing, the video was watched for 3 seconds by 21,793 users; 10 seconds by 5,641 users; 1 minute by 969 users; and for 95% of its entirety by 497 users.  The pre-ad campaign survey of 2746 Facebook users found that 72.4% comprehensively identified levonorgestrel as an EC method, while only 3.3% for ulipristal and 3.9% for copper IUD; there was no difference in the post-ad campaign survey of 1276 Facebook users.
Conclusion: The Facebook campaign demonstrated extensive reach to a large audience, however with relatively low engagement and very brief interactions with the video format.  Meaningful impact of the video is difficult to assess, but there was no difference seen in pre- and post-intervention population knowledge of EC.  The ad campaign focused on a small geographical area, was limited in its scope and did not target Facebook users based on specific risk factors for unintended pregnancy, other than age alone.  Ad campaigns that better utilize Facebook’s algorithms to optimize its audience, or feature very brief messaging, may be more effective.