Using Facebook for Health-related Research Study Recruitment and Program Delivery

Abstract

Facebook has become an important tool for recruiting research participants and for program delivery. Given the wide use of Facebook, there is much potential for the site to help with recruitment efforts in both physical and behavioral health care arenas; reaching groups typically difficult to recruit and providing outreach to individuals that may not have received services elsewhere. Health studies using Facebook have generally reported success, including cost-effectiveness, recruitment of samples in brief periods of time, and ability to locate participants for follow-up research. Still, the use of Facebook for research and program delivery is a relatively new area that warrants more research attention and guidance around issues like validity of data, representativeness of samples, and protections of human subjects.

Facebook, a popular social media website that was founded in 2004, is the second most visited website in the world today [1], with 1.44 billion monthly users globally and over 161 million daily active users from North American [2]. The website is used by upwards of three-quarters of online adolescents and adults [3,4] to share pictures and status posts with friends and family, follow content of interest (e.g., organizations, commercial products, news outlets), and stay apprised of current events and popular trends. Given the popularity and reach of Facebook in the population, Facebook represents an important tool to reach individuals for programmatic efforts that may not have sought services otherwise. With paid advertisement space available, it also represents an important, novel mechanism for recruiting participants into research studies. In this review, we discuss the studies that have used Facebook for recruitment and programmatic efforts in areas of physical and behavioral health. Based on this review, we also offer recommendations for using Facebook for these efforts.

Physical Health Research and Program Delivery

For physical health, Facebook has primarily been used for research and program delivery in areas of female sexual, reproductive, and physical health, such as to promote screening for medical concerns like breast cancer and sexually transmitted infections (STIs) and to assess physical activity and nutrition among adolescent girls [5–15]. There are few physical health studies outside of female health that have utilized Facebook for recruitment. Notably, however; researchers used Facebook to recruit boys with the genetic condition Klinefelter syndrome, after traditional recruitment mechanisms were unsuccessful [16]. Facebook campaigns have also been used successfully to track regional estimates of human papillomavirus vaccinations among young adults [17].

Behavioral Health Research and Program Delivery

Behavioral health studies on Facebook have included areas of mental health (e.g., depression, trauma and Posttraumatic Stress Disorder [PTSD]), substance use, and risky sexual behavior. Facebook has been used to recruit groups typically difficult to reach outside of clinical settings or through traditional community recruitment mechanisms (e.g., posting flyers, newspaper advertisements), such as sexual minorities for alcohol research studies [18], immigrant groups not typically captured by available research methods in population estimates [19], youth affected by violence [20], and young adult U.S. veterans [21]. Facebook has also been used in survey research to assess drug use attitudes and behaviors among young adults [22–23] and to recruit participants for interventions; such as U.S. veterans interested in reducing alcohol misuse and PTSD symptoms [24], young adult smokers interested in smoking cessation programs [25], and depressed individuals for an emailed intervention study [26]. It has been used to screen individuals for mental health concerns and suicidal ideation [27]. Intervention content has been delivered successfully on Facebook as well. For example, Facebook pages have been designed to promote condom use and other safe sex behaviors among young adults [28,29] and young urban Black women at high HIV risk [30] Researchers have also used Facebook as an intervention delivery tool for an approach to correct college students’ misperceptions of their peers’ drinking behavior [31].

Recruitment on Facebook

The two main avenues to recruit participants via Facebook are through paid targeted advertisements and peer referral. Paid advertisements can involve direct promotion of study content through text and picture-based advertisements displayed in one’s news feed or Facebook page side panel, promotion of text and picture based status updates posted on a study’s Facebook page, and invitations to “like” (publically endorse) the study’s Facebook page. Advertisements can be tailored towards the interests, demographics (e.g., specifying an age range), and location (e.g., zipcode) of the targeted population. Although these methods are paid initially, once Facebook users see advertisements, they can begin to interact with them in ways that maximize the social network capabilities of Facebook. For example, an interested participant can “like” an advertisement, make a comment on it, or share it with friends, which are actions that allow the Facebook users’ friends to be alerted to the advertisement or study Facebook page to consider the opportunity to participant for themselves.

In addition, there is promise for researchers to recruit participants into behavioral health studies using respondent driven sampling (RDS) or snowball sampling methods, which utilize the site’s inherent peer network structures to have Facebook users recruit other Facebook users (or peers not on Facebook) into studies. RDS has been successful for recruiting adolescents and young adults into studies and programs in areas of smoking cessation [32,33], condom use [28], and delivery of mental health information and provision of online support [34]. Facebook also represents an avenue for gathering information about the population to generate alcohol and drug behavior prevalence estimates through RDS; for example, researchers recruited 22 initial “seeds” and expanded the sample of 18–24 year olds to 3,426 via Facebook friend referral [35]. Of note, Facebook recruitment has typically focused on younger samples, with some work finding age differences between Facebook and other website-based recruitment mechanisms [36] and others struggling with recruiting some groups such as middle aged women [8].

Recommendations for Using Facebook in Health-related Studies and Programs

Consider the Limits of Facebook

Facebook research excludes those individuals who do not have Facebook accounts and Internet access on a computer or phone. Although most Facebook campaigns discussed in the literature have been very successful, others have struggled. For example, Close and colleagues [16] found their highest number of recruited participants (boys with Klinefelter syndrome) came during a brief one-week Facebook advertisement campaign, where visits to a study website increased from an average of two to three visits to the website per day to an average of 63 visits per day during the campaign. However, Kapp and colleagues [8] reported no success after an 11 day, $300 advertisement campaign targeted toward women aged 35 to 49 years. Many factors will play into the success of a Facebook recruitment campaign, such as the population targeted (e.g., age, gender, race/ethnicity, national/global versus regional, specific group versus general population), incentives offered, and funding allotted to an advertisement campaign. In addition, one needs to understand the Facebook interests of the targeted population as advertisements are targeted on users’ endorsed content (i.e., “likes”). However, some groups may not have “liked” content that would help identify them as appropriate research targets (e.g. a spouse of a problem drinker may not have “liked” Al-Anon groups on Facebook since doing so may reveal to his/her Facebook friends that their partner was a problem drinker).

Consider benefits and costs associated with Facebook versus other methods

Studies in both physical and behavioral health areas that have directly compared Facebook recruitment to traditional recruitment methods (e.g., flyers, approaching potential participants in clinics, email invitations) have found Facebook to be more successful in initial recruitment of participants [15,16,30], as well as in finding participants already enrolled in studies that were otherwise not able to be located for follow-up research [7,37]. Some studies suggest Facebook is more cost effective than postal recruitment [16,27] and other Internet-based recruitment methods [15,38]. However, researchers have reported less success with recruiting depressed individuals on Facebook as compared to more cost-effective advertisements elsewhere [26]. Like with any recruitment strategy, costs will vary depending on the population targeted and the nature of the study (e.g., one time survey or more time-intensive intervention study), but researchers have successfully recruited participants ranging from no cost or just a few dollars per recruited young adult participants (e.g., [7,17,21,23] to upwards of $10 to $30 per recruited adult participant for more intensive physical or behavioral health research (e.g., [10,24,26]. Still, Facebook may not always be the most cost-effective approach [26,38] and researchers report variations in successful recruitment rates even within Facebook advertisement campaigns based on content (e.g., wording, pictures) [25, 27].

Examine and report how samples compare to the targeted population

In order to establish Facebook as a viable recruitment mechanism and a legitimate population-based data source, it is important for researchers to report how their Facebook samples compare to available data sources that are accepted within the research community as adequate. For example, prevalence estimates of young adult substance use collected via RDS on Facebook have been comparable to estimates obtained by the National Survey on Drug Use and Health [35], and young adult veterans recruited from Facebook were similar in most demographic factors to the broader population of young veterans reported by the American Community Survey and the Department of Defense (e.g., education level, income, age, gender; but not race/ethnicity or branch of service) [21]. For regional samples, comparisons with U.S. Census records are also important, such as research from the Minneapolis-St. Paul area that found a sample recruited from Facebook matched the race/ethnicity, but not the education level, of the regional population [17]. Even though population estimates may not match, Facebook can be useful in identifying and recruiting members of a specific group. For example, researchers targeting young Australian adults for mental health and suicidal ideation screening recruited individuals with higher rates of mental health problems than expected in the population, which was the group of interest targeted in the study [27]. Still, reporting comparisons with population estimates, and employing methods to improve representativeness of data such as use of post-stratification weights, can help establish Facebook as a viable approach to reach populations traditionally captured by more labor-intensive methods (e.g., phone-based random-digit dialing; mailed population-based surveys).

Check for misrepresentation and data validity

Since there is no face-to-face contact with participants recruited via online methods, there is potential for people to misrepresent themselves in order to receive study incentives. Researchers have attempted to reduce misrepresentation by checking their data thoroughly post-collection, not offering any incentives, requiring identifiable information only accessible to people within a certain group (e.g., requiring college student participants to have an “.edu” email address), verifying consistent responses across similar or repeated screening questions, and asking “insider knowledge” screening questions known only to people within the targeted group [21,22]. These and other key methods for limiting misrepresentation and validating data in Facebook and general Internet-based research studies are outlined in detail by Kramer and colleagues [39].

Prepare for low recruitment rates and have methods to enhance retention

Since Facebook displays advertisements to a targeted but still large population, researchers should expect that many individuals who see advertisements and potentially click on them will not go on to complete surveys or programs. Although Facebook advertisements can be targeted to the population of interest, studies typically find recruitment rates as low as less than 1% of the targeted Facebook population, generally because advertisements are shown to hundreds of thousands to millions of Facebook users. This is evident at the global level as well [40]. Thus, consideration of resources, costs available, and staff effort is needed prior to a campaign launch. In addition, while many potential participants may initially click on advertisements, most will drop-out before completing screeners [8,11,12,21,24]. Then, even after expressing interest and completing screeners, steep drop-out may again occur if they are asked to engage in intensive programs. For example, about half of the participants recruited for an HIV prevention study from Facebook dropped out of the study after being invited for the full study post-screening [30]. However, in this study, all but one of the participants continuing past screening returned to complete a one-month post-program follow-up survey; an impressive retention rate similar to that of college students completing a Facebook-delivered alcohol intervention [31]. Yet, other studies suggest low completion rates of intensive online interventions following Facebook recruitment [24, 28]. Thus, other methods to enhance retention (e.g., reminders via text message, phone, email) have been helpful in studies [15] and may be needed.

Conclusions

There are many benefits to using Facebook for research and programmatic efforts. As mentioned, Facebook’s popularity greatly expands the reach to individuals on a national or global scale and can access those not likely to see advertisements in clinics or elsewhere, such as stigmatized groups or those in remote or rural settings. Targeted advertisements help reach specific populations of interest to reduce costs. Internet-based programs can be accessed immediately from Facebook advertisements, which reduce the delay between recruitment and receipt of services seen in traditional studies. The social aspect inherent to Facebook allows for enrolled participants to recruit others for research studies. Challenges exist, however; and much is still to be learned about how to address potential drawbacks of Facebook research (and Internet-based research in general). There is also a lack of clear guidance around human subjects issues that may arise, such as handling of misrepresentation and management of posted responses to public advertisements that may identify individuals as study participants. In addition, Facebook is not the sole answer to finding “hard-to-reach” groups; indeed, most notably, Facebook cannot reach those with no Internet access. Other reviews exist that further detail the benefits and drawbacks of Facebook [41**,42]. Papers also discuss how to overcome drawbacks to recruitment, such as misrepresentation and retention, using Facebook and other Internet sites [39*,43–45].

Highlights.

• There is much potential for use of Facebook in health-related research and program delivery.

• We review studies that have used Facebook to recruit diverse samples and deliver programs.

• Facebook studies have generally reported success with recruitment.

• Recommendations for researchers interested in using Facebook are discussed.