key: cord-0808088-vtsv7fyd
authors: Skeens, Micah A.; Sutherland-Foggio, Malcolm; Damman, Callista; Gerhardt, Cynthia A.; Akard, Terrah Foster
title: Facebook recruitment for research of children and parents during the COVID-19 pandemic
date: 2022-03-16
journal: Appl Nurs Res
DOI: 10.1016/j.apnr.2022.151574
sha: 06c4a00b5cfe37068c87454a3c024353c4060c3c
doc_id: 808088
cord_uid: vtsv7fyd

The COVID-19 pandemic has created unique challenges for recruitment of adults and children into clinical research. The sudden onset of stay-at-home orders and social distancing enacted in much of the United States created sudden barriers for researchers to recruit participants in-person. Recognizing the critical need to understand the impact of COVID-19 on children and families in real time, studies required an alternative approach. The present study sought to develop methods and establish the feasibility of utilizing Facebook's targeted advertising to enroll schoolaged children and their parents for a study examining the impact of the COVID-19 pandemic on families. This study used an 8 week pay-per-click advertisement approach via Facebook for research recruitment. Parents of children age 8 to 17 were invited and asked to include their child. Standardized measures were included for parents and children. Zip code targeting was used to increase diversity in participants. The ad campaign reached 213,120, yielding 3563 clicks, 684 parent participants, 494 child participants and a 26% conversion rate over eight weeks. The cost-per-click was $0.64, and cost-per-participant was $3.30 and $4.60 for parents and children, respectively. This nationwide study successfully used social media to recruit a robust nationwide sample of parent-child dyads during the COVID-19 pandemic. Social media recruitment mitigated typical time and engagement barriers for participants while also circumventing social and physical distancing orders due to the pandemic which allowed for real time assessment of the pandemic's effects on families. Future consideration should be given.to social media as a research recruitment methodology.

details eligibility and enrollments. A total of 912 parents completed screening questions. Of those, 842 were eligible, and the majority (82%, n=684) agreed to participate. Table 2 depicts the demographic characteristics of the 684 parent participants, and Table 3 describes the 494 child participants. Most of the parent participants were female (94%, n = 645) and White (90%, n = 624). While 25% (n = 167) of parents were unemployed, almost half of the participants reported an annual income prior to COVID-19 of greater than $100,000. Child participants were equally distributed by sex with a mean age of 12 years old (SD = 2.7, range 8-17). The majority (60%, n = 330) of participants were geographically located in the Midwest, however, all 50 U.S states were represented within the sample.

On average, 26% of individuals that clicked on the ad completed eligibility screening.

Those that did not meet study criteria included; non-English speaking (n=2), child age (n=62), home schooled (n=5). The overall campaign had a total of 213,120 impressions and received a total of 3,563 advertisement clicks in an 8-week period with an average cost per click of $0.64.

The general marketing campaign received 2,348 clicks in 6 weeks. Examples of Facebook analytics by age group are included in Figure 3 

This study reports on methods and feasibility of Facebook targeted recruitment to enroll a national sample of school-aged children and their parents during the COVID-19 pandemic. The findings demonstrate Facebook to a feasible, efficient recruitment method. We enrolled a total of 1,178 participants, almost 500 of which were children. The recruitment rate for enrollment was also rapid, 85.5 adults per week and 61.75 children per week. This is a considerable increase over traditional methods. In addition to rapid, robust recruitment rates, this approach was also cost effective. The total cost of the study was $2,280, and approximately $0.64 per participant.

As a relatively new communication medium, many researchers are skeptical of the utility of social media for participant recruitment in health care research (Shere, Zhao, & Koren, 2014) . Concerns include the assumption that samples recruited online, and particularly through social media, may result in selection bias and an unrepresentative sample, attributable to the exclusion of individuals that are not using the Internet or social media (Khatri et al., 2015; VanGeest & Johnson, 2011) . While we initially shared these concerns, given the benefits of rapid recruitment (Godino, Turchetti, & Skirton, 2013) , snowballing effects (Child, Mentes, Pavlish, & Phillips, 2014) , and minimal required resources (Mannix, Wilkes, & Daly, 2014) , this method demonstrated efficiency and effectiveness.

Recruiting research participants can be an extremely costly aspect of the research process (Carlini, Safioti, Rue, & Miles, 2015) . As new media for advertisement arise, they may provide new avenues and potentially more effective methods for researchers to reach J o u r n a l P r e -p r o o f participants. Additionally, assessing the impact of the COVID-19 pandemic required an effective way to access individuals' experiences in real time that would not be hindered by public health guidelines. With close to two thirds of the USA population now using Facebook, the platform holds promise in recruiting large, diverse samples for research at minimal cost to the researcher and participants (Clement J., 2020; Shaver et al., 2019) .

This novel study successfully used a social media recruitment methodology to enroll parent-child dyads in research during the COVID-19 pandemic. Recruitment over eight weeks, with a small budget, resulted in a robust sample of over one thousand participants from 684 families. Social media recruitment mitigated typical time and engagement barriers for participants while also circumventing social and physical distancing orders due to the pandemic which allowed for real time assessment of the pandemic's effects on families over a short period of a few weeks. In fact, Facebook has been used to recruit samples close to 100 in as little as three days (Child et al., 2014) . Facebook advertising methodology resulted in rapid recruitment of a large, geographically diverse sample, which would otherwise be difficult to access at low cost when compared to traditional recruitment strategies (Whitaker et al., 2017) .

The use of Facebook recruitment expedited remote recruitment during the pandemic.

Most participants were non-Hispanic white females, similar to other studies utilizing Facebook recruitment (Akard et al., 2015; Chu & Snider, 2013) . Females represented the majority of the parental sample in this study. This is consistent with Facebook user data indicating 77% of women in the U.S use Facebook, compared with 61% of men (Gramlich, 2019) . In our study, 80% of users were college educated. Facebook demographic data indicates that three-quarters (73%) of adults with at least a college degree are users, compared with 64% J o u r n a l P r e -p r o o f of those who have a high school diploma or less (Statista, 2021) . It is unclear exactly why college graduates are so overrepresented in this sample. A degree of ascertainment bias may have influenced the recruitment, as college graduates are more likely to be Facebook users, however it could not explain this effect entirely.

With zip code specific targeting over a 2-week period, the diverse population in our study increased to 26% and 11% respectively during those weeks, as compared to less than 5% in previous weeks. While zip code targeted ads have previously been successful in improving diversity (Pechmann et al., 2020) , the majority of our participants were still White (90.6%).

However, our increased diversity suggest that zip code targeting can effectively increase diversity and inclusion of underrepresented populations. Alternative methods to increase representation include health system recruitment and referral recruitment (UyBico, Pavel, & Gross, 2007) . Community involvement and improved communications have also been shown to lead to increased participation by minoritized populations (Kraft & Doerr, 2018; Vuong et al., 2020) . The authors attempted to work with a hospital based inner city organization to increase enrollment via a newsletter, however metrics related to this approach could not be captured.

National community organizations with diverse populations or target of populations of interest could also be a way to reach a more diverse sample. Sharing ads and information related to the study via those mechanisms have the potential to increase recruitment. Alternatively, partnering with a church the target population attends or creating a community advisory board of local leaders may be effective strategies to increase recruitment. To ensure full representation, future studies should consider combining a Facebook based approach with J o u r n a l P r e -p r o o f alternative methods of recruitment that have been effective in increasing research participation from diverse groups.

Over 491,833 impressions occurred for this campaign. Impressions are views and considered the number of times an ad appears on the screen. The number of times the ad was seen at least once by a user or reach was 213,120 (Carter-Harris, 2016). Facebook reach varies based on budget, target audience, and bids. The cost per click for the overall study was $0.64. This is similar to cumulative means ranging from $0.54 to $0.71 reported in recent systematic reviews of Facebook recruitment research (Reagan et al., 2019; Whitaker et al., 2017) . Cost per click is defined as the amount the advertiser pays for each click on the ad. Typically advertisers set a maximum cost-per-click bid that they are willing to pay (Ads, 2021b) . Conversion rate is the percentage of people who click on the ad and then become participants (eligibility is not excluded) (Ads, 2021a) . Conversion rates in other studies have been reported as high as 48%, but the mean conversion rate is closer to 4% (Whitaker et al., 2017) . The conversion rate in our study is unusually high at 26% for parents. While this may be due to the salience of the topic during a national pandemic, the increase in conversion needs to be replicated or examined in future Facebook recruitment studies. Additionally, the cost per participant (CPP) of our study was considerably low. The cost to recruit each parent was $3.30 and $4.60 for children. Reviews of studies utilizing Facebook recruitment have reported a median cost per participant between $6.18 and $14.40 (Thornton et al., 2016; Whitaker et al., 2017) . Regardless, rates in these studies are considerably lower than traditional recruitment methods that range from $13.12 to $250.00 (Carlini et al., 2015) . Low costs in our study could have been due in part to the broad inclusion criteria targeting a healthy child-parent population, as opposed to vulnerable hard to J o u r n a l P r e -p r o o f reach populations that typically have higher costs (Cho et al., 2021) . Thus, target population is an important consideration to determine if Facebook recruitment is a cost-effective recruitment strategy.

While a potentially effective recruitment method for pediatric nursing research, limitations exist. Accuracy of the data could be questioned and is frequently considered a limitation in anonymous online research. This can occur if someone entered data on behalf of the child, someone repeatedly completed the survey, or even created a computer program to repeatedly complete the survey in an attempt to increase the chance of receiving compensation from the study. Some methods to prevent these issues are to block individuals from filling out the survey multiple times from the same IP address, use spam prevention software (e.g. reCAPTCHA), and monitoring survey participants for abnormal responses (Bybee et al., 2021) . Even using these strategies, researchers should still be wary of potentially inaccurate data. An additional limitation is the constrained time of recruitment. However, due to the sensitive and rapidly changing nature of the pandemic, the shorter time frame was necessary to evaluate the early impact of COVID-19 on school-aged children. This sample presented an overrepresentation of participants from the Midwest. This outcome is likely because Facebook advertisements were posted by a children's hospital located in the Midwest.

As forementioned, there was also an underrepresentation of diverse participants, which is similar to several studies using Facebook recruitment (Akard et al., 2015; Bauermeister et al., 2012; Ramo & Prochaska, 2012; Ramo, Rodriguez, Chavez, Sommer, & Prochaska, 2014) . The authors attempted to increase diversity by using zip code targeting but this approach is not J o u r n a l P r e -p r o o f without limitations. The use of zip codes to specifically target advertisements to diverse racial and lower income groups is limited by the data available via the U.S. Census Bureau. This data may not be fully representative of the population within a given zip code as data may be out of date or otherwise uncharacteristic of the true population within a zip code. Thus, careful consideration is warranted when using zip code advertisement targeting. However, this issue presents through other recruitment methods as well. Similarly, father participants were significantly outnumbered by mother participants. Reasons for this may be two-fold. First, women are more likely to be users on Facebook (Gramlich, 2019) . Second, mothers are often the primary caregivers for their children (Livingston & Parker, 2019) . There is also some evidence to suggest that females may be more likely to respond to a survey than males (Smith, 2008 ). Thus, a mother might be more likely to respond to a survey ad. Underrepresentation of some participants may demonstrate a volunteer bias with non-Hispanic white women being the most likely to volunteer for this study. Finally, participants were limited to those with internet access and Facebook accounts. Despite limitations, this study still successfully recruited a robust geographically diverse sample early in the pandemic.

Pediatric nurses are in ideal roles to consider innovative strategies to advance nursing science and ultimately improve care for children and their families. Social media is changing the paradigm for future nursing communication, practice, education and research to a model that encompasses a variety of Web-based social media applications (Peck, 2014) . The benefits of technology are well-known and provide nursing the opportunity to provide a real-time exchange of health information, attaining timely information, the potential of maintaining J o u r n a l P r e -p r o o f contact with patients and their families, extending the reach for education and services as well as dissemination of knowledge and findings. More globally, the societal change in digital use also requires nursing to acquire effective digital literacy skills (Ross & Cross, 2019) . From a practice standpoint, and more broadly, these skills will create a networked nursing workforce that can influence and create collaborative communities to develop and shape healthcare from a clinical and nursing education perspective. Social media use in some chronic illness has improved communication and support for patients and caregivers. Nurses can carefully and ethically use social media to enhance the patient experience and improve outcomes. Nurse researchers should consider Facebook and other social media as a potential tool in the research toolbox to not only recruit, but also effectively deliver interventions. It is necessary to identify best practices and learn how to use these tools so that they can take full advantage of these new communication platforms (Miller, 2013) . Despite the positive aspects, pediatric nurses and beyond should give careful ethical considerations to the use of social media in practice, education and research.

An emerging body of evidence suggests Facebook is a valuable recruitment tool (Foster Akard et al., 2016; Pechmann et al., 2020; Shaver et al., 2019) for pediatric nursing research.

Social media recruitment can be cost effective and provide rapid access to large, sometimes difficult to reach samples. Facebook provided specific benefits over other social media platforms because of use by parents with children in our target population. Active users on other platforms, such as Instagram, tended to be younger and, thus, have younger children that did not qualify for the study. Furthermore, Twitter has also been used for recruitment, but J o u r n a l P r e -p r o o f

Facebook has shown to be a more effective recruitment tool (Whitaker et al., 2017) . With this in mind, it would be useful for future researchers to consider the characteristics of typical active users on a given platform to best utilize advertising funds. The underrepresentation of marginalized groups in this and other studies does suggest that Facebook has some limitation in recruiting diverse samples. Future research in other methodologies to ensure diversity within the sample would benefit understanding of Facebook and social media as a recruitment tool.

Moreover, there are other recruitment strategies that have proven to be effective in recruiting ethnically diverse populations, despite a higher cost per participant to recruit. It may be the case that Facebook is best used in conjunction with other strategies to optimize sample diversity. Although not appropriate in all settings, Facebook and social media should be considered as an acceptable recruitment method for parents and healthy school-aged children in future pediatric nursing research.

J o u r n a l P r e -p r o o f Table 1 Facebook Analytics Terms and Measures Definition Impressions

The number of times the advertisement appeared in a news feed Reach

The number of times the advertisement was seen at least once by a user Click

An interaction with the advertisement Cost Per Click The cost of the ad divided by the number of times the ad was clicked Conversion Rate

The number of people who click on the ad and then become participants (eligibility is not excluded) Cost Per Participant

The cost of advertising divided by the eligible recruited participants Frequency

Average number of times each person saw the ad Result Rate

The number of clicks divided by the number of impressions J o u r n a l P r e -p r o o f 

Conversion rate: Definition

Cost-per-click (CPC) Definition

Facebook advertisements recruit parents of children with cancer for an online survey of web-based research preferences

Pediatric cancer research: Surviving COVID-19

Innovative recruitment using online networks: lessons learned from an online study of alcohol and other drug use utilizing a web-based, respondent-driven sampling (webRDS) strategy

Recruitment and retention of youth for research using social media: Expereinces from the Just/Us study

Bots and nots: safeguarding online survey research with underrepresented and diverse populations

Using Internet to recruit immigrants with language and culture barriers for tobacco and alcohol use screening: a study among Brazilians

Facebook targeted advertisement for research recruitment: A primer for nurse researchers

Using Facebook and participant information clips to recruit emergency nurses for research

Facebook Recruitment for Children with Advanced Cancer and Their Parents: Lessons from a Web-based Pediatric Palliative Intervention Study

Use of a social networking web site for recruiting Canadian youth for medical research

Number of Facebook users in the United States from 2017 to 2025

The Decade in Broadband: 2020 Statistics and Predictions

Facebook recruitment of nurses as research participants: Methodological considerations

Web-Based Recruiting for Health Research Using a Social Networking Site: An Exploratory

Facebook Advertising To Recruit Pediatric Populations

Knowledge of genetics and the role of the nurse in genetic health care: A survey of Italian nurses

10 facts about Americans and Facebook

Recruiting adolescent girls into a follow-up study: benefits of using a social networking website

Social Media and Internet Driven Study Recruitment: Evaluating a New Model for Promoting Collaborator Engagemnt and Participation

Engaging populations underrepresented in research through novel approaches to consent

8 Facts about American dads

Pragmatism, persistence and patience: A use perspective on strategies for data collection using popular online social networks

How social media affects our practice

Recruiting Adolescent Research Participants: In-Person Compared to Social Media Approahces

Conducting Clinical Research During the COVID-19 Pandemic: Investigator and Participant Perspectives

Facebook Recruitment Using Zip Codes to Improve Diversity in Health Research: Longitudinal Observational Study

Social media in nursing education: responsible integration for meaningful use

Broad reach and targeted recruitment using Facebook for an online survey of young adult substance use

Facebook Recruitment of Young Adult Smokers for a Cessation Trial

Integrative Review of Recruitment of Research Participants Through Facebook

Rise of the e-Nurse: the power of social media in nursing

Using Facebook ads with traditional paper mailiings to recruit adolescent girls for a clinical trial

Using Facebook Advertising to Recruit Representative Samples: Feasibility Assessment of a Cross-Sectional Survey

The role of social media in recruiting for clinical trials in pregnancy

Percentage of U.S. adults who use Facebook as of February 2021 by educational background

Recruiting for health, medical or psychosocial research using Facebook: Systematic review

Recruiting Vulnerable Populations into Research: A Systematic Review of Recruitment Interventions

Surveying Nurses: Identifying Strategies to Improve Participation

Overcoming Barriers: Evidence-Based Strategies to Increase Enrollment of Underrepresented Populations in Cancer Therapeutic Clinical Trials-a Narrative Review

The Use of Facebook in Recruiting Participants for Health Research Purposes: A Systematic Review

Annual Income Prior to COVID-19 Under -$