Ethical considerations of social media to recruit caregivers of children with cancer

Abstract

Background

Social media platforms are useful for recruiting hard-to-reach populations, such as caregivers of children with cancer, for research. However, there are unique ethical considerations in using social media.

Objectives

To describe the methods used to recruit hard-to-reach caregivers (parents of children with cancer) for research and related ethical considerations.

Methods

We used The Belmont Report tenets (respect for persons, beneficence, and justice) as a guiding framework to identify issues relevant to social media recruitment of hard-to-reach populations and to describe how we addressed these issues in our study.

Results

We engaged leaders of two online communities that offer peer support for caregivers of children with cancer to help with recruitment to our study on financial effect of pediatric cancer. We identified issues in using social media for recruiting hard-to-reach populations in alignment with The Belmont Report including: risk for subject-selection bias, privacy rights, protecting identity of participants, data security issues, and access to research. We addressed issues by deliberate study design decisions and engagement with online community advocates.

Discussion

Using social media to recruit hard-to-reach populations may be a successful way to engage them in research. While researchers may remain compliant with the institutional review board of their facilities and are faithful to the tenets of The Belmont Report, unanticipated ethical issues may arise directly or indirectly as a result of using social media. This paper identifies these issues and provides suggestions for dealing with them.

Social media use has the potential to enhance recruitment of populations hard to reach because of its ease of use and availability. Certain populations may be difficult to reach using traditional recruitment efforts due to psychological factors and circumstances (Sokol & Fisher, 2016). Caregivers (e.g., parents or legal guardians) of children diagnosed with cancer are hard to reach for study participation for myriad reasons. A cancer diagnosis can have far-reaching disruptive effects on the lives of the diagnosed child, their parents, and other family members. Prior research has shown that affected caregivers may be overwhelmed by juggling multiple caregiving and life-related responsibilities including complex caregiving, maintaining a job, and managing the househould (Mu et al., 2015). In particular, fathers of children with cancer may be more difficult to reach due to “inaccessibility or lack of invitation” (Davison et al., 2017). Caregivers and health care providers prioritize medical care over research participation during clinical visits. These circumstances are difficult to modify and make it challenging for caregivers to participate in clinical research. Flexible recruitment and study participation methods are needed to increase gender diversity of samples in caregiving research so that study findings can be reflective of the population of caregivers of children with cancer. Allowing for flexibility through online participation and social media recruitment is one way to improve the reach of recruitment efforts to caregivers of children with cancer and other underrepresented populations.

Over the past decade, social media has increasingly created a space for connection and social support in the general population (Paulus & Wise, 2019; Topolovec-Vranic & Natarajan, 2016; Whitaker et al., 2017). Moreover, given its far-reaching capabilities and generally low-economic cost, social media has been used successfully for research recruitment (Paulus & Wise, 2019; Whitaker et al., 2017). Many caregivers of children with cancer choose to use semipublic social media sites to access emotional support and information regarding caring for their child with cancer (Nagelhout et al., 2018). Semipublic sites require users to create profiles to engage in the public site (Elm, 2009). Although social media is used by caregivers of children with cancer, inherent in using social media for recruiting this population for research are evolving ethical issues. Comprehensive protocols regarding research using social media have not been empirically developed and have failed to match the rapidly increasing use of social media groups (Lee, 2017). The Belmont Report, a widely recognized and helpful guiding framework, addresses ethical issues relative to its three tenets: respect for persons, beneficence, and justice (National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research, 1979). Respect for persons concerns ensuring persons’ autonomy in deciding whether to enter into research and providing them with adequate information about how their data will be used; beneficence involves minimizing risks and maximizing the benefits inherent in the research; and justice refers to equality in the participation of research and that all participants can benefit from the research taking place (National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research, 1979).

While social media may be a useful tool to recruit and collect data from hard-to-reach populations, consideration should be given to address ethical concerns. In comparison to in-person recruitment and recruitment flyers, social media also has the potential to be seen by more caregivers than those who are currently caring for a child in active treatment. Therefore, the purpose of our paper is to describe the methods used to recruit caregivers of children with cancer to a study using online, community-based social media platforms by: (a) describing the methods we used to recruit caregivers; and (b) discussing ethical considerations relevant to each of the three Belmont Report tenets and how we addressed them.

Methods

The primary study was an institutional review board (IRB)-approved descriptive cross-sectional survey in which we asked caregivers of children diagnosed with cancer about the effect of pediatric cancer-related costs on themselves and their families (Santacroce & Kneipp, 2020). The primary study used a convenience sample of members of two online social media communities whose leaders were acquaintances of the primary study’s principal investigator (PI). Members were eligible to participate if they were the legal guardian or parent of a child with cancer. Although not required by the IRB, participants were asked to consent to participate on the first page of the online survey. We consulted with two gatekeepers of two parent-run organizations and asked them about the survey’s relevance and importance to their members. Both gatekeepers indicated that the survey would be a topic of interest for their organization members and appreciated that not all questions were forced responses.

Researchers are responsible for protecting human subjects from the potential risks associated with participating in research where individuals are recruited through social media. Unlike in-person approaches, recruitment via social media platforms holds unique potential risks related to maintaining respect for persons (issues with autonomy and data use), beneficence (minimizing risk, issues with data security, privacy, and confidentiality), and justice (subject-selection bias; Kamp et al., 2019). The coauthors engaged in analytic discussion with the primary study’s PI about how The Belmont Report tenets would be addressed in the primary study and reflected back on how these methods were actualized. In the following sections, we briefly describe results of using social media for recruitment, and present unique risks in using social media and the approaches we used to protect against them.

Results

Brief Description of Results Using Social Media for Recruitment

The primary study sample was recruited from Momcology® and Childhood Cancer Dads Support private Facebook group. Each group had around 200–400 members. Recruitment efforts resulted in a total of 545 active clicks on the primary survey between August 12, 2018 and September 17, 2018, and a final sample of 413 participants (Santacroce & Kneipp, 2020). Recruitment efforts resulted in a geographically and educationally diverse sample across North America; 45 of 50 states and 4 of 10 Canadian provices were represented (Figure 1). Of participants who responded to the survey, 48% were recruited through Momcology®, 16.6% through Cancer Dad’s Facebook group, 34.1% through a spouse/partner/another parent, 1.3% chose not to respond (Tan et al., 2019). Please see manuscript by Santacroce and Kneipp (2020) for more detailed demographics and primary study findings.

Figure 1.

Map of final participation by State (Note: Four Canadian provinces not shown)

Description of Ethical Considerations and Protective Approaches

Tenet 1: Respect for Persons

As with in-person recruitment, participants are free to decide whether to engage in research. Those who do may change their mind at any time. To accommodate for all potential participants, all information should be provided at a 7th- or 8th-grade level, free of medical and legal jargon (Ferrigno & Sade, 2019; Pandiya, 2010). Additionally, social media may not provide the participant the ability to meet face to face with the researchers during the consent process. Since there is no opportunity to gauge facial expressions and body language of the participant, including informed consent in online research is necessary (Ferrigno & Sade, 2019).

We were cognizant of alerting participants that they were entering into the study voluntarily by including a consent page at the beginning of the survey that reminded participants that the study was voluntary, could be stopped at any time, and that their data would be used for research purposes only and not shared with any third parties. We eliminated direct contact with participants by asking gatekeepers to distribute study information to them. In addition, having the gatekeeper post the link to the social media page or email gave caregivers the option of ignoring the survey and not participating if he or she did not have an interest in sharing information about their family. As part of the decision process for participating in research, we communicated that there was no compensation tied to the survey; thus potential participants were informed there was no explicit benefit to participating.

Tenet 2: Beneficence

We identified two issues in regards to beneficence when using social media for recruitment: breaches in confidentiality and privacy, and breaches to data security. Breaches in confidentiality and privacy are difficult to address because social media is inherently nonprivate. Caregivers may unknowingly reveal sensitive personal information regarding their child’s cancer and their own experiences, with no guarantee that the information is not shared outside of the group (Bender et al., 2017). Although researchers may not be able to control this aspect of social media support groups, privacy and confidentiality must be taken into consideration as these social media groups are used to recruit participants (Bender et al., 2017). We anticipated several risks associated with participant privacy including: maintaining privacy rights and protecting the identity of participants. We addressed these issues by not using personal Facebook accounts or emails to engage with groups, by not collecting identifiable data (e.g., first 3 digits of zip code instead of 5 digits), and by using aggregate group level data from Facebook (e.g., number of post views) and list-serv email (e.g. number of email reads). These measures reduced public and researcher’s chances of identifying participants.

Data security issues regarding how, when, and where researchers are monitoring, interacting, and assuring security with participants can be complicated by using social media for recruiting (Kamp et al., 2019). In order to protect human subjects, researchers should be aware of the potential risks to participants no matter how hidden. In regards to data security issues, we used a survey link that would bring participants to an anonymous secure survey website that is housed on a secure university server separate from Facebook or the nonprofit group’s webpage. A limited amount of protected health information (PHI) was collected in regards to the child’s cancer diagnosis. All PHI collected was deliberately nonspecific; for example, we asked caregivers to select what types of treatment their child received in terms of category (e.g., chemotherapy, radiation, surgery) instead of specific treatments. By limiting the types of PHI collected, we were better able to maintain a data set that would not allow for the identification of others in the support group or Facebook group.

Mitigation of harm was addressed by the research team by minimizing risk and maximizing benefits. We allowed caregivers to personalize their survies, e.g., not all questions were mandatory, and potential risks this survey could initiate were stated.

Tenet 3: Justice

Two challenges to consider in regard to justice are the appropriateness of accessing a population on social media (i.e., is this the target population driving the research question vs. an easily accessible population) and subject-selection bias using this platform (Kamp et. al, 2019). Ethical considerations surrounding justice reflect the need for recruitment strategies that allow participation in research for those who might not be able to access research.

We deliberately selected two large, preexisting parent-run groups that use social media platforms. Caregivers self-select and join these groups voluntarily. The Facebook group we utilized was a private, closed group in which members are vetted by the group administrator. Although the group’s member list was viewable public information at the time, we chose not to approach individual members about the study to reduce researcher pressure on potential participants. Instead, we partnered with the groups gatekeeper who vetted the survey and recruitment message, and posted and pinned (marked importance, keeping the message at the top of the webpage) the message to the group’s wall. The second group gatekeeper posted the research information on the main website and emailed the group list-serv email. The survey linked to an external secure Qualtrics site that collects data outside of the social media platform.

In our case, barriers to research participation include life circumstances related to having a child with cancer and/or recruitment strategies limit participation. By using social media to recruit participants, greater access was granted to caregivers who may be too busy to engage in in-person research. We were able to recruit a sample of fathers whose experiences with financial costs of cancer were previously unexamined. By selecting two organizations that provided services to our population of interest, we were able to address justice issues related to recruiting individuals who were best suited to answer the research question.

Subject-selection bias refers to issues regarding nonrepresentative sampling or biased sampling (Kamp et al., 2019) and can be a byproduct of voluntary group membership and advertising towards a particular demographic. Participating in groups that use social media could create opportunity for targeted ads that may divulge that the participants are caregivers of children with a serious illness. Subject-selection bias was difficult to address due to the specificity of the population we were targeting: caregivers of children with cancer. Subject-selection bias was reduced by selecting two large organizations with a diverse population of members, but still existed in that the sample was not diverse by ethnicity or race. However, the sample was diverse by caregiver gender, such that we had a substantive number of fathers participating (n = 85; 21%).

To further address justice issues, all participants were given equal weight and will be able to benefit from information gleaned from this survey. We deliberately left inclusion criteria broad so that caregivers at all stages of caregiving could describe how the cost of cancer affected them and their families. All participants were treated equally in terms of no caregiver responses were more highly valued over others nor were any caregivers compensated for completing the survey. Each parent group organization was acknowledged in published manuscripts.

Discussion

Using social media as a tool was a successful way for our research team to engage hard-to-reach groups. However, anticipation of ethical issues is key to participant protection. While researchers may remain compliant with the IRB of their facility and are faithful to the tenets of The Belmont Report (National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research, 1979), ethical issues may arise directly or indirectly as a result of using these online groups to recruit research participants. We recommend considering and addressing ethical issues even if not required by the IRB.

Social media remains a place for individuals or groups to seek help or support. Researchers should communicate with gatekeepers of social media sites instead of posting information themselves, especially if there is a gatekeeper available who has control over what type of information is posted (such as invitations to participate in research). Researchers should also be conservative in their approach and err on the side on consenting participants, especially if research is to occur outside of the social media site (Valdez & Keim-Malpass, 2019).

While many of the circumstances of dealing with childhood cancer, such as treatments and appointments are difficult to modify, using social media platforms to recruit this population to research may allow for more opportunities for these busy caregivers to participate. With further investigation, we can determine what the barriers to use of social media platforms are and whether decreasing these barriers would lead to caregivers more easily entering into research. As we move toward more technology-based research, we must remain vigilant in the ways that we protect participants recruited through social media platforms, making sure that we protect their privacy and confidentiality, and generate research results that are broadly applicable by making access to research more equitable.

Conclusion

Recruitment through preexisting, social media communities presents an innovative and efficient modality to engage hard-to-reach populations, such as caregivers of children with cancer. The authors found that a private Facebook group and an email list-serv were ideal settings to recruit caregivers of children with cancer to discuss their financial concerns. However, using social media platforms to recruit participants required deliberate attention and application of key tenets of The Belmont Report to ensure that the participants are protected. Researchers must anticipate the ethical concerns that could arise from this unique and relatively novel recruitment method. We considered privacy, ethical, and protection of human subject issues and found that investigators can implement the tenets of The Belmont Report as it is easily accessable, well-known, comprehensive, and a highly regarded framework among nurses. With social media being increasingly used for scientific research, researchers must adhere to strict implementation of key ethical tenets and frameworks to protect individuals engaged in research.