Challenges and Possible Solutions to Adapting to Virtual Recruitment: Lessons Learned from a Survey Study during the COVID-19 Pandemic

Abstract

The COVID-19 pandemic required rapid changes to research protocols, including immediate transitions to virtual recruitment and informed consent. This case study details the challenges our research team faced adapting an in-person, behavioral intervention and survey study to virtual recruitment. We reflect on the impact of these rapid changes on recruitment and retention, discuss protocol changes we made to address these challenges and participants’ needs, and propose recommendations for future work. Using computer technology to display professional return phone numbers, flexibility in contacting participants through various means, minimizing email communication due to added regulatory requirements, and partnering with the institutional review board to shorten and improve the consent document and process, were critical to study success. This case example can offer insight to other researchers as they navigate similar processes. Virtual recruitment is likely to continue; it is important to ensure it facilitates, rather than hinders, equitable and just recruitment practices.

Introduction

The COVID-19 pandemic required rapid changes to research protocols, including immediate transitions to virtual participant recruitment and informed consent. In most cases, institutional review boards supported these transitions and reviewed and approved modifications to studies quickly. While restrictions to essential personnel were necessary to keep staff and participants safe, many researchers were not prepared for some of the unique, complex barriers to virtual study recruitment and retention. This paper will illuminate some of these challenges and describe possible solutions and “silver linings” to use for future virtual recruitment and consent moving forward, using a case example from a survey study among newly-diagnosed cancer patients. Given that virtual recruitment is likely to continue going forward, it is important to ensure it facilitates, rather than hinders, equitable and just recruitment. We will provide recommendations to proactively address these recruitment and consent challenges based on our experiences, stakeholder feedback, and institutional changes implemented after working closely with our institutional review board to ensure both short and long term success of remote studies.

Case Example:

This case example of a research study that adapted from in-person to virtual recruitment beginning in the initial COVID peak (April 2020- December 2021) involved a behavioral intervention and survey.¹ The survey assessed health insurance knowledge, decisional conflict, and decision self-efficacy after some participants used a web-based health insurance decision aid and others engaged in usual care.^{2 3} Participants with newly diagnosed gynecologic, colorectal, or lung cancer completed one survey in the usual care group and two surveys in the intervention group: an initial survey and a 3–6 month follow-up survey after receiving the intervention. Outreach and follow-up for this study was primarily intended to be in-person at clinical appointments to build rapport, facilitate understanding of the study, ensure participant engagement with the intervention, and establish workflow with clinicians trained to have cost conversations. However, our team faced multiple barriers as we transitioned to virtual recruitment. We outline these barriers below and provide examples of the protocol changes we made to address the evolving needs of the study, especially our participants, during the pandemic.

Barrier 1: Phone Calls to Recruit Participants

The first barrier we encountered as we transitioned to virtual recruitment involved switching to phone calls rather than a combination of phone and in-person connections for study outreach and follow-up. As the COVID-19 pandemic rapidly changed work locations, our research staff relied on their cellular telephones to make recruitment phone calls from home initially. Many of these numbers had out-of-state area codes or showed up as unknown numbers. Without an in-person, direct avenue to introduce ourselves, build rapport, and provide an overview of the study, and without the caller identification displaying our institution’s name, many potential participants were skeptical of the reason for the call. They often declined to answer calls because of cold calls from unknown numbers or assumed we were telemarketers or bots from political campaigns. This trend mirrors findings that just 19% of Americans answer calls from unknown phone numbers and 14% don’t answer and ignore voicemail.⁴ Overall, these barriers led to initial distrust of research team members.

Protocol Solution

In order to improve study recruitment, our team needed to develop novel strategies to address the challenges of exclusively recruiting and following participants using phone calls. First, our team worked with information technology at our institution to set up a computer software on our desktops that registered our phone calls from the institution instead of from research staff’s personal phone numbers. This process gave more credibility to the research team and allowed us to build rapport with potential participants who did not know the research staff. Moreover, this established better continuity by offering a system for participants to return phone calls and leave voicemails. Next, our team created flyers for the patients’ clinicians and staff to provide to potentially eligible participants. A face-to-face connection with a clinician or clinical staff member allowed the participants to learn about the study and be prepared for a phone call from a research team member. The research team adapted to calling participants at various times throughout the day and even offered to discuss study procedures via videoconferencing so participants could see research coordinators and feel more comfortable asking questions. Flexibility in approaching participants in ways that worked best for them supported this process.

Barrier 2: Email Communication

Restrictions around communicating with potential participants via email were also a barrier to virtual recruitment. As part of standard institutional regulations for electronic communication about research, research team members must get explicit permission from participants to email them for research purposes. Once permission is obtained, the first email sent to a participant needs to be a test message to verify their email address. Participants need to reply to this test email to acknowledge receipt, before researchers can send research-related information. Alternatively, participants need to open a secure encrypted email sent from the researchers.

Although many participants were eager to have study information sent to them via email to review on their own time, they were often very confused by the test email to which they had to reply. They would often say verbally “I got it” or “that’s my address, yes” without understanding the need to electronically acknowledge receipt. The alternative approach, sending a secure encrypted message, appeared to look like spam. Given cybersecurity recommendations not to open unfamiliar emails or click unknown file attachments, many participants understandably worried that the link might contain harmful information because of the way it appeared (see figure 1). Finally, some participants check their email infrequently and missed the test email or the secure message, seeing it after a week or two when they forgot the instructions for the study. Although these restrictions around emails are intended for data security, they added burden by requiring an additional layer of follow-up and were a barrier to recruitment because many participants found them difficult to navigate.

Figure 1:

“Visual Display of Encrypted Email Notification That Confused Potential Participants.”

Protocol Solution

Reaching participants by telephone was key to addressing the email challenges. To avoid missed opportunities for participants to consent and complete surveys prior to their oncology appointments, for those who preferred to view the intervention and fill out the survey electronically, the team asked participants to verify receipt of test emails while participants were on the phone, when possible. This process allowed us to engage eager participants in a timely manner and quickly recruit participants that might have opted not to participate due to the added burden of these additional steps. Next, we also offered to read consent documents and complete surveys over the phone with participants who had difficulty opening email communication.

Despite our best efforts to address these barriers, some larger systemic issues remained. For example, some participants did not have access to email, had slow internet connections, or could not use video conferencing. Others worried about privacy when using videoconferencing because they did not want others to see their house or family members in the video background. These issues could be even more pronounced among older participants who can find technology and internet/email challenges difficult to navigate. In our study, the average age of those who declined participation was 67; of the 207 people who declined, 129 (62%) were 65 years or older. Although our study team was not approved to document race or rural/urban residence of non-participants, privacy concerns or access might have been barriers among those from rural residences who commonly have internet connection issues, and/or people of color who might mistrust research overall. ^{5 6} Protocol modifications centered around being as inclusive and accommodating as possible. Being flexible about approaching participants and meeting them in the most comfortable and convenient manner—in-person with clinical staff, by telephone, by email, or videoconference—was essential to enable recruitment.

Barrier 3: Lengthy and Repetitive Informed Consent Templates and Lack of Suitable Virtual Alternatives

Informed consent is a basic requirement of ethical research. A report on responsible research by the Institute of Medicine in 2002 stated that investigators should ensure that informed consent processes are effective and remain active through the duration of studies. ⁷ Informed consent documents are a key component of consent processes and are used to provide more details on the purpose of the research, the procedures involved, descriptions of risks and benefits, an explanation of compensation, and a statement that participation is voluntary. ⁸ Moreover, informed consent documents detail the rights and safety of participants and explain how ethical processes will be maintained throughout the study.⁸ Although these documents are intended to provide all the information potential participants need to make a decision about participation, participants still often lack basic understanding of research elements and also have inaccurate beliefs about the right to refuse participation, alternatives to treatment, and risks from participation. ^{9 10} Improving informed consent documents has been shown to be an effective way to improve participants’ subjective and objective understanding across levels of health literacy.^{11 12} Such improvements might also help to address reluctance to participate in studies and mistrust of the health care system. ¹³

Lengthy, repetitive informed consent templates, an institutional requirement to read consent forms verbatim for those opting for phone consent, and lack of suitable virtual alternatives to pre-pandemic, in-person discussions were some of the biggest challenges of transitioning to virtual recruitment.

a. Lengthy, complex, and repetitive informed consent template required by our institution

The original information sheet for this minimal risk expedited review study contained 2024 words, 5 pages of text, and had a Flesch Reading Ease score of 50.7 (10th-11th grade reading level and fairly difficult to read). Similarly, the consent form for one of our full board review clinical intervention and survey studies contained 3013 words, 8 pages of text, and had a Flesch Reading Ease score of 55.3 (also 10^th-11^th grade reading level and fairly difficult to read). Despite differences in risks between the two studies, the consent form specifications were similar, demonstrating the extent of regulatory language, even for minimal risk studies.

Due to the burden of regulatory language, we found participants were so inundated with information that they did not understand all relevant study components and overly attended to particular risks, such as data security. An eligible patient declined stating: “I did not like the part of the consent form about risk of confidential information being disclosed because there are people in my family that do not know I have cancer. I do not want them to find out from you.” This quote suggests that participants can misunderstand the technical privacy and disclosure sections of consent forms; this study did not have any aspects that involved disclosure to family members. Moreover, it also highlights that repetitive language can heighten participants’ fears about privacy breaches, data misuse, and anonymity related to e-consent. Our initial consent form referenced disclosure of confidential information seven different times throughout the document including the key information, certificate of confidentiality, HIPAA, and risk of breach of confidentiality sections which inadvertently suggested that breaches are more likely to occur. Even though the risks are present, they are minimal and not likely to occur with protections in place.

b. Lack of suitable virtual alternatives: Requirement to read consent form word-for-word by phone

Although some participants were able to consent electronically and complete the survey, some participants requested phone consent. In these cases, our institution initially required us to read the entire consent form over the phone word-for-word. This process is not unique to our own institution and can lead to participants being passive listeners rather than actively engaging with the consent.¹⁴ When participants opted for this approach, the repetition and legal jargon became even more apparent.

As a result of this substantial barrier, obtaining consent from participants over the phone was extremely difficult compared to in-person recruitment, which allows research staff to summarize key information while giving the participant time to read and review the consent. The latter process can be shorter, but includes exchanging information between researcher and participant, and allows participants time to absorb information at their own speed. By phone, reading the consent word-for-word confused many participants and caused them to scrutinize language in the form. They often asked us questions or declined the study providing reasons that implied misunderstandings. For example, a potential participant declined stating: “…the document that describes the study. It is very confusing for me…I just don’t feel comfortable doing it.” This further illustrates that long and repetitive consent forms do not always facilitate understanding of research protocols, and can be a recruitment barrier. Due to confusion, many participants often opt not to participate out of fear or mistrust.⁵

Secondly, many potential participants requiring phone consent did not have the time or interest in the consent information as we read aloud. In fact, some participants declined to participate in the middle of these phone calls because they were extremely long and described legal and regulatory policies that were not engaging in terms of what would happen during a study. Previously, participants could learn about key information (study requirements, risks, benefits, compensation, confidentiality) during an in-person visit, follow along with the content, and easily reference the consent document in writing during their decision process. By phone, maintaining engagement and interest was difficult. Some potential participants were initially interested in the survey, but not when it required them to be on the phone for up to thirty minutes listening to a research coordinator read a consent form and describe study related information.

Protocol Solution

Our team was fortunate to have an engaged partnership with the institutional review board (IRB). To develop this partnership, we held meetings to communicate the burden on participants and better understand the regulatory restrictions and why specific informed consent template components were non-modifiable. With a better understanding of each party’s goals in the process and a shared desire to reduce participant burden, we worked extensively together to modify recruitment materials to fit the changing needs of participants during virtual recruitment. We used our research studies as case examples to suggest changes to informed consent templates, improve phone consent procedures, and outline new guidance for research studies. Leveraging knowledge gained from recruitment challenges and feedback from eligible patients that declined to participate, the IRB allowed us to modify the consent templates (for our study and for future studies at the institution) in the following ways:

a. Shortening the length and repetition of the consent form

Modifications to the Common Rule in 2019 required informed consent documents to begin with a “concise and focused” presentation of the key information a “reasonable person” would require to have a better “understanding [of] the reasons why one might or might not want to participate in the research.”¹⁵ Following official guidance from the Office for Human Research Protections (OHRP) for how this can be accomplished, recent institutional changes allowed our research team to leverage the IRB’s latitude with the key information section compared to stricter requirements in other parts of the consent document that were created with review from HIPAA regulators and/or general counsel. Given that the revised Common Rule requires that key information be presented in a manner that “facilitates comprehension,” IRBs should have the most flexibility in making changes to this portion of the document. Content recommendations for the key information section by the OHRP include: a statement that research is voluntary, the purpose and duration of the research, the procedures to be followed in the research, an explanation of risks and benefits, and alternatives, if any.¹⁶ With our revised IRB guidelines, if information is presented clearly and thoroughly in the “key information” section of the consent form, researchers do not have to repeat the information in later sections of the form. Focusing on improving the key information section allows research teams to succinctly communicate pertinent study information, risks, and benefits to potential participants. As a result, teams can minimize repetition and redundancy to ease participants’ fears regarding their privacy and reduce mistrust by improving comprehension of study procedures, including processes that protect confidentiality.

b. Incorporating Plain Language Principles

Prior to the pandemic, there was also little to no official institutional guidance on plain language, reading level, or the formatting of information such as font and margin size.¹⁵ However, many of our team’s virtual recruitment challenges highlighted the need to improve the readability of our informed consent documents. In order to address and minimize these barriers, our institution agreed to allow changes to font sizes to improve readability, the addition of colored headers, and even graphics to improve the structure of consent templates. Given that formatting for increased understanding considers the structure and design of the document and its intended purpose, this was a step forward in tailoring our informed consent documents to match participants’ preferences (short, easy to read materials, better meet the needs of our audience, and be more inclusive to those with lower reading comprehension levels). ^{17 18} Moreover, being able to increase white space, shortening documents, lowering reading levels through plain language, and using colored headers can allow researchers to facilitate understanding without compromising content, while adhering to the revised Common Rule requirements. ^{19 20} After modifications, our revised form had a Flesch reading ease score of 57.1 (9^th-10^th grade reading level). Although the length of the final document increased due to non-modifiable regulatory content, the consent document was better formatted and we saw the most improvement in the key information section. Because we had the most flexibility with this section and its content, the key information section in the revised form had a Flesch reading ease score of 71.2 (7^th-8^th grade reading level). This is important given that participants may attend most to the key information.

In addition, the revised IRB guidelines state that as long as a full copy of the consent is provided and participants are able to ask questions, research team members can read the key information section aloud for phone consent and summarize other sections, rather than read the whole consent form aloud. This improved process allows potential participants to be more engaged throughout consent procedures, reduces participant burden, and can increase comprehension of study procedures. The revised processes might even improve recruitment and retention and more closely mirrors procedures for in-person consent.²¹

Barrier 4: Participant payment and personal information needed to process incentives

Finally, our institution’s previously required consent language for participant compensation stated, “you will need to provide your social security number (SSN) in order for us to pay you. If your social security is obtained for payment purposes, it will not be retained for research.” Despite including language that SSNs would not be retained for research, we learned many participants were still uncomfortable sharing their SSN, and some did not have them. As written at the time, the language did not provide alternatives to providing a SSN, and did not explain that for payments less than $600, participants could decline to provide their SSN and still participate and receive payment. The SSN language further added to fears of participation, especially for recruiting non-residents and undocumented immigrants who do not trust organizations to protect this sensitive data. For example, a potential participant declined stating: “I am not going to send in my social over the internet. Not even for $20…My son is a coder, works with Google. No way I’m sending that over the internet.” This further emphasized discomfort surrounding SSN language and even suggests participants might feel like these requests cross a privacy boundary.²²

Protocol Solution

Our institution provided guidance for improving SSN language (see example of new consent language in Table 1, which states that participants will be asked for their SSN, but does not state that it is required in order to be paid). Because institutions must issue tax forms to people who receive more than $600 in a tax year in total, research team members will still have to ask for SSN, but participants can decline to give their SSN without penalty or exclusion from study payment. The new guidance also states that asking for a SSN is not required for payment of $10 or less, token gifts, or participant reimbursements related to study visits (e.g., transportation). Changes in this section’s language can help to find a better balance of tax law risk management and research participants’ preferences. Because most studies compensate less than $600, this process will provide additional privacy protections for participants and might even offer reassurance that no more than the minimal amount of information necessary to conduct the research will be collected.

Table 1:

Examples of Improved Informed Consent from Expedited Review Study

	Original Consent Document and Procedures	Impact: Revised Consent Document and Procedures
Length	• 5 pages • 2,024 words • Required to be read aloud in full, word-for-word • Repetitive sections	• 7 pages due to more white space and bulleted headings • 1,914 words • Required to read “key information” and summarize the rest, if participants had a copy to review on their own • Deleted repetition
Key Information	• 3 paragraphs of text	• Added distinct sections with bullets and white space   ○ “What is the goal of this study?”   ○ “What will I do in this study?” • Increased white space • Clear separation between Key Information & remainder of the consent
SSN Language	• “You will need to provide your social security number (SSN) in order for us to pay you. You may choose to participate without being paid if you do not wish to provide your social security number (SSN) for this purpose.”	• “We will ask for your social security number (SSN) in order to pay you. If we collect your SSN for payment, it will not be kept for research. We will ask for your address if a gift card will be mailed to you.”
Risk Communication	• “ One risk of participating in this study is that confidential information about you may be accidentally disclosed. We will use our best efforts to keep the information about you secure.” (similar language in 7 places)	• “One risk of participating in this study is that confidential information about you may be accidentally disclosed. We will use our best efforts to keep the information about you secure and we do not expect this to happen.” (stated once)

Discussion

Overall, the transition to virtual recruitment during the pandemic emphasized the importance of improving informed consent documents and procedures for virtual recruitment and the essential role of researcher flexibility in reaching participants to assist with technological issues.²³ Despite our initial recruitment challenges, partnering with our institution’s IRB to balance regulatory requirements and the evolving needs of our participants was a helpful process that led to revised institutional policies and implemented procedural changes across other studies to maximize study success and minimize participant and researcher burden. This case example showing challenges, protocol solutions, and our institution’s engagement can offer insight to other researchers as they navigate similar needs. It also highlights that IRBs are willing to adapt their processes to meet the needs of participants and researchers, and can do so in ways that are consistent with regulatory requirements.

Working to find solutions to the above challenges has important consequences for diversity and equity in research studies and clinical trials. Virtual recruitment requires having participants who are able to learn and use the required technology, potentially leading to barriers and exclusion of people who may benefit from these interventions most. It also may potentially increase access for individuals who are hard to recruit, but we must ensure we attend to the potential barriers caused by the method itself. As one group astutely noted, “fully virtual/remote research, whether done out of convenience or necessity, brings with it a significant cost to recruiting and enrolling diverse populations who are often excluded from research to begin with.”²⁴ It is crucial to leverage advances in tele-consent platforms and best practices in electronic informed consent (i.e., incorporating audio, video, or graphics) to directly address the needs of historically underrepresented groups and increase access to research participation for participants that are otherwise hard to reach and/or under-recruited. ^25–27

Improving informed consent processes, in-person or virtually, is critical to the success and inclusion of research studies. Though these changes will require amendments to some IRB processes and guidelines, it can be accomplished through partnering with the IRB to enhance recruitment and retention. Many IRBs may be willing to engage in this process with time, resources, and ongoing investigator collaboration.

Finally, as these much-needed changes are made to the structure of consent forms, it is important to understand how they might impact participant comprehension and perceptions. Future collaborations might include working closely with community partners, institutional research and outreach coordinators and/or stakeholder advisory boards. Through these networks, researchers, institutional officials with regulatory oversight, community members, and the public can provide an additional opportunity to revise and refine consent and study procedures to better fit the needs of diverse participants.