Survey design

The surveys were designed to be brief to reduce participant burden. The researcher survey had 12-items, and the IRB professional survey had 9-items. The surveys took approximately 5 to 15 minutes to complete. The surveys included open-ended and demographic items. The online surveys were administered using Qualtrics online survey system, which is HIPAA compliant. Participants were not paid for participating.

Participants who completed the Researcher survey were asked to describe their COVID-19 research, any ethical, regulatory or oversight obstacles or challenges they experienced, to propose solutions and provide their recommendations for overcoming these challenges. We collected demographic data from participants about the type of COVID-19 research study they proposed or conducted, the participant population being studied, whether they worked as a clinician, their role on the research team, and if they sought IRB approval for their research or medical innovation. The IRB professional survey asked participants to describe the challenges they faced when reviewing or overseeing COVID-19 research, new ethical or regulatory challenges they experienced, accommodations made for COVID-19 research studies, and advice for IRBs or researchers involved in COVID-19 research. Demographic items asked about the type of IRB they serve, and to describe their role on the IRB.

Recruitment of survey participants

We used publicly available contact information to identify COVID-19 researchers and IRB professionals in the United States. The Becker Research Library at Washington University assisted our team, building contact lists for researchers using publicly available contact information from public sources (e.g., ClinicalTrials.gov, Scopus, and Embase Conference Abstracts). Researchers’ contact information, including names and email addresses were added to our recruitment list containing 7,106 researcher contacts. We completed an application process with Public Responsibility in Medicine and Research (PRIM&R), to share our project and recruitment information with their members who are IRB professionals. PRIM&R supported our recruitment efforts by posting our recruitment email and project information on their IRB Forum blog page, on their dedicated COVID-19 webpage, and a brief note about our project with a link to the blog post appeared in their electronic newsletter, PRIM&R This Week.

Data collection occurred during the pandemic, between April 21, 2020 and July 30, 2020. First, we sent the recruitment email to Washington University colleagues as part of a pilot study (n = 12 researchers, and n = 6 IRB professionals). The pilot study did not identify any notable areas of concern and we maintained the same survey and recruitment strategy. Following the pilot study, we sent individuals in the researcher recruitment list an initial recruitment email with an anonymous link to the online survey and three recruitment email reminders. We sent individuals in the IRB professionals recruitment list an initial recruitment email with an anonymous link to the survey and two recruitment reminder emails.

To ensure our recruitment invitation reached as many COVID-19 researchers and IRB professionals as possible, we utilized additional research strategies including posting the recruitment invitation with the link to the survey on social media (e.g., Twitter and LinkedIn), asking colleagues and participants to forward the recruitment invitation to interested individuals, and announcing the project on our Bioethics Research Center website.

We removed eight survey responses from the Researcher survey database because they reported they did not conduct or propose COVID-19 research in the United States, and were ineligible to participate. We added two responses from researchers who mistakenly used the IRB professional version of survey to provide information on their experiences. After removing ineligibles, incomplete responses in Qualtrics, and adding the two surveys from researchers who completed the IRB professional version of the survey, N = 211 for the Researcher survey.

We removed one duplicate response from the IRB professional survey database. We removed eleven ineligible survey responses from respondents who reported they had not reviewed COVID-19 research, or who were not IRB professionals who reviewed COVID-19 research in the United States. After removing partials, ineligibles, one duplicate survey, and moving two responses to the researcher survey database, N = 143 for the IRB professional survey.

We cleaned the open-ended text data and removed the extra spaces in sentences, fixed typos and misspelled words, removed bullet and number formatting, spelled out the abbreviations of some words, and added punctuation and capitalization to ensure responses were in sentence format for data analysis. We had initially proposed to analyze these data using SPSS Modeler’s Text Analysis software. However, after 20 hours of consultation and 11 hours of additional work, we determined that a machine-learning approach would not be time efficient absent a previously established vocabulary and hierarchical ontology. As such, we proceeded with manual coding, as described below.

Data analysis

Open-ended survey responses were exported from Qualtrics into Dedoose qualitative data analysis software. In consultation with all authors, the Principal Investigator and co-investigator (DuBois, Sisk) led codebook development for both stakeholder groups through iterative consensus coding of 69 transcripts (21 IRBs, 48 researchers). Researcher characteristics relevant to codebook development include: One investigator (DuBois) has served on two IRBs (biomedical and social/behavioral) and trains investigators referred for diverse kinds of compliance failures [29, 30]; another investigator (Sisk) is an oncology clinical investigator and expert on communication in clinical encounters [31–33]. Our coding approach employed inductive content analysis to generate an inventory of key categories of challenges experienced by researchers and IRB professionals while conducting or reviewing research on COVID-19 during the early months of the pandemic, as well as recommended accommodations or solutions [34, 35]. Two authors (Baldwin, Parsons) then coded all transcripts in both stakeholder groups. Coders first consensus-coded 90 (42 IRBs, 48 researchers) transcripts for training purposes, meaning that they each coded the same 90 transcripts and identified any disagreements in coding. These coders then resolved disagreements through discussion with the full authorship team. All transcripts were then split between the two coders, who independently coded their assigned transcripts. To ensure ongoing quality of data coding, coders conducted multiple iterative rounds of coding checks by reviewing each other’s coded transcripts. With these iterative rounds of coding, the authors further refined definitions of themes to ensure clarity in application of codes. This process did not lead to major changes in themes, but rather refinement of codebook definitions and code application. Coders kept detailed records throughout coding and held weekly meetings to resolve questions or discrepancies via consensus.

This study was funded by the National Institute of Aging as a supplement to an R01 implementation trial focused on implementing evidence-based informed consent practices; accordingly, during coding, the team paid special attention to challenges that arose with regard to the informed consent process.

Participant characteristics

IRB professionals who participated in this study were predominantly from academic centers (87%). Most respondents were IRB directors (53%) or IRB chairs (22%). Researchers who participated were predominantly physicians (63%) and principal investigators (75%). Researchers were leading various types of COVID-19 research projects, including basic science, diagnostic, treatment, epidemiology, and prevention studies (Table 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Participant characteristics.

https://doi.org/10.1371/journal.pone.0265252.t001

Ethical, regulatory, and practical barriers to COVID-19 research

IRB professionals and Researchers described 19 types of barriers to COVID-19 research, with the majority of barriers described by both groups of participants (Table 2). Because our purpose was to generate an inventory of considerations generated by key informants and our sampling method does not permit generalizations about the frequency of barriers or solutions, we do not report counts in any tables. However, every theme was mentioned by at least 3 participants, and most themes were mentioned by more than 15 participants in each cohort.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Identification of challenges and barriers.

https://doi.org/10.1371/journal.pone.0265252.t002

These barriers related to 5 overarching categories: policy and regulatory, biases and misperceptions, institutional and inter-institutional conflicts, risks of harm, and pressure of the pandemic. We present barriers separately for IRB professionals and researchers in Tables 3 and 4, which include excerpts that illustrate the barriers. However, we discuss findings for both groups together in our results narrative to illustrate connections.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Barriers to research—Researcher excerpts.

https://doi.org/10.1371/journal.pone.0265252.t003

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Barriers to research—IRB professional excerpts.

https://doi.org/10.1371/journal.pone.0265252.t004

Although participants identified myriad barriers, 45% of researchers (94/211) and 8% of IRB professionals (16/143) reported that they did not experience any difficulties. The remaining researchers and IRB professionals identified a multitude of barriers described below.

Potential responses, adaptations, and solutions to challenges.

Researchers and IRB professionals described 8 categories of adaptations and solutions to the challenges that arise when trying to conduct research on a pandemic during a pandemic. Their proposed responses were frequently general and could be mapped onto more than one barrier or challenge (Table 5). Additionally, 25% (53/211) of researchers provided praise for IRBs during this challenging time. Researchers described how IRBs were prompt, responsive, adaptive, and easy to work with. Several researchers noted how IRBs provided essential resources and guidance that supported their protocol development and submission.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. Proposed adaptations and solutions.

https://doi.org/10.1371/journal.pone.0265252.t005

Enacting technological solutions. IRB professionals and researchers described the importance of developing electronic consent processes due to infectious risks. IRB professionals also described the importance of electronic/online IRB systems to avoid reliance on printed documents or wet signatures. Furthermore, some IRBs utilized videoconferencing to offer researchers virtual office hours.

Developing protocol-based solutions. IRB professionals and researchers described the importance of providing standard language to expedite drafting of protocols and approvals, such as consent documents, concurrence letters, and methods language around consent processes. One researcher suggested these standard documents should be posted on the IRB website. IRB professionals and researchers also described the utility of minimizing specificity in protocols to allow flexibility and adaptation in the implementation of a protocol. They also described the importance of planning for future data uses in the initial protocol to support data sharing in the future. IRB professionals encouraged researchers to include contingency plans in their protocols to minimize the number of revisions prior to approval. Researchers also recommended that multisite studies utilize a central IRB and centralized resources whenever possible.

Maintaining high standards. Despite this need for adaptability and flexibility, IRB professionals and researchers described the need to maintain ethical and scientific integrity. While both groups recommended bureaucratic short cuts, they were adamant that “flexibility” must not be allowed to increase harms to research participants, diminish the quality of the research, or lead to inaccurate or poor quality publications.

Managing team well-being and communication. IRB professionals and researchers both encouraged active and open communication between research and IRB teams, as well as effective communication within teams. IRB professionals also described the importance of clarifying workload expectations and monitoring for burnout within the team. One group strongly encouraged IRB professionals to take time off if they demonstrated signs of burnout given the heavy workload.

Establishing and communicating appropriate standards. IRB professionals described the need to interpret guidelines and regulations in a manner that is favorable to expediting research, so long as this approach does not harm participants. However, IRB professionals also described how this flexibility must be balanced with consistency so researchers have clear understandings of expectations. Given the rapid changes in guidance during a pandemic, IRB professionals also described the need to publish guidance routinely to educate researchers about updates. Researchers described the need to remove unnecessary steps in the approval process and to provide clear guidance prior to review to improve the quality of the proposal.

Developing national guidance and leadership. IRB professionals described the exemplary responsiveness of FDA in providing guidance during the early pandemic. For example, the FDA posted guidance on how to obtain electronic consent. Researchers described the need for stronger federal or national leadership in developing and prioritizing research protocols. For example, one researcher called for a national repository of protocols and more centralized leadership in prioritizing and funding high-priority multisite trials.

Appropriately screening and prioritizing COVID-19 studies. IRB professionals and researchers described the need for multiple levels of prioritization early in a pandemic. At some centers, scientific oversight/advisory boards worked with researchers early in the planning stages of studies, and also reviewed full proposal prior to submission to the IRB. One researcher agreed with the importance of scientific review prior to IRB review, but also called for clearer standards for evaluation and an appeals process. Within the IRB, some centers created subgroups of individuals with regulatory and clinical expertise to prioritize and/or review COVID-related protocols. One site developed a new workflow that expedited modification approvals for COVID-19 protocols.

Identifying and incorporating experts. IRB professionals and researchers described the importance of engaging clinical, methodological, and regulatory experts into the protocol development and review processes.

Shared missions and effective collaboration of IRBs and researchers

IRB participants and researchers generally praised the way that so many IRB personnel rose to the occasion. We learned of offices that expanded hours, gave priority to the review of COVID-19 protocols, added experts in infectious disease and related matters to provide competent reviews, exercised discretion with protocol deviations and modified consent procedures, and in many other ways facilitated research while enforcing regulations meant to protect research participants. They often accomplished these feats without any additional funding or staffing.

Notably, researchers and IRB professionals identified 13 overlapping challenges, and only 6 categories were unique to one group. Given the common perception of adversarial relationships between IRBs and researchers, this amount of agreement was unexpected and encouraging. Overall, researchers and IRB professionals strove for efficient, safe, and effective research to help the medical community and society combat the COVID-19 pandemic.

Policy and regulatory barriers

Since research on COVID-19 began in early 2020, significant progress has been made to facilitate responsible research. IRBs and researchers have worked together to find solutions for remote informed consent. IRB professionals have debated and disseminated best practices through webinars and blogs, such as Amp&rsand, the blog for Public Responsibility in Medicine and Research (PRIM&R) [36]. Many sites instituted COVID research committees to screen and prioritize COVID studies, even though the prioritization process created tensions and frustrations at times. Furthermore, in April 2020, FDA created guidance on how to balance safety with protocol deviations [37] and OHRP clarified the scope of public health vs clinical activities [38].

However, IRB professionals and researchers agreed on the need for clear guidance—something IRBs did not always feel they could provide in the earliest days of the pandemic. We recommend that IRBs and federal oversight bodies continue to develop clear policies about allowable protocol deviations, standards for review of expanded access program protocols, and acceptability of electronic platforms for documentation of informed consent.

Biases and misperceptions

Many IRB professionals and researchers described how public and professional misperceptions impeded research on COVID-19. Many in the US public held misperceptions about COVID-19 diagnosis, severity, and treatment that affected the progress of research. Beliefs about the risks and severity of COVID-19 were partly affected by political ideology, as different information ecosystems provided different analyses and sometimes even different facts. The media likely also contributed to public misperceptions about the efficacy of unproven drugs. Over time, this faith in unproven therapies was compounded by a skepticism about vaccines despite rich data demonstrating safety and efficacy. Some clinicians and researchers also believed in the efficacy of unproven treatments, which led some clinicians to withdraw patients from studies if they were enrolled on the placebo arm [39]. Our results support previous findings that political pressures were influencing COVID-19 research agendas. When President Trump touted the efficacy of hydroxychloroquine to treat and prevent serious COVID-19 infections (without sufficient evidence), there developed an urgency to prove the efficacy (or lack thereof) of this drug. An analysis of data from ClinicalTrials.gov in June 2020 found that 237,000 patients were to be enrolled in trials of hydroxychloroquine, which represented 35% of all participants in COVID-19 clinical trials [40]. Eventually, the RECOVERY study in UK found that hydroxychloroquine was ineffective, although some in the media continue to extol the benefits of this medication for COVID-19.

Researchers might feel powerless when they encounter this multitude of biases, misperceptions, and political pressures that drive research agendas. This experience can also frustrate IRB members who are trying to provide sufficient protections for participants. Given the current media environment and information silos within social media, information and misinformation have equal access to enormous platforms. Many of these problems seem intractable, especially for individuals operating within the research system. Any attempts to limit the spread of misinformation will raise concerns of infringing on freedom of speech. Any centralized national oversight to address institutional biases and politics will raise concerns of government overreach or misplaced authority. Also, centralized authority might simply concentrate political pressures and exacerbate the problem.

Political pressures and misinformation are inevitable in any society. Rather than preventing these pressures, perhaps researchers and IRB members should identify ways to leverage these pressures to positive ends. For example, rather than refusing to develop or approve another study of hydroxychloroquine, perhaps researchers should include hydroxychloroquine as one of several treatment arms to encourage public interest and enhance recruitment. Rather than arguing that IRBs need more time to evaluate proposals, perhaps IRBs can utilize the urgency to argue for more staffing or additional incentives to encourage IRB members to expedite processes to the fullest extent possible. Redirection, rather than resistance, could be a useful approach to counteracting misinformation and political pressures.

Institutional and inter-institutional conflicts

Some COVID-19 research was impeded because of conflicting interests within or across institutions. Within institutions, the decentralized structure of academic institutions created barriers to collaboration and communication between departments. These conflicting interests were also present across institutions, related to intellectual property, access to patients, invitations to join multisite clinical trials, and access to databases and biorepositories.

Given these myriad barriers, it is unclear if the US academic and healthcare systems can facilitate a timely, coordinated response to future public health emergencies. One could argue for more investment of resources to facilitate coordination, but how should these resources should be allocated? NIH leveraged existing Clinical and Translational Science Award (CTSA) hubs in 2016 to develop Trial Innovation Network with the goal of expediting clinical trials by creating new alliances and minimizing roadblocks to research. This investment in CTSAs and research infrastructure likely improved coordination and prioritization of studies during this pandemic [41], but IRB members and researchers still identified major challenges. The US was flooded with small studies and investigators experienced difficulties joining multi-institutional studies. Addressing these barriers will require regulatory reform and stronger incentive structures for collaborations. Perhaps the NIH could develop a panel that prioritizes a small number of experimental therapies and creates standardized research protocols. The NIH could host a central IRB and pay for experimental agents and staff effort. Any site that demonstrates capacity to perform these studies and recruit sufficient patients could join these clinical trials and access government funding. This funding could be tied to a mandate that sites agree to standard data-sharing agreements and utilize a central IRB at the NIH. Perhaps models could also be developed for patent-sharing among collaborating institutions, as we believe that protecting intellectual property impedes the open sharing of data and resources. Such an approach would address some barriers identified by our respondents while adapting to the decentralized nature of the US academic and healthcare systems.

Risks of harm

When COVID-19 emerged, scientists had very limited knowledge about how to prevent or treat COVID infections. These unknowns created the potential for novel risks of harm to research participants and research staff. Shortages of personal protective equipment further exacerbated this challenge. However, IRBs and researchers across the country seemed to have taken sensible approaches to ensuring safety of all involved, such as permitting teleconferencing rather than in person visits, accepting e-signatures for informed consent, and making determinations about allowable protocol deviations. Furthermore, federal guidance has encouraged this flexibility and adaptability [37, 38]. Protecting research participants is the primary responsibility of all researchers and IRB professionals [42, 43]. As COVID-19 surged, IRB members and researchers seemed to respond quickly and effectively to ensure the safety of participants and staff while supporting the continuation of essential research. This response represents another major successful collaboration between IRBs and researchers.

Pressure of the pandemic

The urgent need for data in the pandemic created expectations for IRB professionals and researchers that were difficult to sustain. IRBs experienced challenges with sufficient staffing, coordinating remote work, and adapting to significant increases in workload. Researchers similarly felt an urgency to write and submit protocols to answer essential questions about the virus. These pressures also led to researcher opportunism, where some individuals used the opportunity to publish low-quality studies that might have been driven by self-interest. One group performed a Medline search for COVID-related studies in May 2020 (5 months after the first US case of COVID), identifying more than 15,500 articles [24]. Another group estimated that 137 COVID-related papers had been published daily since February 2020 [26]. This vast amount of publications did not necessarily provide clarity or definite answers. These studies often were poorly designed, observational, anecdotal, and written in short format [24].

The potential harms of expediting publication are evidenced by notable retractions from major journals. By late 2020, 33 COVID-related studies had been retracted, withdrawn, or noted with concern because of either data falsification, methodological concerns, inaccurate interpretation of results, authorship disputes, or participant privacy issues [22]. Beyond personal responsibility of researchers and editors, perhaps journals could create an “expedited review process” that involves review by a smaller collection of standing reviewers or editorial board members. To incentivize these expedited reviews, perhaps content experts could receive remuneration for high-quality, expedited reviews. Furthermore, journals might publish articles with reviewer criticisms. This approach could expedite publication while maintaining transparency about study limitations. Furthermore, perhaps the National Library of Medicine (NLM) might consider funding high-quality, ongoing systematic reviews of data during public health emergencies. Such structured, ongoing evaluations could provide up-to-date understanding of collective knowledge while also highlighting major gaps in the literature. Furthermore, this dedicated funding might expedite the completion of these reviews by providing staff support. NLM could also partner with a high-profile journal to expedite review, publication, and dissemination of the current state of evidence at routine intervals.

Preparing for the next global emergency

The COVID-19 pandemic has served as a test case for whether the US is prepared to handle a major public health emergency. There have certainly been areas of success. For example, pharmaceutical companies and the federal government collaborated to develop remarkably effective vaccines in record time. Our study also showed how IRBs and researchers overcame some of these barriers by collaborating effectively and taking on enormous workloads and responsibilities for the benefit of society. However, we cannot necessarily rely on heroic individual efforts to save us in the future. These efforts were possible because most IRB members were not infected and did not require hospitalization. We did not require emergency IRB review procedures, but it would be reassuring if the OHRP worked with IRBs to establish procedures for delegating review when necessary and for supporting rapid review of greater than minimal risk research.

However, our study and review of literature also identified several areas of unresolved barriers and challenges that might hamper future responses to public health emergencies, ranging from individual behaviors to structural and regulatory issues. These unresolved questions pose an ongoing threat to the health and safety of the US in future public health emergencies. Efforts to address these challenges must account for multiple levels of influence and contributing factors. Implementation science is a field of study that explores the “constellation of processes intended to get an intervention into use within an organization; it is the means by which an intervention is assimilated into an organization. Implementation is the critical gateway between an organizational decision to adopt an intervention and the routine use of that intervention” [44]. Previously, we argued that using the lens of implementation science in ethics is essential to develop actionable, specific ethical norms [45]. In proposing and developing solutions to this inventory of challenges, these solutions should be evaluated using an implementation framework (such as CFIR) [44] and with ongoing stakeholder engagement.

Study limitations

This study focused on the U.S. context. While some challenges are likely to generalize to other national contexts (e.g., triaging access to COVID-19 patients), many issues will not. Our open-ended survey approach enabled us to get to the field quickly, by April 21, 2020. However, because such a format is more demanding on participants, we kept the survey very brief, perhaps limiting the amount of information captured. Further, our criterion-based sampling gave us a large sample of key informants; however, it does not allow us to state in a generalizable manner the percentage of people who experience a challenge or endorse a response.