Patient- reported reasons for non-participation in a COVID-19 therapeutics clinical trial: Findings from a multi-center investigation

Abstract

Background

Early in the pandemic, there were no evidence-based treatments for SARS-CoV-2, creating an urgent need to identify effective therapeutics. However, public participation in medical research is low; trial enrollment in the US is typically 10–20%. Thus, the aim of this study was to identify common themes underpinning patient reasons to decline participation and evaluate the impact of specific contextual factors.

Methods

This sub-study was conducted in five VISN-1 Clinical Trials Network participating facilities from 4/10/2020–2/3/2021. The trial evaluated the addition of the IL-6-inhibitor, Sarilumab, to the current standard of care for inpatients with moderate-to-severe SARS-CoV-2. Consent procedures varied by site and included fully in-person and fully remote processes. Reasons for declining enrollment were collected among eligible patients who declined to participate but agreed to answer a short follow-up question. Qualitative data were analyzed using directed content analysis. Enrollment rates were assessed using simple, descriptive statistics.

Results

N = 417 COVID-19 positive inpatients were screened and 53/162 eligible patients enrolled. Enrollment varied across study sites and by study period. Prior to identification of effective treatment, the enrollment rate was 10/11 (91%) versus 43/144 (30%) during the later period of the study. N = 85/102 patients who did not enroll answered the follow-up question. The most commonly reported responses were: concerns about the study drug and participation in clinical research in general, comorbidity concerns, competing priorities, external factors, and external advice and influence from family members and clinicians.

Conclusions

Identifying reasons behind declining to enroll may help investigators develop strategies to increase research participation.

1. Background

Since the identification of COVID-19 in early 2020, over 500 million people have become infected, and an estimated 6.2 million deaths have occurred worldwide [1]. Given rapid spread and high case fatality rate early in the pandemic, there was an urgent need to identify effective therapeutics to improve clinical outcomes. Due to the effectiveness of IL-6 blockade for other inflammatory states [2], Sarilumab, an IL-6 receptor inhibitor, was of great interest as a potential therapeutic option for management of severe COVID-19 [3]; however, early evidence supporting the potential effectiveness of the intervention was limited to case-series level-data.

Insufficiently powered trials and low patient enrollment remain a concern despite the recognized need to improve enrollment [4]. Pre-pandemic, and outside of a state of emergency, public participation in medical research was already low in the United States, estimated at 10% [5]. Low recruitment rates are the most common reason for terminating clinical trials in the United States, with typical enrollment rates ranging from 6 to 17% among eligible patients approached for participation [[6], [7], [8]]. Prior investigations have found that patient-driven reasons for low enrollment are varied and include limited perceived benefit, fear of the unknown or adverse effects, physician influence and presence of other comorbidities, inconvenience, and access to research information [9,10]. Other reasons include limited health literacy, distrust of medical system, access to transportation and loss of work required to participate, which are more marked in minority groups, limiting participation in clinical trials and leading to biased inclusion of populations with higher socioeconomic status. [9,11,12] Non-patient factors that impact enrollment include physician attitudes about the intervention being tested, and existing research infrastructure is necessary for clinical trials that require documentation of written informed consent.

The Designing for Accelerated Translation (DART) of emerging innovations in health Framework postulates uptake of novel interventions is a function of effectiveness multiplied by demand divided by the sum of risks and costs [13]. Based on this theoretical framework, in the setting of a global emergency with limited treatment options available, perceived demand for novel therapies would be high and perceived risks and costs low, leading to higher than typical engagement in clinical trials to access therapeutic options. However, while perceived demand was high, during the pandemic, patient enrollment also faced a new set of challenges, including limited in-person contact between clinical investigators and potential participants, lower number of on-site research staff, potentially impacting communication, and a shift to remote and electronic informed consent processes, all of which may have introduced additional barriers to participation [14].

During the conduct of a COVID-19 clinical trial, investigators noted that patient enrollment was lower than initially expected. Given the substantial barriers introduced by the pandemic, team members hypothesized that reasons for non-enrollment may have been driven by challenges with complexity and communication, rather than factors more traditionally associated with limited engagement and enrollment. Thus, given the need to identify strategies to improve participation in COVID-19 clinical trials and overall in the United States, we report qualitative responses from patients who were pre-screened and potentially eligible for enrollment in a Sarilumab COVID-19 Therapeutics Trial conducted in five Veterans Affairs (VA) Medical Centers in the Veterans Integrated Service Network (VISN)-1 Clinical Trials Network but who declined to participate in the trial after they were given the opportunity to review the study consent form [3]. Our aims were to identify common themes driving patient decision making with regard to enrollment and to determine how specific contextual factors (e.g., availability of different treatments at different time periods) impacted enrollment rates.

2. Methods

2.1. Study sites, study design, intervention

This was an observational sub-study that was conducted in parallel with an open-label, adaptive, pragmatic, embedded randomized clinical trial across five Veterans Affairs (VA) Medical Centers in the Northeast from April 10, 2020 and February 3, 2021 (NCT04359901) [3]. The trial was initially conceived and implemented during a 10-day period from 4/1/2020–4/10/2020, with first enrollment 10 days after study conception. The five VISN-1 Clinical Trial sites included low and high complexity urban and rural inpatient facilities ranging in size from 67 to 549 inpatient beds. VISN-1 Clinical Trials Network full-time research staff supported this study at all five facilities, with participation at times limited due to research staff and clinician availability.

The intervention in the trial was a one-time dose of a medication (sarilumab, delivered subcutaneously). Clinical outcomes assessments were measured using electronic health record data only; no additional participation was required of patients after the one-time intervention, thus there was no need for ongoing patient participation. During the conduct of the trial, hospital inpatients were screened for potential eligibility and then approached for possible enrollment.

2.2. Trial eligibility criteria

Patients with a positive SARS-CoV-2 diagnostic test (PCR or antigen testing) within 4 weeks of hospitalization for moderate to severe COVID-19 and with symptom duration of <14 days were eligible. As with other trials initiated early in the pandemic, eligibility criteria changed over time. As described in detail in Branch-Elliman et al, during the initial study period, only inpatients with severe but non-critical COVID-19 disease based on a clinical risk score (high respiratory rate, hypoxemia, radiographic changes) were eligible. As more data emerged about COVID-19 therapeutics, eligibility criteria were expanded to include inpatients with more moderate disease. Most notably, criteria for hypoxemia were expanded from SpO2 < 90% to ≤94% to align with other early trials. Throughout the trial, use of additional medications to treat COVID-19 was permitted. The trial initially opened at one site on April 6, 2020 and expanded to include additional sites during the period from between April 24,2020 and May 13, 2020.

2.3. Recruitment and consent processes

Per VA guidance, research study staff were not permitted to contact potentially eligible patients without approval of the inpatient provider and the patient's expressed willingness to meet with the study team following initial introduction of the study by the clinical team; thus, all patients had initial contact from their clinical team, and agreed to meet with study staff prior to contact from the research team. Consent processes varied by trial site and ranged from fully in-person to fully remote processes.

During the period in which the study was enrolling, electronic signature documentation, for example via DocuSign, was not available. Thus, the remote consent process was highly resource intensive for the clinical research team, but was designed to minimize patient burden [15]. Full details are previously published. In short, potentially eligible patients were visited in-person by a clinical staff member and were given a paper copy of an informed consent packet. Potential participants were then contacted by a member of the study team, typically via phone (video calling was also approved, although generally not used). The research team then discussed the study in more detail over the phone after the patient had ample time to read the consent form in the packet, which was provided in paper format. If the patient or LAR opted to participate in the clinical trial, a witness was contacted and asked to join the call for the official consent process. The study team then reviewed the informed consent documents with the patient, or legally authorized representative, provided answers to patient queries, and confirmed intent to participate. After confirmation, the patient was instructed to sign the paper form. For research compliance purposes, documentation of the wet signature was then obtained via a variety of mechanisms that did not involve the patient.

2.4. Collection of reasons for non-participation

Patients or their LAR who indicated that they were not interested in enrolling in the trial were then invited to answer a semi-structured, open-ended question about their reason(s) for opting not to enroll. If the patient did not offer a response to an open-ended question and agreed to provide additional information, the patient was offered several categories of options (e.g., risks of study drug, unwillingness to be randomized to control arm, family members/relative do not support participation, unwillingness to participate in clinical research, patient satisfied with current treatment plan, and other) and asked if their reason fell into any of the broad categories.

Due to the rapidly changing nature of the pandemic, and the speed at which this trial was opened, qualitative data collection strategies about patient-reported reasons for non-enrollment were expanded over time. Early in the trial, the data collection process for reasons for non-enrollment was not formalized, thus study team members may or may not have documented reasons for non-participation. During the conduct of the trial, it became clear that participation rates were lower than were initially estimated by study team members and questions about how enrollment rates could be improved arose. Thus, the main study site requested that reasons why patients opted to not enroll be formally collected and documented in study logs. Because collection of this data was not required or recommended during the early months of the trial, more complete data are available for later time periods.

3. Data analysis

3.1. Quantitative

Trial enrollment rates by time period were assessed using simple descriptive statistics and compared via Chi-squared test. Study time periods were defined in the following way: time period 1 (before 06/25/2020), before any evidence-based treatments were available [16]; time period 2 (6/25/2020–8/26/2020), after dexamethasone, a corticosteroid anti-inflammatory, was recommended but before tocilizumab was recommended against; and time period 3 (8/27/2020–3/5/2021, the study end date), the period after IL-6 receptor inhibition was recommended against (this recommendation was changed later in the pandemic, based on additional study findings) [17]. Given the very low number of eligible patients during time period 2 due to low regional case rates of COVID-19 during this time period, only time periods 1 and 3 were compared. Of note, for the majority of time period 1, the study was only open and enrolling at 1 of the 5 study sites. During time periods 2 and 3, the study was open and enrolling at all 5 study sites.

3.2. Qualitative

Qualitative responses for opting not to enroll were analyzed using directed content analysis, which is a deductive method for evaluating qualitative data [18]. In directed content analyses, responses are categorized based on previously defined key concepts demonstrated in other contexts to contribute to non-enrollment [12]. In addition to previously-defined themes, directed content analysis also allows for the emergence of new themes that are identified during data analysis. After qualitative responses were coded according to the defined themes, two analysts (WBE, SRD) met to review categorization, and differences in opinion were adjudicated through discussion. Responses from patients who were screened but not eligible for participation were not included in the analysis.

3.3. Ethical considerations

The trial was approved by the VA Boston IRB (#3305) prior to data collection and analysis. This ancillary study was conducted with a HIPAA waiver and with a waiver of informed consent. It was determined that the HIPAA waiver approved for the purpose of recruitment allowed data on dates, sites, and reasons for non-participation to be used for the current study, without allowing any other information about the patients to be obtained.

4. Results

N = 417 COVID-19 positive inpatients were screened during the entire conduct of the trial, and 162 patients met eligibility criteria for potential enrollment. Of these 162 potentially eligible patients, 53 (32.7%) consented to enroll. A total of 102 (62.9%) patients or their legally authorized representatives declined to participate in the trial, and 7 were unable to give consent and were excluded. Among the 155 patients who either declined or consented to enrollment, 11 were approached in time period 1 and 144 were approached in time period 2. Enrollment varied across study sites, ranging from 5 to 100%. Two larger study sites with substantial existing research infrastructure and support accounted for the majority of screened and enrolled patients (Table 1 ).

Table 1.

Comparison of Patients who opted to enroll versus declined to enroll.

Variable	Screened (N = 417)	Eligible (N = 155)	Enrolled (%) (N = 53)	Declined to Enroll (%) (N = 102)	p-value⁎
Site
Site 1	11	10	4 (40%)	6 (60%)	0.220
Site 2	13	8	8 (100%)	0
Site 3	235	75	26 (35%)	49 (65%)
Site 4	114	42	14 (33%)	28 (67%)
Site 5	44	20	1 (5%)	19 (95%)
Study period
Pre-dexamethasone guideline-recommendation (Study period 1: <6/25/2020)		11	10 (91%)	1 (9%)	<0.0001
Post-dexamethasone guideline-recommendation (Study period 2 and 3 (after 6/25/2020–3/5/2021)		144	43 (30%)	101 (70%)

^⁎Pearson chi-squared.

4.1. Enrollment rates by time period and study site (quantitative)

During time period 1, before the first effective treatment (dexamethasone, a corticosteroid) was identified in clinical trials, the enrollment rate was 10/11 (91%). During time period 3, from August 25 until the end of the trial, after initial reports suggested lack of effectiveness of IL-6 receptor inhibition for management of severe COVID-19 and the treatment received a “recommendation against, except in the setting of a clinical trial,” the enrollment rate fell to 43/144 (30%) (p-value, <0.0001).

4.2. Site-specific contextual factors

The study sites with the highest rates of enrollment were urban facilities with substantial research programs and clinical research infrastructure; one of these sites utilized an entirely remote process, and the other a primarily in-person informed consent process. During Time Period 3, enrollment rates among these two facilities were similar (16/64, 25% versus 14/42, 33%). A third, relatively large, urban facility had substantially lower enrollment rates attributed to lack of availability of research staff to assist with study screening processes. Additionally, the site PI of this study reported that negative physician attitudes about the study impaired the ability of research team members to initiate outreach. Two of the study sites with lower enrollment were smaller, rural facilities with lower facility complexity. Lack of intensive care unit availability in these sites also reduced the number of patients with severe COVID-19 who were admitted and potentially eligible for enrollment.

4.3. Patient-reported reasons for non-participation (qualitative)

In total, 85/102 (83%) patients who did not enroll agreed to offer a reason for their decision. A summary of qualitative responses to the open-ended question are presented in Table 2 . Inter-rater agreement regarding the category for reason to decline participation among a sample of reviewed responses was high (87%).

Table 2.

Key themes and examples reported by patients who opted to decline clinical trial participation.

Themes	Examples/Patient Statements
Perceived limited benefit/Distrust	Patient factors •Didn't want to be a guinea pig (n = 1) •Internet research and thinks it's just another thing we are trying that doesn't work based on what he read (n = 1) •Patient does not believe in COVID and that it causes pneumonia and does not think it was cause for admission (n = 1) •Don't want to risk getting worse (n = 1) •Doesn't like needles (n = 1) External factors •Risk of getting more medication (n = 2) •Risks of study drug (n = 13) •It is experimental and don't know if it works (n = 9) •Been in too many research trials (n = 1)
Competing priorities/External influences and stressors	Patient factors •Worried about wife, who was admitted to hospital with COVID (n = 1) •Brother died of COVID 2 weeks prior and too overwhelmed to participate (n = 1) •A lot going on and very stressed and doesn't want another variable to that stress (n = 3) •Already enrolled in two other trials (n = 1)
Clinician/Relative recommendation	External factors •Patient/MD satisfied with current treatment plan (n = 27) •ID physician recommended not to participate (n = 1) •Family member against patient's participation (n = 6) •Doctor told him that he was doing better and he did not need to participate (n = 1) •Daughter felt trial would be too much (n = 1)
Comorbidities/Concern about impact of study drug by treating physicians/MD satisfied with current treatment plan	Patient factors •Schizophrenia with history of psychiatric decompensations (n = 2) •Recent aspiration pneumonia (n = 1) •History of coccidioidomycosis (n = 1) •History of psychosis⁎ (n = 3) •Concern about kidney issues (n = 1) •Advanced age (n = 2) •Admitted for stroke⁎⁎ (n = 1) External factors •Per infectious diseases, patient has sepsis/open wound (n = 1) •On medications for previous kidney transplant, physician determined not a good idea (n = 1)

^⁎Concern that the patient would not be able to consent given underlying psychiatric diagnoses.

^⁎⁎Patient had mild COVID-19 and was primarily admitted for management of stroke and felt not to be eligible for the study.

Patient reported reasons for non-participation fell into two broad categories: patient factors and external factors, such as physician influence. Under these two major categories, several sub-themes were identified, including patient concerns about the study drug and participation in clinical research in general, concerns about how underlying health and comorbidities might impact safety of participation, competing priorities, and external advice and influence from family members and clinicians.

Many patients or their legal representatives reported a perception that the treatment being tested in the trial offered limited benefit and distrust about participation in clinical research. This was expressed as concerns about not wanting to “be a guinea pig” and perceived potential harms of the medication without proven benefit. Presence of comorbidities that were actively complicating their clinical course or posing higher risk for decompensation should they receive sarilumab also impacted patient decision making. Competing priorities included outside stressors that were unrelated to the current disease status of the patient, such as illness in family members, that made consideration of participation in a clinical trial overwhelming.

In addition to patient-driven factors, patients reported that physician or family recommendation to decline was a major driver of their decision making and attitude regarding clinical trial participation. Patients reported being influenced by their clinicians, who relayed concerns to patients about the presence of comorbidities or and perceived lack of benefit of treatment for various reasons, including clinical improvement rendering additional treatment unnecessary.

5. Discussion

During the early phase of the COVID-19 pandemic, which was characterized by high mortality rates and a high proportion of severe disease, no effective treatments or preventative interventions were available. Due to these specific circumstances, authors hypothesized that patients and family members would be highly motivated to enroll in clinical therapeutics trials and typical barriers to participation and enrollment would be overcome, as approved alternatives were not available. The DART framework highlights the role of demand for an intervention for driving uptake of novel interventions [13]; empirical evidence supports that high-risk clinical situations with limited treatment options drove the speed and scope of uptake of interventions during the COVID-19 pandemic [19]. Study investigators theorized that demand would drive high enrollment rates in a COVID-19 therapeutics trial, but found that recruitment was lower than was initially anticipated. Thus, in the current study, we report an analysis of patient-reported reasons for non-enrollment in a COVID-19 clinical therapeutics trial to identify factors that drove decision making in the setting of a world-wide emergency. Findings are informative about the importance of a changing clinical context, specifically with regard to availability of medications and conflicting evidence about the study drug, on driving enrollment decisions; participation rates fell substantially after other interventions were found to be effective, likely reducing the perceived need and increasing the perceived risks of clinical trial enrollment.

During the period prior to availability of evidence-based treatments, enrollment rates were lower than initially expected. In this point-of-care clinical trial with limited need for follow up and relatively low burden on patient participation, the most common patient- or legally-authorized representative-reported reasons for opting not to participate included limited perceived benefit, competing priorities, physician or family influence, and presence of comorbidities leading to perception of increased risk of participation. Additional factors that may have impacted enrollment include site-specific factors, such as existing research infrastructure and support, which impacted ability to screen patients and conduct outreach to potential participants, physician attitudes about treatments, which may have changed over time, and the complexity of the informed consent process in the United States, which included a requirement for signed documentation of informed consent [15,20]. Interestingly, the fully remote consent process did not appear to play a major role in impacting patient enrollment decisions, potentially because complexity of the remote process primarily created barriers for the clinical teams and research staff, with relatively minor impacts on patients.

These findings suggest several potential mechanisms to improve enrollment directed at patients, providers, and administrative requirements and processes. Identifying reasons behind clinical trial non-participation and mapping these factors to strategies that may address these patient concerns may help improve participation in trials; based on the findings of our study and others, multi-faceted strategies targeting recipients likely to have different concerns will be needed. Simplifying administrative processes, integrating remote consenting processes to expand trial access, and reducing regulatory complexity are likely to be effective for improving enrollment by reducing burden on clinical and research study staff and patients.

Our findings about reasons for non-enrollment may be used to develop informed consent strategies and clinical trial designs that facilitate participation. Our findings suggest that some drivers of non-participation are actionable with interventions from research study teams, whereas others will require external influences to impact change. For example, beliefs that COVID-19 is not a real entity are unlikely to be changed by study research teams. However, we also found that information given to the patient by their clinical provider and from trusted friends and relatives were major drivers of patient decision making. Thus, it is likely that strategies that address provider perceptions about clinical need (e.g., demand) and appropriateness of the study intervention (e.g., through educational strategies to promote understanding of the research question and study drug) may be effective for improving patient participation. Additionally, increasing engagement with clinical teams about the study in real time, and providing clinical teams the opportunity to directly ask questions to the study team, may increase trust in the research process, thereby impacting provider perceptions about the study, and ultimately their recommendations to their patients. While not cited by patients, another factor that may have impacted enrollment was the complicated documentation of informed consent processes required; identifying ways to reduce complexity, for example by allowing witnessed phone consent and allowing electronic consenting solutions may reduce clinician perceptions about complexity and barriers to participation and thereby impact their recommendations to patients. While not specifically reported by patients, we also found that providers were impacted by traditional and social media influencers, and that their positions about the study question impacted enrollment. Addressing communications ecosystems and culture about participation in clinical trials may yield positive changes in the long-term.

Notably, this trial only required a one-time intervention delivery without post-discharge follow up, and thus the burden on patients for ongoing trial activities was extremely low; this low burden may have contributed to the relatively high enrollment rate compared to other studies in the United States, where clinical trial enrollment is typically under 20% [6,8,21]. Another novel feature of the study was the inclusion of sites that did and did not employ fully remote consenting processes; although these remote consenting procedures were quite time consuming for investigators, impact on patients was minimal. Notably, the site with fully remote consenting processes had the highest levels of enrollment throughout the trial, and also the highest level of support from study staff. It is notable that the fully remote interaction with research staff did not clearly negatively impact enrollment, suggesting that providing remote consent may be a strategy for increasing outreach and access to clinical trials, particularly among facilities with limited existing research support, experience, and infrastructure. Specifically, these findings suggest that emerging clinical research models with a centralized study team that provides remote research administrative support, for example through tele-health platforms, to institutions without their own established research program may be a successful strategy for increasing clinical trials access to small facilities without experience or support for clinical research investigations [22]. Simplification of administrative burdens, such as need for clinical research training to enroll patients in trials, and waiver of complex consent documentation practices, would also likely expand clinical trials access and thereby improve uptake [20].

Data on patient-reported reasons in the US for not participating in SARS-CoV-2 clinical trials is limited, and most of the existing data is focused on clinical trials in oncology, outpatient SARS-CoV-2 investigations, or studies assessing patient participation outside of the United States [[23], [24], [25]]. These prior studies suggest that physician attitude and enthusiasm about the trial, and about the therapeutic agent in question, are important drivers of patient decision-making, either by not offering the trial to patients or expressing concerns about comorbidities that impact the safety of participation. [23,26,27] These themes were also found in our study. Limited perceived benefit or distrust has previously been shown to limit enrollment in clinical trials, particularly in underrepresented groups, [9,10] which may have been amplified given the rapidly changing landscape of available COVID-19 therapeutic agents. Our clinical trial was numerically under-represented for Black/African-American (8%) and Hispanic/Latino (4%) patients relative to the total US population (12% and 18%, respectively), but with only 52 patients enrolled overall, it is not clear whether such under-representation was a true finding or due to the demographics of the region in which it was conducted. The nature of this sub-study (retrospective analysis of patients who did not consent to enroll in the trial but who voluntarily answered a follow-up question) means we are unable to determine demographic information about patients who declined enrollment; thus we cannot compare the demographics of those who enrolled versus those who did not [3]. Findings from other studies and ours suggest that additional research into health communications interventions that identify strategies to promote trust are needed to improve clinical trial enrollment in the US.

To our knowledge, patient reasons for opting not to participate in inpatient COVID-19 therapeutics trials including those with fully remote consenting processes have not been previously reported. Fewer patients have enrolled in US clinical trials than in clinical trials in other settings and countries, most notably the UK [28,29]. Although not specifically addressed in this qualitative study of patient-reported factors, our experience running the clinical trial suggests that the long and cumbersome, often remote, informed consent documentation processes may have served as a deterrent, particularly for clinical providers. In total, the remote consent process was estimated to take 6–8 h of research staff time due delays in receiving approval from clinical staff and, due to infection prevention measures, required multiple complicated document transfers to obtain consent and enroll, in a clinical trial involving a single therapeutic agent. This process is in stark contrast to the process in the UK RECOVERY Trial, in which a simplified consenting process was permitted, and one enrollment process was used for multiple agents [29]. Recently published pragmatic designs with limited patient and provider burden for participation, such as the ADAPTABLE trial, also suggest that simplifying processes for both patients and providers may also improve enrollment [30,31].

Based on enrollment rates before and after publication of the clinical trial demonstrating the effectiveness of dexamethasone and after release of early studies that suggested a lack of benefit to IL6-blocking therapies, the availability of other medications and the national treatment guidelines also impacted decision making, despite allowance for use of the medication in the setting of a clinical trial in the official recommendations. In this study, we are not able to determine whether the availability of other medications impacted patient attitudes about enrollment, physician attitudes about enrollment which then impacted patient attitudes, or both, but prior work suggests that the availability of other therapeutic options and conflicting evidence were both strong determinants of physician prescribing behaviors during the pandemic [19]. However, in our experience, we found that there was an impact of the availability of other evidence-based treatments on the decision making of both patients and providers. Anecdotally, early in the pandemic, when no evidence-based treatments were available, providers were more enthusiastic, and this translated into higher patient participation. As more treatment options became available, and early trials about other IL6-blocking therapies became available, attitudes and enthusiasm changed. In addition, some site principal investigators and our own anecdotal discussions with providers suggested that information disseminated on social media impacted their clinical decision making and counseling of patients. At least one patient also reported that he read about the treatment on social media and that was a major driver of his decision not to enroll. Future studies should investigate these factors more formally; as information is increasingly shared through non-traditional settings, these external factors may only grow in importance in the future.

The limitations of this report include the setting, which was conducted in the VA VISN-1 Clinical Trials Network. Given the nature of the HIPAA waiver, information regarding ethnicity, education and socioeconomic status was not available, which are known factors that impact clinical trial enrollment that we were not able to assess in this study. [9,11,32] Reasons for non-participation in this population may be different than in others, particularly younger populations and populations that include more women. Insufficient data were available to assess the impact of a legally authorized representative versus the patient providing consent; it is possible that factors driving non-participation varied between these two groups. It is also possible attitudes and beliefs about clinical trial participation may vary by region. In addition, early in the study, only patients with more severe disease were eligible; it is possible that disease severity was a factor that drove high enrollment rates early on. Although physician attitudes are known to be an important factor in patient decision making about clinical trials, these were not assessed in this study, and thus only patient perceptions of provider attitudes are reflected.

6. Conclusions

Increasing enrollment into COVID-19 therapeutics trials and therapeutic trials more broadly remains of highest priority as patient enrollment remains one of the biggest barriers for conducting adequately randomized and powered trials. To our knowledge, this is the first attempt to qualitatively identify barriers to COVID-19 therapeutic trial participation and to include an assessment of how remote consenting processes impacted enrollment. Understanding actionable reasons and attitudes behind declining to enroll may help investigators address them during the recruitment and consenting process, for example through educational and engagement strategies directed at clinical providers, and thereby increase enrollment and retention rates in future studies.

Funding

This material is the result of work supported with resources and the use of facilities the VISN-1 Clinical Trials Network and the VA Boston Healthcare System. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Conflicts

WBE, PM, and JMS were site investigators for a study funded by Gilead Sciences (funds to institution). WBE was supported by NIH NHLBI 1K12HL138049–01 during the conduct of the clinical trial and reports funding from the VA Health Services Research and Development Service. All other authors report no conflicts of interest to report.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

Data will be made available on request.