Resilient design: decentralized trials recovered faster from the impact of COVID-19 than traditional site-based designs

1

1. Introduction

Conducting a successful clinical trial is arduous at the best of times. Protocol development, site selection, contracting, and regulatory compliance all take significant time and attention from multidisciplinary teams before a single participant is enrolled [1]. Once active, many trials face challenges recruiting sufficient participants in a timely fashion and maintaining their engagement throughout the study [2]. Study participants often report their trial experience is burdensome, confusing, and via time lost working or travel expenses even costly, despite their strong motivation to contribute to science and their hope to receive a beneficial treatment [3]. Unfortunately, the COVID-19 pandemic exacerbated this situation. Participants following social distancing guidance missed numerous visits. Site staff were absent or worked remotely, leading to delays in data processing, limited sourced data verification, and a disrupted vendor supply chain [4]. As a result of the pandemic, up to 80% of all non-COVID-19 research in the US was paused or halted [1], only 57% of expected trials were started in 2020 [5], and cancer patients who would have been eligible to take part in studies became significantly less likely to enroll while the virus persisted [6].

From March 2020 onwards, regulatory bodies including the FDA, EMA, and MHRA issued guidance to investigators which outlined mitigation strategies to maintain patient safety, data quality, and trial integrity [7]. In practical terms, these included reducing the number of face-to-face visits, slowing recruitment, and changing where lab tests were processed. They also opened the possibility of using a range of technology approaches to replace or augment traditionally designed data collection. Rules changed to allow electronic consent, telehealth video visits to replace specified site visits, electronic distribution of patient reported outcomes (ePROs), and shipping trial medication directly to participants’ homes rather than hospital dispensaries. These new recommendations align with a preexisting but under-utilized approach to clinical research; what used to be known as “remote trials’ or ‘virtual trials,’ but what the FDA (and industry) now refers to as ‘decentralized clinical trials’ (DCTs) [8].

Traditionally designed, exclusively site-based studies recruit, screen, and enroll participants face-to-face in a healthcare setting. This is typically an urban tertiary clinical facility associated with an academic institution [9]. According to the scheduled protocol, participants travel on a regular basis to sites to complete assessments, questionnaires, lab tests, and take home their study medication. Such studies generate vast quantities of paper and require careful source data verification (SDV) to maximize the quality of data and prevent abuse or fraud. By contrast, fully decentralized trials use software and devices to manage nearly all aspects of consent, data collection, patient engagement, authentication, and data verification, typically in the participant’s home rather than a traditionally designed study site [10]. Data that can only be collected by a healthcare professional can be gleaned from ‘home health’ visits by trained workers. In between these two extremes lie hybrid decentralized clinical trials. Hybrid trials may take aspects of an site-based design but add decentralized digital elements such as extending study visits by health professionals into the home, collecting continuous data from wearable devices or replacing a notoriously unreliable paper diary with an electronic diary which can improve adherence and data quality [11], for example by authenticating participant identity, including time stamps, having more reliable scoring than paper-based forms, or by offering participants the opportunity to respond in their own language.

2. Decentralized trial resilience during COVID-19

As researchers and applied practitioners, we have been actively supporting decentralized clinical trials since 2015 through a cloud-based digital research infrastructure called THREAD, which has been described previously [12]. We compared the subject enrollment rate and subject discontinuation rate performance of over 1,300 deidentified traditionally designed site-based clinical trials using the proprietary Lokavant industry benchmarking dataset across a range of study phases, geographies, and therapeutic areas, and compared these to the performance of decentralized trials conducted on THREAD across the first three quarters of 2020. The datasets were min-maxed normalized (Figure 1) relative to respective reference points from January 2020. We noted a steep decline in participant enrollment from February to April 2020 for both traditionally designed and decentralized studies [13]. However, by May it was only the decentralized trials that had recovered and exceeded pre-COVID recruitment rates, whereas traditionally designed trials had a slower recovery across the rest of the year and never came back to pre-COVID levels. A similar pattern was found for study withdrawal by participants. Both traditionally designed and decentralized studies saw a rapid increase in discontinuation in March 2020 which never recovered for traditionally designed site studies, but which returned to baseline for decentralized trials within 2–4 months.

Figure 1. Normalized participant enrollment for decentralized trials (top line) compared to industry benchmarks during the initial COVID-19 pandemic outbreak (Image reproduced with permission from Lokavant [13]).

Evidence suggests that participants will become more interested in becoming participants in clinical trials once the pandemic is over, with 62% saying they have increased awareness of clinical research because of COVID-19 [14]. However, today fear of exposure at medical facilities has directly driven significant declines in participant study involvement [6]. Decentralized studies take advantage of increased research interest while providing participants a mechanism of decreased risk by conducting some or all of the study from home.

For example, in one fully decentralized registry for Duchenne Muscular Dystrophy conducted by THREAD, participant registration increased 50% in 2020 compared to previous years, driving recruitment for over a dozen clinical trials [15]. The use of eConsent, eCOA, surveys, and remote health solutions alleviate participant concerns while providing sponsors with robust data collection. While studies running when the pandemic struck may have struggled to adjust to changing circumstances, these data suggest a higher degree of resilience for decentralized trials.

It remains to be seen whether COVID-19 will serve as the watershed for the normalization of decentralization in clinical trials, but industry data indicates a continuing trend in that direction. A 2020 report from the research firm Jefferies projected that 25% of all clinical trials conducted will be decentralized by 2025 (a 60% compound annual growth rate from 2020–2025) [16]. A 2020 Survey from the same group found that half of the clinical research organizations (CROs) surveyed expected decentralized trial elements would be used for most new studies [16]. Research sites and investigator staff are expecting the shift as well. In their 2020 Global Survey of Clinical Trial Investigators, McKinsey & Co found a large majority of investigators representing researchers in the UK, US, and France expected to see increased use of telemedicine and remote visits in research after the COVID-19 pandemic [17]. Perhaps the most compelling evidence for the future of decentralization in clinical trials comes from broader trends in how consumers are receiving care. An internal CMS analysis of Medicare Claims data from March through June of 2020 found that claims for telemedicine-based healthcare interactions climbed significantly from a pre-COVID 19 average in the tens of thousands to well over 1.6 million a week [18]. This fundamental shift in the way patients receive care in routine practice is one of the most significant impacts of COVID-19 in healthcare. We expect this trend will reset expectations for how clinical trial participants engage with investigators in a research context: in-person visits with paper consents and COAs will no longer be acceptable to a participant and decentralized trial elements will become essential to recruitment and retention.

These industry trends mirror our experience where we have seen a significant increase in decentralized study interest from both sponsors and CROs. This includes the use of eConsent, telehealth, home health, and remote data collection methods. Studies specifically targeting COVID-19 have incorporated several decentralized elements. For example, the University of Cambridge’s HEAL-COVID study is a pragmatic platform trial of widely-available generic medications attempting to reduce mortality and hospital readmissions in the post-hospitalization period of COVID-19 [19]. The study takes a hybrid approach with participants recruited and consented on discharge from hospital with a smartphone app to capture ePROs on breathlessness, fatigue, and other symptoms in seven languages, along with analysis of linked real-world data from NHS digital. Such rapid adoption of digital technologies in a traditionally conservative health system like the UK’s NHS demonstrates that many of the barriers to adoption had always been surmountable with sufficient motivation.

3. Limitations to decentralized approaches

Digitization is rarely a panacea, and any innovative approach has limitations. Patient safety is the paramount concern in any investigational study and adopting novel approaches requires complex safety management processes. The lack of face-to-face interaction may be a disadvantage for those participants who enjoy clinical contact or may get better care as a result [3]. The disparities of the ‘digital divide’ may be felt most by those who are older, less educated, or who lack access to high-speed internet connections, particularly in rural areas [20]. Adopting decentralized approaches to clinical research requires adjustment for professionals, who may resist losing valuable contact time with patients. Clinicians in some countries like the US are licensed to practice only regionally (e.g. for specific US states). While some degree of state-to-state license reciprocity was initiated during the initial wave of the pandemic, these are being re-restricted. Academic clinical trial units are typically reimbursed for use of their facilities and staff time, so financial changes will need to be made to avoid penalizing investigators who wish to adopt decentralized methods. Finally, not every study medication will be suitable to be taken orally or shipped to patients’ homes, with oncology trials a particular area for development [7].

4. Expert opinion

The devastating human, political, and economic impact of COVID-19 will be felt for decades to come. While the speed with which researchers have pivoted their trials to adopt decentralized methods to keep their studies going has been both impressive and admirable, patients may wonder why it has taken so long to embed technological features which make their participation less burdensome? One potential answer comes from the former FDA Commissioner Scott Gottlieb who commented that while decentralized approaches may generate reliable evidence more efficiently, they were potentially incompatible with the business models of clinical research organizations. With billions of dollars of revenue predicated on billing for the human hours to administer or check paper questionnaires [8] there is a risk that fiscal rather than scientific or participant interests may drive a return to traditionally designed studies. Innovative digital approaches have benefitted from a relaxation of regulations in research-intensive geographies, but in their guidance the EMA specifically state ‘the regulatory flexibilities are not intended to be kept once the COVID-19 crisis is over’ [21].

We believe going back to the traditionally designed approaches would be a mistake for three reasons: inequity, efficiency, and failure to adopt innovation. At the time of writing, the COVID-19 pandemic is still causing disproportionate mortality, morbidity, and disruption to the most disadvantaged people globally. Traditionally designed clinical trials have been most accessible to white, educated, younger, healthier patients in rich industrialized countries. Decentralized protocols have the potential to incorporate a wider and more representative population [9]. In terms of efficiency, there is early data to suggest that decentralized approaches are faster, cheaper, and produce higher quality data than traditionally designed approaches – however there is a paucity of systematic approaches to study this such as randomizing arms of large trials. While this might be burdensome to implement there should be sufficient interest from national funding bodies such as NIH or NIHR to study decentralized approaches from a neutral perspective. Finally, the potential to go far beyond the simple questionnaires and clinical examinations of traditionally designed approaches is a real opportunity for innovation. The explosion in wearable devices and digital sensors with medical-grade capabilities represents a trove of valid and reliable data capture systems available in the setting that matters most to patients living with illness: their homes. After a century of dragging in research participants to wait patiently for us on hard chairs in cold waiting rooms perhaps it is time we came out to meet them where they live.

Declaration of Interest

P Wicks is a consultant to THREAD; has received research funding from the University of Cambridge for work on a decentralized trial, HEAL-COVID; is an associate editor at the Journal of Medical Internet Research and is on the editorial advisory boards of The BMJ, BMC Medicine, The Patient, and Digital Biomarkers; is employed by Wicks Digital Health Ltd, which has received funding from Ada Health, AstraZeneca, Baillie Gifford, Bold Health, Camoni, Compass Pathways, Coronna, EIT, Endava, Happify, HealthUnlocked, Inbeeo, Kheiron Medical, Lindus Health, Sano Genetics, Self Care Catalysts, The Learning Corp, The Wellcome Trust, THREAD Research, VeraSci, and Woebot. J Price, N Goodson, J Warren, & J Reites are employees of THREAD The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.