Abstract
While research oversight is necessary and desirable for some types and elements of research, we must also recognise that research review is itself a healthcare intervention and should therefore be subject to the same evidence-based requirements demanded of other healthcare interventions.1 Inefficiencies in any part of the research process – including its regulation, governance and ethics review – are harmful to human health. It is one of five elements1,2 that contribute to an estimated 85% of avoidable health research waste, which costs over US $100 billion per year.3,4 The inefficiencies are caused by complex administrative processes that lead to delays, submission of multiple forms (many of which seem to have no clear purpose) and requirements that much low- or negligible-risk research undergoes disproportionate review, often by large ethics committees constituted with higher risk projects in mind. In addition to the financial costs, these inefficiencies can impose considerable burdens (in terms of time) on researchers and members of research ethics committees. 5
Delayed or altogether abandoned research impedes the identification of effective and ineffective and harmful treatments, thus compromising the health and well being of patients.6,7 For example, by comparing US and UK requirements for informed consent for participants in the ISIS-2 trial, Collins estimated that about 10,000 unnecessary deaths were directly attributable to whatever it was that slowed recruitment into the trial in the United States, slowing results, publication and uptake. 8 Conversely, high-quality and timely research can rapidly improve health and reduce mortality, as has been illustrated by the vaccine research and some drug treatment trials done during the COVID-19 pandemic.
Variations among countries can both illustrate the problems and provide potential solutions by examining the different processes used. For example, in a recent pseudonymous, non-interventional, observational European study, the time between application and first approval in the 16 countries varied between 7 and 300 days, with a median of 60 days. 9 There was also considerable variation in whether informed consent was required (9/16) and from whom. A qualitative survey of researchers in human and organisational factors identified several inefficiencies and potential solutions to this ‘ethics review roulette’, 10 and proposed a two-step review process that included pre-application (does the study need an approval?) and for studies that need approval, a separation into low risk (expedited process) or high risk (regular process). Several countries, such as the UK and USA, do provide two-tiered processes, and for many countries this would represent a substantial improvement. For all countries, there are likely many further possible efficiencies. Hence, these different processes and rules used across and within countries are worth a closer examination, to learn how to streamline ethics review processes so that they are proportionate to risk, timely and efficient. This article sets out some options for improvement, with examples from different countries. While we focus on ethics reviews in particular, similar issues and solutions will apply to research governance procedures. In an accompanying article, 1 we discuss some of the consequences of adverse effects of confused research ethics.
Methods to reduce the waste
Examples of the ‘4Rs’.
Reduce
The ideal change is to reduce, which achieved be made in several ways:
This variation creates inefficiency for studies conducted within countries, but it is an even greater problem across national boundaries.
To address the problem of unnecessary submissions, the UK has developed a self-certification scheme to identify research that does not require a Research Ethics Committee review. This involves answering four sets of questions about the proposed project. Once complete, the system automatically issues a document confirming that the project does not need an NHS Research Ethics Committee review.
11
A similar self-certification system operates in the United States,12,13 for research that is eligible for exemption under the United States regulation 45 CFR 46 (e.g. secondary use of biospecimens, educational tests, etc.). An ‘exempt study submission form’ is filed by the researchers with their institutional review board. So-called ‘n-of-1’ trials are usually done for the benefit of individual patients and delaying this with unnecessary ethics reviews harms patients. To guide when n-of-1 trials should require ethics review, a group has developed a simple flowchart,
14
but we are not aware of jurisdictions currently using this.
Review turnaround time
During the COVID-19 pandemic, a number of trials have received expedited review. The three-day turnaround was crucial to the RECOVERY trial’s success in recruiting in the first wave of COVID-19. 15 Following on, the UK Health Research Authority is now piloting a fast-track governance and ethics review process for medicinal trials. This aims to provide a final opinion in less than 15 days from submission.16,17 Similarly, the UK’s Ministry of Defence Research Ethics Committee (MODREC) uses an expedited review based on ‘Urgent Operation Requirements’ to review projects within a few days. This process has been used for both Ebola and COVID studies. 18
Documentation requirements
The required paperwork expands, burdening both applicants and Research Ethics Committees and causing inevitable delays. What is absolutely necessary for review must be agreed and then this, and this alone, provided for review.
Application pathways
The UK Health Research Authority is piloting a Combined Ways of Working process for clinical trials of medicinal products. This brings together a single application for the Clinical Trial Authorisation and Research Ethics Committee opinion. 19
Duplication of ethics review
The number of bodies involved in review increase decade by decade, with consequent duplication. Their roles and responsibilities must be defined and agreed, with mechanisms to resolve dispute without burdening the applicant. The UK Health Research Authority has established partnerships with others involved in research review. 20 For example, research ethics committees often (re)-review the study design and statistical issues. 21 Such review should be done once, by those with training and experience in statistics and design. The Research Ethics Committee should ask for evidence that such a review has been done, ensure it meets acceptable standards and then accept its result.
Unnecessary research by requiring systematic reviews of relevant existing research showing justification for new primary research
From 2006 to 2014, a 35% reduction was achieved at Radboud University by requiring systematic reviews before approving new research. 22 To facilitate this, ethics review should require appropriate registration and reporting of research, as is being attempted by the UK’s Health Research Authority through their ‘Make It Public’ campaign. 23
Re-use
Re-use or adapt applications or components of applications previously approved by a research ethics committee. For example, the methodological and data analysis components of many studies can be very similar, so subsequent ethics reviews might be more efficient if they focus on relevant changes such as the justification for using different populations, recruitment processes and consent, rather than re-reviewing all aspects of the project. Research ethics committees often re-review design and statistical issues, 21 and also risk assessments or safety arrangements. However, such reviews should be done once and preferably by those with training and experience in these areas. Ethics committees should merely ask for evidence that such review has been done rather than attempting a review of something they are not qualified to judge. Some other re-use options are:
Promoting research platform design
Many studies have very similar protocols, and review might then be confined to the changes made, similar to a protocol amendment process – as was the case of the RECOVERY trial. 15 The UK’s Health Research Authority is currently developing guidance for Research Ethics Committees on such platform design. 24
Requiring only a single Research Ethics Committee review for multi-site research
Perhaps the single most important ‘re-use’ is to prohibit requirements for multiple ethics applications. For example, one reason the UK’s RECOVERY trial was able to rapidly expand to 100 recruiting centres was the single approval process in the UK, 15 whereas the SOLIDARITY trial’s recruitment was slowed or even thwarted by the multiple ethics committee requirements within and between the 30 countries involved. While single ethics review models have become more common (e.g. in the United Kingdom), particularly for clinical trials, they are still far from universal in all countries and for all types of research.
Recycle
Recycle review elements by creating standard requirements and components for different types of research. Research ethics review currently treats each research project as if it is entirely new. This ignores the spectrum of research – from direct replication of a published study to innovative interventions and designs. Direct replications are akin to multicentre studies and could just require submission of the previous approval. We could also re-use standard trial or survey consent forms, which, if already approved, would be noted in the submission. Near identical protocols might also be given generic approval.
Pre-approval of generic emergency protocols: For example, the Medecins Sans Frontiers Ethics Review Board approves generic protocols within the context of infectious disease outbreaks to make research in emergencies feasible. This involves a decision by researchers about what topic to research in the next emergency, and submission of a ‘generic’ protocol for pre-approval before the outbreak occurs/location is known. Once outbreak location is known, a finalised protocol is submitted for approval (local committee and Medecins Sans Frontiers).
25
Blanket approval for randomised ‘n-of-1’ trials that compared two approved treatments requiring only a standard consent form was given by Queensland University (Glasziou, personal communication). This concept could be extended to many other types of research.
Replace
Replace the current ethics review processes by alternative processes. For example, low-risk research guidance should recognise that a ‘one-size-of-ethics-review-fits-all’ is unnecessary and wasteful. Ethics review could be streamlined by a proportionate and flexible process that uses different approaches depending on the likelihood of benefits, harms and costs of different types of research and evaluation. Countries vary substantially in how they manage and classify research projects as at low risk or exempt from ethics review. In a recent comparison of research ethics review requirements in four countries, all countries exempted proposed research involving re-use of publicly available datasets, such as registry data. Although surveys of professional staff aimed at improving professional practice were exempt in many countries, this was not universal. Questionnaire or survey research was exempt in three of the four countries compared, albeit subject to stipulated conditions, such as minimal inconvenience.
26
Methods might include:
Proportionate review: The UK Health Research Authority’s Proportionate Review defines research studies judged to pose minimal risk, burden or intrusion for research participants, and provides an accelerated, proportionate review. A final decision is emailed to applicants within 21 calendar days. Self-certification: An alternative is to allow self-certification of low-risk research (e.g. as in the UK and the US, see above), with random audits in place to discourage inappropriate self-certifications. One recent estimate is that if Australia shifted from the current ethics review system for low-risk research to the model widely used in the USA, Australian research could save $160 million per year.
27
Chair’s discretion: Allow the committee chair to decide whether full review is needed. That is an important advance but can still require full application and delays.
Discussion
To improve on each of these ‘Rs’ will require dedicated and sustained work to design, implement and evaluate innovations. This requires a better understanding of which processes are necessary for research to proceed, and which are sufficient depending on the risk profile of different projects. Distinguishing clearly between important governance processes and proportionate ethics review processes is vitally important. Any changes implemented will need to include audit and monitoring of research ethics processes, such as timelines, proportion exempted, proportion not registered, proportion that are multicentre (and whether only one submission was required), etc. Regular reporting on these will be useful for researchers but also necessary as quality improvement measures.
To foster the needed changes, we need high-level policy changes recommending their introduction. 33 However, the changes need to be done in stages – such as the steps suggested by Brian Nosek: (1) make it possible (appropriate infrastructure and procedures); (2) make it easy (user interface and experience); (3) make it normative (dissemination in the researcher and review communities); (4) make it rewarding (appropriate incentives); (5) make it mandatory. 34 For example, to introduce proportionate review an agreed taxonomy of risk with clear boundaries is needed, illustrated with examples. That taxonomy needs to be made easy for researchers and REC staff to use, and for all those in the research process to become accustomed to its use. Finally, it might be used as part of a mandatory process which allowed low-risk research to be readily exempted.
Whatever changes are set up, there will be inevitable disagreements. Hence, organisations and national authorities should consider an ‘appeals process’ to rule on disputes between researchers and individual research ethics committees. As one example, the UK’s Health Research Authority has an ‘Appeals Manager’. 35 Such rulings could then be used in further refinements to the regulation systems.
The needed changes will take some years to achieve, but the potential gains for research, and more importantly, for patients, are great, as has been illustrated by a few exceptional examples during the COVID-19 pandemic.
Historically, pandemics have forced humans to break with the past and imagine their world anew. This one is no different. It is a portal, a gateway between one world and the next. We can choose to walk through it, dragging the carcasses of our prejudice and hatred, our avarice, our data banks and dead ideas, our dead rivers and smoky skies behind us. Or we can walk through lightly, with little luggage, ready to imagine another world. And ready to fight for it.
36
Let us recognise that for the benefit of all, research and care should be conjoined partners. This, the pandemic can teach us.
