The gaze of scrutiny in the relationship between research studies and Research Ethics Committees (REC), Institutional Review Boards (IRB) and other bodies which grant permissions for (funded) research to proceed is typically unidirectional. Such organisations and their members review research studies for the ethical, legal and governance issues they raise. Occasionally, roles are reversed: research studies engage in probing and critical exploration of the work of those gatekeepers. This paper represents one such example. We build on existing literature and the striking findings of an audit-type study of the ethics and governance approvals of one of our recent projects (“Prepared to Share?: a study of patient data sharing in complex conditions and at the end of life” [1]) and offer proposals for change.
The paper is structured as follows:
An introduction sets the scene of researchers’ attitudes towards ethics and governance approvals processes.
We then present an extensive review of the empirical and critical literature on RECs, IRBs and ethics and governance approvals for health research (here understood to include biomedical research, health services research, and health-related research informed by the behavioural and social sciences and the humanities). We present a more detailed account of the background literature than typical of an empirical paper, as the literature is dispersed and little known. We include scholarly work which goes further than the mainstream storyline around the ethical regulation of research, namely how vital the work of ethics committees is to prevent a repetition of, for instance, the Nazi atrocities in the name of science, the Tuskegee syphilis experiment or experiments like those of Milgram and Zimbardo. The publications that transcend this narrative do not (obviously) question the need for ethical research, but probe into or challenge the extent to which the current systems are achieving that goal.
The respective scholarly field is far broader than appears at first sight, yet it lacks both ‘self-awareness’ as a distinctive area of study with core assumptions, questions, typical methods, exemplar studies, etc. and meaningful visibility to readers, at least outside the realm of bioethics. At the same time the work of RECs, IRBs and other permissions granting bodies affects all “human subjects research”. It is an oft-forgotten determinant of the contents and boundaries of our scientific knowledge.
Following the literature review, we present our empirical findings. These are from an audit-type study of the ethics and governance approvals of a project on patient data sharing at the end of life in a single locality in England. The trigger words of ‘data sharing’ and ‘end of life’ notwithstanding, this was, in the words of a research governance officer, a “low-key” (low-risk) study in terms of the ethical and governance issues it raised.
A further clarification of scope and terminology may be helpful. Most of the background literature addresses the ethics approvals of studies. However, study approvals are a broader enterprise, variably termed (at least in the UK) “research ethics and governance approvals”, “ethics and R&D (research and development) approvals”, “study assurances”, etc. Additional sign-offs are needed by R&D departments of participating organisations or university research offices (if the ethics approval has been granted by a committee unattached to a university) around compliance with regulations or the capacity of an organisation to host a project. Researchers involved in human subjects research often use “ethics approvals” as a synecdoche for this broader class of research ethics and governance approvals, not least because the latter are typically contingent on the former and because the bulk of the documentation is first prepared for the ethics review. Our paper concerns this broader class of approvals.
Finally, we outline six proposals for change in the systems of ethics and governance approvals, with potentially global application. Key features of these proposals are that they emphasise the enrichment of the ethical debate at least as much as the efficiency gains and that they rely strongly on opportunities provided by the growing sophistication and use of IT.
The negative set point of researchers’ expectations
Most researchers conducting human subjects research soon internalise an emotional response to the REC/IRB process and the broader system of research approvals. While researchers variably like or dislike different phases of a study, we are still to meet one looking forward to ethics and governance approvals applications. More accurately, we are still to meet a researcher mentioning that they are about to embark on those processes without pulling a face or changing their tone of voice. They/we may latter comment that the process went smoothly, was not as bad as expected, and was even quite useful. In some cases, they/we may feel resentful about being ‘told’ what we can or cannot do when we think we ‘know better’. For fully valid or rather suspect reasons, the set point of researchers’ expectations of the ethics and governance approvals for research studies is often remarkably negative.
This emotional response may be more local, less generalisable than it appears to us, fostered by the systems and institutions we have been working in. We hope readers will be willing to join a debate, both on the scope of problems of research approvals and on new solutions to them from a variety of perspectives and experiences. It is, nonetheless, a sad state of affairs to be experienced even in pockets of the academic world. First, in research as in daily life, ethical dilemmas (typically the starting point of approvals after research has been funded) are amongst the most engaging and enlivening topics of conversation. Then, the people behind the bureaucracy are often thoughtful, well-intentioned individuals willing to make a difference to research and research subjects. REC and IRB board members are generally volunteers. Some of them are, or have been, researchers themselves. Many members of staff working on broader governance approvals, for instance in clinical research networks, individual healthcare organisations or universities, are consistently going the extra mile to support researchers in navigating complex systems. They can easily introduce inordinate delays or even bring the whole enterprise down if only ‘working to rule’ and within the explicit remit of their (paid) roles.
Yet for many researchers with no insider experience in such processes, ethics and governance approvals are hardly ever about ethical deliberation and motivating exchanges with people supporting you to conduct a better study. More often than not, researchers experience them as exercises of form filling; of engaging with a tiresome and wasteful bureaucracy; of learning the rules and language of a game and sticking with your prescribed roles in it; and, more cynically, of somebody out there caring about ethical conduct only insofar as litigation and reputational damage can be avoided. As a result, many of us working in areas or with methods that may be “exempt” from REC/IRB review would consider fitting a study into an exempt category and avoiding the process altogether, while upholding ethical research standards as an integral part of taking professional responsibility (more or less successfully).
Such impressionistic accounts of researchers’ discontent are borne out by formal studies. In a survey of social researchers working on health and applying for NHS ethics approval (UK system under the National Health Service), Richardson and McMullan [2] found that 78% of their respondents (101) agreed or strongly agreed that the time for preparing an application inhibited social research in health and 75% (98) either disagreed or strongly disagreed that the amount of paperwork required was reasonable. Over half (51%, 64 individuals) reported they had modified their research design for the worse in order to secure ethical approval; 45% (58) had modified their research design to avoid the process; 30% (40) had abandoned their research before applying and 12% (16) during the application process (though we cannot assume that the ethics approval process was the sole reason). Twenty-one percent (26) revised their research design after obtaining approval and did not seek further permissions. Fifteen percent (19) had started research needing NHS research ethics approval without applying and 12% (14) had commenced their research before the formal approval was given. More positively, 32% (41) of respondents had made changes for the better as a result of the ethics approval process [2].
Interestingly, for those of us who internalise such negativism, this happens almost entirely through socialisation in our research community and overlaying our personal experiences on the socialisation process. We do not learn to be critical of the ethics and governance approvals of human subjects research by reading academic and professional literature presenting critical arguments or worrying empirical findings. Such literature is sparse, dispersed and hardly ever a syllabus item.
The sparse and fragmented literature on ethics and governance approvals
To our knowledge, there are only two well-known book-length critical explorations of IRBs (with the popularity still confined to bioethics circles rather than those of medical and health researchers): Zachary Schrag’s Ethical Imperialism: Institutional Review Boards and the Social Sciences, 1965–2009 (2010) [3] and Laura Stark’s Behind Closed Doors: IRBs and the Making of Ethical Research (2011) [4]. Schrag’s Ethical Imperialism explores the regulation of research in the social sciences and humanities by IRBs and how derivative this regulation is of the concerns, ethical codes, rules and blueprints of legislation and regulation pertaining to biomedical and, to some extent, psychological research. It illustrates how limited or unsuccessful the involvement of social scientists in the debates was (with them practically “left howling outside the door” and becoming “objects of policy rather than its authors”, op.cit., p. 4, p. 8). As any good history book, it shows the contingency of much of what we now take for granted and ethical in human subjects research (e.g. the very concept of prior/prospective review; the risks-benefits framework; even the “do not harm” principle once applied outside of patient populations, such as in studies aiming to increase public accountability). Much of the regulation wielding “totemic influence over the practice of research ethics” (op. cit., p. 78) has also been shaped by time and staff limitations; the moment-to-moment availability or lack of particular types of expertise; by public and political pressures; by tactics and personalities; by the supposedly neutral work of redrafting and editing; by the gradualism of introducing changes, which keeps them off the radar of those most affected by them. Schrag’s book is an unlikely page-turner for work so meticulous in tracing and representing historical detail, so careful in crafting precise and information-dense sentences, and so concerned with getting the history right as opposed to advancing a righteous argument.
Published shortly after Schrag’s monograph, Laura Stark’s Behind Closed Doors is an ethnographic study (again from the US) based on observations of IRB meetings and interviews with IRB members, with some elements of a historical exploration. The historical account focuses on the origins of the method for making decisions about research ethics: group review by experts required to reach a consensus. After over 60 years of using this approach, it may appear to be the most reliable and overall ‘best’ way of making research ethics decisions. Yet group reviews are also a consequence of a particular historical configuration: IRBs are “direct descendants of the Clinical Research Committee that met inside NIH’s Clinical Center starting in the 1950s” [op. cit., p. 157]. Stark also argues that rather than having group review imposed on researchers by outsiders (such as bioethicists or activists, as the story often goes), it was primarily researchers and NIH leaders who developed, justified and expanded this mechanism “as a technique for promoting research and preventing lawsuits” [op.cit., p. 8].
Core findings from Stark’s ethnographic work concern the IRBs’ methods for making decisions. For example, board members appeared to be reading “researchers’ application documents like tea leaves for signs of good character” (p. 6). Most notably, the researcher’s apparent attention to detail was translated into judgments of his or her professional competence and trustworthiness (pp. 17–18). Board members used “warrants” to justify their recommendations and although some pertained to matters of fact and professional expertise, a surprising number came from their private lives. The people who featured there came to serve “as stand-ins for research participants”, thus reinforcing the biases of IRB membership (p. 14). IRBs typically invoked local precedents (p. 47 onwards) in their decision-making processes as opposed to abstract ethical rules. This, Stark argued, ensured efficiency and internal consistency of decisions over time, but was also the source of the much-lamented variability of decisions across IRBs. (Of course, we need to acknowledge the critical stance of “Behind Closed Doors”, meaning that it draws more attention to unacknowledged and problematic aspects of the system as opposed to its strengths and good processes.)
The empirical and critical explorations of research approvals in published papers have a rather limited overlap with the work of Scrhag and Stark. No relevant literature review seems to have been conducted, whether systematic or traditional. We offer a preliminary summary of key themes and approaches, with a focus on the last 20 years.
“Horror stories” appear to be one of two main ‘genres’ in the literature, with the term first used by Bradford Gray in 1978 of the “sad tales told by scholars in their letters to the National Commission and in their testimony at the IRB hearings of 1977” (p. 161) [3]. Mary Dixon-Woods et al. call them “wounding encounters” [5]. More neutrally, these are case studies of challenging ethics review processes, due to time delays and inconsistencies of decisions across committees; [6,7,8,9,10,11] bureaucracy, form filling, paperwork and over-concern with minor language and grammar issues; [12,13,14] lack of appropriate expertise, which may lead to inappropriate rejections, criticisms and recommendations; [15] blocking of important research [16] or compromising its rigour, e.g. through negative effects on recruitment [17]. Mounting costs are also a source of frustration. Most strikingly, Kano et al. estimate personnel costs of obtaining the 84 approvals for their multi-site project at US$121,344 [18] (note, however, that the system for multisite projects in the US has now changed).
Some of those publications are strongly emotionally charged. Mentions of “horror stories” and “wounding encounters” are also typical elements in polemical texts of scathing criticism and radical dismissal of the value of the existing systems of ethics review, especially if applied outside of biomedicine, such as to social sciences research (e.g. see Dingwall and to some extent Derbyshire [19, 20]). Many publications describing significant challenges, however, aim to remain objective and constructive, giving specific recommendations for improvements rather than questioning the overall value of the system.
There are also studies which, while arising from a background awareness of the challenges around ethics approvals, go straight into exploring and evaluating models for optimising the work involved. Recent research of this type concerns, for instance, the introduction of an ethics officer, [21] the use of structured tools to evaluate the work of RECs/IRBs, whether through self-assessment or researcher assessment, [22, 23] models for simplifying standard operating procedures, [24] and approaches built around specific conditions and enabling better separation between the scientific and ethical appraisal within the ethics review [25].
More broadly, audits and evaluations of RECs/IRBs, typically survey-based, appear to be the other leading genre in the literature. This may be a historical path dependence dating back to 1975, when the National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research (the commission which created most of the foundations for regulating human subjects research) commissioned the original survey of researchers and IRB members (see p. 63 onwards in Schrag [3]). In the last decade, topics include the overall availability and profile (composition, arrangements for their work, etc.) of IRBs and RECs, primarily in areas where ethical regulation of research is new; [26,27,28,29,30,31,32] comparisons of committee members’ and researchers’ views on ethical issues in complex areas; [33,34,35] practices, policies and requirements of IRBs/RECs; [36,37,38] committee members’ knowledge of and attitudes to ethically and legally complex topics [39, 40] and researchers’ perceptions of the review process, outcomes and impact [2, 41, 42].
The literature described above makes little explicit use of theory. In contrast, there is an identifiable niche of work on ethics approvals drawing on theory from the social sciences [5, 9, 43,44,45]. It problematises the foundations of the new regimes of control and warns against their dangers, particularly in the context of the social sciences. Narrower topics are also explored. For instance, Dixon-Woods et al. look into the social functions of REC letters, including ‘stabilising “what is ethical” in a world where ethical matters are in fact indeterminate’; providing institutional displays of “care, thoroughness and thoughtfulness”; and structuring the relationship between the REC and applicants, reinforcing a ritualised supplicant-authority relationship [5]. Or, against expectations that REC decision-making is yet another impersonal process of a modern bureaucracy, Hedgecoe argues that local knowledge, trust and “facework” (roughly, the direct face-to-face interactions between applicants and committee members which serve to build trust) are surprisingly important for the decision-making processes of local RECs [44].
Some recent trends in the literature can also be identified. There is a fast growing global dimension to it (see, for instance, [46,47,48,49]; about a fifth of the 100 most recent publications retrieved by “ethics research committees” as a major heading in PubMed, Jan 2018, were from developing world countries or countries with smaller research communities). There is also attention to ethics approvals challenges and solutions for newer or newly popular study types, such as participatory research or co-design studies, [50,51,52,53] biospecimens and biobanking studies, [54,55,56] digital health research [57,58,59] and research in disaster and developing world contexts [60,61,62,63,64,65].
Finally, there are red threads of issues which do not become the focus of texts but are picked and interweaved in much of the critical literature. It often reminds us that ‘there is no single “right” answer to (most) ethical issues’ (p. 795); [5] that what constitutes ethics, ethical practice and ethical decisions is contested, contextual, theory-dependent, subjective, personal, relational, evolving … (e.g. in 2, 9, 43). In contrast, what filters down to ethics review forms and processes is often an “impoverished ethics” (Carl Schneider quoted in Schrag [3], p.3). The shift of power from scholars to administrators and the rise of a bureaucratic and self-serving “machine” are lamented. Appeals are made to trust researchers and their professionalism more and enable them to take more responsibility for the ethical decisions they make. Attention is repeatedly drawn to serious flaws in or lack of relevant definitions, classifications and boundaries, which have profound impact on what is being regulated and resulting, for instance, in journalists conducting (and being praised for) investigations which social scientists would never be permitted to do or the state performing observations that are far more secretive and sinister that any researcher would ever consider engaging in.
Against this backdrop, the empirical aspect of our paper is a case study of ethics and governance approvals, with a focus on quantifying aspects of the process, one of which does not seem to have been quantified before (the number of “approvers” involved). Although some of the findings are striking, we did not have the experience of a “horror story” or “wounding encounter”. We did not start an audit because we were frustrated. Our study arose out of a table we kept in the project Master File to avoid excessive printing and to justify not printing (communication concerning project approvals must be printed and stored in a physical Master File). If it were not for this table, we would not have paid any further attention to the approvals process once we had gone through it.
This is one of the important points we want to make: much of the burden of ethics and governance approvals and the complexity of the system remain too vague or subliminal to be perceived clearly, like the underwater part of an iceberg. We fail to change course in steering ahead, as we can hardly discern what lies beneath, and we end up damaging the integrity of our research.
Inevitably, by virtue of the position we are taking, as researchers who have found that, largely unknowingly, we have been part of a rather inefficient process, our presentation of issues may feel or be one-sided. We acknowledge that there are many more sides to the story that need to be heard before we arrive at a balanced view of the advantages and disadvantages of current systems. This is one of them.