Do we need research ethics committees?

This issue of the journal sees a number of exchanges on significant ethical problems. ‘Nudges’ have attracted a good deal of attention recently in the context of the ethics of public health interventions. Martin Wilkinson (see page 486) writes a guest editorial introducing important debate on Yashar Saghai’s featured article, Salvaging the concept of nudge (see page 487, Editor’s choice). Also, Timothy Murphy (see page 529) locks horns with Katrien Devolder (see page 533) and Ezio Di Nucci (see page 537) on the doctrine of double effect as it applies to research on embryos. One of the exchanges published here involves the legitimacy of research ethics review. Murray Dyck and Gary Allen (see page 517) claim that only in a small minority of cases is research ethics review warranted and that, in the main, responsibility for the ethical conduct of research should lie with the researchers themselves. However, David Hunter (see page 521) Mark Israel (see page 525) and Michael Dunn, (see page 527) in different ways, take issue with the claims made by Dyck and Allen. Hunter challenges their arguments, Israel criticises their distinction between research that requires review and that which does not, and Dunn supports the legitimacy of research ethics committees (RECs) by undermining some of the assumptions about the review process made by Dyck and Allen. Two distinct kinds of criticism can be identified in the many lines that have been written on the shortcomings of RECs. First, there are criticisms of the research governance system and the way it is constructed and functions in practice. These criticisms range from over-bureaucratisation and inconsistency to actual failures to prevent harm to vulnerable research subjects. Second, broader theoretical questions are raised about the need for RECs at all. These sceptical claims go to the heart of the ethical issue here, raising questions about the right of society to decide what research should and should not be permitted. Not all authors always distinguish between these two kinds of criticism, but they are so obviously different and have such radically different consequences that they must be clearly separated. First, criticisms of the functioning of one system do not necessarily apply to other systems and generalisations across systems run the risk of simply failing to be accurate. Claims about, for example, a one-size-fits-all application of ethical principles by RECs need careful and thorough evidence if they are to hold any weight in arguments for any kind of change. Second, there is always a possibility that any faults in the way in which a system functions can be corrected within the system or are the product of natural variability or human error. Systems can evolve and develop in important ways to redress inefficiencies and relevant inconsistencies. The establishment of multi-centre research ethic committees (MRECs) in the UK in the mid-2000s is good example of a system adjusting to certain kinds of inefficiencies. Those who argue against RECs in general on the basis of specific criticisms of a system need to show why we should think that these problems cannot be dealt with by adjustments in the system. Finally, scepticism about whether research ethics reviews are warranted at all requires a special set of claims. It requires a presumptive, libertarian-style argument about the illegitimacy of the intervention of ‘the state’ or ‘the public’ into the researcher’s domain. If the space of research is the private realm of the researcher, then the instruments of the state, in the form of RECs, have no place. That this space is private is far from obvious. A common defence of the general scepticism is to suggest that risk of harm should be the criterion which determines the need for review: only if your research poses significant risk of harm to subjects should it be reviewed. But clearly this kind of proposal needs work: (i) who determines the appropriate level of risk? (ii) who determines whether a proposed piece of research falls above or below the required level? (iii) why think that exposing the subject to risk of harm is the only way in which research can be problematic? Without substantial argument, it is not at all clear that these questions should be answered by the researcher or can be decided in advance or according to a pre-set schema. The papers by Dyck and Allen and their three commentators all engage directly with these issues and should be judged in the light of these distinctions and the arguments that are given for the relationships between them.