The overall efficacy and performance of ethics review systems have not been investigated thoroughly in spite of their important role of protecting research subjects and ensuring the continuity of the research enterprise. Canada has a long history of research ethics governance and has inspired research ethics review in other countries thus evidence on its workings may have national and broader international relevance. The results of our analysis of approved consent forms for MRI and fMRI research in the Canadian context provides some perspectives on the ethics review process, a process that has been reported as challenging for neuroscientists to understand and navigate.[15, 16] There appears to be significant variability and inconsistency in important areas in neuroimaging research such as risk reporting and the management of incidental findings. For example, we found evidence of substantial between-site and within-site variability in the strategies approved to handle incidental findings. There were also significant variations in the disclosure of risks associated with MRI and fMRI consent forms, especially with respect to psychological risks and risks associated with dizziness, discomfort, stress, and fatigue. We also encountered a few forms where the risks of MRI were downplayed (e.g., "no-risk" statements) and incidental findings were described as a benefit.
There are important limitations to keep in mind in this retrospective study of approved consent forms. First, the study is based on a convenience sample and we relied on researchers to send us their most recent approved consent forms. This could have induced some biases if researchers did not conform to this instruction (e.g., if they sent us their "best" or "worse" consent forms). Getting access directly to REB-approved files would have been another route but this has been reported to be difficult in other studies on MRI research. Second, we did not manage to collect forms from all sites in representative proportions and therefore our results should not be interpreted as an exact representation of the Canadian situation. Third, our focus in this paper is to report observations on the research ethics process based on the examination of informed consent forms. This could be criticized since research ethics, and even informed consent, involves much more than the informed consent form itself. The presentation of the consent form is only a portion of the consent process and the verbal exchange between the participant and the researcher introduces several other variables that are unaccounted for in this specific study. However, a focus on this aspect of ethics review allows us to gain insights on a deliverable of ethics review that has perhaps been vested with the most practical and academic energies. Finally, another limitation consists of our focus on a typically minimal-risk research area and the potential generalizability of our findings. On the one hand, REBs may not consider as extensively minimal risk research as they do with invasive and riskier research. This would be consistent with the proportionate approach of the TCPS which advocates adjusting the level of ethics review scrutiny to the level of risk of a study. On the other hand, MRI and fMRI make compelling cases to examine if a relatively safe and minimal risk area is dealt with adequately or not.
Although it's easy to conclude that REBs could be faulty for any variability and inconsistency observed in the process or final approved ethics documents, it is worth reiterating that research ethics and ethics review are shared responsibilities and that ethics governance systems are comprised of distinct key stakeholders such as researchers, REBs, funding bodies, and academic institutions.
One possible partial explanation for the variability and apparent inconsistency we observed could be explained by sensitivity of REBs to local concerns. REBs were created to respond to local concerns of patient and subject communities as well as local administrations and cultures. Accordingly, reported variability could reflect the fact that the institutional and research context varies between sites (e.g., having access to clinicians or not in strategies deployed to handle incidental findings). However, variability is somewhat hard to justify on this basis, given that different strategies for incidental findings were approved by the same institution. Further, some variability in the reporting of risks seems hard to understand on the basis of local concern, considering subject protection and information are key goals of all REBs and research ethics policy and that the basic risks of MRI and fMRI do not vary in essence between sites.
Another partial explanation relates to lack of cohesion and consensus, across Canadian researchers and REBs, about the risks involved in MRI and fMRI studies.[17, 18] The causes of this could be manifold: lack of a common and standard understanding of the risks related to MRI by researchers and REBs, lack of access to knowledge of the risks of MRI (e.g., review papers on risks of MRI, MRI guidelines of the Food and Drug Administration or Health Canada); inconsistent application of commonly accepted knowledge of the risks of MRI; lack of REB resources to review the specificities of the risks of novel technologies; lack of consensus between researchers and REBs on the risks of MRI that should be reported to subjects. A previous study using a mock review of fMRI results suggested that although REBs tend to use similar criteria, they can arrive at radically different decisions.
From a research ethics standpoint, the variability and inconsistencies we observed are important because the full disclosure of risks remains both legally mandatory and ethically required in Canada. At least two Canadian court decisions Halushka v. University of Saskatchewan in 1965 and Weiss v. Solomon in Quebec have highlighted these obligations.[7, 20, 21] The variability we encountered regarding some basic risks of MRI and fMRI suggest that not all subjects receive similar information to consent to research participation (e.g., variability in reporting physical and psychological risks). This means that subjects may not be fully informed based on Canadian standards and that protocols and informed consent forms are being accepted with some lacunae. The amount of variation found in risk disclosure of a common research tool points to potential problems with the ability of current research ethics governance systems to deal with risks in novel neuroscience techniques like fMRI. Additionally, there is now mounting evidence of challenges plaguing ethics governance based on local ethics review processes like those in place in Canada, the USA, the UK, and Australia.[22–26] For example, some studies and analyses have suggested that practical shortcomings and variability in the process undermine the trust that researchers place in REBs and creates difficulties for collaboration and open communication.[5, 16] REBs are also struggling to deal with demands of the research community for both comprehensive in-depth and expedient review.
Administrative hurdles coupled with an increasingly large burden of responsibility placed on researchers are not reasons to take ethics short-cuts or to dismantle research ethics as we know them. But the combination of sub-optimal performance in dealing with ethics issues clearly and consistently along with a significant practical burden should make us reassess if our energies are well served and if other ways of dealing with the requirement of a proper assessment of ethical issues should be considered. Based on several reports about the challenges encountered in current research ethics governance systems, the status quo option appears hard to defend in research-intensive countries with similar ethics systems like the US, Canada, and the UK. Several reports on ethics governance in these countries indicate structural problems that will not be resolved by only more of the same practice and policy. Different strategies could be adopted to revise current research ethics governance. In Canada, a revision process of the TCPS is taking place. However, little in the proposed revisions seems to address the practical or administrative issues that may make the ethics review process difficult or inconsistent. The focus is on reshaping the policy itself, not its implementation by REBs and research institutions. Additionally, we lack an evidence-based approach to making revisions that would be commanded by both evidence about the existence of problems and evidence about the efficiency of proposed revisions.