Health research guidelines consistently rule out retrospective ethics approval1, and it is easy to see why. The possibility of retrospective approval may generate perverse incentives for researchers, tempting them to start research without approval and to initiate the ethics process at their convenience, perhaps after satisfying themselves that their project looked likely to produce results that warranted doing so.
However, it is also easy to see that a blanket rule against retrospective review can itself generate ethically undesirable results. Suppose a researcher failed to obtain ethical approval before carrying out research which produced valuable information, while imposing some burdens on participants (e.g. blood tests), and which would have been approved had it gone through the review process. The rule against retrospective review effectively means that such research must either be left unpublished or repeated.
Neither outcome seems ethically desirable. Indeed both appear to contravene other plausible principles of ethical health research, principles that require the results of research be made public, the burdens of research on participants to be limited to those necessary to produce worthwhile results, and that respect for research participants requires not wasting their contributions.
What, then, should we think of the rule against retrospective review? It seems unclear that the rule should have the priority it is typically given. We should ask whether a ‘default rule’, which made clear that retrospective review would only be available in rare cases, would provide the right kind of incentives to seek ethical review while also allowing committees to avoid the ethical costs of a more absolute rule. Perhaps cases in which there were genuine but merely clerical failures, or when a committee accepted that there had been genuine uncertainty about whether the project was research or audit.
We should not assume too quickly that some such rule would be too ‘fuzzy’ or unclear. Ethics committee are, after all, in the business of case-by-case consideration of research applications and are familiar with the sort of balancing and weighing that comes with non-absolute standards. While, on the other side of the process, researchers are a relatively small and sophisticated group of regulatory ‘subjects’ who might be expected to learn the boundaries of a discretionary rule fairly quickly.
In summary, there are good ethical reasons to favour a more flexible rule, and few obvious ethical or administrative costs to doing so.
News article written by Associate Professor Tim Dare, The University of Auckland, for the HRC publication, Ethics Notes (PDF 627kb).
1 Section 24 of the New Zealand HRC Guidelines on Ethics in Health Research (HRC, 2005) is typical, providing simply that "[n]o retrospective approval for any study shall be given by an accredited ethics committee". The reference to "retrospective approval" seems not quite right in these rules: what ought to be ruled out is retrospective ethical review, whatever its outcome.