Improving transparency and ethical accountability in animal studies

The use of animals for research remains a highly controversial issue that receives much public attention and comment. This was illustrated, for example, by the participation in the European Commission's internet consultation on the revision of Directive 86/609/EEC, which regulates their use in this way. Moreover, scientists using animals in their work still face fierce, sometimes violent opposition that includes death threats or the vandalism of research facilities by anti-vivisectionist and animal rights groups who want to see all use of animals in research completely banned. Given the polarizing nature of the issue, it is not always easy for the public and lawmakers to effectively weigh the benefits of biomedical research against the deaths of the animals used.

To address public resistance to animal experiments, governments and scientific entities have called for increased transparency and an informed dialogue between regulators, scientists and the concerned public. As a part of this endeavour, several European countries, notably Denmark and the UK, have begun to post summaries of approved animal experiments on the websites of responsible authorities. The rationale is that more transparency will increase public confidence in the appropriate conduct and regulation of animal research and therefore help to maintain public acceptance.

However, this approach towards greater transparency leaves out one crucial piece of information, which is of both ethical and scientific relevance: the link between approved experiments and the results thus generated. The scientific community also misses an important opportunity to bring together information from prospective ethics reviews—which contain detailed information about animal procedures and ‘3R' considerations (replacement, refinement and reduction), but little scientific information—with retrospective peer-reviewed results that provide extensive scientific information but say very little about the animal procedures (Olsson et al, 2007). Furthermore, it would also make it easier for reviewers and journal editors to check the validity of authors' claims to the effect that animal experimentation has been conducted in accordance with existing legislation.

There are at least three different ways by which to establish a connection between published results and their previous approval by review boards. Two of these involve relatively simple changes to present publishing policies, while the third implies the need to create a more complex infrastructure. The first method is a requirement in journal articles to provide the complete legal reference, including the name of the ethical entity that gave authorization and the project reference, which would allow other parties to contact the authorization entity and obtain the approval documents for each study. The second method is to include these approval documents as part of the supplementary online material. The third strategy would involve setting up a comprehensive, prospective database of animal studies that includes details of the ethical approval, subsequently published articles and other material reporting the results of the study. Here, we discuss these three options with a special focus on the associated practical and legal issues.

Journals typically require a ‘legal statement' from authors to confirm that the research was conducted in accordance with the relevant legislation and that the research protocol was approved by the appropriate institutional review boards or ethics committees for human or animal experiments (Scott-Lichter et al, 2009). The format and content of these legal statements, however, vary so much that they are of limited use. While preparing a review on animal models used for metabolic syndrome research, we examined 87 ‘legal statements' appearing in 137 articles from 75 journals. We distinguished five main types of statement: a partial reference to the approving body (34 cases); the ‘title' of the regulation relevant to the particular study (15 cases); partial reference to the approving body and the legal regulation (31 cases); statements that the laboratory in which the particular study was conducted was fully accredited (3 cases); any of the above in combination with the authorization number of the study (4 cases). None of the cases provided a complete legal reference.

By requiring authors to provide and publish complete legal references as part of scientific papers, scientific journals would make public the necessary information to establish the connection between the published work and the background ethical information in a way that would not increase the administrative burden for the scientific community. For those interested in obtaining details on ethical approvals, however, this solution would still require active involvement; that is, the need to request the ethical approval information from the relevant authorities. Although these details are not made public, freedom of information acts in European Union member states, in principle, would enable citizens to obtain this information.

The second option again postulates the involvement of scientific journals by requesting that authors provide complete information on ethical approvals—including all relevant documents from the approval process—and that this package of information be made available to readers as online supplementary material. For anyone interested in this information, such a system would have the great advantage of providing access to ethical approval information without the need to contact national authorities or university review boards. The great disadvantage, of course, is that much of the material would probably be behind journal firewalls and access would be restricted to those with a subscription to the journal. In addition, searching for this information would be neither easy nor robust given that it would be spread across many journal sites.

The third option, the registration of animal studies in an online database, is the most complex and presents several challenges. This system would, in some ways, be similar to the current agreements for human clinical trials, as suggested by Roberts et al (2002): all studies using human subjects are, in principle, identifiable through freely accessible registration systems such as the International Clinical Trials Registry Platform (WHO, 2009). The establishment of human trial registers was first required by Articles 19 and 30 of the Declaration of Helsinki (WMA, 1964), which states that every clinical trial must be registered in a publicly accessible database. The International Committee of Medical Journal Editors announced that, from July 2005, they will not publish articles based on unregistered clinical trials (ICMJE, 2005). Today, human clinical trials must therefore be registered before the first participant is recruited, and studies based on trials cannot be published without a registration number.

Although journal-based approaches to providing information about animal research would incur some costs in terms of online storage space and bandwidth, the establishment of a database of animal studies by comparison implies far greater legal, administrative and financial challenges. The first question to discuss would be what kind of information to include in such a database. Research with animals is more heterogeneous than clinical trials with human subjects and relevant legislation varies between countries. In terms of the different species covered, the national and federal legislation of most countries covers vertebrates and a few invertebrate species—it seems neither reasonable nor practical to expand coverage to lower species, at least initially. However, given the large number of heterogeneous experiments, the inclusion of all vertebrates might even be too ambitious an approach at the outset.

Second, establishing and maintaining the database would cost money and manpower. If the database—similarly to the human registry—were to be available free of charge or at cost price, it would require additional funding. An administrative body would also need to ensure that an entry is ‘sealed' after the corresponding ethical approval has been given, to prevent further modification of the data. Moreover, scientists are unlikely to be enthusiastic about demands for additional paperwork: typically research governance measures are considered as unwelcome burdens (Dixon-Woods & Ashcroft, 2008).

Of course, there would also be advantages compared with the previous approaches. Information would be made available in a standardized format and in English, making it internationally relevant. The database could provide an interesting way of avoiding publication bias—that is, the fact that negative results are often not published, coupled with the tendency to report results that show only significant findings (Dickersin, 1990)—a factor known to interfere significantly with the translation of results from animal studies into clinical application (van der Worp et al, 2010). Prospective registration would overcome this problem if the database were complemented with a requirement to report results from all work with animals, by reference to the journal publication in the case of published studies and by uploading a standardized research report in the case of unpublished results.

By now, many scientists reading this article will question whether this information is really suitable to be made public. The problem of violent animal rights extremists (Council of Europe, 2004) has made scientists increasingly reluctant to talk about their work in public. This, of course, is the main reason why the UK Home Office publishes only an anonymous summary of the research; the Danish Ministry of Justice publishes detailed, albeit anonymous information about the applications. However, refraining from publishing information does not necessarily mean that the information is not available for anyone with the determination to find it.

Most democratic countries regard the ‘freedom of information' as a fundamental right for their citizens (Council of Europe, 1950). Practically, this means an individual can request and receive any public information held by governmental bodies that is not classified for reasons of ‘national security' or otherwise. In regard to information about the ethical approval of animal studies, the most relevant exceptions are economical, such as proprietary results, and especially privacy interests. These might pose obstacles for the disclosure of all details of animal studies, which are otherwise included in the ethical approvals. In addition, it might be relevant to exclude the identity of the applicant to protect him or her from terrorist actions by opponents of animal research. The People for the Ethical Treatment of Animals Foundation, for example, requested information about animal studies from research institutes and was refused on the basis of legally protected interests such as personal information and proprietary data (Grant, 2010).

Such concerns are even more pertinent for a prospective registry of animal experiments; by making information easily available in a standard format and in English, such a registry would effectively increase public access to this type of information. In the scientific community using animals, there is a fear that facilitating information might lead to undesired consequences in terms of public opposition and even to increased and better targeted attacks from anti-vivisectionists. This might be used as an argument for justifying the restriction of human rights, including the freedom of information (Brown, 2004).

Restricting access to information about animal experiments would be acceptable within existing legal freedom of information frameworks if it were to be expected that such information could lead to violent actions against scientists and their laboratories (Court of First Instance of the European Communities, 2009) and if there is no other effective way to protect them (Council of Europe, 2009). However, the causality between transparency and terrorism against scientists and labs seems to be rather indirect. It is obvious that, if a database of this information were introduced, the protection of privacy of scientists should be taken into account in the design of the database. For example, an obligatory registration of users accessing the database could reduce the risk of abuse, and access to the identity of applicants could be limited. Thus, even though the interests that must be protected might serve as an argument for limiting access to the database, they are hardly sufficient arguments against the creation of such a database. Furthermore, as the background of terrorism conducted by militant animal rights activists is complex, a simple restriction on freedom of information can hardly solve the problem (Council of Europe, 2004).

Nevertheless, it is to be hoped that increased transparency would increase levels of public sympathy and understanding towards animal research, as the public would be more aware of the progress made possible through the use of animals. However, one has to accept also that transparency alone will not appease, nor prevent, the actions of extremists, whose goal is the complete abolition of all animal experiments (O'Sullivan, 2006).

This article has described three potential methods that might establish a link between the published—and unpublished—results of animal experiments and the processes that granted them ethical approval. The involvement of journals is the simplest to implement, but perhaps not the most effective in communicating the information to non-scientists. The establishment of a database has legal and cost implications, but might prove a vital resource for members of the public, lawmakers and scientists. Thus, at the very least, we would argue that the obligatory reporting of an authorization number and approving entity in scientific articles should be an important first step towards improved transparency in regard to ethical compliance in the use of animals in research.