Comment: Reviewing Existing Knowledge Prior to Conducting Animal Studies
Highly polarised viewpoints about animal experimentation have often prevented agreement. However, important common ground between advocates and opponents was demonstrated within a discussion forum hosted at www.research-methodology.org.uk in July–August 2008, by the independent charity, SABRE Research UK. Agreement existed that many animal studies have methodological flaws — such as inappropriate sample sizes, lack of randomised treatments, and unblinded outcome assessments — that may introduce bias and limit statistical validity. There was also agreement that systematic reviews of the human utility of animal models yield the highest quality of evidence, as their reliance on methodical and impartial methods to select significant numbers of animal studies for review, serves to minimise bias. Unfortunately, disagreement remained that animal experimental licence applications should reference systematic reviews of existing studies, before approval. The UK Medical Research Council requires that researchers planning human clinical trials must reference such reviews of related previous work. Existing knowledge is thereby fully and appropriately utilised, and redundant experimentation is avoided. However, objections were raised that a similar requirement would interfere with animal experimental licensing, because, to date, there have been very few systematic reviews of animal studies. In fact, the relative dearth of such reviews is a matter of considerable concern, and may partially explain the very poor human success rates of drugs that appear safe and/or efficacious in animal trials. Nevertheless, the disturbing number of human trials which have proceeded concurrently with, or prior to, animal studies, or have continued despite equivocal evidence of efficacy in animals, clearly demonstrate that many researchers fail to conduct adequate prior reviews of existing evidence. Where neither sufficient primary studies, nor systematic reviews of such studies, exist, for citation within a licence application, researchers should be able to provide evidence of this shortcoming, and, concurrently, demonstrate that the available literature and evidence have been adequately reviewed. This should also enable them to clearly demonstrate the need and scientific appropriateness of their proposed study, the validity of its design, and — importantly — that the benefits are reasonably likely to exceed the animal welfare, bioethical and financial costs. Invasive animal studies should never be permitted solely on the basis of less probable, speculative or intangible human benefits, or the mere satisfaction of scientific curiosity.