Medical progress depends on animal models – doesn't it?

With regard to the JRSM article ‘Medical progress depends on animal models – doesn't it?’ (JRSM 2008;101:95–8):1 there is misunderstanding in the preclinical science field about applying bias-reducing principles to animal research.2,3 Animal project licence applications should be supported by references to systematic reviews of all the existing relevant human and animal studies, documenting the materials and methods used to show how the reviewer attempted to minimize bias. The Home Office Inspectorate should not be satisfied by reference to bibliographic databases and statistical analyses alone because these do not take into account publication bias and variability in quality of bibliographic searching skills. The reviews need to take account of completeness and quality of research and address the probability that null results had not been reported as equally as were the positive results. The applicant needs to show that they have searched for and assessed what relevant research has already been done and, of equal importance, to show what relevant research has not been done. Failure to apply these principles was highlighted by Kenter and Cohen in their assessment of the TGN1412 drug trial.4 The authors demonstrated that if there had been access to and synthesis of all the research data, the trial disaster and consequent harm to the research volunteers could have been avoided. Systematic reviews would tell us more about confounding variables and other characteristics of animal research that need to be accounted for (e.g. genetics, species differences, variations in feeding, location and laboratory practice),5 the validity of animal models and whether a rigorous search for alternatives had been conducted. We are petitioning6 the Prime Minister to make improvements to animal research which would have relevance for ‘a systematic study of the use of animal models’. It is time that animal research was made more evidence-based.