Limitations of the Animal Model
Scientists often cite the similarities between animals and humans as the rationale for animal experimentation. Their premise is based on the presumption that because we share similar biological processes, and because animals are whole, complex organisms, data obtained from an animal model can be extrapolated and applied to humans. Although animal experimentation has become the norm in research labs around the world, a closer look at the data makes it clear that even species who share close evolutionary relationships or anatomical similarities differ in important ways—and these differences greatly impact the usefulness of using an animal model as a stand-in for a human.
Humans are not mice or rats or dogs or any of the other animals that are used to mimic us in the lab. Nonhuman animals, by their nature, have different genetic backgrounds that make them unique—and ill-suited to serve as predictive models for people. Even when researchers have discovered an animal model which appears to mirror what occurs in humans, the response observed in animals can occur via different mechanisms due to the intrinsic differences between species.
Diseases that develop in people can differ significantly from the artificial ways that they are imposed on animal models in a laboratory setting. And some drugs show toxic effects only within certain genetic backgrounds. Does it make sense, then, to expect an animal model, with a different genetic background, to predict what happens in people? No. When you need medical treatment, you want what works for you, not what works in a completely different species. While no one model should be expected to recapitulate every aspect of the human condition, it certainly makes the most sense to work with models that have the most human relevance.
One of the most striking examples of how reliance on the animal model proved to be misleading and detrimental to human health was its failure to predict the negative effects of thalidomide. This drug was prescribed in the 1950s as a sedative and a treatment for morning sickness in pregnant women. Use of thalidomide caused birth defects in thousands of children whose mothers used the drug, although such defects were not observed in many animal models given thalidomide. Thalidomide is not the only example. Numerous compounds which have shown promise in animal models have not been beneficial in humans ,,,.
Even an animal model system that is more closely related to humans, such as primates, has failed to be predictive. One of the best examples of this occurred when monkeys treated with a therapeutic antibody (anti-CD28 monoclonal antibody TGN1412) did not predict the potentially fatal immune response that was triggered in humans.
Advances in science have helped us appreciate how even subtle genetic variations among humans can affect many aspects of disease risk and progression, as well as response to various treatments. When you consider that genetic differences among humans can generate such varied responses, it becomes abundantly clear that animals, being so genetically different from humans, should not be expected to accurately predict what might happen in people.
Moreover, even if animals could be predictive on any level, the better question to ask is to which humans the data could be applied, considering the genetic variability among our own species (as well as environmental differences such as diet and lifestyle, which may further complicate extrapolation).
The animal model too often fails to translate into cures and treatments for people. NAVS believes that science can do better. Greater emphasis should be placed on developing better and more sophisticated human-relevant models that advance science without harming animals.
 Gilbert, S. “Development Biology, 10th Edition, Sinauer Associates, 2010