RDS is an organisation which seeks wide consensus about the need for well-justified, properly regulated, scientifically valid and humanely conducted animal research. For this reason we have no problem with the suggestion by Robert Matthews in the latest edition of the Journal of the Royal Society of Medicine that:
‘Animal models can and have provided many crucial insights that have led to major advances in medicine and surgery’.
This is very close to our position in any case. Our keynote article in the Journal of the European Molecular Biology Organisation (EMBO) setting out the ethical aspects of animal research makes the similar assertion that:
‘Animal research has had a vital role in many scientific and medical advances of the past century and continues to aid our understanding of various diseases’.
An article to be published next week setting out our views on the direction of the debate about animal research (in ATLA - Alternatives to Laboratory Animals) likewise points out that:
‘Many medical advances are still likely to depend to some extent on animal-based research’.
And our advice in our Communications Handbook to research institutions considering their own position statements is that:
‘Research using animals has made a huge contribution to advances in medicine and surgery, which have brought major improvements in the health and well-being of humans’.
We will return later to the rest of Robert Matthews article (available online only by payment or membership of the Royal Society of Medicine), which is a deeply flawed critique of those who defend animal research. In the meantime, there is little point in getting dragged into his pointless and pedantic arguments about exactly which statements are right or wrong, while patients still need vital research to be done.
Animal research is morally and scientifically defensible whether it has contributed to many, most or just a few medical advances. Permission to carry out animal research (through the project licence) is made on a case-by-case basis only after the potential benefits are weighed against the likely harms to the animals.
There are many important issues to be debated about animal research. We need to concentrate on progressing the 3Rs, improving experimental design, better analysing the findings of the research (with more systematic reviews), improving animal models, critically challenging the applicability of animal models in certain fields of research, and finding replacement alternatives. Let’s face up to the real challenges of today. Is Matthews up for that?
I wonder if Robert Matthews would confine biologists to stating that evolution is only a theory, because it cannot be formally substantiated, however much evidence there is in its favour, and despite the fact that there is no evidence against it.
Matthews believes that it is not possible to substantiate the statement that virtually all medical achievements of the last century depended on the use of animals. Prima facie evidence to support his belief, in the form of specific instances of achievements which did not depend on animal research, is absent. He says that the statement is lacking in logical or evidential support, but this is not true. An inability to provide any evidence to the contrary of the statement, is logically supportive of it, if not complete proof. A list of achievements in which animal research undoubtedly played an important part, without a single example of an achievement in which they undoubtedly did not, also constitutes prima facie evidential support for the statement.
He does not say that the statement is false or misleading; he says it would be best to confine statements regarding the value of animal research to those which can be formally validated.
Matthews identifies animal research with animal models, and animal models with quantifiable predictivity. The reality is more complex than that. We cannot predict the results of basic research, nor precisely where those results will lead, and basic research is the area where the greatest number of experimental procedures take place. It is misleading to imply that there is only one way of using animals in medical research - as predictive models - and misleading to imply that all animal models are, or should be, used to make specific predictions. In many cases they only provide tentative clues to guide the direction of research; but a tentative clue to investigate further, is better than being clueless.
After questioning the validity or usefulness of the small amount of existing data available to quantify the predictivity of animal toxicity-testing, (whilst not questioning the validity or usefulness of animal models themselves, and acknowledging the serious ethical obstacles in the way of gaining human data in this field), he asks whether animal models do have predictive value. He answers by saying that there is a wealth of evidence to show that animal models have provided many crucial insights that have led to major advances in medicine and surgery - a fact, he says, that can be formally validated.
After going all around the houses Matthews arrives at a position less than a hair’s width from the one he says we ought to move away from. He questions whether a statement can be formally substantiated, which if untrue, could easily be shown to be so. There is a wealth of evidence to show the great value of animal models - and a dearth of specific instances of medical achievements which did not depend on the use of animal research. This is essentially the meaning of the statement to which he objects. It is not saying that because we see animal research everywhere, it must be vital in all cases; it is saying that nearly everywhere we look, we see vital animal research. Matthews misrepresents the meaning of the statement and fails to substantiate his argument.
I’d like to pick up on a few points being made and suggest that Matthews critique is not “highly flawed”.
I wonder if Robert Matthews would confine biologists to stating that evolution is only a theory, because it cannot be formally substantiated, however much evidence there is in its favour, and despite the fact that there is no evidence against it.
As I understand it, this is not what he’s saying. He presents a method for formally substantiating predictive models, but says cannot find any published studies where this has been done. As such there is no evidence, either in support or rebuttal, of their worth, but there could be (and in fact, he concludes that we should undertake these validation studies).
He says that the statement is lacking in logical or evidential support, but this is not true.
I have not been able to find any studies which would provide evidential support for the statements he is questioning, when considered statistically. Could you provide a reference to any references? (But not just a list of studies involving animals - see next point).
A list of achievements in which animal research undoubtedly played an important part, without a single example of an achievement in which they undoubtedly did not...
Animal research is currently required by law (due to lack of validated alternatives) in the development process of all NCEs, therefore all successfull NCEs will have been tested on animals. That does not mean the animals provided evidential weight in support of the NCE. Therefore, listing medical advances in which animal models were used does not prove that they helped. However, it is possible to prove it (or, as the case may be, disprove it), using statistics like Likelihood Ratios. There are undoubtably problems in doing so, he says (quoting the Nuffield Report), but not insurmountable ones.
Matthews is not denying that animal research has not had a role in medical advancements. His article focuses on predictive animal models (specifically for toxicological testing of NCEs), and his claim is that within that field there have not yet been any (large?) statistically valid analyses of whether those models are any good. We are hampered by a lack of human corollaries, but it is possible to generate the sensitivity and specificity scores required for his suggested statistical method of analysis, likelihood ratios (allbeit using the biased ‘side-effect’ dataset for sensitivity). As such, he does not say that these predictive animal models cannot be validated, or that they are invalid, he merely says that they have not yet been validated (or invalidated). The point he’s making is these studies could be done, but until they are claims that these predictive models are either valid or invalid cannot be substantiated. Am I wrong in my interpretation? Or is his statement wrong? I’d be very interested to know, as I think this is a critical area for debate and research.
Intelligent design advocates offer the bacterial flagellum as evidence against evolution; perhaps they should also be asking for a likelihood ratio. I am not aware of any study in evolutionary biology demonstrating this. By way of corollary, one or two examples of major medical achievements of the last century in which animal research made no useful contribution would go some way to demonstrating that Matthews has a legitimate argument.
He says that there is a wealth of evidence showing that animal models have played an important role in advancing medicine and surgery. That said, I fail to see the relevance to his argument of the ethical difficulty of obtaining selectivity data in toxicity testing. Sanity demands that we take the conservative approach of not testing drugs and other chemicals on humans that have been found to be highly toxic in animals .
Furthermore, if some way could be found of obtaining toxicity selectivity data without putting human lives at risk, there still remains a problem. A statistically significant positive likelihood ratio, (ie having a 95% confidence interval), is not absolute proof of a true effect. There remains a 1 in 20 chance that the effect is the result of chance. Absolute proof is scientifically impossible, a fact which will always allow scope for denialists to do what they do - simple deny what they do not wish to believe.
It is the combined weight of many different kinds of empirical evidence that constitutes the scientific case for the medical value of animal research - just as in the case of evolution. This value is most easily communicated by providing examples where animals were vital to medical advancement.
