The number of scientific experiments conducted on animals has declined considerably over the past 30 years. The trend, however, has been reversed recently. The total has risen in each of the past five years and new data released by the Home Office this week show that the 2006 figure exceeded three million for the first time since 1991.

This has angered even the more considered elements of the animal rights lobby. The RSPCA pronounced itself furious and shocked, while the Dr Hadwen Trust, which supports medical research with nonanimal methods, blamed the Government’s “ethical negligence”. Its message was clear: scientists might talk about replacing, reducing and refining animal experiments, but this is mere lip service. The statistics tell a tale of more animal suffering.

This view might look compelling, but it is not founded in logic. A rise in the raw number of animal procedures does not necessarily mean that medical researchers are being cavalier. As it happens, the upward trend has a perfectly reasonable explanation that has nothing to do with callous indifference to animal welfare.

A close look at the Home Office figures makes this plain. The recent rise in animal use is almost entirely explained by the growing importance to science of genetically modified mice. The number of experiments that use these has more than quadrupled since 1995, to reach 1.04 million last year. One in three animal procedures now involves a GM mouse.

This headline figure, though, is a little misleading. The birth of every GM animal must be recorded as a scientific procedure in the Home Office statistics, even if it is never used in an experiment. Two-thirds are created purely to maintain breeding colonies or to provide cells, and are never given drugs or surgery. Many suffer no ill-effects from being genetically altered. Take them out of the equation and animal experiments would have continued to fall.

That said, it is beyond dispute that the number of GM animals used actively in research is rising and will continue to do so as more genes that influence disease come to light. But that is because these mice – and 97 per cent of GM animals are mice – allow scientists to answer medical questions that could not even have been asked a decade ago.

Conditions with a genetic contribution, such as diabetes, can now be modelled effectively by manipulating the genes of laboratory mice. These animals can then be used both to understand the disease process and to test new drugs. Such work is already having important results: treatments for incurable disorders such as muscular dystrophy that have been developed using GM mice are close to beginning clinical trials.

Such insights, regrettably, cannot be obtained in any other way. Scientists are using more GM mice not because they have become hard-hearted but because they are the best available tools for a certain kind of research of exceptional medical promise. From a patient perspective, the increasing number of GM mouse experiments is something to be welcomed. It means that science is closing in on the genetic origins of disease and thus on new approaches to therapy.

The development of nonanimal methods is of course welcome, and when such techniques have been validated it is right to use them. The number of nonGM animal procedures in research, indeed, has come down from 2.27 million in 1995 to 1.65 million last year. Further investment is appropriate, but too narrow a focus on reduction would mean abandoning new animal models just as they are becoming most useful. Science must be serious about both medical progress and animal welfare, but that may mean using more animals when necessary, and fewer when it is not.

Mark Henderson is Science Editor of The Times

Times recommends

• It is possible to look stylish while pregnant
• Britain’s tot models
• Who and what benefits from the heat wave?