The pressure group GeneWatch is not normally among the more constructive critics of biotechnology. Its very name implies that genes are something to be afraid of and it has run scare-mongering campaigns on GM crops, cloning and gene therapy. Its most recent report, however, has raised legitimate concern about an emerging branch of medicine.

Since the Human Genome Project, several companies have begun to sell genetic tests to the public. Plenty are founded on sound science, allowing, for example, women with a family history of breast cancer to check for susceptibility genes. Other services, however, are of more dubious validity. Many expensive tests have benefits that are questionable at best, and are barely regulated.

Particular disquiet surrounds a field called nutrigenomics. Customers swab a sample of cheek cells and send them off for analysis. A company tests these for variants of certain genes and draws up a personalised diet. The idea is to identify those at most risk of conditions such as heart disease and hypertension, so that they can modify their lifestyle accordingly.

It sounds good in principle, but there are grave doubts about the science. Most geneticists think that the interaction between genes, lifestyle and disease just isn’t sufficiently well understood to justify selling such advice. Much of the public has yet to grasp that the links between genes and health are usually complex and far from causal. The fear is that people are buying a service they think is predictive, which is not.

Some critics have dubbed such diets “health horoscopes” — not a bad metaphor. The prescriptions they offer have much in common with the vague platitudes dispensed by astrologers. They might do little harm but neither do they provide insights of value. They also come at a hefty price.

A £300 test from a British company, Sciona, gives people “science-based health and lifestyle advice matched to their genetic profiles.” Though the DNA test is not its only element — a detailed diet, exercise and general health questionnaire is also involved — it is its unique selling point. When Keith Grimaldi, Sciona’s chief scientist, tried the service himself, it suggested a diet rich in green vegetables, wholegrain cereals and garlic to cut his risk of heart disease and stroke. Good counsel for sure, but there are few genetic profiles it would not match. It looks more like good general nutritional advice than bespoke preventive medicine.

There may be risks in this, as the GeneWatch report points out: “The most useful preventive measures — quitting smoking, eating healthily, exercising, living in a healthy environment — are of benefit to everyone. Genetic testing would make these interventions only more expensive and less effective — by restricting help to the ‘genetically susceptible’.”

Chris Martin, Sciona’s chief executive, accepts that the tests have limitations. “Is it personalised? Yes, compared to something like the Atkins diet,” he says. “Is it as personalised as science could make it one day? Absolutely not.”

Most experts agree that genomic insights will eventually lead to tailor-made diets and drugs. But there’s a risk that making overblown claims for genetic testing, before the science can back them up, will undermine trust in something that may ultimately have much to offer. When the public feels a technology’s benefits have been exaggerated, they often turn against it, regardless of potential. GM crops and nuclear power are good examples. It would be sad if this happened to nutrigen- omics and pharmacogenetics.

That is why GeneWatch’s call for tougher regulation is welcome. Companies selling genetic tests do not have to supply data on their predictive power or value for health. That ought to change. A more rigorous attitude to evidence will protect not only consumers, but also a promising area of medical science.

Mark Henderson is the Times science correspondent