Fizzy thinking

Observations on scientific research

Parents fret about the health of their kids, so it's nice when impartial folk come up with reassuring advice. Should I let the children drink that syrupy gloop they're demanding? Ask a qualified scientist. Some public servant out there will have studied the product in question and can give us an answer to trust.

Except that very often we can't, because the public is not paying the scientists. As a consequence, the results may not be impartial. There are scientific papers on the health effects of soft drinks written by researchers who appear to be independent - they work in public universities, for example, rather than for drinks companies. Yet many of those researchers receive grants from drinks firms. And when industry-funded experiments are compared with ones that the public has paid for, the independence of the scientists looks less clear.

According to one recent survey, the first of its kind to look at the beverage industry, scientists are up to eight times more likely to reach positive conclusions about a drink when they have been funded by the beverage firm. In fact, in one study examined by the authors, they could not find a single industry experiment that concluded with bad news for the sponsors.

Equivalent experiments funded by non-industry sources resulted in more than a third having critical things to say about the drinks being studied.

Yet the corrupting influence of big business on scientific results has been documented in detail in the pharmaceuticals industry. We know that regulatory decisions about drugs are routinely undermined by the big companies. There is no conspiracy or fraud here, just a systematic distortion.

But as with the drinks studies, commercially funded drugs trials are more likely to reach positive conclusions. Corporate researchers are also more prone to exaggerating beneficial effects. And negative results, which could signal the end for a product that has absorbed millions of pounds of development money, tend to be watered down or not published at all.

Some drugs get licensed and are subsequently shown to be dangerous, such as the painkiller Vioxx, which was pulled from the market in 2004 after being shown to increase the risk of heart attack. The Vioxx problem might have been spotted sooner had some negative data been published earlier, but such cases are rare. The drugs companies are not setting out to sell dangerous products. They are merely trying to maintain a competitive edge.

That does not make the industry's behaviour acceptable. Patients sign up for clinical trials thinking they are helping medical science, not to become a number in a marketing exercise. And drugs are expensive, so it matters which ones governments buy.

To rein in the pharmaceuticals companies, medical research needs to be transparent. Clinical trials should be registered in advance, so the results cannot be hidden and the data should be made freely available, so that a statistical sleight of hand cannot be used to conjure good results from bad.

And, as we become more aware of the impact on health of the chemicals used in food processing, similar standards should be applied to the food and drinks industry.

So why aren't scientists in both industries already abiding by such standards? The World Health Organisation has a plan in place, but no power to enforce it. That rests with governments, and ours is ambivalent. New science institutions are springing up in China and India. Pharmaceutical companies will not hesitate to head east should it make economic sense.

In the meantime, we should be wary of "independent" scientific research findings.

This article first appeared in the 22 January 2007 issue of the New Statesman, Sex and politics