The science behind food

What they may not realize is that a clinical trial can be anything from ten people using a product for a week and self-reporting their observations, to a large-scale placebo-controlled studylasting several years under rigorously monitored conditions.

There is a huge gulf between the two, and a company that tries to pass off its research as weightier than it is risks the wrath of the authorities descending upon it.

Last month, L'Oreal was forced to withdraw two advertisements in the UK featuring Claudia Schiffer, after the Advertising Standards Agency ruled that the claims made for the skin care productswent beyond what the company's tests had shown.

In the United States, PepsiCo-owned Tropicana Products reached a settlement with the US Federal Trade Commission earlier this year over claims that its Healthy Heart orangejuice could reduce blood pressure, cholesterol and homocysteine levels. The FTC alleged that there was no clinical data to back up these claims.

Essentially, there is nothing wrong with companies carrying out their own research and communicating their findings in their marketing efforts. After all, while scientists only really respectstudies published in peer-reviewed journals, such journals would be busting out of their bindings if they were to publish the results of every single worthy study.

But companies should never take advantage of their target audience's respect for science by using slapdash methodology or over-egging unremarkable results. Rather, they need to carry out theresearch as rigorously as if the consumer had all the know-how of the leading scientist in their field.

If companies do not do this of their own volition, there is a chance that legislators may wrest control away from them and prescribe a certain level of scientific enquiry before any marketingclaims can be made.

At the very least, then, the results for the substance under trial should be compared with those produced by an inactive placebo. Keeping both participants and administrators in the dark as towhich is which ('double blind') can help prevent human expectations interfering with the findings. And randomly assigning participants to groups and switching them midway though can reduce thelikelihood of bias.

If any bias or confounding factors could have come into play, declaring and discussing them will win greater respect than glossing them over and opening the entire study to criticism.

Budgetary constrains might limit the number of participants and time period, but in general, the larger and longer the trial, the more likely the findings are to be accurate.

Such rigour might be difficult to convey in the glamorous marketing hype that swathes a new product. But if a company cannot stand up and say, honestly and truly, it has done the very best by itsscience, it has no-one to blame but itself when consumers lose faith in its so-called scientific proof.

For when it comes to sales success, the proof is in the pudding. Health claims are best reserved for products that really are proven to do what they say they do.

Jess Halliday is editor of NutraIngredients-USA.com. She has worked across broadcast, print and online media in the United States and Europe.​