A new report suggests traffic-light labels have not affected how consumers choose their food - in contrast to the FSA's own study. Nick Hughes reports

With the Food Standards Agency on the verge of delivering its advice on nutritional labelling policy to ministers, the publication of a new report questioning the effectiveness of traffic lights puts the agency in a quandary.

In May this year, the FSA published what it billed as the "most comprehensive and robust study of front-of-pack nutrition labelling published to date", which concluded that the two labelling schemes favoured by consumers both involve colour-coded labels.

Yet new research published by the Oxford University Press in the Health Promotion International journal suggests the FSA's support for traffic lights is informed by flawed research based on what consumers say they do as opposed to what they do in practice. The independent research is based on actual sales data and concludes that there was no association between changes in product sales and the healthiness of selected ready meals and sandwiches following the introduction of traffic-light labels.

The results are "a little surprising", says the report's co-author Dr Mike Rayner. Part of his surprise may stem from the fact that in 2008 he co-authored a European Heart Network study that showed consumers found traffic lights more helpful than GDAs. But, like all previous studies, the EHN research was based on anecdotal rather than physical evidence.

Rayner is at a loss to explain why, three years after traffic lights entered the mainstream, FSA advice is still being informed by declared rather than actual effects. "It's rather to our shame that we've lost the opportunity to do some more sophisticated research into these schemes in the UK," he says.

Rayner, who played a pivotal role in developing the Nutrient Profiling Model and is a firm advocate of traffic-light labels, says the FSA spurned the chance to produce a piece of solid, academic work on consumer responses to nutritional labelling. "Interesting as its research was, it didn't actually get to grips with what consumers do."

In contrast to the FSA study findings, the Oxford report found that in the four weeks after traffic-light labels were introduced to ready meals and sandwiches in a major retailer in 2007, consumers did not start to make healthier choices. In fact, the two healthiest sandwiches actually saw the largest sales decline.

The FSA contends that the Oxford study relates to information supplied by just one supermarket, adding that "comprehensive sales data would be required from all supermarkets tracked over time before any assessment could be made of how it influences purchasing choices". It claims to be in talks with retailers about acquiring such data, but maintains that its own study "looked in detail at how people actually use front-of-pack labels when shopping", despite it eschewing sales data in favour of shopper interviews.

The multiple retailers, with their myriad consumer insights, hold the key to unlocking the truth behind labelling schemes, says Rayner. Many claim to have evidence that their respective labelling schemes have led consumers to make healthier choices. Tesco, for example, claims on its website that sales of lower-fat ready meals rose 7% when GDA signposts were introduced, based on eight-week sales data. As yet, however, no retailer has published research into nutritional labelling in a peer review journal. "One of the problems is getting the retailer to let you have their data, because they treat it as confidential," says Rayner.

In the meantime, the only published studies remain unsubstantiated by physical sales. The FSA maintains that traffic-light labels have helped consumers choose healthier options. But until it demands supermarkets relinquish their sales data in the interests of public health, the FSA will always be taking a shot in the dark about what system really works best.