Fat chance of cutting calories

“Given the lack of evidence that calorie posting reduces calorie intake, why is the enthusiasm for the policy so pervasive”, asked an editorial in the American Journal of Clinical Nutrition in February this year.

It’s a good question, considering that mandatory nutrition menu labelling in chain restaurants is being rolled out in New South Wales and Victoria this year. Major fast food outlets, pizzerias, cafes, bakeries, and juice bars will have to show the amount of kilojoules in each item on their menu.

Mandatory menu labelling is a useful case study in the lack of good evidence behind much public health paternalism. “Useful” because it has a stronger evidentiary base than most. But that’s not saying much.

New York City began mandating nutrition labels in chain restaurants in 2008. The New York Department of Health predicted big things – 150,000 people would be saved from obesity within five years, preventing “more than” 30,000 cases of diabetes.

Preventative health policy is contagious. Menu labelling was picked up by Australia’s Preventative Health Taskforce, and included in its paper on obesity released in October 2008. The New York Health Department’s ambitious but well-publicised claims were quoted.

The critical piece of scholarly evidence that backers of this policy have relied upon is a 2008 study published in the American Journal of Public Health (it predates New York’s mandatory measures).

The study’s authors looked at a sample of customers from 11 fast food chains across New York’s five boroughs. All provided nutritional information to customers. But only one provided that information at the point of sale. The rest simply displayed it on a website or tucked away in the store.

No shock then only 4 per cent of customers reported seeing it in those latter stores, as opposed to 32 per cent who saw it when displayed on the menu.

A third of those who reported seeing the information purchased items with fewer calories.

Pretty conclusive, you’d think. Except for the fact that the restaurant which displayed the calories was Subway – a chain which has deliberately marketed itself to the health conscious, not least of all by displaying calories on its menus.

That’s no incidental detail. The big issue in obesity policy is that many interventions require a pre-existing preference to eat healthy. Subway regulars are likely pickier about their food intake than, say, McDonald’s regulars. There’s a reason their mascot is a guy who lost of lot of weight.

Even then, the overwhelming majority of Subway customers did not reduce calories even after they saw the information.

Still, as far as evidence goes the paper isn’t bad – equivocal in the details, but better than nothing. So Kevin Rudd’s Preventative Health Taskforce cited the paper, and recommended menu labelling (Of course, the taskforce didn’t mention the chain in question was Subway).

But evidence for the effectiveness of menu labelling has slid backwards fast since the policy was made mandatory for all New York chains in 2008.

A 2009 paper in Health Affairs found that while 27 per cent of people self-reported that the calorie labelling influenced their purchase decisions, analysis of actual purchase receipts did not bear this out. People were saying one thing, but doing another. Just one more example of why we shouldn’t put much stock in social science surveys.

The Health Affairs study looked at New York’s low income communities. Calories consumed had actually gone up slightly in some restaurants. Some academics have supposed that menu labelling helps a subset of consumers to calculate the most calories they can purchase for their buck.

A similar phenomenon was detected in a 2009 study in the American Economic Review, which found that “providing calorie information may have small effects on food choices, but may also produce perverse effects.”

A study published this year in BMJ also found that, while chain by chain the response could vary significantly, mandatory labelling led to no overall decline in calories purchased.

Of course, the New York City Health Department now says it has new – and unpublished – evidence which apparently proves the regulation works in four of the 13 chains it surveyed (alternatively: it doesn’t work in nine).

So it’s fair to say that in 2011 there is no good reason to suppose that mandatory menu labelling will have a discernable effect on obesity rates.

That the one real-world example of mandatory labelling appears to have failed should give pause.

But that failure has not stopped public health lobbyists from working hard to impose them. Or governments from pressing ahead.

The Australian Obesity Policy Coalition’s policy brief on menu labelling cites the Subway research, without naming Subway, and without mentioning the easily accessible recent studies of mandatory menu labelling’s real-world effectiveness. The brief was written this year, so no excuse.

Newspaper articles announcing the Victorian decision also only referred to the Subway research (again, without naming the chain), but not the subsequent studies. Politicians leant on the once-over-lightly – and now outdated – Preventative Health Taskforce report, which, as we’ve seen, also relied on this 2008 paper.

The Heart Foundation, to its credit, did a proper literature review up to 2010. And having done so, it could conclude no more than mandatory menu labelling should be trialled to see if more concrete evidence could be found.

Nevertheless, joining John Brumby at the policy’s announcement, the co-chair of the Heart Foundation claimed implementing menu labelling permanently was a “fantastic initiative” and “definitely will raise the awareness around what people are eating”. No calls for a limited trial. No equivocation. No scholarly dispassion there.

It’s almost as if the evidence is beside the point. The public health community have their mind fixed on mandatory menu labelling.

Menu labelling is now going to be rolled out across the United States. In Australia, activists are talking about the need for a “national approach”.

Sure, as far as regulations go, it’s a relatively minor one. Yes it is expensive to implement. But many restaurants are doing it anyway (Subway was merely first). The legislation is limited to large chains which can spread the cost. And if it saves just one calorie…

But the published data can be very inaccurate. Sixteen-year-old pizza chefs don’t exactly measure pepperoni by the gram. The McDonald’s auditors can’t control exactly how many fries constitute “medium” fries. There’s a lot of variation in even the most regimented cooking.

It should be concerning that labelling could result in poor people consuming more calories rather than less – a result which nobody predicted before the New York regulations were imposed. Unintended consequences are like that.

Nevertheless, as a case study in public health paternalism, it should be more concerning how policies which have little evidence to support them gather an unstoppable inertia.