Calories on menus are already a fact of life in New York City and were set to appear in a handful of states like California and Oregon in 2011. Instead, thanks to a provision in the health care legislation Obama signed in March, they’ll be required nationwide. The policy calls for all restaurant chains with 20 or more locations to publish calorie counts for all items on all menus. The policy also applies to vending machines, buffets, and bars.

The policy’s advocates and authors claim that it will reduce obesity rates and improve public health. In a press release from The Center for Science in the Public Interest, Margo Wootan, a nutritionist who helped write the calorie count part of the bill said:

"Congress is giving Americans easy access to the most critical piece of nutrition information they need when eating out…. It’s just one of dozens of things we will need to do to reduce rates of obesity and diet-related disease in this country…. Menu labeling at restaurants will help make First Lady Michelle Obama’s mission to reduce childhood obesity just a little bit easier.” (CSPI press release)

In an interview with the LA Times, she expanded on the logic of the claim: 

"People will be able to see that the order of chili cheese fries they are considering will be 3,000 calories.”

Well, probably more like 400-500. But how could she be expected to know that before the law goes into effect?

Kelly Brownell, director of the Rudd Center for Food Policy and Obesity at Yale told the NYTimes that even if some consumers ignore the information, it will affect enough people to create a public health benefit. However, he also hedged his bet—saying that even if it doesn’t make people eat better, it’s an issue of rights as much as an issue of health:

“You don’t need a study that proves anything,” Mr. Brownell said. “You just have a right to know.”

Proof? Who needs proof? His disclaimer is savvy, because now in 5 or 10 years if obesity rates are still the same* or higher and there’s been no significant decrease in cardiovascular disease, diabetes, cancer, or any of the other conditions correlated (albeit often weakly) with obesity, Brownell can claim we’re still better off knowing than not knowing.

I’m not so sure. While I don’t think posting the number of calories is likely to have a significant, negative impact on public health, nutrition is one realm where more information isn’t always better. The usefulness of information always depends on its reliability, relevance, and people’s ability to place it in meaningful context. Calorie counts fail on all three measures, which is why I suspect the new policy isn’t going to have the desired effect on obesity rates or public health.**

Here are a few of the problems with calorie counts I’ll address in this series:

1) The number posted is often wrong (a problem for reliability)

2) Most people don’t know how many calories they burn (a problem for meaningful context)

3) Even though calorie restriction is a highly effective short-term weight loss strategy, it doesn’t work long-term (at least for 90% of dieters) (a problem for relevance)

4) Not all calories are equal (another problem for relevance)

Furthermore, the limited evidence available so far about how calorie counts on menus affect purchasing decisions based on the New York City law is mixed. That calls into question the mechanism by which the policy is supposed to improve public health. Apparently, knowing the calorie content of menu items doesn’t necessarily reduce the number of calories people purchase. And that’s before even beginning to try to measure whether purchasing fewer calories on single visits to restaurants actually leads to weight loss or if people just compensate by eating more on other occasions or eating more often.

One response might be: well, it can’t hurt. I’m also not so sure about that. While I don’t think it’s likely to make public health worse, by reinforcing the idea that your health (or your weight) is based on the number of calories you eat, it may prevent people from taking steps that would actually improve their health, which the preponderance of evidence suggests that calorie-restriction dieting will not.

Part I in this series, on why the number posted is often wrong, coming later today.

*The rate of increase in obesity has already been slowing down so even if it plateaus, that’s not necessarily evidence this or anything else is “working,” it may simply mean that obesity rates have reached an upper limit.

**Two separate issues which are often unjustly conflated. For more on that, see Paul Campos’ The Obesity Myth, J. Eric Oliver’s Fat Politics, Glen Gaesser’s Big Fat Lies, or Michael Gard and Jan Wright’s The Obesity Epidemic: Science, Morality, and Ideology—if you feel like I’ve said that before, it’s because I have. The reason I bring them up again and again is that they completely changed my thinking about nutrition, fatness, and health. The authors of those books all—independently—examined the evidence for the argument that obesity is dangerous and all reached the same conclusion: it’s not, and the belief that it is is based on some shockingly bad science. They also argue convincingly that the actual increases in Americans’ weight in the last few decades are actually quite modest (it’s the rate of people being defined as obese that’s trumpeted, not the amount of weight people have gained on average and some of the increase is based on changes in the definition of “normal” or “healthy” with no medical justification); that the correlations between obesity and disease or early mortality—many of which are quite weak—can be entirely explained by other factors that also happen to be correlated with BMI like differences in physical activity, income, and insurance status; and that weight-loss dieting, especially low-fat and calorie-restrictive dieting, do more harm than good. You don’t have to take my word for it. Substantial portions of the books are available for free online, as are many of the studies they cite (including the CDC study that revised the widely-cited statistic that overweight and obesity causes 300,000 deaths per year in the U.S. and said, effectively, “Actually make that 26,000 and by causes we mean correlates with.”)