The following originally appeared on The Upshot (copyright 2015, The New York Times Company).
Americans remain very overweight. According to the Centers for Disease Control and Prevention, about 38 percent of adults were obese in 2013-14, compared with 32 percent just 10 years ago. This is in spite of huge efforts to get people in the United States to eat more healthily.
Policy makers continue to believe that the problem is people’s lack of knowledge that they are wolfing down calorie-rich foods. It is assumed that once Americans know what they are eating, they will eat less, or at least with health in mind. For this reason, many health advocates have called for restaurants to provide people with calorie counts of what they are ordering. Recent mandates mean that by the end of next year, calorie labeling will be required on all menus in chain restaurants and establishments selling food in the United States.
Because many restaurants are already trying menu labeling, we can look at how they have worked, or haven’t, and begin to predict how this widespread regulation might function. For instance, researchers looked at data from 66 of the largest chains — those that posted calories and those that didn’t — and found that average calories per item were 139 calories lower in restaurants that posted their nutritional information.
But this doesn’t mean that calorie counting forces restaurants to reduce the calories they offer. It could just as easily mean that restaurants with lower-calorie offerings are more eager to let you know.
A better understanding of how initiatives could change behavior might come from using the methods of scientists, like prospective trials. Another study did that when it examined how Walmart’s healthier food initiative changed how people bought food. One of Walmart’s methods was adding a front-of-package labeling system indicating items that met certain nutritional criteria. The retailer also offered price reductions on some healthier items.
Over the years of the study, customers bought significantly fewer calories. But most of the changes happened before the start of the program. The conclusions of the researchers were that retailer-based initiatives may not be enough to change behavior.
Another study from this month was also on point. In New York City, menu label mandates began in 2008. Back then, people reported that they saw and used calorie counts more often than people did in restaurants without labels. However, every year after that, fewer and fewer people reported noticing them or considering them. Over time, customers started to ignore the labels. More significantly, at no time did the labels lead to a reduction in the calories of what diners ordered. Even if people noticed the calorie counts, they did not change their behavior.
The problem here is that while all of these studies are being packaged as new, we’ve known about much of this for years. The New York City program was studied after it first went into effect. A 2012 publication in The American Journal of Public Health noted that menu labeling seemed to lead to an increase in calorie intake because people bought more higher-calorie entrees, not fewer. No one should be surprised that three years later the program still doesn’t work. A different menu labeling program in Seattle, whose impact was investigated in a study published in The American Journal of Preventive Medicine in 2011, led to “increased nutrition information awareness, but no decrease in calories bought by parents or children.”
A systematic review, published this year, reviewed all the existing studies looking at menu labeling in the medical literature through October 2013. It found a lot of variation in how the studies fared; many of the studies didn’t occur in real-world environments. But the six controlled studies that took place in restaurant settings showed that over all, menu labeling did not produce any significant changes in what people ordered. The Department of Agriculture’s Nutrition Evidence Library concluded that “limited and inconsistent evidence exists to support an association between menu calorie labels and food selection or consumption.”
Further, it’s important to remember that the labels are only as good as the individual cooks. In a 2011 study published in JAMA, researchers went to restaurants that had menu labeling, bought food and took it to the lab to measure its calorie content. They found that while menus were, on average, pretty accurate, substantial variation existed. About 20 percent of foods purchased had at least 100 more calories than what was reported. This discrepancy continued when researchers repeated the purchases and measurements later.
Worse, the food items with the lowest reported calories — the so-called healthiest — were the most likely to be incorrect.
Menu labeling may be taking our eye off the ball. By offering us what seems to be a solution, it may prevent us from trying other things that might work better. Previous work in Health Affairs showed that training servers to ask if customers might like to downsize three starchy sides induced up to a third of customers to order and eat 200 fewer calories per meal. More recent work in the journal showed that changing the “prevalence, prominence and default nature of healthy options” on children’s menus led to sustained changes in what people ordered.
We can help people make healthier choices in restaurants. Menu labeling might be the method of policy makers’ choice, but it might not be the proper prescription.