Restaurant Menus for Fewer Cancer Deaths?

Breathtaking. That’s the only word we can find to describe the claims coming from a cost effectiveness study of calorie labeling on restaurant menus for preventing cancer deaths. Published yesterday in BMJ Open, this study is already generating headlines like this one:

“Thanks to calorie-counting menus,
fewer Americans are dying
of obesity-related cancers”

Making an Argument Instead of Finding Evidence

Rather than look for actual evidence of an effect on health outcomes, these investigators constructed a model to demonstrate the possibility that calorie labeling on restaurant menus might be cost-effective for preventing cancer deaths. With the right assumptions, they found they could argue it might work. So their press release gives fans of this policy something to like with this headline:

“Menu calorie labels estimated
to save U.S. billions on cancer care”

To their credit, the investigators acknowledged the limitations of this work. They tell us:

“Given the nature of modeling research, this study does not provide a real-world evaluation of the impact of policy implementation on health and economic outcomes.”

In other words, this is a feasibility study, answering the question of what effect these policies might have. Not a study of the effects they are having in the real world after five years of implementation. Deep within their paper, they tell us their modeling is most sensitive to assumptions about the translation of changes in diet to changes in BMI.

This is important because the very best evidence on the effects of menu labeling tells us only that it leads to small, short-term improvements in diet quality. Changes in BMI? No evidence exists. Sustainable long-term effects? Unknown.

Ought to Work ≠ Actually Works

The human power for rationalization is something no one should underestimate. The authors of this modeling exercise clearly believe in menu labeling as an article of faith. They tell us that testing the effects of nutrition policies “is extremely difficult and often implausible,” so simulation modeling can provide complimentary evidence and flexibility for informing policy.

However, we’re with W. Edwards Deming on this one: “In God we trust. All others must bring data.” Will small changes in diet quality from menu labeling be lasting and beneficial to health? Or will they be transient and washed out by compensatory changes in other aspects of diet? We need real answers to these questions. Assumptions are not good enough.

We simply do not know, nor do we have compelling evidence to suggest that menu labeling will have a net positive effect on population health. What we do know is that decades of efforts to educate, nudge, cajole, coerce, and tax people into more healthful eating patterns have not put a dent in the prevalence of obesity.

Maybe we should take the need for real evidence more seriously and test some new ideas for obesity prevention. In both science and public policy, curiosity is a virtue.

Click here for the new modeling study in BMJ Open and here for further perspective on menu calorie labeling.

Menu, illustration by Coreen Mary Spellman / Wikimedia Commons

Subscribe by email to follow the accumulating evidence and observations that shape our view of health, obesity, and policy.


April 19, 2023