A Model for Curing Childhood Obesity in Mexico
When reality is filled with pesky problems, people need an escape. With a new paper in Pediatric Obesity Rossana Torres‐Álvarez et al offer us such an escape. They describe a model for the effectiveness of beverage taxes to reduce childhood obesity in Mexico. Based on assumptions that these taxes will work, they built their model. Unsurprisingly, their model shows that the taxes will work.
Unfortunately, though, reality is on a different track. Six years have passed since Mexico started taxing both sugary drinks and junk foods. But childhood obesity in Mexico is as high or higher than ever. In fact, few countries have higher rates of childhood obesity than Mexico.
Assumptions Make the Model
Any model is only as good as the assumptions behind it. A core principle of good modeling practices is to disclose assumptions explicitly. On this measure we give the current paper low marks. You will not find the word assumption anywhere in this paper. It includes no list of critical assumptions.
However, the authors do mention that they assume no energy compensation in their model. In other words, if children stop drinking sugar-sweetened beverages, they will not compensate at all with calories from anything else.
If you are utterly convinced that sugary beverage taxes work, then this is a reasonable assumption. If not, then it makes no sense.
Disconnected from Reality
Modeling assumptions should always be checked against empiric evidence. However, we can find no discussion of how this model lines up with the health outcomes in Mexico after six years of experience with beverage taxes. Instead, the authors present this as a closed case. They conclude:
The tax represents an effective component of any child or adolescent weight control program, and must be considered as part of any integrated population-level program for children and adolescent obesity prevention.
That is not how models work. A model does not tell us what actually works to reduce childhood obesity. Writing in a landmark paper on best modeling practices, Milton Weinstein et al explain:
In our view, the most important thing to keep in mind in evaluating a health-care evaluation model is that its outputs must not be regarded as claims about the facts or as predictions about the future.
In God we trust. All others must bring data. So we eagerly await real data to support or disprove the speculation that beverage taxes will work to reduce obesity.
Click here for the paper in Pediatric Obesity and here for the Weinstein paper on best modeling practices.
Children in Plaza Hidalgo, photograph by Adam Jones, licensed under CC BY-SA 3.0
Subscribe by email to follow the accumulating evidence and observations that shape our view of health, obesity, and policy.
April 26, 2020
April 26, 2020 at 2:07 pm, Donna Ryan said:
Ted, you have a bee in your bonnet about modeling! Maybe you have been seeing all the different models predicting US response to COVID-19 and it made you peckish.
I think its become clear that the taxing of sugary beverages in Mexico decreased consumption. The taxes were just part of a public health campaign on the importance of avoiding those to improve nutrition. That consumption is sustained.
Of course there are many other drivers of obesity in kids and we can’t expect one thing – a tax on sugary drinks – to reverse obesity in children.
I am worried that you are sending the wrong message. In critiquing the model you seem to be saying that the taxes are useless. You have a lot of political capital, so please blog some more about what we need to do as a society to address childhood obesity. Reducing sugary drink consumption is just one brick in a very very large wall.
April 26, 2020 at 3:58 pm, Ted said:
Thank you, Donna, for your thoughtful comments. You’re right. I am a bit cranky about the misuse of models as proof that an intervention is “effective” to reduce obesity. Certainly, taxes are an effective way to reduce the consumption of anything we tax. I hope that taxing sugary beverages will result in a reduction in obesity. But I’m not aware of any evidence to prove that this is true. Just lots of reasons to think that it might.
You’re right that it’s important to be clear about what we need to be doing to overcome obesity. My belief is that we need more of three things. We need more curiosity about obesity and the dynamics that are driving it. We also need more objectivity about what we know to be true or false and what we suspect might be true. And finally, we need more care extended to the people affected by obesity. This means more access to better care and less bias – either explicit or implicit.
I’ve written about this here and in many other places.
Again, thanks for your thoughtful comments.
April 28, 2020 at 4:30 pm, Simon Barquera said:
“I hope that taxing sugary beverages will result in a reduction in obesity. But I’m not aware of any evidence to prove that this is true.” – Ted, this remind me of similar responses from tobacco companies in the 70s, and is the typical response of soda companies today.
We know there was a reduction of 69,000 tons of sugar/year after the tax. Do you think this is not enough and we need to wait for more evidence? I am glad scientists from 20 countries disagree and have used evaluations and modeling tools to implement soda taxes as part of the policies to improve the food environment.
Added sugar consumption has additional independent effects and there is plenty of evidence of this, so even after modelling full caloric compensation frm other sources, models predict health benefits. (Sanchez LM, et al PLOS Med 2014).
I do not think someone can disqualify a policy that represents a tiny part of a complex problem for not solving it. Particularly with the evidence of reduced consumption. It would make more sense to recommend increasing it if you want more rapid results. – Greetings from Mexico
Simon Barquera
April 29, 2020 at 4:22 am, Ted said:
Thank you, Simon, for taking time to add your perspective. I hope that your belief in the value of taxing beverages will prove to be justified by actual evidence.
It’s worth noting that smoking is not the same thing as drinking sweetened beverages. We have good evidence that tobacco taxes caused fewer people to smoke. Fewer people smoking has led to less lung disease.
On the other hand, Americans have been reducing their consumption of SSBs for two decades now. (https://doi.org/10.1002/oby.22056) So far, this has not been sufficient to cause a reversal in obesity trends. We still have much to learn about this.
May 01, 2020 at 6:39 pm, Patrick Bradley said:
Of course there has been no or little reduction in childhood obesity after 6 years of taxing sugar sweetened beverages in Mexico. However, it took 20 years for the benefit of reduced cigarette consumption following increased taxes on cigarettes to lead to a decline in rates of lung cancer.
The fact that obesity is highly resistant to change and highly heritable indicates that we are witnessing the evolution of a new phenotype resulting from the increased consumption of sugar sweetened beverages and other junk foods.
Thus it may possibly take also at least 20 years to see the benefits of reduced consumption of junk foods in our children because the damage done to their health like the damage done to the health of adults from cigarette smoking persists long after ceasing consumption of the causative agent.
May 02, 2020 at 5:02 am, Ted said:
Thanks for sharing your perspective, Patrick.
The logic of this argument is not strong because:
1. Cigarettes and sweetened beverages are very different.
2. The promise for the tax was to save many lives within ten years.
https://doi.org/10.1377/hlthaff.2018.05469
Over-promising the effects of a policy from the start and then trying to move the goal posts brings serious questions about objectivity.