Repeat after me: all models only say what they are told to say.
NEW: As a result of our junk food advertising ban on @TfL, nearly 100,000 cases of obesity have been prevented since 2019.
It’s expected to save the NHS over £200 million. An incredible result. https://t.co/SIUi25RG95
— Mayor of London (@MayorofLondon) August 2, 2022
An incredible result, says the mayor of London. Let’s see how this incredible result came to be. From a news story:
Junk food advertising restrictions on Transport for London (TfL) networks have prevented almost 100,000 obesity cases, research suggests.
The advertising policy, which has been in place since 2019, could save the NHS more than 200 million, researchers claim.
They estimate the policy has directly led to 94,867 fewer cases of obesity than expected (a 4.8 per cent decrease), 2,857 fewer cases of diabetes, and 1,915 fewer cases of cardiovascular disease.
Interesting they call these “obesity cases”. Cases. Let that pass and note that this is 94,867 fewer cases, and not 94,866, as some sources earlier reported.
The peer-reviewed paper is “The health, cost and equity impacts of restrictions on the advertisement of high fat, salt and sugar products across the transport for London network: a health economic modelling study” by Chloe Thomas and others in the International Journal of Behavioral Nutrition and Physical Activity.
The Abstract begins with this sentence, starting off on a false footing: “Policies aimed at restricting the marketing of high fat, salt and sugar products have been proposed as one way of improving population diet and reducing obesity.”
Eliminate processed sugar if you like, but decreasing fat is the opposite of “solving” (notice how we’re always discussing government “solutions”?) the problem of too many fat people.
Never mind that. Here’s what they did.
In 2019, Transport for London implemented advertising restrictions on high fat, salt and sugar products. A controlled interrupted time-series analysis comparing London with a north of England control, suggested that the advertising restrictions had resulted in a reduction in household energy [calories] purchases.
Now, most don’t know this, but London is a huge heterogeneous city, and the north of England, north of a land called Yorkshire, is not much like London at all. For instance, in Yorkshire they speak mainly English. Comparing the two areas is thus likely to confirm the two areas aren’t the same.
Here are their Methods:
A diabetes prevention microsimulation model was modified to incorporate the London population and Transport for London advertising intervention. Conversion of calorie to body mass index reduction was mediated through an approximation of a mathematical model estimating weight loss. Outcomes gathered included incremental obesity, long-term diabetes and cardiovascular disease events, quality-adjusted life years, healthcare costs saved and net monetary benefit. Slope index of inequality was calculated for proportion of people with obesity across socioeconomic groups to assess equity impacts.
A microsimulation model. Maybe this microsimulation had something to do with wee p-values? Too good to check.
Anyway, they stopped showing forbidden ads about bad foods in one location. Who noticed? Who knows. But
The study [by a survey firm] found that purchases of HFSS [government-defined bad] food increased over time in both locations. However; following the TfL intervention in London, relative purchases of energy decreased by an average of 1,001 (95% CI: 456 to 1,546) calories per household each week compared to the counterfactual, which was constructed using the pre-intervention trend in London and incorporating the changes seen in the North of England (to account for seasonal and secular changes common to both areas).
See that? Models of models of models. Based on sketchy data—the food bought was determined by food-recall surveys. All of which conclude what was desired: that the lack of ads caused people to buy less of certain foods.
And even if it’s all fine, which it isn’t, that’s a thousand calories less per household each week, which is to say on average 143 or so calories a day, per household. I don’t know how many are in an average household, but suppose it’s 3. Then you have about 50 calories of the “bad” food reduced per person. Plus or minus. How much more in calories people got from non-government defined “bad” food we don’t know.
An extra 50 calories a day—a cookie?—isn’t going to cause obesity, especially from only eliminating “bad” food ads in train stations.
And we don’t know what people ate, only what they bought. Our authors solve this by assuming “that reductions in weekly calorie purchase could be directly equated with reductions in weekly calorie consumption”. (I did the calculation above without having read the whole paper: when I did, I was happy to discover the authors came to an estimate of 55 calories per day.)
To determine eliminating the cookie is going to prevent obesity requires yet another model. And to determine that reduced obesity leads to fewer cases of diabetes and heart disease requires other models.
This paper is therefore a modelapalooza. All in service of providing rulers with The Science they need to justify their policies.
Buy my new book and learn to argue against the regime: Everything You Believe Is Wrong.
Subscribe or donate to support this site and its wholly independent host using credit card or PayPal click here; Or go to PayPal directly. For Zelle, use my email.
Data, we don’t need data.
We have a model, correction, we have a microsimulation model.
{What a travesty of the word ‘research}
Not supporting the stupid “study”, but 55 calories per day is four pounds per year. Which also seems tiny, unless you keep doing it for 10 or 20 years. Habits are hard to break.
They could just close London transport completely and make everyone walk or bike to their destinations which would be much more effective at eliminating obesity. And think of the ancillary benefits for climate change by eliminating all the fossil fuel usage of London transport. My model says there would be no negative side effects to this plan worth worrying about and therefore is totally safe and effective.
Losing weight by government. It’s so easy!
Unless you factor in that overweight people are also people that like to eat. A lot. And they tend to get hungry. All the time.
The stupidity of Experts is truly mind-boggling.
Implicit in this research scenario is the vile concept of “Nudging.”
That is, government can (and should) manipulate the Eloi by subtle actions (ads in underground stations, for example). These manipulations are done by Experts for the benefit of the Eloi.
Nudging is the velvet glove covering the iron-mailed hand of the State. Studies like this are bricks in the wall of State justification for actions/policies that manipulate its subjects.
Wonder if these modelers had to do a Human Subjects Protection form to experiment on the population of an entire city.
That tub Boris Johnson is not the only political porker in Albion but, that aside- how many of these researchers are fat, not the real Governor of Georgia fat, but so fat they’d call themselves unhealthy?
Modelapalooza! My favorite new word.
Now that “Data Science” is a recognized discipline, it’s sure to follow.
“How can I help you?”
“I’d like one family size bucket of energy and two medium energy drinks please.”
“That will be £12.50 please.”
“Here you go”
“Sorry, we don’t accept cash. Credit or debit only”
Food recall surveys? Isn’t that a subtle form of confirmation bias?
Most everyone is familiar with the food “police” ranting about snack food, junk food, convenience food, and fast food being high in fat, sugar, and salt, therefore no one readily admits to consuming them, or admits to as much as ctually consumed. It’s fashionable to hedge on that topic, so it’s likely whatever is reported is vastly under-reported. People are only human.
I guess it should be expected: a junk study on junk food.
Food recall surveys are notoriously inaccurate.
Sugar, salt, and fat! One can forgive imbecility when it is congenital and cannot be helped. Learned imbecility is unforgivable.
Sugar: definitely to be minimised, but harmless in even large doses as long as the average intake is kept low.
Salt: essential to life; the correct intake can, with practice, be left to the appetite.
Fat: For a start, what kind of fat? Animal fat from healthy animals, which is not only a vital nutrient but the only nutrient that cannot be overdone – try to eat too much fat and your stomach rebels. On the other hand, vegetable oils are mostly harmful to a greater or lesser extent.
What about refined carbohydrates? Those, and vegetable oils, are the worst offenders.
Nice one, the tweet and the study are really unscientific nonsense if anything. On top of what you’re writing, the attribution of causation to the ad-intervention is at the butterfly effect level.
That a reduction of 50-60 calories per day, in a context of readily available high-calorie foods, would effectively disrupt the behavioral patterns behind developing obesity – to the exclusion of ALL other potential social and environmental factors? It’s just amazing. Mind-boggling. And add to that, almost 100 000 “cases” of obesity prevented – in three years? Obesity generally develops over many years, and serious researchers in the field are well aware of the vast number of causes behind the condition. This is just mocking them, basically.