Ultra-processed food - my comments to the House of Lords

The House of Lords Food, Diet and Obesity Committee started its pointless proceedings in February and is still slogging on. It has managed to avoid asking anybody who is sceptical of the ultra-processed food panic from appearing (last week it heard from three 'ambassadors' from the Food Foundation and three 'ambassadors' from Jamie Oliver's BiteBack front group). The committee did, however, ask for written submissions so this is what I sent them back in February after Dimbleby, Spector and Coco Pops guy kicked things off....

I was concerned to see that six out of the seven people who appeared at your first three oral evidence sessions were activists and/or authors who subscribe to a particular view of the food system and favour heavy regulation. Whilst I am sure you plan to invite some mainstream scientists, economists and consumers to future sessions, I felt it might be of use to respond to some of the questions in your call for written evidence below.

The definition of a) ultra-processed food (UPF) and b) foods high in fat, sugar and salt (HFSS) and their usefulness as terminologies for describing and assessing such products.

The official HFSS definition, which campaigners use as the implicit definition of ‘junk food’, creates many absurdities, partly because the limits laid out in the Nutrient Profiling Model (NPM) are so low. HFSS food includes raisins, sultanas, most tinned fruit, most yoghurts, two-thirds of morning goods, nearly all cheese, cream crackers, tomato soup, hummus, ham, pesto, cereal bars, olive bread and salami. Whether these foods are ‘high’ in fat, sugar and/or salt is a matter of opinion, but few people would view all of them as ‘junk food’ and want to ban them from being advertised. It is doubtful whether the general public considers even very high sugar foods like jam or very high fat foods like butter to be ‘junk food’.

Campaigners such as Jamie Oliver have exploited the confusion about what ‘junk food’ is (it has no scientific definition and is legally meaningless) to push for legislation against the much broader category of HFSS food. Having capitulated to these campaigners in 2020, the government has tacitly acknowledged that it bought a pig in a poke by exempting a number of HFSS foods from its list of ‘less healthy’ food in the Health and Care Act (2022). There is no scientific justification for exempting these foods - the NPM has the virtue of at least being consistent - but they are so far removed from the popular idea of junk food that the government would look ridiculous if it included them.

‘Ultra-processed food’ is an even broader category which demonises the vast majority of food that isn’t eaten raw or prepared by hand in a domestic kitchen. As a phrase, it was never uttered in the House of Lords until 2020 but there has been a flurry of mentions and several specific debates in the chamber since Chris van Tulleken published his book Ultra-Processed People in April 2023. Van Tulleken is responsible for introducing many people to the concept of UPFs. He makes many claims in his book and in his media appearances that are at best contentious and at worst demonstrably false. For example, based on a misreading of an academic study, he falsely claims in Ultra-Processed People that Chilean miners burn the same number of calories as sedentary office workers and that exercise therefore does not help people lose weight. He has falsely claimed on Twitter/X that the very idea that exercise assists weight loss was invented by the Coca-Cola company.

UPF is an even worse classification system than HFSS because whereas fat, salt and sugar are known to be harmful to health when consumed in large quantities, UPF portrays all but the most basic ingredients as harmful. UPF is characterised by the presence of emulsifiers, preservatives, artificial flavourings and additives, all of which are generally recognised as safe by scientists but which are portrayed as somehow dangerous (carcinogenic or obesity-causing) when used in industrial food processing. Anti-UPF activists have been unable to show that any of these ingredients - let alone all of them - are individually harmful. Their evidence comes from observational studies in which people who eat UPFs appear to have worse health outcomes than people who do not. Leaving aside the well known weaknesses of nutritional epidemiology, the use of such a broad category as UPF limits what can be learned from such research. It zooms out when we need to zoom in. As a category, UPF may include foods that have a causal relationship with certain diseases, but lumping them together with everything made in a factory does not help us tease out which ones or why. Critiquing a recent study which claimed that UPF consumption causes cancer and heart disease, Visioli et al. (2024) say:

 

‘the association between UPF consumption and the risk of multimorbidity would disappear if the data were adjusted not only for the consumption of sugary or artificially sweetened beverages, but also for foods of animal origin at the same time. Indeed, in our opinion, the article underlines the absolute need to return to the evaluation of foods on the basis of their nutritional role (including their nutrient composition, quantities consumed, metabolic effects, etc.) and not on the basis of their degree of processing.’

 

I agree.

One peculiarity about the UPF system is that some of the few ingredients that are not portrayed as inherently dangerous are fat, sugar and salt. So long as they are used to make cakes, biscuits, etc. in a domestic kitchen, they are de facto healthy. It is the industrial processing, not the nutrients, that somehow creates the risk. This not only turns the scientific consensus of recent decades on its head but is palpable nonsense and it is a sign of how quickly the West is retreating from Enlightenment values that anyone is taking it seriously.

Growing fanaticism from certain policy entrepreneurs would lead us to portray a large number of food products as inherently dangerous and urge us to abstain from them all. This is not only unscientific but useless in terms of practical public policy because people are never going to abstain from eating chocolate and bacon let alone sliced bread.

The cost and availability of a) UPF and b) HFSS foods and their impact on health outcomes.

Processed/HFSS/UPF food is typically more expensive than what the government defines as healthy food (fruit, vegetables, starchy carbohydrates, white meat, etc.). Campaigners often claim the opposite by comparing food prices using the cost-per-calorie method. This turns the virtue of being low in calories into a vice since the cost-per-calorie naturally rises as the number of calories in the product declines. A better method is to measure the cost per meal or the cost per kilogram. Under these measures, the cost of healthy eating is relatively low and is far lower than the cost of eating takeaways and fast food (Snowdon 2017).

The impact on heath of HFSS and UPF has been partially addressed in the previous answer, but it is worth mentioning the study by Kevin Hall et al. (2019) which will doubtless be cited a lot in the written and oral evidence to your inquiry. By the low standards of nutritional science, this was a well conducted randomised controlled trial. It found that the people who ate the UPF diets consumed significantly more calories than those on the non-UPF diet and that the former gained weight while the later lost weight. Unfortunately, due to cost pressures, the researchers were not able to offer UPF and non-UPF versions of the same meal. Instead, the UPF group was given a diet that differed more fundamentally than in the way it was processed. For example, one lunch consisted of burger and fries for the UPF group while the non-UPF group got salmon and green beans. The non-UPF food was not merely non-UPF, but was minimally processed which perhaps explains why those who ate it lost weight. As a control group, it would have been more useful (and more realistic) to give them processed (but not ultra-processed) food. The hypothesis that ultra-processing causes over-eating/obesity can only really be tested by giving people processed and ultra-processed versions of the same meal, but no RCT has yet done this to my knowledge.

Lessons learned from international policy and practice, and from the devolved administrations, on diet-related obesity prevention.

I am aware of no country that has reduced obesity rates through deliberate government action (although a few socialist countries have done so as an unintended consequence of its economic policies, e.g. Venezuela). An obvious approach would be to look at comparable countries that have significantly lower rates of obesity than the UK and emulate their policies. This is rarely suggested by ‘public health’ campaigners, presumably because countries with lower obesity rates do not have the kind of taxes and bans they support.

The impact of recent policy tools and legislative measures intended to prevent obesity.

The most significant measure in recent years has been the sugar tax which plainly had no impact on childhood obesity (rates of which rose for three successive years after it was introduced). The ban on HFSS food advertising on television during children’s programmes also failed to have any impact on rates of obesity. Food reformulation didn’t work because consumers couldn’t be forced to eat the reformulated products. It is highly doubtful that forthcoming bans on advertising and promotion will lead to a decline in obesity. Even the government’s own Impact Assessment, based on a starry-eyed interpretation of weak and biased research, only predicts a trivial effect on calorie consumption.

The persistent failure of anti-obesity policies to reduce obesity never leads to any apologies or resignations from the activists and academics who campaign for them. Rather than repealing unsuccessful policies, governments tend to double down with further restrictions, thereby creating a runaway train of regressive, anti-competitive policies that logically leads to some form of prohibition (as we are now seeing with tobacco). Policy entrepreneurs who seek to regulate the food supply are in the privileged position of never having to show that their policies work. This sets them apart from doctors and nurses who cannot, in the modern era, get away with conducting unnecessary and harmful operations or prescribing useless drugs.

No one expects a single policy to reduce obesity rates to zero, but it should not be too much to ask for them to reduce them by a little bit, even if only a fraction of one per cent. This, however, has so far been beyond the purported experts of Britain’s lavishly funded ‘public health’ institutions. To disguise their failings, they have invented something called the ‘whole systems approach’ to obesity which amounts to throwing as many untested and unproven policies at the problem as possible, regardless of cost and unintended consequences, in the hope that one or more of them might work. When each of them predictably fails to move the dial, they mutter something about there being ‘no silver bullet’ and move on to the next idea. This, I submit, is neither a wise nor evidence-based approach to policy-making.