- Aug 15, 2018
- 602
- 346
- 63
Following up on our earlier discussion about ultra-processed foods, here's something I think deserves more attention: **the way food is built and eaten may matter alongside the ingredients on its nutrition label.** That doesn't mean every packaged food behaves the same way, and it certainly doesn't mean a single experiment settled every question about diet.
We've all watched these threads split into two camps. One side acts as if “ultra-processed” automatically explains everything. The other says calories are calories, end of discussion. The published research gives us a more interesting middle ground. It asks how the food environment affects the amount people eat when they're free to decide.
**The inpatient study that changed the conversation**
Kevin Hall and colleagues published a randomized crossover trial in *Cell Metabolism* in 2019. Twenty adults stayed at a research facility and experienced two different diets: two weeks of ultra-processed foods and two weeks of unprocessed foods, in randomized order. The researchers offered meals designed to be comparable in presented calories, energy density, macronutrients, sugar, sodium and fibre, while participants could eat as much or as little as they wanted.
On the ultra-processed diet, participants consumed an average of **508 more kilocalories per day**, with the paper reporting an uncertainty estimate of ±106 kilocalories per day. Weight rose during that period and fell during the unprocessed period. That's a striking finding because the laboratory setting reduced many everyday sources of dietary variation. The study measured what happened when people had access to these particular menus, not merely what they said they intended to eat.
But let's be equally clear about its limits. Twenty participants over four weeks isn't a population-wide lifetime experiment. A difference between two menus also doesn't identify which feature of those menus caused the outcome. Texture, eating rate, energy intake rate, food form and other properties may interact. The researchers themselves discussed possible mechanisms rather than declaring one ingredient guilty.
**Eating rate is a serious research question**
Hall's team found that people ate faster during the ultra-processed phase. That matters because eating speed can change how much energy enters the body during an eating occasion before fullness-related feedback has fully developed. It is one plausible pathway, not a completed explanation of the entire study.
There's independent evidence here too. Eric Robinson and colleagues published a systematic review and meta-analysis in the *American Journal of Clinical Nutrition* in 2014 examining eating rate, energy intake and hunger. Across the studies available, slower eating was associated with lower energy intake in experimental comparisons, although study methods and findings varied. The review didn't establish that every slow-eating manipulation produces the same result in every setting.
I also found a particularly relevant experiment by Forde and colleagues published in *The American Journal of Clinical Nutrition* in 2017. The researchers compared different textures of rice porridge in controlled breakfast studies. In one experiment involving **61 adults**, the thin and thick versions produced markedly different eating rates: **140 grams per minute versus 77 grams per minute**. Texture and energy density were studied together, illustrating how physical food properties can influence eating behaviour.
This is the kind of detail that gets lost when online arguments reduce everything to an ingredient list. Physical structure changes the pace and experience of eating. The research isn't an instruction to anybody about what to eat; it's a reminder that researchers measure more than the nutrient totals printed on paper.
**What exactly does “ultra-processed” mean?**
Carlos Monteiro and colleagues explained the NOVA classification in *Public Health Nutrition* in 2019. NOVA groups foods by the extent and purpose of industrial processing. The ultra-processed category generally describes industrial formulations with substances and additives that are uncommon in ordinary kitchens. It's a classification framework intended for research and public-health analysis.
The useful point is that **NOVA is not itself an appetite measurement**. It doesn't tell us a person's eating rate, the sensory qualities of a specific meal, or how much they actually consumed on a given day. Items within a broad processing category can differ considerably. That makes it possible to take the classification seriously while still asking whether its categories are too heterogeneous for certain scientific questions.
Processing can also have different purposes. Preservation, safety, convenience, texture and palatability are not identical concepts. It's possible to discuss the effects of food structure without treating every technique as interchangeable. The real scientific argument is about what variables best explain observed outcomes.
**The food environment is bigger than willpower**
Now zoom out beyond the controlled facility. Portion presentation, eating opportunities, social context, availability, time pressure and food texture can all influence the circumstances in which people eat. Studies of one isolated factor are easier to interpret experimentally, but real environments combine them. That is both why tightly controlled experiments are valuable and why their findings shouldn't be copied directly into sweeping claims about daily life.
Palatability deserves care as a term. It refers to how pleasant a food is perceived to be, but “tastes good” isn't a scientific synonym for “automatically causes overeating.” Palatability may interact with texture, energy density, variety and prior experience. Separating those relationships requires experiments that deliberately vary one feature while holding others as steady as possible.
What I take from the Hall trial is a credible demonstration that diet patterns can change spontaneous energy intake under controlled conditions. What I take from the eating-rate literature is a plausible set of mechanisms worth testing further. Neither point requires moralizing foods or assuming that one person's experience overrides a randomized comparison.
If anyone wants the original papers, the names to search are **Hall et al., *Cell Metabolism* (2019)**; **Robinson et al., *American Journal of Clinical Nutrition* (2014)**; **Forde et al., *American Journal of Clinical Nutrition* (2017)**; and **Monteiro et al., *Public Health Nutrition* (2019)**. They ask related but distinct questions, which is exactly why reading them together is useful.
There's room here for a real forum conversation that isn't a food-label shouting match. Let's compare how we interpret the methods and the unanswered questions, rather than turning an interesting body of research into another all-or-nothing camp.
1. In the Hall inpatient trial, which possible explanation interests you most—eating rate, texture, energy intake rate, or some combination—and what experiment would best separate them?
2. Do you think the broad NOVA categories are useful for population research while being too coarse to explain differences between individual foods?
3. When you read a food-environment study, what matters more to your confidence: tight laboratory control, realistic everyday conditions, or independent replication across both?
— OLYMPIC
We've all watched these threads split into two camps. One side acts as if “ultra-processed” automatically explains everything. The other says calories are calories, end of discussion. The published research gives us a more interesting middle ground. It asks how the food environment affects the amount people eat when they're free to decide.
**The inpatient study that changed the conversation**
Kevin Hall and colleagues published a randomized crossover trial in *Cell Metabolism* in 2019. Twenty adults stayed at a research facility and experienced two different diets: two weeks of ultra-processed foods and two weeks of unprocessed foods, in randomized order. The researchers offered meals designed to be comparable in presented calories, energy density, macronutrients, sugar, sodium and fibre, while participants could eat as much or as little as they wanted.
On the ultra-processed diet, participants consumed an average of **508 more kilocalories per day**, with the paper reporting an uncertainty estimate of ±106 kilocalories per day. Weight rose during that period and fell during the unprocessed period. That's a striking finding because the laboratory setting reduced many everyday sources of dietary variation. The study measured what happened when people had access to these particular menus, not merely what they said they intended to eat.
But let's be equally clear about its limits. Twenty participants over four weeks isn't a population-wide lifetime experiment. A difference between two menus also doesn't identify which feature of those menus caused the outcome. Texture, eating rate, energy intake rate, food form and other properties may interact. The researchers themselves discussed possible mechanisms rather than declaring one ingredient guilty.
**Eating rate is a serious research question**
Hall's team found that people ate faster during the ultra-processed phase. That matters because eating speed can change how much energy enters the body during an eating occasion before fullness-related feedback has fully developed. It is one plausible pathway, not a completed explanation of the entire study.
There's independent evidence here too. Eric Robinson and colleagues published a systematic review and meta-analysis in the *American Journal of Clinical Nutrition* in 2014 examining eating rate, energy intake and hunger. Across the studies available, slower eating was associated with lower energy intake in experimental comparisons, although study methods and findings varied. The review didn't establish that every slow-eating manipulation produces the same result in every setting.
I also found a particularly relevant experiment by Forde and colleagues published in *The American Journal of Clinical Nutrition* in 2017. The researchers compared different textures of rice porridge in controlled breakfast studies. In one experiment involving **61 adults**, the thin and thick versions produced markedly different eating rates: **140 grams per minute versus 77 grams per minute**. Texture and energy density were studied together, illustrating how physical food properties can influence eating behaviour.
This is the kind of detail that gets lost when online arguments reduce everything to an ingredient list. Physical structure changes the pace and experience of eating. The research isn't an instruction to anybody about what to eat; it's a reminder that researchers measure more than the nutrient totals printed on paper.
**What exactly does “ultra-processed” mean?**
Carlos Monteiro and colleagues explained the NOVA classification in *Public Health Nutrition* in 2019. NOVA groups foods by the extent and purpose of industrial processing. The ultra-processed category generally describes industrial formulations with substances and additives that are uncommon in ordinary kitchens. It's a classification framework intended for research and public-health analysis.
The useful point is that **NOVA is not itself an appetite measurement**. It doesn't tell us a person's eating rate, the sensory qualities of a specific meal, or how much they actually consumed on a given day. Items within a broad processing category can differ considerably. That makes it possible to take the classification seriously while still asking whether its categories are too heterogeneous for certain scientific questions.
Processing can also have different purposes. Preservation, safety, convenience, texture and palatability are not identical concepts. It's possible to discuss the effects of food structure without treating every technique as interchangeable. The real scientific argument is about what variables best explain observed outcomes.
**The food environment is bigger than willpower**
Now zoom out beyond the controlled facility. Portion presentation, eating opportunities, social context, availability, time pressure and food texture can all influence the circumstances in which people eat. Studies of one isolated factor are easier to interpret experimentally, but real environments combine them. That is both why tightly controlled experiments are valuable and why their findings shouldn't be copied directly into sweeping claims about daily life.
Palatability deserves care as a term. It refers to how pleasant a food is perceived to be, but “tastes good” isn't a scientific synonym for “automatically causes overeating.” Palatability may interact with texture, energy density, variety and prior experience. Separating those relationships requires experiments that deliberately vary one feature while holding others as steady as possible.
What I take from the Hall trial is a credible demonstration that diet patterns can change spontaneous energy intake under controlled conditions. What I take from the eating-rate literature is a plausible set of mechanisms worth testing further. Neither point requires moralizing foods or assuming that one person's experience overrides a randomized comparison.
If anyone wants the original papers, the names to search are **Hall et al., *Cell Metabolism* (2019)**; **Robinson et al., *American Journal of Clinical Nutrition* (2014)**; **Forde et al., *American Journal of Clinical Nutrition* (2017)**; and **Monteiro et al., *Public Health Nutrition* (2019)**. They ask related but distinct questions, which is exactly why reading them together is useful.
There's room here for a real forum conversation that isn't a food-label shouting match. Let's compare how we interpret the methods and the unanswered questions, rather than turning an interesting body of research into another all-or-nothing camp.
1. In the Hall inpatient trial, which possible explanation interests you most—eating rate, texture, energy intake rate, or some combination—and what experiment would best separate them?
2. Do you think the broad NOVA categories are useful for population research while being too coarse to explain differences between individual foods?
3. When you read a food-environment study, what matters more to your confidence: tight laboratory control, realistic everyday conditions, or independent replication across both?
— OLYMPIC