Contact The Editor
Articles By This Author
Everyone's telling you to worry about how processed your food is.
I'm telling you not to.
It's not because processed foods are good for you! They're not, mostly. But the processing isn't the reason, and if you're scanning ingredient lists to find the ice cream that doesn't have cellulose gum, you've lost the thread.
I think we need to pay attention to what actually makes ultra-processed food bad for you. If you come here often, you know I've been shouting this from the rooftops, but now we have a brand-spanking-new data point that adds to the evidence, so I'm going to shout some more.
The data comes from Kevin Hall, a scientist at the National Institutes of Health. If you follow this issue, you probably know him as the lead author of what is arguably the most important and influential processed-food study, which found that people given a diet of ultra-processed foods ate about 500 calories more per day than those fed a minimally processed diet.
I talked with Hall about it when the study came out in 2019, and asked him what he thought was driving the increased consumption. His No. 1 culprit: energy density.
Wait! Don't turn the page! I know energy density is wonky and boring, but it's one of the main drivers of obesity, and we need to care about it.
Energy density is a measure of how many calories are packed into a certain weight or volume of food. The more calories are in each bite, the more calories you're likely to consume overall. Foods that have a lot of bulk for the calories they deliver can satiate better and lead to lower consumption.
Ultra-processed foods are typically more energy-dense, and the diet in Hall's experiment was no exception. But the study wasn't designed to figure out why people ate more, and Hall emphasized that pointing the finger at energy density was speculation.
Now, though, he's partway through an experiment that is designed to figure that out, and it looks like he might have been right all along.
It also means - whew! - that I have been right all along.
Like the original study, the new one pits an ultra-processed diet against a minimally processed one, but this time, the two diets are matched for energy density and hyperpalatability (a prescribed combination of salt, sugar and fat that's supposed to be a proxy for deliciousness).
And guess what? Once you match the diets for those two items, people eat about the same number of calories. Reducing hyperpalatability alone (in a third diet) didn't reduce consumption much. This certainly suggests that energy density is what drives overeating - but doesn't prove it. Hall told me that, because they didn't include a diet that was low in energy density but high in hyperpalatability, they can't rule out a synergistic effect.
This new result jibes with decades of work demonstrating how effectively energy density drives overeating. Penn State's Barbara Rolls, who has probably done more work on this issue than any other scientist, co-authored her first paper on it in 1998. Over and over, study after study found that the more calories are packed into a bite of food, the more calories people eat. The results are consistent, Rolls told me, and the effect is larger than that of other kinds of differences among foods (looking at you, macronutrient ratio!).
For some reason, it's a lot harder to get people interested in energy density than, say, gut bacteria or insulin excursions. This is a prosaic metric, mostly driven by water, fiber and air, which decrease energy density (by either weight or volume); and fat, which increases it. There was a slew of research on energy density in the '90s and '00s, but then it fell out of favor.
"The thing about doing studies on energy density is that they're kind of boring." Rolls said. "You know it's going to have an effect." But the focus has changed, and "it's the food, stupid" studies don't attract funding. If you want to get a research grant, Rolls said, "you have to follow people over a year, you have to do MRIs." Or test for glucose excursions and microbiome changes.
I asked Rolls if she felt vindicated. "I think we knew it was going to turn out this way," she said. Hall also suspected energy density was a prime culprit after his first study. Anyone in this space is familiar with this old-school metric and its importance.
Energy density, which can drive obesity, isn't the only problem with processed foods. Low levels of nutrients combined with high levels of sugar and salt (also saturated fat, but let's not have that fight again) can have health consequences unrelated to obesity. But, like energy density, we've known about the problems of nutrients, sugar and salt for a long time.
Why are we looking at the level of processing - a reasonable but imperfect proxy for badness - rather than the things that cause the actual badness, which are staring us in the face?
And a related question: If processed foods are mostly bad, what's the harm in focusing on those foods and recommending that people eat fewer of them?
Because it gives people who are trying to eat better yet another reason to put their common sense on a shelf. When fat was the enemy, we got low-fat everything. When it was gluten, we got gluten-free baked goods. When it was high-fructose corn syrup, manufacturers switched to cane sugar.
And it worked. Because people, being human, are looking for a reason to buy the things they want to eat. Give me a reason to pay attention to the "cellulose gum" in fine print on the ingredient list, and I can ignore the "ice cream" in big letters on the front of the package. The brouhaha over processing is a distraction from the everyday ingredients - the "clean" ones, not the additives - that make foods unhealthful and easy to overeat. (My favorite example is Domino Golden Sugar. It's granulated sugar that's somewhere between white and brown, and the label says "Less Processed." And it is less processed than white sugar. But it's sugar.)
Hall expresses frustration about "strong opinions" on processed foods, in light of incomplete evidence. And he has an excellent point! We don't know everything we should about every additive, and we don't have many rigorous trials like Hall's that compare ultra-processed and minimally processed diets that are controlled for everything but processing. And this new study does find a difference in body composition; only the people on the minimally processed diet lost body fat. (Although that loss was less than a pound, and in Hall's 2019 study, the weight loss from the minimally processed diet came mostly from fat-free mass.)
I, nevertheless, have a strong opinion (in my defense, that's my job). We have overwhelming evidence about the features of processed food, and the food environment in general, that drive overeating of unhealthful food. And this tantalizing data point from Hall's new study is another piece of evidence that the straightforward, well-understood aspects of food are at the root of obesity. Energy density, this time around, but also portion size, convenience, low price and proximity.
Macronutrient ratios, hormones and microbiome changes don't have anything close to the same impact on overeating.
If you're trying to make better choices in our ridiculously obesogenic food environment, don't bother trying to figure out if something is ultra-processed. Instead, think about energy density. Rolls recommends a quick check of the nutrition facts label: "If a food has more grams than calories in a serving, it is relatively low in energy density," and vice versa.
If you cruise the supermarket aisles, you find that ultra-processed foods run the energy-density gamut. Sure, Doritos come in at 5.2 calories per gram, and hot dogs are 3.2. But Ragu tomato sauce is a mere 0.6, and Progresso Minestrone Soup is 0.5.
And that ice cream? A two-thirds-cup serving of Häagen-Dazs Vanilla Bean has 350 calories, but none of the additives you're supposed to avoid. Turkey Hill Original Vanilla, meanwhile, has the cellulose gum, mono- and diglycerides, and caramel color to make it ultra-processed, but a serving has less than half the calories of the "clean" ice cream. I'm perfectly capable of eating too much of either of them, but I'll definitely overeat less of the brand with half the calories.
But don't forget to consult your inner grandmother. Ask her what makes ice cream unhealthful, and I guarantee she's not gonna say "cellulose gum."
Tamar Haspel is an American columnist who "writes on the intersection of food and science". Her column, "Unearthed", has twice been nominated for the James Beard Foundation Award, which she won in 2015.
(COMMENT, BELOW)