Why NOVA Misses the Mark

Dire warnings to avoid processed foods no longer grace the internet. Turns out, processed foods are not so bad after all.

Instead, now we need to watch out for heavily processed foods, hyper-processed foods, or ultra-processed foods. Michael Pollan warns us to avoid any processed foods with more than five ingredients. Carlos Monteiro codified these additive-filled foods as ultra-processed in 2009 and declared them unhealthy in a classification scheme called NOVA.

I don’t argue that all processed foods are good for us. Overeating these products can lead to obesity and other health issues. I do object to the arbitrariness of the NOVA classes. It makes no sense to write off 60% of the American food supply as unhealthy and declare the other 40% healthy.

For example, a popular commercial cupcake, a high-fiber breakfast cereal, a plant-based burger, and a shot of rum are all ultra-processed and equally unhealthy under NOVA. Not to worry, as a homemade brownie, New England clam chowder, a 16-ounce T-bone steak, and as much beer as you can drink are not ultra-processed and equally healthy.

How is a food classified as ultra-processed in addition to the five-ingredient rule? It contains an unacceptable additive. These additives are listed on the NOVA website.

How do we know that ultra-processed products are unhealthy? For one thing, many additives sound like chemicals and are hard to pronounce. Which beverage would you prefer? A fresh mango juice with added caffeine, trimethylpurine dione, or 1,3,7 trimethylxanthine? Spoiler alert: they are all the same chemical, with differing levels of pronounceability.

Let’s say we lived in a world where coffee was the only caffeinated beverage, but the population didn’t know which molecule was the stimulant. Some wise food chemist discovers that it is a molecule named 1,3,7 trimethylxanthine. She then develops a beverage that contains this useful, novel compound. Coffee would still be an acceptable drink, but many consumers would shy away from a product with such an unpronounceable chemical on the ingredients list.

Ultra-processed foods are more about additives than about processes. One of the few processes forbidden is extrusion. But again, the rules are arbitrary. Theoretically, an extruded product without forbidden additives would be acceptable. Extruded cheesy snacks fit the pattern. But pasta is also extruded. Are all pastas, even homemade ones, therefore ultra-processed?

Some additives, like sugar and salt, are acceptable to NOVA if added by home cooks or chefs as culinary ingredients. But when added to a commercial product, sugar and salt are among the ingredients condemned by ultra-processed critics.

In a former life, I was a research scientist grubbing for funds from federal agencies. One line of study was understanding how certain fruits and vegetables became injured at low-temperature storage above the freezing point. Any hope of receiving funds for chilling injury of bell peppers required a detailed description of a possible biological mechanism.

No such reference to mechanism is required of ultra-processed investigators. All they have to do is to correlate the entire mix of ultra-processed foods with a host of chronic diseases in massive datasets. Back when I was in school, correlation did not imply causation. Has that rule been repealed?

Are some products high in sugar, salt, or fat more lethal than others low in these components? Maybe. But it’s just as likely some of these products are healthy and misclassified.

Take plant-based meat products. Most of these foods appeared in the market after the development of the NOVA concept. They were not coded in these studies. They are guilty by association simply because they contain unauthorized additives.

Are all so-called ultra-processed foods toxic, addictive, and unhealthy? I don’t think so. Are some unhealthy when eaten to excess? Probably. Before we condemn 60% of the U.S. food supply, let’s sort out those that pose a danger from those that don’t. And let’s find a true biological mechanism rather than relying on questionable studies of mass correlation. FT

The opinions expressed in Dialogue are those of the author.