Not All Processing Is the Same

We heard Robert Lustig talk about ultra-processed food. It led us to a surprisingly complicated question: what does “processed” actually mean?

Rice is processed when its husk is removed.

Ragi can be cleaned, sprouted, roasted and ground.

Milk can be pasteurised.

Dosa batter can be ground and fermented.

So when we're told to “avoid processed food,” what exactly are we supposed to avoid?

That question came back to us while listening to Dr. Robert Lustig discuss ultra-processed food on The Diary of a CEO. His argument made us curious about a food-classification system increasingly used by researchers around the world: Nova.

And Nova starts with an important idea:

Not all processing is the same.

Four ways of looking at food

Developed by researchers at the University of São Paulo in Brazil, Nova classifies food according to the nature, extent and purpose of processing, rather than simply looking at calories, carbohydrates or fat.

It places foods into four broad groups.

Group 1: Unprocessed or minimally processed foods

Vegetables, fruits, pulses, grains, milk and other foods that are fresh or have undergone processes such as drying, grinding, roasting, refrigeration, freezing, pasteurisation or fermentation.

That last part is worth noticing.

Grinding something doesn't automatically make it an ultra-processed food. Neither does roasting or freezing it.

Group 2: Processed culinary ingredients

These are ingredients generally used to prepare other foods — things such as oils, butter, sugar and salt.

Group 3: Processed foods

Relatively simple foods generally created by combining foods from Group 1 with ingredients such as salt, sugar or oil. Traditional breads and cheeses, for example, can fall here depending on how they're made.

Group 4: Ultra-processed foods

This is the controversial category.

Nova describes these as industrial formulations that often contain substances rarely used in a home kitchen — such as modified starches, protein isolates and hydrogenated oils — or certain additives used for flavour, colour, texture and other purposes. Examples can include soft drinks, confectionery, instant noodles, some breakfast cereals and many ready-to-eat products.

Why are researchers paying attention?

Because something interesting keeps appearing in the research.

A major 2024 umbrella review in The BMJ examined 45 pooled analyses involving almost 10 million people. Greater exposure to ultra-processed foods was associated with higher risks across a range of health outcomes, with stronger evidence for some outcomes including type 2 diabetes, cardiovascular-related mortality and all-cause mortality.

That's significant.

But there's an equally important word in that sentence:

associated.

Much of this evidence is observational. People who consume more ultra-processed foods can differ in many other ways, and researchers are still working to understand exactly which characteristics of these foods — their nutrient profiles, energy density, additives, physical structure, processing methods or combinations of these factors — may explain the associations.

So the research gives us reasons to pay attention.

It doesn't make every question settled.

And Nova has its critics

Here's where the story becomes more interesting.

Not every scientist agrees that placing such a diverse collection of foods into one category tells us enough about whether an individual food is healthy.

A 2025 critical review pointed out that health outcomes appear to differ substantially between types of ultra-processed foods. Sugar-sweetened beverages, for example, do not necessarily show the same associations as some fortified cereals or dairy products that may also be classified as ultra-processed.

The criticism is essentially:

Can one category be too broad?

Supporters of Nova argue that this misses the point. The framework isn't intended simply as another nutrient score. It is trying to capture something broader about the transformation of our food supply and the increasing prevalence of industrial formulations. A 2026 Nature Food paper defending Nova argues that variation within the category doesn't necessarily make the concept scientifically useless.

The debate continues.

And perhaps that's okay.

Maybe the useful question is simpler

We don't think the lesson from Nova needs to be:

Memorise four categories.

Nor should it become:

Anything in a packet is bad.

That would make little sense.

Grinding grain into flour is processing. Fermentation is processing. Pasteurisation is processing. Cooking itself transforms food.

Traditional kitchens have always processed food.

Perhaps Nova offers something more useful than another rigid food rule.

It encourages us to become curious about what happened to a food before it reached us.

Was a grain simply cleaned and ground?

Was a vegetable frozen?

Were a few familiar ingredients combined?

Or was the original food substantially transformed into a formulation containing ingredients we would rarely encounter in a kitchen?

None of those questions alone can tell us whether a food is “good” or “bad.”

But they can tell us something that the calorie number on the back of a packet cannot.

So perhaps the next time we pick something from the shelf, we can ask not only:

What's in this food?

But also:

What was done to it?

Sometimes understanding our food begins with asking a better question.

Item added to your cart