What the Word 'Ultra-Processed' Hides

"Ultra-processed" has become a moral category. We sort foods into it almost reflexively, then assume the label settles the question of whether something is good for us. But a recent study asked a sharper question: when you already know a diet is plant-based, does the processing tell you anything more about how healthy it is? The answer is less obvious than the label suggests.

The idea that processing determines healthfulness has a clean, satisfying logic. Whole foods good, factory foods bad. It fits how we like to think about eating — as a series of clear-cut choices — and it has quietly reshaped dietary guidance in several countries. So it is worth sitting with a recent analysis that pressures that logic from the inside.

On a July episode of Sigma Nutrition Radio, Dr. Alysha Thompson discussed a large prospective cohort study she led, which examined plant-based dietary patterns through the lens of food processing [3]. The design is what makes it interesting. Rather than lumping all plant-based eating together, the researchers separated healthful plant-based patterns from unhealthful ones — because a diet can be entirely plant-based and still be built largely on refined grains, sugar, and fried foods [3]. Then they did something most of the processing conversation skips: they stratified the contributing foods by whether they were classified as ultra-processed or not [3].

That second step is the crux. If processing is the thing driving health outcomes, you would expect the ultra-processed foods within a plant-based diet to explain most of the relationship with long-term health. If instead the overall quality of the pattern — how healthful versus unhealthful the foods are — is doing the work, then the processing classification adds little once you already know what someone is actually eating [3]. The question Thompson posed is whether calling a food ultra-processed "meaningfully change[s] how we interpret the relationship between plant-based diets and long-term health outcomes" [3]. Framed that way, the label stops being an answer and becomes a hypothesis to test.

This matters because the two framings lead to different behaviour. If processing is the master variable, the sensible move is to minimise anything from a factory, full stop — a rule that would flag fortified soy milk, canned beans, and wholegrain bread alongside sweets and refined snacks. If dietary quality is the master variable, the sensible move is to look at what the food is made of and what it displaces, which lets some processed foods stay and pushes out some "natural" ones. The distinction is not academic. It changes your grocery list.

There is a broader thread here that connects to how nutrition science is arguing with itself right now. A companion strand of the field keeps returning to the same corrective question — compared to what? A food or a supplement or a screening test is rarely good or bad in isolation; it is better or worse than the realistic alternative. Austin Baraki made a structurally identical point about medical screening on the same podcast: framing something as an uncomplicated good ignores the fuller picture of what it displaces and produces downstream [2]. The ultra-processed debate has the same shape. The label feels like an uncomplicated warning, but a warning is only useful if it tracks the outcome you care about.

None of this is a licence to fill your cart with processed food. Thompson's point is narrower and more useful: the processing category may be riding on top of dietary quality rather than adding independent information [3]. A plant-based pattern built on vegetables, legumes, whole grains, nuts and fruit tends to be healthful whether or not some of its components were processed; a plant-based pattern built on refined and sugary foods tends not to be, and the processing label mostly restates that.

What is still contested is how far this generalises. The extract describes direction, not magnitude — we are told the study stratified by processing and examined both healthful and unhealthful patterns, but not the size of any effect [3]. And it is one cohort in a literature that has, elsewhere, found associations between ultra-processed intake and worse outcomes. The honest reading is not "processing doesn't matter" but "processing may not be the variable you should organise your eating around." That is a quieter claim, and a more durable one. The theme of this newsletter is eating without extremes or guilt. Retiring a label that generates a lot of guilt for possibly little signal is very much in that spirit.

RESEARCH RADAR

• A large prospective cohort analysis separated healthful from unhealthful plant-based dietary patterns and then stratified the foods by whether they were ultra-processed, testing whether the processing classification changes how we read the diet–health relationship [3]. It reframes processing as a hypothesis rather than a verdict.
• A narrative review in the Strength and Conditioning Journal notes that many dietary and nutrient recommendations were built on studies conducted only in men, despite physiological, neurological and hormonal changes across a woman's life cycle that shift nutritional needs [1]. It highlights the absence of sex-specific guidelines in current frameworks.
• A review in Annals of Medicine points to women's lower peak bone mass and higher fracture risk, especially after menopause, and argues resistance and weight training are effective for maintaining bone health — yet societal biases still deter many women from it [4]. The barrier here is cultural, not physiological.

ONE THING TO TRY

Next time you reach for a food and the word "processed" flashes a warning, ask a different question: compared to what I'd otherwise eat, is this better or worse? Judge the item against its real alternative, not against an ideal [2].

WORTH YOUR ATTENTION

• Sigma Nutrition Radio #615 with Dr. Alysha Thompson — the study behind today's issue, on why processing may not determine the healthfulness of plant-based diets [3].
• Sigma Nutrition Radio #616 with Dr. Austin Baraki — a clear-eyed look at how "more testing, better health" can quietly go wrong through overdiagnosis and overtreatment [2].
• Strength and Conditioning Journal review on female athletes — a useful map of where sex-specific nutrition guidance is still missing [1].
• Annals of Medicine, "Strength unseen" — on the prejudices that keep women out of the weight room, and the bone-health cost [4].

We like labels because they let us stop thinking. "Ultra-processed" promised to do the deciding for us. The more careful work suggests the label mostly restates what the diet already is, which means the thinking was never really outsourced. Eat the mostly-plants, keep the canned beans, and spend your worry somewhere it earns its keep.