Nutritional noise: why what you read about food is probably wrong

You read something about coffee on Tuesday.

On Thursday, you read the opposite.

Both articles cite studies. Both sound convincing. And by Friday, you know slightly less about coffee than you did at the start of the week.

That is not a personal failure.

It is a structural problem with the way nutrition information travels from a research lab to your screen.

The scale of the problem

A 2024 study by Denniss and colleagues at Deakin University examined 676 nutrition-related Instagram posts from accounts with at least 100,000 followers. 45% of the posts contained inaccurate nutrition information, and 9 out of 10 posts were rated as low quality when factors such as author credentials, evidence base, advertising and commercial interests were considered.

A separate analysis of TikTok content found that 97% of popular videos on weight loss, muscle building and supplements offered no scientific evidence to support their claims.

This is not noise at the margins. It is the dominant signal.

Three forces that create it

The first force is the structure of scientific publishing itself.

Nutrition studies are written for other researchers, not for the people the findings are meant to help.

A scientific paper is full of context: methodology, population, dose, duration, limitations, confidence intervals, subgroup analyses. Often the most important part of a study is not the result itself, but the conditions under which that result was found.

Most of that disappears the moment a study leaves the journal.

What remains is a headline.

The second force is how media selects what to amplify.

Algorithms do not reward consensus. They reward novelty, surprise, anxiety and contradiction.

A nuanced conclusion rarely travels far. A dramatic reversal does.

"Scientists confirm that vegetables are probably still a good idea" is not a viral message. "Everything you know about vegetables is wrong" has a much better chance.

This creates a structural pressure toward the exception and away from the consensus, even when the consensus is well-established and the exception is a single small study.

The third force is commercial interest.

The wellness industry is not financially neutral in this environment. Confusion creates demand. Every new fear makes room for another supplement. Every new trend makes room for another protocol. Every contradiction creates space for another expert who promises certainty.

This does not mean that everyone selling something is dishonest. That would be too simple.

But it does mean that financial incentives shape the information environment. We should not pretend otherwise.

These three forces reinforce each other.

More reading does not necessarily help when the selection mechanism is broken. Adding more input may simply add more noise.

What actually changes this

The problem is not a lack of information.

It is the absence of a reliable way to evaluate it.

Over the years, combining clinical medicine with nutrition science, I noticed that almost every health claim can be examined through three questions.

I call this The Filter. Not a list of correct answers. A way to ask better questions

A claim can be scientifically interesting but practically irrelevant. It can hold up in a controlled study and fall apart in a real life with work, family, poor sleep and competing demands.

So the better question is not simply: is this true?

The better question is: is this reliable enough, relevant enough, and valuable enough to act on?

That is what The Filter helps you decide.

Is it backed by strong evidence?

Not one study, not one personal success story, not a mechanism that sounds plausible on paper. And certainly not one confident person on social media quoting one paper.

What matters is the type and quality of evidence.

Has it been studied in humans? Has it been replicated? Is there a systematic review? Are the studies large enough, long enough and relevant enough? Who was studied? And who funded the research?

A study in mice is not evidence for humans. A study in elderly men with a vitamin deficiency is not evidence for healthy women in their forties.

That sounds obvious once you say it. But a remarkable amount of nutrition advice online ignores exactly this distinction.

Is it applicable in real daily life?

A finding can be true in a controlled study and still not be very useful in ordinary life.

The dose matters. The timing matters. The population matters. Your physiology matters.

So does the difference between a tightly controlled intervention and a normal week with work, family, poor sleep, stress and meals that do not always look like a research protocol.

A supplement that produced a measurable effect at a clinical dose, studied over six months, does not automatically become effective at the amount found in a capsule sold at a pharmacy.

Applicability is not a side issue. It is often where the interpretation either becomes useful or starts to mislead.

Does it deliver a meaningful health benefit?

Many claims stop too early. They point to a biomarker.

"This raises antioxidant levels." "This reduces inflammation." "This activates longevity pathways."

Interesting, perhaps.

But does it improve health? Does it reduce disease risk? Does it change cardiovascular outcomes, metabolic health, cognition, frailty, quality of life or mortality?

The gap between "this changes a marker" and "this improves your health" is large.

A lot of wellness marketing lives in that gap.

What this looks like in practice

Take fibre

This is currently circulating on social media under the label "fibermaxxing".

The word is new. The science is not.

The underlying claim - that dietary fibre supports gut health, helps regulate blood glucose and is associated with lower cardiovascular risk - is one of the more consistent findings in nutrition science.

So if we run this through The Filter, the answer is fairly clear. The evidence is strong. The mechanism is well-established. The real-life applicability is high, provided the increase is gradual and comes mainly from whole foods rather than suddenly adding large amounts of fibre supplements.

This is a case where the trend language is noisy, but the direction of the advice is broadly aligned with the evidence.

Now compare that with

The carnivore diet.

There is a kernel of truth here. Reducing refined carbohydrates and ultra-processed foods can improve metabolic health, especially in people whose previous diet was built around those foods.

But carnivore content tends to leave out a separate finding that the same evidence base supports: the International Agency for Research on Cancer classifies processed meat as carcinogenic to humans, and red meat as probably carcinogenic, based primarily on consistent associations with colorectal cancer. That is not a fringe claim or an outdated panic. It is a formal classification built on accumulated epidemiological and mechanistic evidence.

The nuance matters here too. The classification reflects the strength of the evidence that a risk exists, not the size of that risk. For most people, the absolute increase in lifetime cancer risk from moderate meat consumption is modest. But a diet built entirely around red and processed meat, by design, removes the moderation that keeps that risk small.

This is exactly the kind of detail that gets lost when a trend simplifies a complex evidence base into a single confident claim. The wellness world wants you to think cutting carbs always wins or meat always wins. Both miss what the data actually shows: dose, pattern and what is being replaced all matter more than a single ingredient being declared good or bad.

So here the kernel of truth is real, but it is being used to carry a claim the evidence does not fully support, while a separate, well-established risk is quietly left out of the conversation.

That distinction is not visible from the outside. Fibremaxxing and carnivore content may appear in the same feed, in the same format, with the same confidence, using the same visual language.

The Filter is what makes the difference visible.

The skill that compounds

Once you know these questions, you cannot unknow them.

Every health claim you encounter, for the rest of your life, becomes a little easier to read.

The noise does not disappear. Instagram remains Instagram. Headlines still simplify. Wellness marketing still sells certainty, because certainty sells better than nuance.

But your relationship to it changes.

You stop looking for the one right source to trust. You start asking better questions of any source.

That is already a powerful shift.

But there is also a practical problem.

Applying The Filter rigorously takes time. It requires access to information that most people do not have in front of them while they are scrolling.

Was the study randomised? Was it controlled? Who was included? Was the outcome a real health endpoint or only a surrogate marker? Was the dose realistic? Was the population anything like you?

Those are not small questions. They require training, time or both. Most people have neither, certainly not for every health claim they encounter in a normal week.

That is why I built The Filter as an app.

It uses the same clinical reasoning process: evidence quality, applicability to your situation and meaningful health benefit.

But it also goes further than the three questions alone.

It looks for the reasoning errors inside the claim itself: the leap from mechanism to cure, the invisibility of dose, the transfer of results from one population to another, the use of a biomarker as if it were a health outcome, the confidence that exceeds the evidence.

It flags the red flags.

It tells you what the evidence actually shows, what remains uncertain and what that means in practice.

Not in order to replace your thinking. That is not the goal.

The goal is to do the groundwork, so your thinking can start from a better place.

Because medical literacy is not memorising the correct answer to every health question.

It is learning how to evaluate health information. How to pause. How to ask better questions. How to recognise when a claim is stronger than the science behind it.

And how to make decisions without being pulled from one headline to the next.

---

The Filter is currently in development as an educational tool to help people evaluate health claims using scientific evidence, personal applicability and meaningful health outcomes. It supports informed decision-making. It does not replace professional medical advice.

drfilter.app

A note on terminology: in this article, "nutritional noise" describes the overload of inaccurate, commercially motivated or poorly contextualised health information in our information environment. This is different from "food noise", a clinical term used to describe intrusive and persistent preoccupation with food, often discussed in the context of obesity, eating behaviour or disordered eating. That is not what this article is about.

If you enjoy reading at your own pace, you can explore more of my work here — calm, evidence-based, and free of wellness noise. → Human Recipe Library

Human Recipe is a science-driven wellness platform

that combines medical expertise with intelligent, future-focused tools, helping people turn

clinical insights into sustainable daily habits for a longer, healthier life.

---

Dr. Caroline De Graeve, MD, is a physician and founder of Human Recipe. She trained in medicine and nutrition science (Stanford University’s Nutrition Science program). Her work focuses on translating evidence-based lifestyle medicine into clear, practical guidance for women in midlife and beyond.

About the author →

Scientific References

Denniss E, Lindberg R, Marchese LE, McNaughton SA. #Fail: the quality and accuracy of nutrition-related information by influential Australian Instagram accounts. Int J Behav Nutr Phys Act. 2024;21:16. https://doi.org/10.1186/s12966-024-01565-y

Okafor PN et al. Diet pills and deception: A content analysis of weight-loss, muscle-building, and cleanse and detox supplements videos on TikTok. Patient Educ Couns. 2024. https://doi.org/10.1016/j.pec.2024.108285