Case Study 2 — Nineteen Days: The Life Cycle of a Nutrition Headline

A media case. The study, the outlet names, and Coach Brandt are illustrative composites assembled from patterns that recur constantly; the mechanics of each stage are not.

Setup

In Chapter 1 I described the pipeline in the abstract. This case study puts a clock on it.

I've tracked this cycle informally for years — partly out of professional interest, partly because patients bring me the output and I want to know where it came from. What follows is a representative timeline, compressed from several real instances into one, with the stages and the reach figures kept at the scale I typically observe.

The thing to watch is not the distortion. It's the speed, and the fact that the correction — when one comes — arrives after the belief has already set.

The timeline

Day 0 — Publication

A research group publishes a study in a mid-tier nutrition journal. The design: a randomized crossover trial in eighteen healthy young adults, each consuming a high dose of a particular polyphenol-rich extract for four weeks, then crossing over to placebo. The measured outcome: several markers of vascular function and inflammation.

One marker moved in a favorable direction. The others didn't.

The paper is honest. The abstract says the finding is preliminary, notes the small sample, notes that the dose used exceeds what would be obtained from ordinary dietary intake, and calls for larger trials with clinical endpoints. The discussion section is appropriately deflationary.

What was established: in eighteen young people, a supraphysiological dose moved one marker over four weeks. This is a hypothesis. It is rung 6 of the evidence ladder in a technical sense — randomized! — but with a sample of eighteen, a surrogate endpoint, and multiple outcomes measured, it is a weak rung 6.

Day 2 — The press release

The university's communications office issues a release.

Headline: "Compound Found in Common Berry May Protect Blood Vessels, Study Suggests."

Everything in it is technically defensible. "May." "Suggests." The sample size appears in paragraph five. The dose issue appears in paragraph seven. The senior author is quoted saying the results are "encouraging but early," which is accurate and which will not survive the next stage.

The press officer is not a villain. Their performance is measured in media pickups. A release titled "Small Preliminary Study Finds Modest Change In One Of Several Markers" generates zero pickups, and the officer knows it, and so does everyone above them.

Day 3 — First coverage

Two health outlets run it. One is careful and includes the sample size in paragraph three. The other is not.

The careless headline: "This Berry Could Be Your Heart's Best Friend, Scientists Say."

"May protect blood vessels" has become "your heart's best friend." "Scientists say" has done the work of converting eighteen people into an institution.

Days 4–7 — Aggregation

Roughly a dozen outlets rewrite the story, mostly from each other rather than from the paper. Each rewrite is a degraded copy of a degraded copy.

By day 6, a version exists that says the compound "reduces inflammation and improves cardiovascular health." The conditional is gone. The sample size is gone. The dose is gone. The word "marker" is gone. Two of the twelve mention that the study was in eighteen people; none mention the dose.

Also by day 6: the first listicle. "Seven Foods Cardiologists Want You Eating." The berry is number three. No cardiologist was consulted.

Day 9 — The reel

Coach Brandt makes a forty-one-second video.

Fast cuts. He's holding the berry. Incredulous face. "There's a compound in this that's protecting your blood vessels — and the supplement industry has known for years."

That last clause is invented. Not maliciously; it's a rhetorical habit, the ambient grammar of the genre. He didn't read the paper. He read the day-6 aggregate.

11 million views over the following eleven days. Ninety thousand shares.

His caption includes a discount code for a supplement retailer.

Days 10–16 — Product

Three supplement companies push ads featuring extract of the berry. One quotes the day-3 careless headline directly in the ad creative, which is legal because they're quoting a news outlet rather than making the claim themselves.

Price for a month's supply: $34 to $79, depending on brand. None of them are dosed anywhere near the trial's dose, which is a detail that would only matter to someone who had read the trial.

Day 14 — The clinic

A patient asks me about it. She has read that it's good for the heart. She wants to know which brand.

I have not yet heard of the study. I go and find it that evening, and it takes me about twenty minutes to work out what actually happened, and I have a doctorate in this and institutional journal access.

This is the part I want you to sit with. The information reached my patient on day 14 with total confidence, and reached me on day 14 with none, and the only reason I could resolve it at all is professional infrastructure most people don't have.

Day 19 — The correction that isn't

A science journalist writes a good, careful piece: "That Berry Study Everyone's Talking About Is Much Smaller Than You've Heard." It's accurate, well-sourced, and fair to the researchers.

Reach: roughly 1/300th of the reel.

It does not appear in the feeds of the people who saw the reel, because the algorithm that surfaced the reel is optimizing for engagement, and a correction is definitionally less engaging than the thing it corrects.

The belief has set. The supplement sales continue. In eighteen months the berry will be replaced by a different berry, and the same cycle will run again with the same structure and different nouns.

Analysis

What the numbers do

Accuracy and reach are inversely related at every single stage. That is not an accident or a failure of individual judgment. It's the system's design: reach is allocated by engagement, and engagement rewards certainty, novelty, and grievance — three things that accurate reporting of a small preliminary trial cannot supply.

Three things worth noticing

1. No stage required dishonesty except one clause. The researchers were careful. The press officer did their job. The first journalist met a deadline. Brandt believed what he said, apart from "the supplement industry has known for years," which he'd absorbed as a genre convention rather than constructed as a lie.

2. The most consequential distortion happened early and quietly. Not at the reel — at day 2 and day 3, where "one marker moved in eighteen people at a high dose" became "may protect blood vessels." Everything downstream was elaboration on a frame that was already set.

3. The correction is structurally guaranteed to lose. Not because people are stupid, but because corrections are distributed by a system optimized for engagement, and they are engineered to be boring. Any plan for fixing nutrition misinformation that depends on corrections reaching people is not a plan.

What this means for you

You cannot out-read this pipeline. There isn't enough time in a life, and by day 14 it has already reached your kitchen.

What you can do is change the unit of analysis. Stop evaluating individual studies and start evaluating bodies of evidence. A single study — even a randomized one — is close to worthless as a basis for changing your behavior, and this case study is why: by the time it reaches you, you are not receiving the study. You are receiving the seventh-generation copy of a press release about it.

That's the entire argument for Chapter 2.

Discussion Questions

1. The press officer at day 2 wrote a technically accurate release with a shifted emphasis. Were they wrong to do so? What would have happened to them, professionally, if they'd written the accurate- and-boring version instead? Does that change your answer?

2. Coach Brandt added a clause that wasn't true ("the supplement industry has known for years") and attached an affiliate code. Does the affiliate code change your moral assessment of the video? Would your answer differ if he made no money from it?

3. The correction reached roughly 1/300th of the audience of the claim. Design an intervention that would meaningfully change that ratio. Then identify why your intervention won't be implemented.

4. My patient asked which brand to buy on day 14, and I had never heard of the study. What should a clinician say in that moment — before they've looked it up? Write the actual sentence.

5. The chapter argues that accuracy and reach are inversely related by design rather than by accident. If that's right, is there any version of nutrition communication that works at scale? Or is individual evidence literacy — this book's bet — the only available strategy?

Your Turn

Pick a nutrition claim currently circulating in your own feed and reconstruct its timeline as far back as you can. Try to find:

• The original study (or discover that there isn't one)
• The press release, if the institution issued one
• The first mainstream coverage
• The version that reached you

For each stage, write down one sentence: what was claimed, and what was dropped.

Then answer the uncomfortable question: at which stage did you enter? Most of us enter at stage 4 or 5, which means the version we hold is several generations from the source — and it feels identical, from the inside, to knowing something.