The most persuasive health misinformation is not a lie. It is a true mechanism carrying a conclusion that cannot be tested — which is why checking the first half never catches it. There is one question that does, and the uncomfortable part is that it works just as well on people you agree with.
A lie is easy to dismiss. You check it, it fails, you move on.
What is much harder is a claim where the first half is entirely correct. You check that part — because that is the part with the citation attached — and it holds up. And having verified it, you extend your confidence to everything that followed.
That is the structure of nearly all the health content that goes badly wrong. Not fabrication. A true mechanism used as scaffolding for a conclusion nobody could test.
“Chronic inflammation drives most modern disease, therefore this anti-inflammatory protocol will resolve your condition.” The first clause is broadly true and well supported. The second does not follow from it, and is not tested by it.
“Your gut microbiome affects your immune system, therefore your symptoms are caused by gut dysbiosis.” First clause: solid immunology. Second: an unfalsifiable diagnosis, because there is no result that would count against it.
“Cortisol affects nearly every system in the body, therefore your fatigue is adrenal.” First clause: textbook endocrinology. Second: a conclusion that no test can refute, since a normal cortisol becomes evidence of a different stage rather than evidence against.
In each case the mechanism is real. The mechanism is also doing rhetorical work it cannot bear — it is being used to make a specific claim about a specific person feel like established science.
What result would prove this wrong?
That is it. If there is no answer — if every possible finding can be absorbed into the theory — then what you are looking at is not a hypothesis. It is a framework, and frameworks cannot be wrong, which is exactly why they should not be trusted.
A real hypothesis is vulnerable. “This person’s fatigue is caused by iron deficiency” can be killed by a ferritin of 90. “This person’s fatigue is caused by hidden toxicity” cannot be killed by anything, because a normal result becomes evidence the toxins are sequestered in tissue.
Watch what happens when a test comes back normal. If a normal result is treated as information — the hypothesis weakens, something else moves up the list — you are dealing with someone reasoning. If a normal result becomes evidence for the theory (the ranges are wrong, the test is insensitive, it is in the tissues, you are in the compensated phase), the theory has become unfalsifiable and you should stop paying.
This would be a comfortable article if I stopped there, so let me not.
Functional medicine is structurally prone to this. Our whole premise is that meaningful change happens before conventional thresholds are crossed — which is well supported, and which also means that a normal result is never quite disconfirming. If the argument is "the drift starts early," then a result inside the range is exactly what the theory predicts. That is a genuinely useful insight and a permanent open door to unfalsifiability, and it is the same door.
Here is how I try to keep it shut, and it is not a technique so much as a habit of writing things down.
That methylfolate could replace folic acid for preventing neural tube defects. It raises red cell folate as well or better. Every trial establishing that prevention used folic acid. I was reasoning from a surrogate and I was wrong. The full correction is here.
That preparing properly prevents running injuries. Thirty prospective studies, 3,404 runners, over a hundred variables tested, essentially nothing predictive. I held that position for most of my career. I published the retraction rather than quietly editing it.
That astragalus supports telomere length. The single human trial found significance at the low dose and not the high dose, which is the signature of a chance finding, and every author was commercially connected to the product.
That alkaline phosphatase indicates zinc status, and that a TSH inside the reference range can be a diagnosis. Both taught, both dropped.
I list these not as penance but because they are the only evidence that the falsification test is actually being run. A practitioner with no corrections has either never been wrong or has never checked, and only one of those is possible.
Because the test is most needed precisely where you least want to apply it.
It is easy to demand falsifiability from someone whose conclusions you dislike. The discipline is applying it to the practitioner who confirmed what you already suspected, the podcast that finally explained your symptoms, the framework that made a decade of feeling unwell make sense at last.
That relief is the most dangerous moment in the whole process, and it is not a character flaw — it is what being unwell and dismissed for years does to a person. Which is exactly why the people most vulnerable to unfalsifiable health claims are the ones who have been failed by ordinary care, and why I think that failure is partly medicine’s responsibility rather than theirs.
Demanding falsifiability is not the same as demanding certainty, and it is not a licence to dismiss anything unproven.
Plenty of legitimate clinical reasoning is provisional. “I think this is the most likely explanation, here is what would change my mind, let us test and see” is a perfectly good position. So is “we do not know.”
The failure is not uncertainty. It is confidence without vulnerability — a claim stated firmly that has been constructed so nothing could dent it.
Including me, and including yourself.
What result would change your mind? If nothing would, the conversation is over.
What have you been wrong about? A specific answer with a date is worth more than any credential. A vague one, or none, tells you the checking is not happening.
Is the mechanism doing work the evidence should be doing? A long explanation of how something works, with no outcome data, is a substitution.
Does this explain everything? A framework that accounts for fatigue and anxiety and joint pain and reflux and hair loss is not powerful. It is unfalsifiable.
One last thing, because dismissal is its own laziness.
The people described here are usually not frauds, and the first half of what they say is frequently correct. Inflammation does drive chronic disease. The microbiome does modulate immunity. Cortisol does affect nearly everything. Conventional care does dismiss people whose results are normal and who are plainly unwell.
Being half right is not nothing. It means there is something worth keeping, and the useful move is to take the mechanism and leave the conclusion — then go and find out whether the conclusion has ever actually been tested.
Usually it has not. Occasionally it has, and it holds, and you have learned something. That is the whole job.
Educational content, not medical advice. If a practitioner’s explanation cannot be disproved by any test, that is worth noticing — but it is not a reason to stop seeking help, and it is certainly not a reason to stop investigating symptoms that concern you.
Which only means anything if a result is allowed to say no.
Ask AIdan → Book a discovery call