Most of us meet a health announcement already finished, with someone else having done the science and handed us the conclusion. There is a way to check the working, and it needs no laboratory.
That approach belongs to Dr. Mark Bailey, whose A Farewell to Virology appeared in September 2022. He is a medical doctor, and what makes his work unusual is how little of it happens at a bench. He reads replies, letters and footnotes, and pays attention to what they avoid saying. Asked directly, the evidence questions run like this:
- Request the record showing a particle was lifted clean out of a sick person.
- Spot the moment a reply quietly answers an easier question than the one you asked.
- Keep a written "we hold nothing" as a finding worth quoting later.
- Tell apart an institution that has no records from one that has them and says no.
- Email the scientists themselves, since letters tend to be franker than published papers.
None of it requires training in biology. It requires knowing what to ask for, and noticing when you have been handed something else. When somebody comes to me shaken by a confident headline, we start by finding the single question the claim must answer. Naming it is usually what makes the fear manageable.
Why one refusal proves nothing and two hundred proves a lot
Try proving something does not exist. You cannot. Ask one hospital for a document, get nothing back, and you have learned almost nothing. Perhaps you asked the wrong hospital.
What breaks that deadlock is volume. Freedom of Information law lets any citizen demand government-held records. Researcher Christine Massey used it to put an identical request to 209 health and science bodies in over 35 countries. Not opinions. Documents.
By Bailey's account not one produced what was asked for. That is worth sitting with, because it is not the same as saying the science is shaky. It is a written statement, from the organisations with the most to gain by contradicting it, that a particular experiment was never done and written up.
The reply that answers a different question
Refusals are rare. Substitutions are constant, and they are much harder to notice.
You ask whether anything was ever physically extracted from a patient. Back comes a description of laboratory dish work. It sounds responsive. It is not. Extraction asks whether a thing can be obtained on its own. Dish work asks whether a sample changes some cells.
Only one of those bears on whether the thing is there at all. So the file arrives, it is thick and technical, and your actual question is still sitting there unanswered. One university skipped the manoeuvre entirely and said it had never done such work. Blunt, and far more useful.
Ask the authors what the paper left out
Papers are edited. Authors are not, at least not in email.
Bailey describes chasing one heavily cited study. Its images showed damaged cells in the test group and healthy ones in the comparison group, which is exactly what you would hope to see. What the study never spelled out was how that comparison group had been handled.
A co-author explained when asked. No positive controls existed. And the treated cells had received twice the antibiotic dose, including one compound known to injure kidney tissue. These were kidney cells. Change two things at once and the comparison stops meaning anything. He also mentioned, almost in passing, that the effect turned up in one dish out of twenty-four.
Walk a genome backwards until it runs out of road
Here is a test anyone with a browser can run.
Genetic sequences get built by lining them up against sequences published earlier. Fine. So ask what that earlier one was lined up against. Then ask again. Bailey follows one such trail from 2018 to 2003, back to 1997, back to a 1987 entry, until it lands on a paper from 1984 whose supporting reference is itself.
Nowhere along that road does anyone hold up something they physically separated. Every rung is a comparison resting on the rung below. That is why the trail is worth walking yourself rather than taking anyone's word for where it ends.
What anyone can rerun from a laptop
When researchers deposit their raw data publicly, their conclusions become checkable by strangers. That is the point of depositing it.
Somebody did exactly that with the data behind a reference sequence used worldwide. Feeding it through the same software produced 459 fragments rather than the 384,096 reported, and the longest piece came out shorter than published. It also came back 98.85 percent similar to ordinary human ribosomal material, which is a routine component of anybody's cells. For the correspondence behind that attempt you can put a plain question to this and other sources and follow it as far as you like.
There is a detail underneath that one. Two different programs were run originally, and the longer output was the one published. The shorter program would not have yielded a usable sequence at all.
A test can be excellent and still tell you nothing
Two questions hide inside every test result, and mixing them up causes most of the confusion.
Does the method reliably find the genetic sequence it hunts for? That is one question, and a good test answers it well. Does finding that sequence mean this person is ill? Entirely separate, and it needs proving on its own.
Only the first was ever settled for the protocol sent around the world. Bailey quotes Professor Stephen Bustin, among the foremost authorities on the technique, dismissing positives declared at 36 to 37 amplification rounds as absolute nonsense. Every round doubles what is there, so a signal appearing that late started from almost nothing. And in 2006 a whooping cough test returned false positives every single time, for a bacterium that can actually be grown and used to check the test.
Look at how an experiment was built, not what it concluded
Animal studies feel conclusive, and a work through a claim that has unsettled you helps. That feeling is precisely why the setup deserves your attention before the findings do.
Bailey lists four conditions. Whatever goes in should have been separated out beforehand. It should enter the way the illness supposedly spreads. The quantity should be believable. And a second group should get everything except the one ingredient under test.
A primate study from 2020 satisfied none of them. Anaesthetised animals had fluid poured into their windpipes, scaled to human size roughly a third of a cup. They did not get noticeably sick. Some lung damage showed up, with no second group to compare against, which leaves no way to separate the effect of the fluid from the effect of pouring it in.
When a refusal is the most revealing answer
Occasionally an organisation tells you more by declining than it would have by disclosing.
A national health body was asked how its culture experiments were run. It acknowledged possessing the answer. It then declined to share it, citing national security, and that decision survived review.
Follow the logic and it doubles back on itself. Protecting a method because the organism is dangerous assumes the organism is there, which is the very thing the method was requested to test. The same body had previously answered a request for evidence with a computer-generated picture.
Settle existence before arguing about origins
One argument keeps swallowing the other, and separating them saves enormous effort.
Did it come from an animal or escape a laboratory? Both answers take for granted that the thing is real and behaves as described. Whoever wins that argument, nobody examines the assumption underneath it.
Bailey offers this as his own critical position, not as settled fact, and his method survives whether or not you accept where he lands. Asking about existence first reorders what you read for. It also frees you from having to propose a replacement theory, since a claim that fails a test is not rescued by nobody having a better one.
Take the questions with you
What lasts here is not an opinion about any particular illness. It is five habits you can carry into the next announcement. Ask for the extraction record. Watch for a swapped question. Write to the authors. Walk the references backwards. Separate what a test finds from what it means.
This is not medical advice and it disputes no ordinary care. Health decisions belong with your own doctor, and everything above concerns how evidence gets established rather than how anyone should be treated.
If something specific is nagging at you, perhaps a reply you have received and cannot interpret, you need not untangle it alone. You are welcome to bring the exact wording that is troubling you and we can work out together what it does and does not concede. If steadiness feels like the thing to build first, the free tryit.tv course begins with a first lesson on steadying your attention before you judge. That is the same quality of attention that makes any document easier to read honestly.
