Why We Share These Stories

There is a reason patient stories, misdiagnosis stories and clinical cases are shared here. It is not about winning an argument or proving doctors wrong. People are ill and want to get better. Their families want them to get better. When the explanation they have been given is not getting them there, they need to be able to ask harder questions and look at other possibilities.

Uncertainty should remain uncertainty

One of my concerns with FND is how easily a clinical interpretation can become treated as established fact. Current approaches to functional movement disorder emphasise positive examination signs such as internal inconsistency, variability, distractibility and entrainment. The TASMAN study, for example, describes these as positive signs used in current diagnostic criteria. The observation may be real, but deciding what caused it is another step. Once that interpretation enters the medical record as FND, the distinction can disappear. Later symptoms and findings can then be viewed through the diagnosis rather than being allowed to challenge it.

This is one reason diagnostic reversals interest me. We have cases where signs considered supportive of FND appeared in people subsequently diagnosed with other neurological or medical conditions. These cases cannot tell us how often FND is wrong, but they can tell us something about certainty. If a sign that looked convincingly functional also occurs in somebody eventually found to have another disease, I want to know how specific that sign really is. Observing a pattern and knowing what caused it are not the same thing.

The low reversal rate deserves scrutiny

We keep hearing that FND has a very low misdiagnosis or diagnostic reversal rate. There is published evidence behind that claim. A review cited in the FND literature found a 4% misdiagnosis rate in studies conducted since 1970 after a mean follow-up of around five years, while a later prospective cohort found that 4 of 1,030 patients initially judged to have symptoms not explained by recognised neurological disease had acquired or been revised to a recognised neurological diagnosis at follow-up. A review of current FND concepts summarises these findings, while a later systematic review examining FND and multiple sclerosis provides a useful table of these diagnostic-outcome studies.

Clinicians who have visited this site have made the same point quite bluntly: another explanation is rarely found and FND is rarely overturned. Fine, but before treating that as proof of accuracy, I want to know how easily the diagnosis can actually be overturned. How long are these patients followed? What happens to the person who leaves one service and receives another diagnosis somewhere else? Does that information ever make it back? If another neurological disease is discovered but FND remains as a comorbid diagnosis, has anything technically been overturned? If treatment failure, deterioration or new symptoms do not count strongly against the diagnosis, then what does?

There is a circularity I worry about here. FND is considered unlikely to be wrong because studies report that it is rarely overturned, and the belief that it is rarely wrong can itself become a reason not to reopen the diagnosis. Low observed misdiagnosis and high diagnostic accuracy are not automatically the same thing. Before being impressed by the number, I want to know whether the system producing it is actually good at discovering and recording when it got the diagnosis wrong.

“Misdiagnosis happens in every condition” misses the point

A common response to these cases is that misdiagnosis happens everywhere in medicine. Of course it does. I have never understood why that is supposed to settle anything. There is even evidence that diagnostic error can occur in both directions: a 2019 study found patients initially diagnosed with neurological disease who were later judged to have functional disorders, as well as patients whose initially unexplained symptoms were later better explained by neurological disease. The authors argued that misdiagnosis can cause harm in either direction.

But that is not the question I am asking. The question is whether there is something about this particular diagnosis that can make an error difficult to recognise once it has entered the record. Symptoms can change without necessarily excluding FND. Treatment can fail without necessarily excluding it. Another neurological disease can be discovered and FND can remain as a comorbid diagnosis. Each of those may have a perfectly reasonable explanation in an individual case, but put them together and the question of falsifiability becomes difficult to avoid. If a diagnosis can survive most of the things that might ordinarily make us question a diagnosis, then the fact that it rarely gets overturned becomes less impressive.

“Parkinson's is diagnosed clinically too”

Parkinson's disease is another comparison that comes up. The argument seems to be that Parkinson's is diagnosed clinically, FND is diagnosed clinically, therefore there is nothing particularly concerning about relying on clinical signs for FND. But that only tells us that both diagnoses involve clinical judgement. It tells us nothing about whether the inferences being made from those examinations have the same reliability.

Parkinson's also has something rather important that makes the comparison interesting for a different reason. Its clinical diagnosis can ultimately be compared with neuropathology. A 2021 clinicopathological study explicitly used neuropathological diagnosis as its gold standard and found that the accuracy of the clinical diagnosis varied substantially with diagnostic confidence and disease duration. In that study, 84.7% of patients classified as probable Parkinson's at the first visit had neuropathologically confirmed Parkinson's, rising to 89.1% among probable cases with at least five years of disease duration.

This is not merely historical. A large 2025 UK Brain Bank study again compared clinical Parkinson's diagnoses with post-mortem pathological examination and treated pathology as the gold standard. The authors reported overall clinical diagnostic accuracy of about 91% in their pathology-validated analysis and noted that some people diagnosed with Parkinson's during life had a different diagnosis at post-mortem examination. The point is not that Parkinson's diagnosis is poor. The point is that there is an external reference capable of showing where the clinical judgement succeeded and where it failed.

So “Parkinson's is diagnosed clinically too” does not answer the concern. At some point it starts sounding like the medical version of “but everyone is doing it, so we can do it too.” Another diagnosis relying on clinical judgement does not validate the inferences being made in FND. You still have to demonstrate that your own method works. And the comparison is particularly strange because Parkinson's gives us something FND does not: a point at which the clinical diagnosis can be independently tested against neuropathology. Parkinson's is not an example of why we should worry less about falsifiability. It is an example of why falsifiability matters.

You don't need another suspect

This reminds me of an argument Frank Turek has used about a detective and a suspect. The detective believes he knows who committed the murder. A sceptic looks at the evidence and says the case against that suspect is not strong enough. The detective responds, “Well, who did it then?” When the sceptic cannot provide another suspect, the detective treats that as support for the suspect he already has. But nothing about the original evidence has changed. Not knowing who committed the murder does not make the detective's suspect more guilty.

The same applies to a patient questioning an FND diagnosis. They should not have to walk into a neurologist's office carrying a better diagnosis before they are allowed to question the one they have. “I don't know what is causing this, but I am not convinced this explanation accounts for it” is a legitimate position. The patient failing to solve the case does not strengthen the evidence for the existing suspect.

Now add diagnostic closure to the detective story. Once the detective has settled on his suspect, every new clue gets viewed through that lens. Things that fit strengthen the case. Things that do not fit are explained away. Another possible suspect appears, but rather than reopening the case, the detective demands that somebody first prove the entire alternative theory. Eventually almost every road leads back to the original suspect, partly because the investigation has been organised around that conclusion.

Sometimes “we don't know yet” really is the most accurate answer available. That can be deeply unsatisfying, particularly when somebody is sick, but giving uncertainty a diagnostic label does not necessarily make the uncertainty disappear. You do not have to solve the mystery before you are allowed to point out weaknesses in the proposed solution.

Why the stories matter

This is where the patient stories and misdiagnosis cases become useful. They cannot give us the true FND misdiagnosis rate because there is no denominator, and pretending otherwise would be bad reasoning. What they can do is preserve diagnostic trajectories that might otherwise disappear. Somebody diagnosed with FND who eventually receives another diagnosis elsewhere may never return to the original service. Unless somebody goes looking for that outcome, it may never feed back into the picture of how accurate the original diagnosis was.

They can also provide clues. A Reddit post is obviously not equivalent to a clinical case report, but both can give somebody an avenue they did not previously know existed. It might be a symptom, an investigation, a specialist, another condition or simply a question worth asking. Sometimes that avenue will lead nowhere. That is fine. When somebody is seriously ill and the current road has reached a dead end, knowing that another road exists can still be useful.

Treatment needs a feedback loop

Patients should give reasonable treatment a fair chance. But there comes a point where repeatedly doing the same thing while somebody remains unwell stops looking like persistence. If treatment repeatedly fails, the person deteriorates, new findings appear or the clinical course does not look like what was expected, that is information. It does not automatically tell us what the correct diagnosis is, but it should be allowed to change how confident we are in the explanation we already have.

Otherwise the process can become self-preserving. The diagnosis determines the treatment, the treatment does not work, and the answer is more treatment based on the same diagnosis. New symptoms are absorbed into the existing framework and eventually very little seems capable of reopening the original question. At some point it is reasonable to stop and ask whether we are still moving towards getting the person better or simply repeating what we have already decided.

Not every difficult question is misinformation

Patients searching Reddit, Facebook, forums and medical literature are often discussed as a misinformation problem. There is plenty of rubbish online, but I think this framing can miss something important. Why is the person searching so desperately in the first place? Often they are still sick. What they have tried has not worked, they do not understand why the diagnosis was made, or they cannot get a clear explanation of why other possibilities have been dismissed. Searching is sometimes less about rejecting medicine than trying to make sense of something medicine has not adequately explained to them.

Patients are also going to get things wrong. They are not neurologists. They will misuse terminology, misunderstand parts of papers, overstate findings and occasionally make connections that make no medical sense at all. Correct them. But do not use a small technical mistake as a scapegoat for avoiding the larger question they are asking. Finding an error in a patient's argument does not somehow make the diagnosis they are questioning correct.

I also dislike the tendency to treat patients like children who should simply accept what they have been told. If you want somebody to trust a diagnosis, explain why you trust it. Explain what supports it, where the limitations are, what else was considered and why. Patients should not have to become amateur neurologists just to understand what is being claimed about their own bodies.

The social media crusade against misinformation

Some clinicians appear to have turned correcting medical misinformation on social media into a crusade. There is real misinformation online and some of it deserves to be challenged, but crusades are not very good at nuance. Once misinformation becomes the enemy, patients asking awkward questions, sharing diagnostic failures or reading outside the accepted explanation can start looking like part of the enemy too.

What bothers me is seeing the epistemic failures of patients dissected in great detail while the same scrutiny is not always applied to the certainty coming from the other direction. If we are going to talk about evidence and scientific reasoning, then falsifiability, diagnostic closure, the limitations of positive signs and uncertainty in clinical inference are fair game as well. Epistemic humility cannot be something demanded only from the patient.

Social media makes this worse because certainty plays well there. Medicine already has a large imbalance of expertise and authority, and it becomes very easy for “you misunderstood this paper” to turn into “you don't understand your illness.” Those are not the same claim. Expertise matters, but it does not make every inference correct. If your crusade against misinformation leaves no room for the possibility that you might also be wrong, you have an epistemic problem of your own.

Keep the doors open

Ultimately, I am not interested in protecting anyone's ego, including my own. I am interested in whether people get better. Nobody should need to defend a diagnosis because they made it, published about it, built a career around it or argued for it publicly. If somebody remains seriously unwell and the evidence starts pointing somewhere else, follow it. If the existing explanation survives serious scrutiny, fine. The point is that the scrutiny should be allowed to happen.

That is why these stories are here. They preserve uncertainty, document failures and occasionally provide another clue when somebody has run out of places to look. This is not about collecting ammunition against doctors or keeping score over who was right. We and the people we love are trying to get better. If the road we are on is not getting us there, it is entirely reasonable to look for another one.

Nobody needs to win. The patient needs to get better. When something is not working, ask harder questions, follow the evidence and keep the doors open.