By David Tuller, DrPH
Last July, I sent a letter to Benjamin Tolchin, a neurologist at Yale, about the statement, in a 2021 paper for which he was the lead author, regarding prevalence rates for functional neurological disorder (FND). Last month, I sent it again. I’ve still had no response.
I am reposting it below because related claims about the prevalence of FND has recently been highlighted by the charity FND Hope UK. The charity organized an FND Parliamentary Awareness Day, which took place on Wednesday of this week. Before the event, the organization sent out an appeal for people to write to their member of Parliament. The sample letter included in the appeal stated categorically that FND is “the second most frequent reason for a neurology outpatient appointment.”
Although the assertion isn’t footnoted or referenced, the Scottish Neurological Symptoms Study (SNSS) is the primary source for the common claims made in both Tolchin’s 2021 paper and the FND Hope UK appeal–that the prevalence of FND at neurology clinics is 16% and that it is “the second-most frequent” diagnosis. The SNSS was a major project that yielded multiple papers about a dozen years ago. With 3781 participants, it is by far the largest, most robust and most authoritative study of FND prevalence, and as such has been cited routinely. (According to Google Scholar’s metrics, the paper linked above–just one of a number from the study– has been cited 404 times.)
Not least, the SNSS reflected the decisions of dozens of neurologists across four separate centers in Scotland. Other research, including some that is more recent, has involved much smaller samples from single clinics or settings. That makes it harder to extrapolate or generalize the resulting findings to other patient populations and neurology practices.
As I explained in the letter to Dr Tolchin, the claim that 16% of patients are diagnosed with FND cannot be sustained by a close scrutiny of the data from the SNSS. In fact, this study found that only 5.5% of the patients had rule-in FND diagnoses, not 16%. Approximately 10% more were identified as having “psychological” diagnoses.
(I want to make clear that I have deep sympathy for everyone with a diagnosis of FND. The questions I am raising are about the integrity of the claims being derived from the research. I recommend, as I have previously, this in-depth and well-written article by someone with FND. Called “Cadenza for Fractured Consciousness: A Personal History of the World’s Most Misunderstood Illness,” it is a moving and sometimes harrowing account of the experience.)
In papers and/or public events, FND experts, including co-authors of the SNSS papers, have referenced the study while mis-citing the data to report FND prevalence rates. They insist that FND is not a wastebasket diagnosis and must be determined clinically by rule-in signs. However, to generate the 16% FND prevalence that they attribute to the SNSS, they have combined the 5.5% with the required FND rule-in signs and the approximately 10% with “psychological” issues.
(In some instances, the same experts have described the SNSS findings more accurately by distinguishing clearly between these two groups, so they understand the difference between them. Why they have chosen at other times to lump them together is unclear.)
In the SNSS, the 10% in the “psychological” group were divided into diagnostic sub-categories, including: “non-organic,” “no diagnosis,” “alcohol excess,” “anxiety and depression,” “spinal pain,” “atypical facial/temporomandibular joint pain,” “post-head injury symptoms,” “repetitive strain injury,” and “hyperventilation,” among others. Presumably, the neurologists could not find “organic” causes that, from their perspective, sufficiently explained the symptoms reported by these patients. However, combining the 5.5% with rule-in FND and the 10% with these various “psychological” diagnoses and then asserting categorically that everyone in the total group has FND is unjustified, given the data.
Language matters here. The decision to mis-cite data in this manner appears to be a strategy designed to inflate perceived FND prevalence. Or maybe it is just extreme sloppiness or incompetence. In any event, it is certainly more impressive-sounding and attention-grabbing to promote a problem as #2 rather than presenting it as #6 or #7 or #8.
Unfortunately, this sort of data-mashing contributes to the ongoing confusion among clinicians, patients and the public over FND, the larger domain of what are being called “functional” disorders, and the critical differences between the two. To mitigate this confusion, issue, the mis-citations of the key SNSS findings in published papers–specifically the erroneous claims that the study found an FND prevalence of 16% and that FND was the second-most common diagnostic outcome–should be corrected.
Below is the letter I recently re-sent to Dr Tolchin.
Dear Dr Tolchin:
In a 2021 paper called “The role of evidence-based guidelines in the diagnosis and treatment of functional neurological disorder” (Tolchin et al), published by Epilepsy & Behavior Reports, you and several colleagues wrote that functional neurological disorder (FND) “is among the most common causes of neurological disability, diagnosed in approximately 16% of outpatient neurology consultations.” The citation for this statement is a 2010 paper called “Who is referred to neurology clinics?—the diagnoses made in 3781 new patients,” which was published by Clinical Neurology and Neurosurgery.
Unfortunately, this paper–one of a number arising from a major research project called the Scottish Neurological Symptoms Study (SNSS)–does not appear to support the 16% prevalence claim.
For years, leading neurologists have noted that FND is not a diagnosis of exclusion but a rule-in diagnosis based on positive signs found during clinical examinations. This approach was enshrined in the fifth iteration of the Diagnostic and Statistical Manual of Mental Disorders, which was published in 2013. In 2016, two experts—including Alan Carson, a co-author of both the 2010 paper and Tolchin et al–wrote the following in the abstract of an authoritative account of the epidemiology of FND, published as a chapter of the. Handbook of Clinical Neurology:
“The recent changes in DSM-5 to a definition based on positive identification of physical symptoms which are incongruent and inconsistent with neurologic disease and the lack of need for any psychopathology represent a significant step forward in clarifying the disorder. On this basis, FND account for approximately 6% of neurology outpatient contacts.”
In the text of this chapter, Carson and his co-author, Alexander Lehn, referenced the findings of the SNSS, which investigated the diagnoses of 3781 outpatients seeking care at four Scottish National Health Service neurology centers. (In fact, the text of the Carson & Lehn chapter indicated that the actual FND prevalence from this Scottish study was 5.4%.)
The prevalence at outpatient clinics cited in Tolchin et al is almost triple that mentioned in the 2016 article on the epidemiology of FND. How could the SNSS have yielded these two disparate prevalence rates for the same illness?
In the 2010 paper, the 16% figure included patients found to have what the authors called “functional” symptoms as well as those found to have “psychological” symptoms—all lumped together into a “functional or psychological diagnoses” category. It is unclear why the authors chose to combine these two groups.
The “functional” group included patients diagnosed with “non-epileptic attacks” as well as “functional sensory” and “functional motor” disorders. In a 2009 paper also based on data from the SNSS, this same group of patients was described as having “conversion symptoms.” These patients were the “approximately 6%” (or 5.4%) identified by Carson & Lehn in 2016 as “typical FND cases” based on “positive identification,” per the DSM-5 definition.
That means the additional 10% were the patients identified as having “psychological” symptoms. In the 2010 paper, this group included those given one of a grab-bag of diagnoses, among them “anxiety and depression,” “pain symptoms,” “fibromyalgia,” “post-head injury symptoms,” “alcohol excess,” and “hyperventilation,” as well as a cohort referred to “non-organic.” Given the prevalence highlighted in the Carson & Lehn chapter, patients in this “psychological” category were apparently not deemed to have “typical FND” based on rule-in signs.
Yet investigators in this domain of inquiry have generally ignored this salient point when they have cited the 16% prevalence figure in papers, including Tolchin et al–even as they have maintained that FND is not a diagnosis of exclusion but one requiring positive identification through rule-in signs.
Jon Stone, the first author of the 2010 paper as well as a co-author of Tolchin et al, noted the following during a 2021 podcast produced by the Encephalitis Society: “Some people think that FND is a condition you diagnose when someone has neurological symptoms but you can’t find a brain disease to go along with it. And that’s absolutely not the case. Some people [i.e. clinicians] do that, but if they’re doing it like that they’re doing it wrong.”
Since FND is now officially a rule-in diagnosis, the appropriate and acceptable prevalence rate to cite from the SNSS is the “approximately 6%” (or 5.4%) noted by Carson & Lehn. Unless unreported data indicate that the 10% with “psychological” symptoms in the SNSS sample could have been diagnosed with FND through rule-in signs, claims that the study found prevalence in outpatient neurology clinics to be 16% are unwarranted.
Perhaps these additional patients with “psychological” symptoms had medical complaints that could be described as “functional.” But “functional” in this context simply means that the symptoms or condition have not been found to conform to standard understandings of known diseases; the term should not be casually conflated with the clinical entity known as FND. It is therefore hard to justify identifying patients as having FND under the DSM-5 diagnostic criteria just because they have symptoms deemed to be “functional.”.
If clinicians who diagnose FND without rule-in signs are “doing it wrong,” per Stone, then those who cite the SNSS data to assert a prevalence rate of 16% are also “doing it wrong.” The field of public health relies on consistency and accuracy in the assessment and determination of disease prevalence. It is confusing, not to mention epidemiologically incoherent, when FND experts seem to overlook their own rule-in rule while tripling the reported rates of this challenging diagnosis.
In a related example, a 2021 paper co-authored by Stone and Carson, among others, noted that “tightly defined FND” represents “at least 5%–10% of new neurological consultations.” A prevalence rate of “approximately 6%” (or 5.4%) obviously falls within that range. However, the authors cite no specific reference for the claim that the prevalence of “tightly defined FND”–presumably FND diagnosed through rule-in signs–ranges as high as 10%, which is almost double the 5.4% cited by Carson & Lehn. And certainly 10% is significantly lower than the 16% mentioned in Tolchin et al and other papers.
Rule in should mean rule in. Absent rule-in signs, patients with neurological symptoms that do not adhere to recognized disease patterns should be given a diagnosis of idiopathic neurological disorder, neurological disorder of unknown etiology, or some related term—not FND. If further information regarding the 10% in the SNSS sample with “psychological diagnoses” is unavailable, the prevalence statement citing that research in Tolchin et al should be corrected to reflect the documented rate of “approximately 6%” (or 5.4%) noted by Carson & Lehn.
Thank you for your attention to this matter.
(I have cc’d William Tatum, the editor of Epilepsy & Behavior Reports, on this letter. Since I plan to post the letter on Virology Blog, a popular science venue, I have also cc’d Vincent Racaniello, the host of the website and a microbiology professor at Columbia University.)
David Tuller, DrPH
Senior Fellow in Public Health and Journalism
Center for Global Public Health
School of Public Health
University of California, Berkeley