By David Tuller, DrPH
The Journal of Psychosomatic Research has just published a paper called “The Impact of Depression, Anxiety and Personality Disorders on the Outcome of Patients with Functional Limb Weakness – Individual Patient Data Meta-Analysis.” As I have previously noted, the journal is something of a house organ for seriously flawed research from adherents of the what is referred to as the “biopsychosocial model,” and experts in functional neurological disorder (FND) fall squarely in that camp.
Those diagnosed with FND suffer from very serious and complex disabilities. A patient who goes by the moniker @FnDPortal has written a compelling essay, Cadenza for Fractured Consciousness: A Personal History of the World’s Most Misunderstood Illness, that is well worth a read. I continue to reference this excellent piece even though its author has blocked me on X, the decaying social media platform formerly known as Twitter, presumably because he dislikes that I regularly highlight the way FND experts and investigators misrepresent the research. My empathy for patients does not mean that I need to automatically accept the pronouncements of FND experts, who routinely make claims about prevalence and other issues that are not supported by the evidence they present.
The new systematic review and meta-analysis included the results from eight studies, with a total of 348 participants. Here’s the first sentence: “Functional neurological disorder (FND) is common and accounts for a significant proportion of patients seen in neurological practice.” The sentence is notable for its lack of specificity about how “common” FND actually is. In the last decade, leaders in the field have asserted—in more than 50 papers–that FND is the second-most-common diagnosis at outpatient neurology clinics and/or that the prevalence in these venues is 16%.
To support these data points, investigators have routinely cited Stone et al, a 2010 paper arising from a major research project called the Scottish Neurology Symptoms Study (SNSS). But the paper, and the SNSS overall, do not in fact justify these prevalence claims—as I have pointed out in multiple blog posts and in multiple letters to journals noting these errors. The SNSS, which yielded a number of papers, found that only 209 of 3781 patients—that is, 5.5%–were given diagnoses that could qualify as conversion disorder, the former name for FND. At 5.5%, conversion disorder (or FND) was way down the list of diagnoses—not remotely close to being the second-most-common presentation.
In other words, FND investigators have indisputably misrepresented the data from the SNSS–and they know it. Perhaps the new paper indicates that they have decided to abandon these exaggerated assertions now that they’ve been called out publicly. From my perspective, investigators who deliberately inflate statistics in this manner are engaging in serious research misconduct. The false claim that the FND prevalence in the SNSS was almost triple what it actually was has effectively created the impression that the clinical category of interest to the investigators is much larger than indicated by the most authoritative data. This is an example of what is often referred to as diagnostic creep.
Nonetheless, the first sentence remains highly unsatisfactory. It does not include any references to support the claim that FND is “common.” Perhaps that is because the SNSS, which collated data from 36 neurologists at four different outpatient clinics, was by far the largest and most extensive study of FND prevalence. Some more recent studies have found FND rates higher than 5.5%. However, these studies have all been far less robust than the SNSS; they are much, much smaller and have drawn patients from far fewer neurology practices. Since the authors of the new paper cite no studies, I assume the claim that it is “common” rests largely on their perceived clinical experience—even though their opionion is not really an appropriate scientific basis for making such a blanket statement of purported fact. Nonetheless, the FND experts have asserted this point so often that it has attained the aura of “truth”—which would explain why peer reviewers apparently failed to challenge it by requesting proper referencing. It certainly seems likely that FND is being diagnosed in greater numbers than previously, but that’s generally what happens when the process of diagnostic creep gets underway.
Another interesting aspect of the study is that it failed to document what the authors apparently hoped or expected it would. The introduction includes this statement: “Psychological comorbidities can play an important role as predisposing, precipitating, or perpetuating factors and could shape the course of the subsequent FND.” Of course, it should not be surprising that FND patients would have psychological comorbidities, given the challenges of living with an illness that resists easy etiological explanations and is generally poorly understood. Yet the claims that these co-morbidities are “predisposing, precipitating and perpetuating factors” appears to be based mainly on hypotheses, clinical opinion and interpretations of ambiguous data, which often indicate associations rather than causal relationships.
Moreover, the study itself failed to prove what the authors presumably hoped it would–that anxiety, depression and related mood states have an impact on prognosis. Here’s the bottom line from the study: “There was no clear finding to suggest that the baseline presence of either depression or anxiety in absolute terms was correlated with clinical outcome in these patients. When stratified according to severity of depression and anxiety, there was no significant association between symptom severity and clinical prognosis.”
Given these apparently unsatisfactory negative findings, the investigators are calling for “further exploration in larger scale, prospective studies.” It is perhaps unfortunate that their null findings in this comprehensive review and meta-analysis have not, it seems, led the investigators to question any of their own long-standing assumptions.