Proposals for the classification of somatoform disorders seem to be approaching consensus, but no one knows what to do with conversion disorder. ^{Reference Noyes, Stuart and Watson1–Reference Dimsdale and Creed3} The somatoform disorders in general would be modelled as physical symptoms modified by psychosocial factors, restoring physiology to the diagnoses' centre and dispersing some of their stigma: these are the same physiological processes that afflict us all, the proposals would say, made more problematic by psychosocial factors. Unfortunately, although that might make sense of a range of somatic syndromes from irritable bowel syndrome to fibromyalgia, it does not make sense of conversion disorder. ^{Reference Mayou, Kirmayer, Simon, Kroenke and Sharpe2} It is a problem for DSM–V and ICD–11, for reasons that are informative about the practical and political difficulties of psychiatric classification more generally.

What's special about conversion disorder?

The problem with conversion disorder is not in explaining how physiological symptoms could become such a burden, but in how conversion symptoms could exist in the first place. It is hard to see how a hysterical paralysis, for example, could be the amplification of any ‘normal’ symptom, when what it appears to be is physical dysfunction de novo. There are well-rehearsed arguments as to why this dysfunction arises in a faculty such as volition, which can most plausibly be rendered psychologically. ^{Reference Spence, Halligan, Bass and Marshall4} For two hundred years doctors have argued over the psychological processes – suggestion, hypnosis, dissociation, repression or deception – that are proposed to create these pseudoneurological symptoms despite apparently normal anatomy and physiology. Currently, the criteria for conversion in both DSM–IV and ICD–10 reflect this in two ways: first, they require a psychosocial association, and second, the disorder cannot be merely feigned. ^5,6 In other words, there is an unavoidable although unspecified psychological aetiology, and it cannot be conscious deception. In both of these respects the conversion disorder criteria are unique.

The problem of conversion disorder

Conversion is not only a problem for nosological harmony; it threatens the whole physiological somatoform construct – for if conversion disorder can be purely psychological, why not tension headache? It also mandates explanatory criteria – psychological processes and the absence of feigning – that are unpopular, unproved and hopelessly unreliable. ^{Reference Wessely, Halligan, Bass and Marshall7} Neither of the criteria is formally decidable: there is no plausible clinical investigative system that will tell us whether there is a psychological explanation or whether the patient is feigning. They can sometimes be determined positively: sometimes, of course, a psychological explanation is clear; sometimes the patient is caught in acts of obvious feigning; but that leaves an abundance of cases where neither is shown. Should we conclude that these patients are not feigning because we have not proved it, that they do not have psychological explanations because we have not found them, and send them back to their neurologists?

That response is not hypothetical. Neurologists describe a common scenario in which they demonstrate that there is no neuropathological explanation for a patient's symptoms and refer them to a psychiatrist, who sends the patient back saying that no psychiatric disorder can be found. ^{Reference Espay, Goldenhar, Voon, Schrag, Burton and Lang8} What are we to make of such patients? They are caught between two specialties, explained by neither. Nevertheless, there is a difference: the neurologist will have done more than simply thought the presentation unusual or odd, but will in many cases have made a positive clinical diagnosis of a conversion symptom by finding the symptom inconsistent with neuroanatomy and physiology – showing that the problem cannot be neurological as we currently understand it. The psychiatrist, on the other hand, has merely failed to show that it is psychiatric. There are limits to the neurologists’ certainty, of course: ideally they will have demonstrated an inconsistency within the symptoms themselves, such as a gross difference on and off the examination couch, although often there will only be symptoms incommensurate with investigations, or only incongruity with known disorders. Concerns that this leads to excessive false positives can be laid to rest, however; ^{Reference Stone, Smyth, Carson, Lewis, Prescott and Warlow9} neurologists seem to reliably define their diagnostic group of ‘functional’ patients.

So, conversion disorder exemplifies several problems of classification. The criteria intend to capture a group that they do not, so that diagnoses will be made in spite of the criteria rather than because of them; they employ an aetiology that is presumptive at best and anachronistic at worst, but one that has simply not been replaced despite the ‘decade of the brain’; and although this may all be done with an aspiration to validity, it also preserves a strange political divide, with neurologists playing a supporting, but ultimately toothless, part in diagnosis.

A proposal for conversion disorder

Our proposal is this: the diagnostic criteria for conversion should simply be the following:

1. (a) the patient presents with symptoms suggestive of a motor or sensory neurological deficit of significant severity;

2. (b) neuropathological explanations have been excluded, with a qualifier acknowledging the degree of confidence in that exclusion.

The requirements for a psychological association and the exclusion of feigning should be dropped. Not because they are not relevant, but because they cannot at present be determined. They can be retained as explanatory guides, as exhortations to vigilance, as reminders for therapy or even as (strongly) supportive factors when present. Indeed, the nomenclature should be changed to reflect this, with the diagnosis as a whole relabelled ‘functional neurological symptoms’, with the subgroup with a determinate psychological explanation retaining the name ‘conversion disorder’. Finally, the criteria in ICD and DSM should be fully harmonised, and the diagnosis housed within the somatoform disorders chapter rather than with dissociative disorders, since the process suggested by the latter grouping is also a presumption.

The goals of classification

Although simply stated, the choice of labels here was anything but simple. Diagnostic labels serve partly to communicate with colleagues, partly to communicate with patients and partly as our approximation to the truth. All of the terms available met some and none met all of those goals, but these represent a healthy compromise. The proposals may not seem particularly radical; they merely propose a further downgrading of the psychological assumptions that have been weakened with each iteration of the diagnostic criteria. Indeed, similar proposals were considered for the last round of revisions, as the psychological criteria ‘lacked evidence’ and were conspicuously aetiological rather than descriptive. ^{Reference Martin and Widiger10} Although the psychological criteria are unreliable in application and discrimination, they have a tenacious grip on our sense of validity. No one who has had a patient develop a dysphonia before going on stage or a hand dystonia after signing an unwelcome document is left in any doubt that there is a psychological story that makes sense of their problem; and so compelling is this that we presume a similar story could be told of any other patient if they would only reveal it. Dispensing with this criterion would seem to give up something precious – the last toe-hold of Freudian theory – and one of the few remaining criteria where psychiatrists seem to have something unique to add, at a time when their value is under growing threat; for it would mean neurologists could diagnose the disorder, just as gastroenterologists can diagnose irritable bowel syndrome. In addition, it would mean acknowledging that in some cases conversion really was unexplained by all branches of medicine, including psychiatry. However, that is the reality of our current state of knowledge, notwithstanding renewed – and exciting – research directions.

Funding

R.A.K. was supported by the Wellcome Trust. S.C.W. and A.S.D. were partially funded by the South London and Maudsley National Health Service Foundation Trust/Institute of Psychiatry National Institute of Health Research Biomedical Research Centre. T.R. was supported by the Medical Research Council. S.A. was supported by the Swiss National Research Foundation. A.S.D. has received lecture fees and honoraria from Janssen-Cilag, Eli Lilly and AstraZeneca for work unrelated to conversion disorder.