Article Text

Download PDFPDF

Research paper
Specialist inpatient treatment for severe motor conversion disorder: a retrospective comparative study
  1. Ruaidhri McCormack1,2,
  2. John Moriarty3,
  3. John D Mellers3,
  4. Paul Shotbolt3,
  5. Rosa Pastena3,
  6. Nadine Landes3,
  7. Laura Goldstein4,
  8. Simon Fleminger5,
  9. Anthony S David2
  1. 1National Institute for Health Research (NIHR) Biomedical Research Centre, South London & Maudsley NHS Foundation Trust and Institute of Psychiatry, King's College London, London, UK
  2. 2Section of Cognitive Neuropsychiatry, Institute of Psychiatry, King's College London, London, UK
  3. 3South London & Maudsley NHS Foundation Trust, London, UK
  4. 4Department of Psychology and NIHR Dementia Biomedical Research Unit at South London & Maudsley NHS Foundation Trust and Institute of Psychiatry, King's College London, London, UK
  5. 5South London & Maudsley NHS Foundation Trust and Institute of Psychiatry, King's College London, London, UK
  1. Correspondence to Dr Ruaidhri McCormack, Section of Cognitive Neuropsychiatry, Institute of Psychiatry, 16 De Crespigny Park, London SE5 8AF, UK; r.mccormack{at}


Background Gold standard protocols have yet to be established for the treatment of motor conversion disorder (MCD). There is limited evidence to support inpatient, multidisciplinary intervention in chronic, severe cases.

Aims To evaluate the characteristics and outcomes of MCD patients admitted to a specialist neuropsychiatric inpatient unit.

Methods All patients admitted to the Lishman Unit (years 2007–2011) with a diagnosis of MCD were included. Data relevant to characteristics and status with regard to mobility, activities of daily living (ADLs) and Modified Rankin Scale (MRS) score at admission and discharge were extracted.

Results Thirty-three cases (78.8% female) were included; the median duration of illness was 48 months. In comparison with brain injury patients admitted to the same unit, more cases had histories of childhood sexual abuse (36.4%, n=12), premorbid non-dissociative mental illness (81.1%, n=27) and employment as a healthcare/social-care worker (45.5%, n=15). Cases showed significant improvements in MRS scores (p<0.001), mobility (p<0.001) and ADL (p=0.002) following inpatient treatment.

Conclusions Patients with severe, long-standing MCD can achieve significant improvements in functioning after admission to a neuropsychiatry unit.


Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Motor conversion disorders (MCDs) including functional weakness and movement disorders account for a significant proportion of patients seen in neurology clinics, with up to a third of new referrals presenting with symptoms that are unexplained by organic illness.1 Their aetiology is not well understood and this has led to significant debate regarding nosology; terms such as ‘functional movement disorder’, ‘conversion disorder’, ‘hysteria’, ‘dissociative’, ‘psychogenic’, ‘non-organic’ and ‘medically unexplained’ all have their adherents.2–4 Interest in a neurobiological model is now growing alongside existing aetiological theories focusing on psychological trauma.5–8 The absence of universally accepted diagnostic criteria means that conversion disorders are probably best diagnosed in a multidisciplinary setting by expert physicians (usually neurologists), (neuro-)psychiatrists and (neuro-)radiologists.9–11

In some cases, MCDs can persist for months or years leading to significant functional impairment and distress for both patient and family. Gold-standard treatment protocols do not exist due to the paucity of the literature on the topic and problems in conducting randomised trials on this patient population.

In the case of MCD, there are mostly just case reports and series (maximum n=10) suggesting that a positive rehabilitative approach in the inpatient setting results in positive outcomes.12–16 One recent survey of 26 MCD patients conducted some years after specialist inpatient admission suggested that admission results in improvement in symptoms and function, but this was based on a self-report questionnaire and subject to recall bias.17 One further case–control study (n=60 cases) published in 2012 showed that MCD patients significantly benefit from a week-long intensive outpatient rehabilitative programme.18 Outcomes in this study were general and fairly non-specific—for example, ‘no improvement or worse’ or ‘almost completely normal/in remission’.18 Cases also represented a mix of acute and chronic cases (median 17.5 months).18 The rehabilitative approach adopted in these studies varied in terms of content but invariably involved the multidisciplinary input of at least a psychiatrist, psychologist and physiotherapist.

The purpose of this study was to evaluate outcomes of a group of patients with chronic, severe MCD at the point of discharge from a specialist neuropsychiatric unit; the characteristics of this patient population in comparison with brain injury patients admitted to the same unit were also analysed so that potential risk factors for poor outcomes and increased length of stay could be uncovered.


Sample population

The study included all (n=33) patients discharged from the Lishman Neuropsychiatry Unit (South London & Maudsley NHS Foundation Trust, London, UK) over the 5 years from 2007 to 2011 with an ICD-10 diagnosis of dissociative motor (conversion) disorder F44.4 following multidisciplinary treatment. Although motor symptoms were essential, those with co-morbid sensory or non-epileptic dissociative features (mixed dissociative, F44.7) were not excluded. In cases where there was a documented co-morbid somatoform pain disorder, only those with a clear independent medically unexplained disturbance of motor function were included; patients who primarily attributed their motor dysfunction to pain were excluded.

Nature of the intervention

Inpatient treatment is provided at the Lishman Unit, a specialist neuropsychiatry unit at the Maudsley Hospital in London. The core intervention for MCD involves the offer of treatment from a neuropsychiatrist, psychologist, physiotherapist and occupational therapist. Other professionals (eg, speech and language therapist) are recruited based on individual need. The intervention is tailored to the patient's needs but follows the general principles outlined below.

The psychiatrist liaises with medical colleagues to rule out any possible neurological and medical causes of symptoms. The initial phase of the admission is spent eliciting key elements of the history, symptom evolution and environment that may act as predisposing, precipitating and maintaining factors. Medical and psychiatric co-morbidities are treated and occasionally further investigations done—although more usually the importance of not carrying out further tests is discussed with the patient.

The psychologist pursues a predominantly cognitive behavioural therapy (CBT) model of psychotherapy involving psycho-education, cognitive and behavioural techniques, and relapse prevention strategies. Each patient is first assessed for CBT suitability, a decision usually guided by the ‘Suitability for Short-Term Cognitive Therapy Rating Scale (SRS)’.19 This scale assesses factors relevant to symptom profile, personal responsibility, alliance potential, ability to analyse and differentiate emotions and automatic thoughts, and degree of optimism or pessimism regarding therapy.19 The presence of psychiatric co-morbidities are noted and worked into the proposed treatment plan. The therapy is time-limited and usually delivered on an individual basis. In this study at the Lishman Unit, 28 of 33 cases (84.9%) had CBT—the rest declined treatment or were not suitable.

If considered suitable, the therapy proceeds to challenge any cognitive distortions that might affect a patient's motivation, determination or ability to engage on an interpersonal level. The therapist works to build insight into a more psychological understanding of symptoms and assists in shifting the locus of control from external (eg, a dependence on medications or carers) to internal by fostering insight and assertiveness. If thought important, the patient is assisted to discover links between past or present experience and physical symptoms although this is not essential. Techniques may include mood and thought diaries (the patient then trying to link these moods and thoughts with environmental exposures), relaxation techniques (particularly if the patient has fears and expectations around improvement) and graded exposure (if avoidance is employed as a coping strategy), along with homework and tasks decided upon in collaboration with the patient. Neuropsychological testing is performed if a greater knowledge of a patient's cognitive profile may assist the team in understanding his/her presentation, interpersonal style or functional impairments.

Other professionals like physiotherapists and occupational therapists employ similar strategies; issues concerning insight, motivation, engagement and confidence also apply. A positive rehabilitative approach is adopted with graded steps and positive reinforcement. On the Lishman unit, physiotherapy intervention normally occurs twice weekly and approaches are adapted to the relevant disability. Techniques may include stretching and massage (eg, if dystonia) and exercises targeting posture, stability, balance and strength. The patient may be assisted to recognise inconsistencies in his/her own presentation and work with the physiotherapist to translate this recognition into a different perspective of one's illness, the illness later being perceived as less pervasive, damaging, irreversible or ‘organic’. This is followed by approaches to help ‘relearn control’ over the body, ‘teach the muscles and nerves/brain to reconnect’ and ‘overcome mental blocks’. Techniques employing distraction from the affected limb(s) may lead to demonstrations of preserved function and this can be built upon systematically.

Occupational therapists assess baseline activities of daily living (ADLs) and safety and work to integrate new accumulated motor skills into daily tasks. The patient is encouraged to engage in activities that will test his/her existing skill set. The therapist also looks to assess the patient's perspective on their role and routine; perhaps the patient has come to view himself as ‘disabled’ rather than independent and able. The environment at home (physical environment and social support network) is evaluated. Preliminary discussions may occur regarding a broader integration into society—going back to work, for instance.

All professionals work with the patient to formulate relapse prevention strategies in anticipation of discharge. The patient is assisted to identify triggers for symptoms. The patient is taught to self-monitor in many respects—symptoms, emotions, thoughts—in the hope they can address any problems and apply learned techniques in a timely and autonomous fashion in order to offset resurgence of symptoms and disability.

A pervasive non-judgemental atmosphere and positive reinforcement are crucial and apply across all interventions.

Data extraction and analysis

This study was a retrospective comparative study, with information gained from discharge summaries and electronic records. Data extracted included demographic details, duration of illness prior to admission, length of stay, medical/psychiatric co-morbidity and history, history of abuse, history as a carer or healthcare/social-care professional and status with regard to employment, mobility and ADLs on admission and discharge. Both mobility and ADLs were ranked on ordinal scales agreed by the authors (there are no MCD-specific scales measuring function). For mobility, the categories were ‘walking unaided, ‘walking with aids’, ‘wheelchair or bedbound’, while for ADLs, the categories were ‘largely independent’, ‘somewhat dependent’, ‘mostly/fully dependent’. Modified Rankin Scale (MRS) scores were assigned for each patient at admission and discharge. The MRS score, commonly used in stroke medicine, grades symptoms on a 6-point scale where 0 is ‘no symptoms at all’ and 5 is ‘severe disability: bedridden, incontinent, and requiring constant nursing care and attention’.20 The first author (RMC) assigned these scores; he has worked in stroke medicine and neurology prior to working in psychiatry and has experience in assigning MRS scores.

Statistical analysis was performed using Stata, and the relevant tests were applied to analyse outcomes and predictors of outcomes (eg, Fisher's Exact Test [FET], Wilcoxon, logistic regression). Mobility, ADLs and MRS score were measured per the above ordinal scales at the point of discharge and statistically compared with admission values. Outcomes were defined as ‘improvement in ADLs from admission to discharge or independent at discharge’ (yes/no), ‘improvement in mobility from admission to discharge or walking unaided at discharge’ (yes/no), ‘improvement in MRS from admission to discharge’ (yes/no) and ‘length of stay in hospital’ (above/below median). Patients who self-discharged against medical advice were included in all analyses.


Patient characteristics

The mean age of cases was 40.8 years (SD 12.1, range 20–59) with 78.8% (n=26) being female. All cases saw a neurologist and had appropriate neurological investigations. The median length of illness preadmission was 48 months (IQR 19–72); 97.0% (n=32) were admitted informally. The majority presented with a loss of motor function (n=29, 87.9%), while four (12.1%) had abnormal motor function (eg, tremor, dystonia, ataxia). Almost two-thirds (n=21, 63.6%) had bilateral symptoms. Co-morbid non-epileptic dissociative features occurred in 55% (n=18), while 60.6% (n=20) had a concurrent non-dissociative psychiatric co-morbidity. Of this group with an ICD-10 psychiatric co-morbidity, the most common twin diagnoses were somatoform pain disorder (n=7, 35.0%) and somatisation disorder (n=6, 30.0%). Approximately 18% (n=6, 18.2%) had co-morbid neurological disorder(s) or a particularly relevant neurological history; in each of these cases, the multidisciplinary team decided that the motor symptomatology was not in keeping with the prevailing neurology. Box 1 provides a detailed breakdown of these psychiatric and neurological co-morbidities.

Box 1

Psychiatric and neurological co-morbidities in motor conversion disorder inpatients

Psychiatric co-morbidities

  • No co-morbidity=39.4% (n=13)

  • Co-morbidity=60.6% (n=20) of which...

    • Somatoform pain disorder=35.0% (n=7)

    • Somatisation disorder=30.0% (n=6)

    • Mild depressive episode=5.00% (n=1)

    • Adult ADHD=5.00% (n=1)

    • Specific phobia=5.00% (n=1)

    • Specific developmental disorder (dyslexia)=5.00% (n=1)

    • Multiple=15.0% (n=3)

Neurological co-morbidities

  • No co-morbidity=81.8% (n=27)

  • Co-morbidity=18.2% (n=6) of which...

    • Epilepsy=33.3% (n=2)

    • Neurogenic bladder disorder=33.3% (n=2)

    • Multiple sclerosis=16.7% (n=1)

    • Cerebellar atrophy on MRI Brain=16.7% (n=1)

ADHD, attention-deficit hyperactivity disorder.

One-third (n=11, 33.3%) of cases had a history of medically unexplained symptoms prior to the onset of the current condition, while the most common psychiatric history was a history of one or more depressive episodes (n=21, 63.6%). Patients with MCD had high rates of having histories of all types of abuse, particularly childhood sexual abuse (n=12, 36.4%). We compared the abuse figures for motor conversion cases with age-matched and sex-matched brain injury controls (n=33) treated at the same neuropsychiatry unit using Fisher's exact test. There were significantly higher rates of abuse in the MCD patients versus the brain injury group: (child sexual abuse, n=12 (36.4%) vs 0; p<0.001; child physical abuse, 9 (27.3%) vs 1; p=0.013; adult physical/sexual abuse, 11 (33.0%) vs 2 (6.00%); p=0.011). Table 1 describes cases on a list of other characteristics.

Table 1

Clinical and demographic characteristics: conversion disorder cases

Cases had marked levels of functional impairment on admission with 60.6% (n=20) wheelchair or bedbound, 42.4% (n=14) somewhat/mostly/fully dependent for ADLs and an 87.9% (n=29) unemployment rate. This unemployment rate had risen from 30.3% (n=10) premorbidly, illustrating the severity of conversion disorder. The mean MRS score was 3.64 (SD 0.86, range 2–5).


Table 2 shows the progress in terms of mobility and ADLs for MCD cases. A majority of cases had good outcomes with regard to mobility (n=24, 72.7%), ADLs (n=29, 87.9%) and MRS scores (n=24, 72.7%). The proportion of patients mobilising unaided or with the aid of a stick/crutches rose from 33.3% (n=11) to 72.7% (n=24). The proportion of those who were wheelchair dependent fell from 57.6% (n=19) to 15.2% (n=5), while the proportion of patients largely independent with their ADLs rose from 57.6% (n=19) to 78.8% (n=26). These improvements were statistically significant. Cases also showed a statistically significant improvement in MRS score from admission (mean 3.64, SD 0.86, range 2–5) to discharge (mean 2.82, SD 0.85, range 2–5); p<0.001. No case deteriorated in terms of MRS score.

Table 2

Mobility and activities of daily living (ADLs) of motor conversion disorder patients at admission and discharge

Regarding patients’ ability to be cared for at home, this increased from 90.9% (n=30) to 100% (n=33) (FET, p=0.238). The median length of stay was 101 days (IQR 84–130). Only 3 cases (9.09%) discharged against medical advice before treatment was complete.

Predictors of outcome

There were no significant predictors of good outcomes in terms of mobility or MRS score. Being in a nursing home or hospital, preadmission predicted a poor ADL outcome (OR 28, 95% CI 1.71 to 459, p=0.02); this did not remain significant when adjusted for mobility on admission (p value rose to 0.077). Receiving CBT as part of the treatment package (84.9% received CBT) was associated with a good ADL outcome although CIs were wide and included unity (OR 8.67, 95% CI 0.87 to 86.1, p=0.065).

Having non-epileptic dissociative features in addition to motor features significantly increased length of stay (OR 5.5, 95% CI 1.22 to 24.8, p=0.03). This association grew more significant (OR 9.08, 95% CI 1.45 to 56.7, p=0.018) when adjusted for length of disorder, psychiatric and medical co-morbidity, psychiatric history, place of care and carer status preadmission.


Patient characteristics

The diagnosis of MCD was decided upon in all cases by a multidisciplinary team involving an expert physician (usually a neurologist, stroke-physician or neurosurgeon), a neuropsychiatrist and a neuroradiologist. Most cases of MCD were female (78.8%) and had had a chronic severe clinical course (median 4 years duration of illness with high levels of physical disability, dependence on ADLs and unemployment). Psychiatric co-morbidities were encountered in the majority, particularly somatisation or somatoform pain disorder (n=13, 39.4% overall); a third had a history of medically unexplained symptoms prior to the onset of the current disabling MCD.

The figures in table 1 suggest that patients with MCD have high rates of having histories of all types of abuse, particularly childhood sexual abuse (n=12, 36.4%). The UK Office for National Statistics has only published sexual abuse figures for persons aged 16 to 59,21 but an NSPCC survey has estimated that 7% of girls and 2.8% of boys aged 11 to 17 have a history of contact sexual abuse.22 A total of 6.9% of these children experienced severe physical abuse.22 We also compared the abuse figures for motor conversion cases with a contemporaneous group of age-matched and sex-matched brain injury controls treated in the same unit. When we did this, we showed statistically significant associations between a diagnosis of MCD and all types of abuse. It is recognised, however, that reporting bias may have led to these brain injury controls having relatively lower rates of sexual abuse; this information may have been less actively sought when conducting patient interviews.

The association between MCD and different types of childhood abuse is fairly consistent across the available literature, with rates of abuse being perhaps higher in those with non-epileptic dissociative features.23–26 Rates of abuse are higher in MCD patients even when compared with psychiatric (affective) controls.23 ,24 In our study, those with non-epileptic dissociative features had somewhat higher rates of all types of abuse compared to those without, but the difference between groups was not statistically significant.

Cases were also significantly more likely to have trained/worked as a healthcare/social-care professional (p=0.002). It is accepted in psychological models of conversion disorder that an individual's perception and understanding of illness and disability may influence the nature of manifested conversion symptoms.10 However, we were unable to find any peer-reviewed research supporting the received clinical wisdom that working with the ill/disabled or training as a health professional predisposes to MCD; an association with factitious disorders has, however, been established.27


Cases significantly improved over the course of admission in terms of independence with ADLs, mobility and MRS score. These data suggest that inpatient admission to a specialist neuropsychiatry unit is beneficial in this subset of patients with chronic and severe MCD. There was no significant increase in MCD cases being cared for at home after discharge compared with preadmission but this is probably due to an already high level of home care in these patients at baseline—ultimately 100% of MCD cases were able to return home.

Predictors of outcomes

When one looks at the established literature, few predictors of positive and negative outcomes in MCD patients have been consistently replicated. Duration of symptoms has been consistently shown to affect outcome, with a longer duration of symptoms associated with poorer outcomes.28–30 Co-morbid personality disorder was shown to be predictive of a poor outcome in two studies,30 ,31 but not in a third.28 The second positive study did not do a regression analysis or clearly show personality disorder as an independent predictor.31 Receipt of financial benefits at the time of presentation/admission29 ,32 and negative future expectations31 ,32 have also been reported to be predictive of poor outcomes. A co-morbid axis 1 disorder like depressive/anxiety disorder is predictive of a good outcome, presumably because this axis 1 disorder contributes to MCD symptoms and can be treated.29 ,30

To our knowledge, no systematic review of predictors of outcomes in MCD has been published; reported predictors of a positive outcome include sudden onset of symptoms,28 being female,18 and a change in marital status during the follow-up period,29 the latter perhaps suggesting an improvement in social environment. Reported predictors of a negative outcome include medical comorbidity,31 higher axis 5 functional scores,31 higher age of onset33 and non-attribution of symptoms to a psychological cause.32 Again, one of these studies did not show that the relevant predictors were independent.31

Perception of effective treatment and dissatisfaction with the treating physician are sometimes cited as predictors of outcome. Often the authors simply showed that patients who did badly were more likely to be dissatisfied with their doctor and those who did well felt they had had effective treatment; neither is particularly surprising.34 ,35

In this study, receiving CBT approached significance as a predictor of good ADL outcomes; this suggests that this treatment may be effective for MCD but further research needs to be done; there are no randomised controlled studies in the literature looking at CBT for MCD.

Regarding length of stay, we showed in this study that those with non-epileptic dissociative features were more likely be in hospital for longer (p=0.018). Over half of the MCD patients in this study had co-morbid non-epileptic features while somatoform disorders were the most common non-dissociative co-morbidity; one could speculate that these patients with multiple types of dissociative and/or somatisation symptoms are more difficult to treat or are perhaps more likely to require a specialist inpatient neuropsychiatry admission. We could not find significant independent predictors of improvement in ambulation, ADLs or MRS score in the MCD cases. In particular, we looked for but could not replicate—duration of symptoms preadmission, medical co-morbidity and higher age of onset; perhaps these would have been significant with a longer follow-up period. Only one patient was given a diagnosis of personality disorder, and this probably demonstrates the reluctance of the site's neuropsychiatrists to diagnose personality disorder in the presence of an active axis 1 disorder (ie, MCD±other axis 1).

On the other hand, the lack of strong predictors of good or poor outcome in this study could be interpreted as showing the general applicability of the treatment approach to this group of patients. For example, even very long standing and severe cases may sometimes respond.

Studying the MCD population: limitations

There are considerable problems in studying this patient population. Box 2 lists some of the problems that we have observed from looking at existing studies and conducting our own.

Box 2

Problems with conducting follow-up studies in patients with conversion disorders

  • Evolving diagnostic criteria leading to inconsistent patient selection

  • Lack of uniformity in

    1. Study design (retrospective vs prospective, cohort vs case–control, ‘prospective’ studies without baseline data at the start of the follow-up period to serve as potential predictors of outcome).

    2. Patient characteristics (inpatients vs outpatients, male vs female, type of conversion disorder, etc.).

    3. Duration of follow-up

    4. Accepted outcome measures in conversion disorders—scales validated for use in other psychiatric or neurological disorders are used, or broader outcomes are used (eg, ‘persistence in symptoms’, ‘improvement in symptoms’).

  • Difficulty in powering studies with sufficient numbers

  • Finding a control group that is exposed to the same intervention or treatment as the conversion disorder cases; this is important if one is to statistically compare cases and controls on their outcomes as they relate to treatment efficacy.

Our study was retrospective in design and observational in nature, which means it is not possible to definitively conclude which exact elements broad multidisciplinary treatment package brought about improvement and to quantify their effects. Furthermore, treatment was necessarily individually tailored taking into account the patients’ problems, education, motivation, etc. Further research is necessary to distil the key elements of such interventions. As discussed above, we do know, however, that having CBT approached significance in terms of improving ADL outcome (p=0.065). Generalisability is also limited by the inclusion of a particularly morbid subset of motor conversion cases admitted to a tertiary referral centre. Couprie et al36 have shown that improvement over the course of inpatient admission is predictive of later outcome (risk ratio 3.2, 95% CI 1.8 to 5.6), so although this study did not include a prospective follow-up, the improvements noted here over the course of admission in MCD patients were likely to be sustained; no patient was readmitted to the unit subsequent to the relevant MCD presentation.

In conclusion, we suggest that even very severe cases of MCD are able to respond to a broad multidisciplinary treatment approach delivered in an inpatient setting. The time is now ripe for controlled intervention trials and research aimed at uncovering the key elements of such interventions.


We acknowledge the work of all the staff on the Lishman unit for their contributions to this report.



  • Contributors RMC and ASD conceived and designed the study. RMC conducted data extraction and data analysis and wrote the first draft of the paper. All authors contributed to editing the paper and all authors approved the final manuscript. RMC is guarantor.

  • Funding The first author (RMC) was supported by the National Institute for Health Research (NIHR) Biomedical Research Centre at the South London & Maudsley NHS Trust and Institute of Psychiatry, King's College London.

  • Competing interests None.

  • Ethics approval This study conformed to local approval standards for audit.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Linked Articles