Neuropsychiatric Factors Linked to Adherence and Short-Term Outcome in a U.S. Functional Neurological Disorders Clinic: A Retrospective Cohort Study

The approach to the assessment and management of functional neurological disorder (FND)-conversion disorder has undergone significant changes. The DSM-5 includes FND as a parenthetical term to conversion disorder and emphasizes positive “rule-in” signs on examination as diagnostic features.^1,2 The requirement to exclude feigning has been removed, along with the prerequisite of identifying an acute stressor.^3,4 Patients with FND are commonly diagnosed with affective and trauma-related disorders,^5,6 and the presence of psychiatric comorbidities is linked to prognosis.⁷ Despite advances in diagnosis and treatment,⁸ clinicians across neurology and psychiatry report feeling ill-equipped managing patients with FND.^9,10 From a public health perspective, patients with FND utilize medical resources in the context of repeated emergency department visits, unnecessary diagnostic tests, and frequent shuffling among providers.¹¹

Advancements in the clinical approach to FND emphasize communicating the diagnosis in a straightforward and empathic manner as a critical first step in management.¹² This approach includes providing patients with educational information from available resources (e.g., websites such as neurosymptoms.org or epilepsy.com/learn/types-seizures/nonepileptic-seizures-or-events), as well as showing patients the examination signs used to diagnosis FND.^13,14 Thereafter, research supports roles for physical and occupational therapy (OT) as first-line treatments for functional motor symptoms.^15,16 Consensus guidelines for the physical therapy (PT) treatment of functional motor disorders have helped standardize PT interventions for this population.¹⁷ Several randomized controlled clinical trials have also established a primary role for cognitive-behavioral psychotherapy (CBT) in the treatment of FND.^18–20 Thus, the overall diagnostic evaluation and initial treatments are aimed at appreciating neuropsychiatric and psychosocial factors that may be predisposing, precipitating, or perpetuating factors in a patient’s presentation and addressing modifiable factors.²¹ Notably, despite clinical research studies identifying effective management strategies for FND, many questions remain regarding the feasibility of implementing a new standard of care in an outpatient clinical setting. Furthermore, neuropsychiatric factors linked to outpatient adherence and prognosis are not well understood, although there is emerging evidence that psychiatric comorbidities, illness duration, and psychosocial factors, including pending disability applications, may have prognostic implications.⁷

In this retrospective cohort study, we reviewed the medical records of 100 consecutive outpatients with suspected motor FND evaluated and followed longitudinally in our FND clinic. We have previously published on the baseline characteristics of this cohort and identified neuropsychiatric factors associated with gender, illness duration, work disability status, and motor subtype.^22,23 Here, we investigated neuropsychiatric and psychosocial factors linked to treatment adherence and clinical improvement at 7 months following the use of an updated, individualized standard-of-care treatment approach that included education on FND, physical/occupational therapy, and/or psychotherapy. We hypothesized that psychiatric and neurologic comorbidities would independently relate to outcome and treatment adherence and that poor prognosis would be observed in individuals on or applying for disability, with prolonged illness duration at presentation, past adverse life events, and/or comorbid chronic pain.

Methods

Patient Cohort and Interventions

We reviewed the medical records of the first 100 consecutive outpatients evaluated in the FND clinic at the Massachusetts General Hospital (MGH) for suspected motor FND. Details of patients referred to this clinic have been previously described.^22,23 Patients underwent an initial 1.5-hour semistructured clinical neuropsychiatric interview and physical examination by a dual-trained and board-certified neurologist-psychiatrist (D.L.P.). Nineteen patients were initially evaluated by a clinical fellow or resident supervised by this same neurologist-psychiatrist (D.L.P). Patient interviews focused on any past or ongoing medically unexplained motor and sensory symptoms (including abnormal movements, weakness, gait trouble, seizures, sensory deficits), cognitive complaints, head trauma, mental health, substance use, and medical, family, and psychosocial histories (including developmental trajectories, adverse life events, and work disability status). Neurological examinations focused on detecting “positive” functional neurological signs including collapsing/give-way weakness, Hoover’s sign, variable/distractible abnormal movements, tremor entrainment, astasia-abasia gait, and splitting of the midline nondermatomal sensory deficits.²

Established diagnostic criteria for psychogenic nonepileptic seizures (PNES),²⁴ functional movement disorders (FMD),²⁵ and functional weakness¹ were used. Patients could meet criteria for more than one motor FND subtype.

After the initial assessment, patients were given a positive diagnosis of FND consistent with validated approaches.^12,13 The term “functional neurological disorder” was specifically communicated to the patient, as well as the pertinent subtype: FMD, functional weakness, and/or PNES. Emphasis was given to the diagnosis being both common and real, and the impact that the patient’s symptoms had on his or her daily life was validated. Patients were introduced to the www.neurosymptoms.org website and given printed educational materials from the site to review. The diagnosis was specifically presented as treatable.

Thereafter, treatment plans were individualized, emphasizing physical and occupational therapy^16,17 and CBT.^18–20 For PT, clinicians were asked to follow the consensus PT guidelines for FND.¹⁷ MGH PTs received instruction on the implementation of the published guidelines led by three authors (J.C., J.M., and D.L.P.). For individuals receiving PT outside of MGH, a printout of the guidelines was provided with instructions to provide the article to their local PT provider to inform therapy. For patients with isolated PNES, individuals were commonly referred to the workbook “Taking Control of Your Seizures: Workbook”²⁶ in the context of individual psychotherapy with a therapist familiar with CBT, who was also referred to the workbook “Treating Nonepileptic Seizures: Therapist Guide.”²⁷ For those with other motor FND semiologies (including PNES with mixed symptoms), the guided self-help CBT workbook “Overcoming Functional Neurological Symptoms: A Five Area Approach,”²⁸ based on the clinical trial by Sharpe and colleagues,¹⁸ was frequently recommended, either to be used during the course of individual psychotherapy sessions or to be used by the patient on his or her own in the context of longitudinal FND clinic follow-up. All MGH/outside psychotherapists were encouraged to contact MGH FND clinic providers after their initial outpatient assessment to ensure proficiency in using CBT treatments for FND. Psychopharmacological interventions were based on usual care to treat comorbid psychiatric conditions.

Chart Review and Data Analyses

Chart reviews were performed to collect baseline and follow-up data for patients evaluated at the MGH FND clinic. The coded follow-up visit was the first follow-up occurring after at least 180 days from the baseline diagnostic visit such that the average interval follow-up between visits was 7.0±2.5 months. Chart reviews were performed by two authors (N.M. and S.S.Y.) who were not involved clinically. Institutional review board approval was obtained from the Partners Human Research Committee. Individual informed consents were not required for this chart review study.

The primary outcome (dependent) variables included 1) having at least one no-show appointment, 2) clinical improvement at 7 months, and 3) complete adherence to initial PT/OT and psychotherapy recommendations by those who were newly referred to treatments after the initial consultation (excluding individuals already receiving some form of treatment at the time of their initial visit). “No show” was defined as having at least one missed appointment over the follow-up period. For analyses of improvement at 7 months, patients were included only if they had at least one follow-up visit after their baseline visit. Improvement definitions were operationalized on the basis of retrospective review of the clinician assessment, patient self-report, and employment or disability status at the time of the last follow-up visit in comparison with the patient’s baseline visit. A composite definition for improvement was used to maximize sensitivity for detection of changes in overall clinical status. “Marked improvement” was defined as those individuals who had restarted work and/or exhibited complete or near-complete symptom resolution. “Mild-to-moderate improvement” was defined as individuals reporting some subjective symptom relief yet remaining unable to work and/or endorsing persistent symptoms that impaired daily functioning. “Markedly worse” was defined as those individuals who developed new functional neurological symptoms (by self-report or examination), stopped working (or were newly on or applying for disability), and/or referred for a higher level of psychiatric care (i.e., partial hospital program). “Mildly worse” were those individuals who reported worse symptoms and/or showed exam-based evidence of increased severity without a change in employment or disability status. “Unchanged” was defined as patients reporting similar symptom severity and/or employment or disability status at follow-up compared to the baseline visit.

Outcome was dichotomized, with individuals meeting criteria for either marked or mild-to-moderate improvement categorized as “improved” and with those meeting criteria for unchanged, mildly worse, or markedly worse categorized as “not improved.” “Complete treatment adherence” referred to adherence with all new treatment recommendations, including PT, OT, and/or psychotherapy. Patients who were already in some form of therapeutic intervention at their baseline evaluation were excluded from treatment adherence analyses. Adherence to either PT/OT or psychotherapy was defined as having attended at least three sessions in each discipline recommended during the 7-month interval of follow-up, independent of the number of sessions scheduled or offered.

For baseline variables, education was dichotomized for college graduation. Given the frequent co-occurrence of sexual and physical trauma across psychopathologies, these variables were treated as one factor. History of suicidality included prior suicidal ideation and/or past suicide attempts. Lifetime depression or lifetime anxiety included any past or current diagnoses for these conditions. Independent variables with less than 20% variance across the 100 FND cohort were excluded from analyses.²³

To identify baseline variables associated with outcome variables, we first performed univariate tests (chi-square or unpaired two-tailed t tests). Variables with a statistically significant relationship with one of the dependent variables (p<0.05) were included in second-level multivariate logistic regression analyses. For the regression analysis examining predictors of having at least one no-show appointment, findings were adjusted for the number of outpatient visits attended by each patient. The percentage of variance of the outcome variable explained by each regression model is presented using the Cox and Snell pseudo R² values. Analyses were performed with IBM SPSS (version 23; Chicago).

Results

Table 1 presents the baseline characteristics of patients included in this study.²³ The cohort included 51 patients with PNES (video-electroencephalogram documented, N=32; clinically established, N=3; probable, N=3; and possible, N=13), 41 individuals with functional weakness (36 with “positive” examination findings, four with suspected paroxysmal symptoms, and one with complete bilateral lower-extremity paralysis with unremarkable brain and spine MRI and nerve conduction studies), and 38 patients with FMD (clinically established, N=37; possible, N=1). Twenty-eight patients had mixed motor functional neurological symptoms.

Ninety-nine patients were included for analyses investigating predictors of not showing up for a follow-up appointment, excluding one patient who was not recommended to follow-up in the clinic. A total of 32 of 99 patients (32%) had at least one no-show appointment. Eighty-one patients were included in analyses of improvement at 7 months, excluding 19 individuals who never returned to the clinic following their initial assessment. In addition to the baseline evaluation, patients had 1.8±1.3 outpatient visits (range=0–6). Of those who followed up, 11 (14%) were rated as markedly improved (two went from unemployed to full-time employment; nine reported complete or near-complete symptom resolution), 23 (28%) mild to moderately improved (all reported some, yet incomplete, symptom relief), 36 (44%) unchanged, five (6%) mildly worse (all reported increased symptoms), and six (7%) markedly worse (four developed new functional symptoms; two stopped working) at 7 months. Fifty patients were newly referred for interventions. As shown in Table 2, patients with PNES, compared to other motor FND subtypes, were more commonly in psychotherapy at their initial visit (56% vs. 31%), while patients with functional movement disorders and/or functional weakness, compared to those with PNES, were more commonly in physical therapy at the time of their baseline visit (37% vs. 4%). During the course of treatment, the self-help CBT workbook by Williams and colleagues was recommended to 50 patients, and the PNES workbooks by LaFrance and colleagues were prescribed to 16 patients.

For predictors of no-show appointments, univariate tests revealed that patients with one or more no-show appointments were less likely to be diagnosed with comorbid generalized anxiety disorder (GAD) (18% vs. 40%; χ²=4.53; p=0.033) and were less commonly taking selective serotonin reuptake inhibitors (SSRIs) or serotonin-norepinephrine reuptake inhibitors (SNRIs) (16% vs. 37%; χ²=4.82; p=0.028). Individuals with functional weakness (63% vs. 31%; χ²=8.66; p=0.003) were more likely to have a no-show visit. Controlling for the number of outpatient visits attended, there were no independent factors associated with having at least one no-show appointment in a single logistic regression analysis. Thirty-one percent of the data variance was explained by this model.

Among 81 patients with at least one follow-up visit, univariate tests showed that a history of a chronic pain disorder (12% vs. 34%; χ²=5.27; p=0.022) was inversely associated with clinical improvement. Because only one clinical factor related to improvement, a second-level multivariate regression analysis was not performed.

Among 50 patients newly referred for interventions, 18 (36%) were fully adherent. In univariate analyses, complete adherence was associated with a higher number of self-reported medication allergies (t=2.27; p=0.028) and clinical improvement at 7 months (67% vs. 22%; χ²=9.8; p=0.002). Having abnormal brain imaging (27% vs. 88%; χ²=16.17; p<0.001) was inversely associated with adherence. See Table S1 in the data supplement accompanying the online version of this article for abnormal brain scan findings. A multivariate logistic regression analysis revealed that clinical improvement (odds ratio [OR]=9.7; 95% confidence interval [CI]=2–46; p=0.004) was positively associated with adherence, whereas having abnormal brain imaging (OR=0.3; 95% CI=0.08–.86; p=0.027) was negatively associated with adherence. This model explained 34% of the outcome variable variance. Complete results are shown in Table 3.

Discussion

This tertiary-care FND clinic cohort study found that approximately one-third of patients did not show for at least one follow-up appointment over a 7-month follow-up period. In univariate analyses only, those with a no-show visit were disproportionately represented by patients with functional weakness; conversely, individuals taking antidepressants or those with GAD were less likely to miss appointments. Of the 81 patients who returned for at least one follow-up visit, 42% reported some degree of clinical improvement at 7 months. In the univariate analyses, individuals with a history of a chronic pain disorder endorsed worse outcomes. Thirty-six percent of patients newly referred for PT, OT, and/or psychotherapy were fully adherent to recommendations. Treatment adherence was independently associated with clinical improvement at 7 months, whereas having abnormal brain imaging was inversely associated with adherence.

Predictors of adherence to outpatient clinical care have been minimally studied in FND.^29,30 We have previously shown that individuals referred to our FND clinic from the emergency department were 10 times less likely to show for an initial outpatient consultation than were individuals referred from other settings.²² In the present study, 32% of patients had at least one no-show occurrence during the course of 7 months of longitudinal care. A few studies in the PNES subtype of FND have investigated outpatient adherence. In a study by Mckenzie and colleagues, 28% of patients with PNES failed to show for their second appointment.²⁹ In a study of initial treatment adherence, colocation of medication management and psychotherapy facilitated treatment adherence in PNES.³⁰ In our cohort, patients on an SSRI/SNRI at baseline or those with GAD were less likely to miss an outpatient follow-up visit. These two factors could represent a willingness by this subgroup to engage in an interdisciplinary treatment emphasizing both medical/neurological and mental health components. Notably, although psychiatric comorbidities have been linked to poor outcome in FND populations,⁷ two studies reported a positive association between comorbid affective symptoms and prognosis,^31,32 suggesting that these relationships are complex and multifaceted. Treatment engagement may also vary by motor FND subtype, with our data suggesting that individuals with functional weakness may be particularly at risk for treatment nonadherence. This observation may relate to gender and psychopathology differences in individuals with functional weakness in comparison with other motor FND subtypes, since we have previously shown that individuals with functional weakness are more likely to be male and report lower rates of past psychiatric hospitalizations.²³

Patients with chronic pain were less likely to report clinical improvement at 7 months (only 12% of individuals with comorbid chronic pain improved, whereas 34% of those without chronic pain improved). In general, chronic pain disorders are associated with negative affective disturbances, substance use, unemployment/disability, maladaptive coping, and other psychosocial difficulties.³³ As with FND, treatment for chronic pain involves multidisciplinary approaches, including psychological (i.e., CBT), somatic (i.e., PT), and pharmacologic interventions.^34–36 CBT is a recommended treatment for chronic pain, although CBT may be only modestly effective on pain and pain-related disability.³⁷ In an observational study, individuals who benefited most from CBT were initially less anxious, had stronger beliefs in personal control, were highly educated, and endorsed lower baseline pain and fatigue.³⁶ Associations between FND and chronic pain are also well established.^38–45 Our data suggest that chronic pain in motor FND may represent a particularly treatment-refractory subgroup worthy of additional prospective clinical research.

In the 50 patients newly referred to outpatient PT/OT and/or psychotherapy following our consultation, clinical improvement was positively associated with treatment adherence. Individuals with abnormal brain imaging were also more likely to report a poor outcome at 7 months, which is consistent with the literature.⁴⁶ Nonspecific white matter lesions, observed in 21 of 43 patients with abnormal brain scan findings, may represent a central nervous system vulnerability for the development of FND symptoms, potentially leading to diminished treatment responsiveness. Positive associations between treatment compliance and clinical improvement support the clinical trials literature identifying therapeutic roles for PT and CBT in FND.^16,18,19 Our observations also support the feasibility of implementing new evidence-based interventions in clinical practice.

Evidence for PT as an integral treatment modality for FND has amassed over the last half century. Early work in functional gait disorders detailed a behavioral approach in which desired patterns of movement were positively reinforced, whereas undesirable movements were de-emphasized.⁴⁷ A systematic review identified 28 studies that evaluated the effect of PT in motor FND.¹⁶ Studies described a wide range of potentially efficacious interventions, including motor reprogramming,⁴⁸ walking,⁴⁹ and strengthening exercises.⁵⁰ Generally, more than 50% of individuals experienced partial to complete improvement, which remained at follow-up. In 2015 an expert panel published consensus recommendations for PT treatments. The authors recommended incorporating four components: 1) education to enhance the patient’s understanding, 2) demonstration that normal movements can occur, 3) retraining movements with diverted attention, and 4) changing maladaptive behaviors exacerbating symptoms.¹⁷ Nielsen and colleagues recently completed a randomized study of PT for functional motor symptoms that included 57 subjects randomly assigned to a 5-day day-hospital intervention program or to treatment as usual.¹⁵ Despite aiming to assess feasibility, a moderate-to-large positive treatment effect was observed, and 72% in the intervention group remained improved at six months.

CBT is an effective treatment for FND.^18–20 Sharpe and colleagues conducted a randomized CBT-based guided self-help study consisting of manualized treatment plus four half-hour sessions with a nurse.¹⁸ When compared to usual care, the treatment group improved on a self-rated global improvement scale, as well as on secondary measures of mood/anxiety, physical function, and symptom burden. The therapeutic benefits as measured by the global improvement scale, however, did not persist at six-month follow-up. Goldstein and colleagues demonstrated that 12 weeks of CBT was effective at reducing seizure frequency in patients with PNES at study completion,²⁰ and LaFrance and colleagues, in a multisite, randomized clinical trial, demonstrated that CBT-informed psychotherapy alone and CBT-informed psychotherapy plus sertraline both showed an approximately 50% reduction in seizure frequency in comparison with baseline in patients with PNES.¹⁹ Additional studies are needed to assess long-term outcomes. CBT treatment goals for FND overlap with those developed for depression and anxiety and include: 1) learning to identify symptom triggers; 2) identifying and exploring unhelpful thoughts, illness beliefs, somatic misinterpretations, and behaviors (e.g., avoidance); 3) enhancing one’s tolerance of negative emotions; and 4) learning coping strategies, including relaxation training. CBT workbooks for FND also emphasize education on physiological mechanisms underlying fear/anxiety responses and FND symptoms in general. Our research and clinical experience from neuropsychiatric clinics that evaluate and treat patients with somatic symptom disorders (S.G., D.P., and W.C.L.), along with the findings of this present study, support the finding that patients who adhere to CBT for FND show favorable outcomes, although the interplay between treatment adherence and specific treatment modalities requires additional study.

Whereas clinical interventions have been studied mainly in individual FND subtypes, our group has demonstrated the utility of a transdiagnostic approach to the assessment^22,23 and investigation⁵¹ of the full-motor FND spectrum. In our cohort, one quarter of patients exhibit mixed-motor FND symptoms, highlighting an inherent overlap across FND subtypes.²³ Additionally, patients often develop other neurological symptoms over the course of their illness, suggesting common disease mechanisms across sensory-motor presentations.⁵² A transdiagnostic approach is feasible in FND research, as well as in clinical practice, especially given the many shared psychosocial profiles, health-related quality-of-life impairments, and predisposing/precipitating factors observed across motor FNDs.^45,53–56 Although this unified, interdisciplinary approach may help streamline clinical and research activities, more subtle distinctions across subtypes are also relevant, including differences across subtypes related to dissociation and personality profiles.^55,57,58

This study has several limitations, including the retrospective design and the lack of utilization of objective functional status measurements (i.e., timed walk test) or standardized self-report measures (i.e., quality-of-life scales). However, outcome measures for FND have yet to be standardized, and existing functional movement disorder measures are not generalizable across semiologies.⁵⁹ Future prospective studies, including those underway at our center and elsewhere, should address these limitations by including prospective studies across motor FND populations, using both self-report and objective outcome measures. Lastly, we did not offer group CBT interventions, which have been shown to be efficacious in FND and may improve the cost-effectiveness of treatment delivery⁶⁰; future studies should also examine clinical factors associated with adherence and treatment outcome in patients with FND who are receiving group-CBT interventions.

In conclusion, we report preliminary evidence demonstrating the feasibility of providing individualized treatments for patients with motor FND by using an updated standard of care. This study suggests that patients with FND and comorbid chronic pain may potentially be an indicator of decreased treatment response.