Outpatient Motor Retraining for Functional Movement Disorder: Predictors of a Favorable Short‐Term Response

ABSTRACT

Background

Treating functional movement disorder (FMD) with motor retraining is effective but resource intensive.

Objectives

Identify patient, disease, and program variables associated with favorable treatment outcomes.

Methods

Retrospective review of the 1 week intensive outpatient FMD program at Mayo Clinic in Minnesota from February 2019 to August 2021. Outcomes included patient‐reported measures (Canadian Occupational Performance Measure‐Performance and Satisfaction subscales [COPM‐P and COPM‐S, range 0–10] and Global Rating of Change [GROC, −7 to +7]) and a retrospective investigator‐rated scale (0–3, worse/not improved to significantly improved/resolved). Linear regression models identified variables predicting favorable outcomes.

Results

Participants (n = 201, 74% female, mean age = 46) had median FMD duration of 24 months. The commonest FMD subtypes were gait disorder (65%), tremor (41%) and weakness (17%); 53% had ≥2 subtypes. Most patients (88%) completed a therapeutic screening process before program entry. Patient‐reported outcomes at the end of the week improved substantially (COPM‐P average change 3.8 ± 1.9; GROC post‐program average 5.5 ± 1.7). Available investigator‐rated outcomes from short‐term follow‐up were also positive (102/122 [84%] moderately to significantly improved/resolved). Factors predicting greater improvement in COPM‐P were completing therapeutic screening, higher number of non‐motor symptoms, shorter FMD duration, earlier program entry, lower baseline COPM scores, and (among screened patients) higher GROC between therapeutic screening and program start.

Conclusion

Patients with diverse FMD subtypes improved substantially over a 1 week period. Utilization of therapeutic screening and greater improvement between therapeutic screening and program start were novel predictors of favorable outcomes. Non‐motor symptoms did not preclude positive responses, although patients with predominant non‐motor burden were excluded.

For most of the 20th century, patients with symptoms resembling neurologic disease, without discernible structural damage to the nervous system, were categorized as having purely psychiatric conditions. ¹ Terms including conversion, dissociative, and psychogenic disorders reinforced the notion that these symptoms had no pathophysiological basis in the nervous system. In recent years, however, growing evidence of abnormal brain activity and connectivity and emerging treatment strategies have renewed interest in this condition. ² The term functional neurological disorder (FND) recaptures a concept from the early 19th century that positively identifiable clinical neurological syndromes can arise from changes in functioning of the nervous system apart from structural lesions. ³ The modern reconceptualization of FND further posits that alterations in functioning constitute specific pathophysiologic mechanisms that may operate independently of, or in addition to, deficits in structural integrity and psychopathological processes. ⁴ The subtype of FND that involves movement abnormalities, commonly called functional movement disorder (FMD), includes both hyperkinetic (functional tremor, jerks/myoclonus, dystonia, ataxia) and hypokinetic (weakness, slowing/parkinsonism, posturing) movements.

While research continues to elucidate FMD pathophysiology, effective treatment strategies are being implemented. Motor retraining involves specialized physical and occupational therapy focused on disrupting maladaptive movement patterns and rehabilitating normal patterns. ⁵ , ⁶ , ⁷ , ⁸ Since 2005, the Behavioral Shaping Therapy (BeST) program at the Mayo Clinic in Minnesota has applied a 1 week intensive outpatient motor retraining approach to FMD. Results from an early cohort demonstrated that 44/60 (73%) patients were markedly improved or remitted at completion, and 29/48 (60%) who were re‐evaluated at a median of 25 months after completion of the program had sustained improvement. ⁵ A 2013 systematic review of physiotherapy for FMD found positive results in all eligible studies, with the largest studies reporting long‐term improvement in 58–84% of patients. ⁹ Subsequent cohort studies reported improvement in >80% of patients, ⁶ , ¹⁰ , ¹¹ , ¹² including in the pediatric population ¹³ and inpatient setting. ¹⁴ Randomized controlled trials of functional gait disorder and diverse FMD subtypes have demonstrated robust and cost‐effective benefits. ¹⁵ , ¹⁶

Despite these documented benefits, motor retraining programs are time‐ and resource‐intensive, and do not serve all patients well. Pre‐treatment predictors of favorable responses drawn from relatively small studies (n < 70) include female sex, ⁵ , ⁶ shorter duration of motor symptoms, ¹⁴ , ¹⁷ , ¹⁸ and willingness to consider symptoms in relation to stressors or emotional state. ¹⁹ Within‐treatment positive predictors include concurrent improvement in psychiatric symptoms and social functioning. ¹¹ We sought to improve patient selection and treatment processes for intensive motor retraining by identifying patient, disease, and program characteristics that might predict superior outcomes by analyzing a study cohort three‐fold larger than previous investigations.

Methods

The Mayo Clinic institutional review board (IRB) approved this retrospective review (IRB #19–009967).

Patient Selection

Outcomes data were collected on a consecutive cohort of patients who completed the BeST program at Mayo Clinic in Rochester, Minnesota between February 2019 and August 2021. All patients underwent comprehensive, multidisciplinary evaluations in the Mayo Clinic Departments of Neurology, Physical Medicine and Rehabilitation (PMR), and, if needed, Psychiatry and Psychology, to confirm the diagnosis of FMD in accordance with Gupta‐Lang criteria. ²⁰ Most patients completed therapeutic screening as described below. A minority of patients were enrolled directly into the BeST program without therapeutic screening, typically because the referring movement disorders neurologist or PMR physician considered them to be “good” candidates with few non‐motor symptoms and ready acceptance of the FMD diagnosis.

Therapeutic screening consisted of consultations with a PMR physician, physical therapist (PT), and/or occupational therapist (OT). The objectives were to: (1) Ensure patient understanding and adequate acceptance of the diagnosis; (2) Verify that movements were continuous, frequent, or easily triggerable, as infrequent non‐triggerable movements (eg, FND‐seizures) are not amenable to scheduled motor retraining; (3) Screen for predominant chronic pain, dizziness, fatigue, and other medical, psychiatric, or neurological conditions severe enough to limit intensive motor retraining; (4) Demonstrate and practice initial rehabilitation techniques including diaphragmatic breathing exercises and gradual escalation of high‐quality movements; and (5) Provide educational resources for use prior to starting the program (details in Supplementary 1).

Among 476 patients who underwent therapeutic screening during the time period most relevant to the study cohort (October 2018–July 2021), 277 (58%) were classified as Eligible for the BeST program, 51 (11%) as Conditionally Eligible, and 148 (31%) as Not Eligible. Quantified reasons for each determination are presented in Fig. 1. Of the Eligible patients, 213/277 (77%) completed the BeST program. Among the Eligible patients who did not enroll, 12/64 (20%) improved substantially with the techniques demonstrated during therapeutic screening, to the point that they did not require intensive therapy, while others elected for treatment elsewhere (27%) or faced geographic or financial barriers (11%). Conditionally Eligible patients with potentially remediable problems were referred to appropriate treatment programs/clinicians (including comprehensive pain rehabilitation [51%], standard physical therapy for improving activity tolerance [29%], and vestibular therapy for chronic dizziness [12%]) and invited to contact program staff to report their progress. Twenty of 51 (39%) subsequently completed the BeST program; for the remainder, documentation was not available to determine how many were unable to resolve these problems versus how many had improvement in FMD symptoms during their alternative treatments to the point that they no longer required the BeST program. Not Eligible patients included those with FND‐seizures without FMD (26%), mild FMD symptoms not warranting intensive rehabilitation (10%), rejection of the FMD diagnosis (16%), and predominant chronic pain (34%), fatigue (24%), or comorbidities severe enough to preclude intensive rehabilitation, and were redirected to appropriate treatment programs. Twenty‐nine patients were enrolled directly without therapeutic screening. Exclusion criteria for analysis included declining use of the medical record for research purposes, repeating the BeST program after prior completion, FND‐seizures only without FMD, missing primary outcome data, and completing the program before or after the consecutive cohort. Outcomes data were abstracted from program and electronic medical records for the final cohort (Fig. 1).

FIG. 1.

CONSORT flow diagram for patients included in analysis. *Early Benefit refers to patients who had substantial benefit from the therapeutic screening process, to the point that intensive therapy was no longer needed. **These patients had FND‐seizures without FMD, but completed the BeST program (two via screening as eligible, one direct enrollment without screening) due to events being frequent and/or consistently triggerable, making them amenable to scheduled motor retraining. BeST, Behavioral Shaping Therapy motor retraining program; COPM‐P, Canadian Occupational Performance Measure–Performance rating; Dx, diagnosis; FMD, functional movement disorder; FND, functional neurological disorder; Tx, treatment.

BeST Program

The BeST program consists of 5 consecutive days of intensive outpatient physical and occupational therapy (3–4 1 h sessions daily) focused on deconstructing abnormal movements/postures and retraining normal automatic movement, in line with consensus FMD treatment recommendations ⁷ , ⁸ and refined over 15+ years of program experience. A typical approach progresses through several stages: (1) Education in basic relaxation exercises including diaphragmatic breathing and progressive muscle relaxation; (2) Guidance through small, controlled, high quality movements, beginning as simply as moving one joint or muscle; (3) Gradual addition of more complex movements, building to common natural movements such as walking; (4) Instruction in steps to detect abnormal movements, stop them before they escalate, and restart using effective symptom management strategies, rather than fighting or “pushing through” symptoms. Therapists model calm and controlled reactions to intense/dramatic symptoms. (5) Once normal motor control is re‐established, cognitive and environmental challenges are introduced to promote carryover into real world scenarios; (6) Physician follow‐up is recommended after program completion.

Study Variables

Demographic and clinical data that were collected during the program included age, sex, age at onset, and duration of symptoms. FMD subtype(s) and comorbidities were retrospectively abstracted from the medical record. FMD subtypes included functional gait disorder, tremor, jerks/myoclonus, dystonia, weakness, ataxia, parkinsonism, speech disorder, and FMD not otherwise specified (NOS, including chorea, ballism, writhing, etc.). These were distinguished from FND‐seizures, which manifest as paroxysmal alterations in awareness/cognition with or without sensorimotor dysfunction.

A concise battery of assessments (Table 1) was used during the BeST program to evaluate patients’ clinical status. Baseline assessments collected at program start included: (1) Presence of specific motor and non‐motor symptoms, (2) Presence of spells, and (3) Sheehan Disability Scale (SDS). Patient‐reported outcome measures (PROMs) rated at program start and end included: (1) Canadian Occupational Performance Measure ratings of Performance and Satisfaction (COPM‐P, COPM‐S), (2) Global Rating of Change from screening to program start (GROC‐screen) and program end (GROC‐post), and (3) Patient‐Specific Functional Scale (PSFS). For disorders like FMD that affect physical, psychological, and social well‐being, the use of PROMs ensures a primary focus on patients’ perspectives about symptom severity, performance limitations, disability, and therapeutic outcomes. ²¹ Patient satisfaction was evaluated at program end using the Satisfaction portion of the Focus On Therapeutic Outcomes (FOTO) assessment.

TABLE 1.

Best program assessments

Assessment	Range	Description
Motor symptoms*	0–6	Presence or absence of (1) Gait impairment, (2) Tremors, (3) Involuntary movements, (4) Arm weakness, (5) Leg weakness, and (6) Speech impairment, at program start.
Non‐motor symptoms*	0–6	Presence or absence of (1) Numbness/tingling, (2) Dizziness, (3) Memory/concentration difficulty, (4) Headache, (5) Pain, and (6) Fatigue, at program start.
Sheehan disability scale (SDS)	0–10	Patient‐rated impairment in work (if applicable), home, and social settings, at program start, with higher scores indicating greater disability. Ratings across settings were averaged.
Canadian occupational performance measure (COPM)	0–10	Through a semi‐structured interview with OT, patients identify up to 5 important activities and rate ability to perform (COPM‐P) and satisfaction (COPM‐S), with lower scores indicating worse performance/satisfaction, at program start and end for the same activities. Activity ratings were averaged and change from program start to end was calculated.
Global rating of change (GROC)	−7 to +7	Patient‐rated overall outcome from very much worse (−7) to very much improved (+7), from the time of therapeutic screening (if applicable) to the start of the program (GROC‐screen), as well as at program end (GROC‐post).
Patient specific functional scale (PSFS)	0–30	Patients independently choose 3 activities and rate their ability to perform, each from 0 to 10, with lower scores indicating greater impairment, at program start and end for the same activities. Activity ratings were summed and change from program start to end was calculated.
Satisfaction		Patient rated using the 8‐item questionnaire from the Focus On Therapeutic Outcomes (FOTO) assessment (FOTO ratings of baseline, predicted, and final outcomes were not collected).
Investigator rating of outcome	0–3	Overall outcome rated via retrospective review of post‐program neurology and/or PMR clinician follow‐up visit notes: 0No improvement/Worse than before the program. 1Mildly improved but with ongoing disruptive FMD symptoms. 2Moderately improved with manageable residual FMD symptoms. 3Significantly improved/completely resolved with minimal/no residual FMD symptoms.

^*These motor and non‐motor symptom lists were selected for data collection based on previous observations of common patient‐reported symptoms in the BeST program.

To provide a complementary measure of overall treatment response, two investigators (MC, SB) independently rated treatment outcomes, blinded to PROMs, for all patients who received follow‐up care in our Neurology and/or PMR Clinics (n = 122/201, 61%). This was based on retrospective reviews of motor and non‐motor symptoms and signs and overall functional status as documented in post‐BeST clinic notes using a 0–3 scale (worse/not improved to significantly improved/completely resolved, see Table 1) as previously reported from our program. ⁵ Where raters disagreed (n = 37/122), a third investigator (MM) provided an independent, blinded review.

Statistical Analysis

Continuous variables were expressed using the mean and standard deviation when normally distributed, and median and interquartile range otherwise. Change from program start to end in average COPM‐P, average COPM‐S, and summed PSFS were calculated. Pearson's correlation coefficient was used to estimate association between number of FMD subtypes and number of motor symptoms, and between numbers of motor and non‐motor symptoms. Kendall's τ was used to estimate association between COMP‐P and investigator rating of outcome.

To test potential predictors of treatment results, the change in COPM‐P score was selected as the primary outcome variable. The COPM has high reliability, validity, responsiveness to change, and utility in various clinical settings and populations, although further high‐quality validation has been recommended. ²² , ²³ The minimal clinically important difference is 2.0. ²⁴

A multivariable linear regression model was constructed using change in COPM‐P as the dependent variable, with patient age, sex, duration of symptoms, delay from enrollment to program start, therapeutic screening completion, numbers of motor and non‐motor symptoms, baseline COPM‐P and COPM‐S scores, and averaged SDS as candidate predictor variables. Duration of symptoms and delay from enrollment to program start were converted to log scale due to right‐skewness. Predictor variable selection was done using LASSO. The optimal penalty term was selected using cross‐validation.

A subgroup analysis was conducted for patients who underwent therapeutic screening using a similar multivariable linear regression model, with GROC‐screen as an additional candidate predictor variable.

Results

The cohort of 201 program completers (74% female, mean age 46 years, median duration of FMD 24 months) is described in Table 2. The most common FMD subtypes were gait disorder (n = 130, 65%), tremor (n = 82, 41%) and weakness (n = 34, 17%); 106 (52%) had ≥2 subtypes. Pearson's correlation coefficient suggested a positive linear relationship between the number of FMD subtypes and the number of motor symptoms (0.25, CI 0.11–0.37), and between the number of motor symptoms and the number of non‐motor symptoms (0.50, CI 0.38–0.59).

TABLE 2.

Demographics and clinical characteristics (n = 201)

Sex (Female)	148 (74%)
Age during program (years)	45 (16)
Duration of symptoms (months)	24 (12–54)
Comorbidities: Neurological	128 (64%)
Comorbidities: Psychological	127 (63%)
Comorbidities: Medical	148 (74%)
SDS average score	7.0 (2.2)
Movement neurophysiology study	85 (42%)
Psychiatry consultation	101 (50%)
Therapeutic screening completed	176 (88%)
Delay to program start* (months)	2.2 (1.0–4.8)
Spells (patient‐reported)	37 (18%)
Comorbid FND‐Seizures (chart review)	21 (10%)
Motor symptoms (patient‐reported)
Gait difficulty	160 (80%)
Involuntary movements	124 (62%)
Weakness: Lower extremities	125 (62%)
Weakness: Upper extremities	116 (58%)
Tremors	120 (60%)
Speech difficulty	76 (38%)
Average number of motor symptoms	3.6 (1.6)
Non‐motor symptoms (patient‐reported)
Fatigue	134 (66%)
Memory/Concentration difficulty	113 (56%)
Numbness/Tingling	106 (53%)
Pain	97 (48%)
Headaches	88 (44%)
Dizziness	80 (40%)
Average number of non‐motor symptoms	3.0 (1.9)
FMD subtype (chart review)
Functional gait disorder	130 (65%)
Functional tremor	82 (41%)
Functional weakness	34 (17%)
Functional Jerks/Myoclonus	21 (10%)
Functional Dystonia/Rigidity	20 (10%)
Functional Ataxia/Incoordination	3 (2%)
Functional Parkinsonism	3 (2%)
Functional speech disorder	28 (14%)
FMD NOS	26 (13%)
Average # of FMD subtypes	1.7 (0.8)

Note: Data are given as mean (standard deviation) or absolute number of patients (percentage of cohort), except duration of symptoms and delay from enrollment to program start are given as median (interquartile range) due to right‐skewedness.

^*Delay to program start: measured from the date of therapeutic screening if completed, or from the neurology/PMR consultation if the patient was directly enrolled without therapeutic screening.

Abbreviations: FMD, functional movement disorder; NOS, not otherwise specified; SD, standard deviation; SDS, Sheehan Disability Scale (patient‐rated from 0 to 10 in up to 3 domains and averaged, higher scores indicate greater disability).

The most common neurologic comorbidities on chart review were migraine and headache disorders (n = 62, 31%), peripheral neuropathy (n = 14, 7%), vestibular disorders (n = 13, 7%), and concussion (n = 12, 6%). The most common psychiatric comorbidities were anxiety (n = 95, 47%), depressive (n = 74, 37%), and post‐traumatic stress (n = 24, 12%) disorders. The most common medical comorbidities were back pain/deformity or arthritis (n = 46, 23%), fibromyalgia or chronic fatigue syndrome (n = 45, 22%), and sleep disorders including insomnia and obstructive sleep apnea (n = 32, 16%). Only 5 (2%) patients had no comorbid conditions. Median delay from initial neurologic consultation at our institution to BeST program start was 3.1 months (IQR 1.3–6.1) for screened patients vs 2.8 months (IQR 1.4–4.8) for non‐screened patients. Program completers attended 91% of scheduled BeST sessions. Nine patients (4%) completed <70% of scheduled sessions, most often because of rapid improvement leading to therapists’ recommendations for early discharge.

The GROC‐screen was available for 152/176 (86%) screened patients. Mean improvement was 1.2 (SD = 3.0). Twenty‐four of 152 (16%) screened patients reported GROC‐screen from −1 to −7, indicating worsening from screening to program start.

The PROMs are summarized in Table 3. The minimal clinically important difference in COPM‐P of ≥2.0 points was achieved by 161/201 (80%) patients. Only 1 (0.5%) patient had COPM‐P change <0, indicating worse activity performance at program end. The linear regression model for COPM‐P mean change for the full cohort is summarized in Table 4. Using LASSO, the following variables were identified as predictors: therapeutic screening completion, number of non‐motor symptoms, duration of symptoms, delay from enrollment to program start, and baseline COPM‐P and COPM‐S. COPM‐P improvement increased with completion of therapeutic screening, higher number of non‐motor symptoms (Fig. 2), shorter duration of symptoms, shorter delay from enrollment to program start, and worse baseline COPM‐P and COPM‐S. Patient age, sex, number of motor symptoms, and averaged SDS were not selected by LASSO for inclusion in the model.

TABLE 3.

Patient‐rated outcome measures (PROMs)

	Baseline	Final	Change in score
Screening outcomes (n = 152)
GROC‐screen	1.1 (3.0)
Program outcomes (n = 201)
GROC‐post		5.5 (1.7)
COPM‐P	3.0 (1.2)	6.8 (1.9)	+3.8 (1.9)
COPM‐S	2.0 (1.1)	7.2 (2.3)	+5.2 (2.5)
PSFS	7.6 (4.8)	18.9 (6.9)	+11.2 (7.0)
Satisfaction		97.1% (5.9)

Note: Data are given as mean (standard deviation). Bold value indicates the primary outcome for our retrospective analysis.

Abbreviations: COPM‐P/COPM‐S, Canadian Occupational Performance Measure rating of performance/satisfaction (patient‐rated from 0 to 10 on each of up to 5 activities and averaged); GROC, Global Rating of Change (patient‐rated from +7 to −7); PSFS, Patient Specific Functional Scale (patient‐rated from 0 to 10 on each of 3 activities and summed).

TABLE 4.

Linear regression model of factors which predict mean change in COPM‐P

Predictor variable	Beta	95% CI	P‐value
Number of non‐motor symptoms	0.12	−0.01, 0.25	0.071
Therapeutic screening completed	0.44	−0.29, 1.2	0.2
log(Delay to program start)	−0.33	−0.55, −0.12	0.003
log(Duration of symptoms)	−0.29	−0.51, −0.08	0.008
Baseline COPM‐P	−0.46	−0.73, −0.19	<0.001
Baseline COPM‐S	−0.11	−0.38, 0.17	0.4

Note: Positive beta indicates positive correlation, negative beta indicates inverse correlation.

Abbreviations: CI, confidence interval; COPM, Canadian Occupational Performance Measure; COPM‐P, COPM activity Performance average rating (patient‐rated from 0 to 10); COPM‐S, COPM activity Satisfaction average rating (patient‐rated from 0 to 10).

FIG. 2.

Mean change in activity performance (COPM‐P) as a function of number of non‐motor symptoms. As number of non‐motor symptoms increased, mean change in COPM‐P also increased.

For the subgroup of 152 patients who completed therapeutic screening, LASSO identified GROC‐screen, number of motor symptoms, and age for inclusion in the linear regression model, in addition to the same variables identified for the full cohort model. Higher GROC‐screen predicted greater COPM‐P improvement (beta = 0.11, 95% confidence interval = 0.02, 0.21, P = 0.018). Higher number of motor symptoms and age predicted lower COPM‐P improvement, both with non‐significant P‐values (0.20 and 0.40 respectively). Other predictor variables had similar effect sizes and directions as in the full cohort model (Supplementary 2).

Over half of patients (n = 122, 61%) had a post‐program follow‐up visit with a neurologist or PMR physician. This most commonly occurred on the final day of the program, though 21 (17%) had follow‐up visits >30 days after program completion. Investigator outcome ratings are reported in Table 5; 102 (84%) had moderate improvement to complete resolution of symptoms (rating 2–3), 20 (16%) had mild to no improvement (rating 0–1). Average COPM‐P change did not significantly differ between patients with follow‐up and those without (3.7 vs. 4.0, P = 0.28). Patient and investigator‐rated outcomes were significantly associated (Kendall's τ = 0.37, P < 0.001, Fig. 3).

TABLE 5.

Investigator rating of outcome (n = 122)

3‐Complete/Significant improvement, No/Minimal FMD symptoms	66 (54%)
2‐Moderate improvement, manageable FMD symptoms	36 (30%)
1‐Mild improvement, disruptive FMD symptoms	17 (14%)
0‐No Improvement/Worse	3 (3%)

Abbreviation: FMD, functional movement disorder.

FIG. 3.

Correlation between patient‐reported COPM‐P and investigator rating of outcome. COPM‐P (scale range 0–10, higher values indicating better activity performance) was evaluated at program start and end, resulting in COPM‐P mean change range from −10 to +10. Investigator rating of outcome (range 0–3, higher values indicating improvement/resolution of FMD symptoms) was based on retrospective chart review of clinician post‐program notes. Boxes represent Q1–Q3 interquartile range, line within each box represents median (Q2), whiskers extending out from boxes represent minimum and maximum (Q0 and Q4), except for outliers represented as dots. Positive correlation indicates that patients reporting greater improvement in COPM‐P were independently rated by investigators as having greater improvement in FMD symptoms, supporting the use of COPM‐P as the primary study outcome.

Discussion

More than 4 of 5 patients in this large cohort with diverse FMD subtypes benefitted substantially from motor retraining on patient‐reported (80% COPM‐P change ≥ = 2.0) and investigator‐rated (84% moderately to significantly improved/resolved) outcomes. Only one patient reported worsening symptoms by COPM‐P at program end. Patient satisfaction was high at mean 97%, and the program was well tolerated with high levels of session completion.

The COPM has been utilized in other FMD treatment programs. ¹¹ , ¹² It allows patients to select individualized activity goals, which is valuable in a condition like FMD with extraordinary variability in characteristics and timing. Furthermore, it has wide familiarity among physical and occupational therapists. ²³ Recommendations for development of FND outcome measures that capture the full burden of symptoms emphasize the need for patient self‐reports, given that clinician ratings are typically limited to brief observations and do not necessarily correlate strongly with patient self‐assessments. ²⁵ , ²⁶ , ²⁷ Additionally, patient perception of illness is a key factor in the pathophysiology of FMD. ⁴ Taken together, these factors suggest an essential role for PROMs in assessing and reporting outcomes of treatment for patients with FMD. In this study, the robust correlation between the patient‐rated COPM‐P and retrospective investigator‐rated outcomes (Kendall's τ = 0.37, P < 0.001) lent additional credence to using the COPM‐P as the primary outcome measure.

Our most clinically useful and novel finding was that patients who completed a therapeutic screening prior to the FMD program had better COPM‐P outcomes. This was especially seen in those who reported improvement between therapeutic screening and the start of the 1 week program (GROC‐screen). Beyond screening for eligibility, the therapeutic screening process educated patients about the treatment program and walked them through initial techniques. Patients who embraced a functional formulation of their symptoms and started practicing basic treatment strategies may have been more prepared to apply these techniques during the program. In addition to benefits demonstrated by patients who entered the program, 12 patients from the total of 476 (3%) who were screened improved to such an extent from the therapeutic screening process that they did not require intensive rehabilitation. If validated, therapeutic screening could be incorporated into FMD treatment programs broadly to improve efficiency and outcomes. Furthermore, these results suggest that patients’ initial engagement and treatment responsiveness could be used to triage individuals most likely to benefit for rapid entry into available motor retraining programs, thereby minimizing delay to entry, which was found to be predictive of program success. This is not meant to exclude other patients from motor retraining; those not selected for rapid entry may require more preliminary work to prepare them for success, and others who are unimpressed by their initial exposure to the program's concepts and techniques may voluntarily decline to participate or seek other avenues of care.

Comparisons between screened and non‐screened patients have some limitations in this cohort. Patients were not randomly allocated to a therapeutic screening visit. Moreover, patients with a higher burden of non‐motor symptoms were more likely to be screened to ensure eligibility. However, it is possible that some non‐screened patients might have been found ineligible if they had been screened.

Our finding that the number of motor symptoms did not predict program outcome in the full cohort may be interpreted as an encouraging sign that patients with complex FMD involving multiple motor symptoms are just as likely to benefit from motor retraining as patients with single motor symptoms. Furthermore, patients with a range of FMD subtypes derived substantial benefit. Patient age, sex, and baseline Sheehan disability Score also did not predict program outcome in the full cohort, indicating that the principles and techniques of motor retraining are broadly applicable.

The finding that a higher number of non‐motor symptoms predicted larger improvement in COPM‐P is intriguing, but must be interpreted with caution. Most patients were screened specifically to exclude those with predominant burden of pain, fatigue, or other non‐motor symptoms. If these were severe enough to interfere with intensive motor retraining, then patients were redirected to other resources, including specialized programs focused on chronic pain rehabilitation, fibromyalgia, or persistent postural perceptual dizziness. Thus, a proper interpretation of this result is that patients with a moderate level of non‐motor symptoms can benefit from motor retraining. It is possible that motor retraining may have helped patients understand their motor symptoms in the context of their complex neuropsychiatric presentations. This would be consistent with a growing literature that FMD involves brain networks for voluntary movement being disrupted by emotional and threat processing networks, ² and warrants further investigation. Other interpretations of the correlation between non‐motor symptoms and program benefit are possible; for example, relaxation techniques taught in the program may directly benefit non‐motor symptoms, making these symptoms an additional “target” for program interventions. It is therefore incorrect to presume that the ideal candidate for motor retraining should have absent or minimal non‐motor symptoms. Future investigations should quantify the severity of non‐motor symptoms to investigate for correlation of improvement in motor function and non‐motor symptoms. ²⁸

Our finding that longer duration of symptoms predicts a less favorable outcome is consistent with existing literature. ¹⁷ , ¹⁸ , ²⁹ We also found that greater delay from multidisciplinary evaluation to program start was unfavorable. Delays may reflect multiple factors such as geographic or financial barriers, COVID‐related restrictions during the study period, or patient motivation, in addition to scheduling logistics. Given the demonstrated value of therapeutic screening, it is important that this not significantly delay entry into a full treatment program. Fortunately, in our program, screening was efficiently scheduled and time from initial neurology consultation to FMD program start was similar for screened versus unscreened patients (median 3.1 vs 2.8 months, respectively).

The correlation of high baseline COPM‐P and COMP‐S with lower mean change in these scales may be in part due to central tendency bias; that is, individuals tend to select items in the middle of a scale more often than extremes. ³⁰ However, the ceiling effect also may have played a role given that patients with higher baseline levels of functioning had less room to improve than those with lower baseline scores. This may encourage clinicians to target patients with greater impairment for treatment, as they may achieve relatively greater benefits, as long as they are able to tolerate intensive therapy.

Limitations of this study included a program‐generated checklist of motor and non‐motor symptoms which did not assess details or severity of symptoms, the absence of standardized instruments for psychological symptoms, the lack of clinician‐rated measures completed during the program, and the lack of long‐term follow‐up. While a prior study demonstrated long‐term benefit from the BeST program for a median of 2 years, ⁵ the risk of FMD relapse cannot be ignored. Predictors of short‐term response to therapy would be less valuable if they cannot predict long‐term outcomes, and should be validated in a cohort with longer follow‐up. Our list of predictive factors was by no means complete, and more detailed characterization, including psychological variables such as personality traits, locus of control, agency, coping style, as well as social determinants of health, may uncover other predictive factors for treatment success. Our cohort was limited to a single academic medical center, with program funding coming primarily from patients’ commercial insurance plans, Medicare, or self‐pay (estimated cost $5500–$6500 USD for the program). Results may not be generalizable to programs funded through other sources. Nevertheless, this is the largest patient cohort reported from an established FMD program with standardized PROMs, which allowed us to explore the range of patient, disease, and program variables described in this report. We hope to use these preliminary findings to generate hypotheses for prospective trials to further investigate the effects of motor and non‐motor symptoms, therapeutic screening, and other factors on outcomes from motor retraining for FMD.

Conclusion

Motor retraining represents a powerful intervention for treating FMD. Selecting patients who can tolerate intensive motor retraining and providing initial education and treatment strategies via therapeutic screening may improve efficiency and treatment outcomes. If non‐motor symptoms are not severe enough to disrupt motor retraining, they may not be a disadvantage for treatment response. Earlier enrollment in treatment is important to optimize recovery. We look forward to ongoing ³¹ and future investigations that will illuminate the pathophysiologic mechanisms and optimal treatment approaches for FMD.

Author Roles

(1) Research project: A. Conception, B. Organization, C. Execution; (2) Statistical Analysis: A. Design, B. Execution, C. Review and Critique; (3) Manuscript Preparation: A. Writing of the first draft, B. Review and Critique.

M.C.: 1A, 1B, 1C, 2A, 2C, 3A, 3B

M.K.: 1B, 1C, 2A, 2B, 2C, 3B

S.B.: 1B, 1C, 2C, 3B

M.M.: 1A, 1B, 1C, 2C, 3B

E.G.: 1A, 1B, 1C, 3B

K.T.: 1A, 1B, 1C, 3B

J.S.: 2C, 3B

A.H.: 1A, 1B, 1C, 2A, 2C, 3A, 3B

Disclosures

Ethical Compliance Statement: The Mayo Clinic Institutional Review Board approved this study (IRB no. 19–009967). Patients who declined use of their medical record for research were excluded from this study. Informed patient consent was not necessary for this work. We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this work is consistent with those guidelines.

Funding Sources and Conflicts of Interest: Mayo Clinic Graduate Medical Education funded the statistical analysis for this project. No other specific funding was received for this work. The authors declare there are no conflicts of interest relevant to this work.

Financial Disclosures for the Previous 12 Months: JS reports unrelated disclosures include grant funding from the US Army Medical Research and Development Command, grant number W81XWH1810760, Sleep Number Corporation and Itamar Medical, Ltd. The other authors declare no additional disclosures to report.

Supporting information

Supplementary 1. Therapeutic Screening process description. Detailed description of the therapeutic screening process involving physical medicine and rehabilitation, physical therapy, and occupational therapy.

Supplementary 2. Screened Subgroup Linear Regression Model. Linear regression model for the subgroup of patients who underwent therapeutic screening, constructed using the same candidate predictor variables as the full cohort linear regression.