CC BY-NC-ND
0AC RM Archives of Physical Medicine and Rehabilitation
AMERICAN CONGRESS OF REHABILITATION MEDICINE
journal homepage: www.archives-pmr.org Archives of Physical Medicine and Rehabilitation 2015;■: I
ORIGINAL ARTICLE
Comparison of Robotics, Functional Electrical Stimulation, and Motor Learning Methods for Treatment of Persistent Upper Extremity Dysfunction After Stroke: A Randomized Controlled Trial
Jessica McCabe, MPT,a Michelle Monkiewicz, DPT,a John Holcomb, PhD,b Svetlana Pundik, MD, MS,a Janis J. Daly, PhD, MSa
From the aStroke Motor Control/Motor Learning Laboratory, Louis Stokes Cleveland Department of Veterans Affairs Medical Center, Cleveland, OH; and bDepartment of Mathematics and Statistics, Cleveland State University, Cleveland, OH.
Current affiliations for Daly, Brain Rehabilitation Research Center of Excellence, Malcom Randall Gainesville Department of Veterans Affairs Medical Center, Gainesville, FL; Department of Neurology, College of Medicine, University of Florida, FL; and Brain Rehabilitation Research Program, McKnight Brain Institute, University of Florida, Gainesville FL.
Abstract
Objective: To compare response to upper-limb treatment using robotics plus motor learning (ML) versus functional electrical stimulation (FES) plus ML versus ML alone, according to a measure of complex functional everyday tasks for chronic, severely impaired stroke survivors. Design: Single-blind, randomized trial. Setting: Medical center.
Participants: Enrolled subjects (N= 39) were >1 year postsingle stroke (attrition rate= 10%; 35 completed the study).
Interventions: All groups received treatment 5d/wk for 5h/d (60 sessions), with unique treatment as follows: ML alone (n = 11) (5h/d partial- and whole-task practice of complex functional tasks), robotics plus ML (n= 12) (3.5h/d of ML and 1.5h/d of shoulder/elbow robotics), and FES plus ML (n= 12) (3.5h/d of ML and 1.5h/d of FES wrist/hand coordination training).
Main Outcome Measures: Primary measure: Arm Motor Ability Test (AMAT), with 13 complex functional tasks; secondary measure: upper-limb Fugl-Meyer coordination scale (FM).
Results: There was no significant difference found in treatment response across groups (AMAT: P>.584; FM coordination: P>.590). All 3 treatment groups demonstrated clinically and statistically significant improvement in response to treatment (AMAT and FM coordination: P<.009). A group treatment paradigm of 1:3 (therapist/patient) ratio proved feasible for provision of the intensive treatment. No adverse effects. Conclusions: Severely impaired stroke survivors with persistent (>1y) upper-extremity dysfunction can make clinically and statistically significant gains in coordination and functional task performance in response to robotics plus ML, FES plus ML, and ML alone in an intensive and long-duration intervention; no group differences were found. Additional studies are warranted to determine the effectiveness of these methods in the clinical setting. Archives of Physical Medicine and Rehabilitation 2015; ■ :■■■-■■■ © 2015 by the American Congress of Rehabilitation Medicine
Treatment methods using motor learning (ML) principles for the treatment of persistent upper-limb dysfunction after stroke have been reported in the literature.1-11 Some have compared the application of ML principles with neurorehabilitation methods
Supported by the Department of Veterans Affairs (grant nos. B2801R, B9024-S, and B5080S). Clinical Trial Registration No.: NCT01725659. Disclosures: none.
(eg, Bobath concept, neurodevelopmental treatment).1'2 Still, others have used bilateral upper-extremity exercise3 or constraint-induced motor therapies for mild/moderate upper-extremity dysfunction.4-11 Most of these studies showed promising results, but stroke survivors did not recover normative function, and gains were statistically significant but small.
In addition to ML-based treatment strategies, technology-based upper-limb therapies (eg, robotics training, functional electrical
0003-9993/14/$36 - see front matter © 2015 by the American Congress of Rehabilitation Medicine http://dx.doi.org/10.1016/j.apmr.2014.10.022
stimulation [FES]) have also produced some positive results. Treatment with robotics has shown statistically significant gains in measures of impairment for chronic stroke survivors,12-17 but some reported gains were not considered clinically significant according to an outcome measure of coordination.18 At the same time, surface FES was reported beneficial with chronic stroke subjects according to measures of impairment.19'20 Although robotics and FES have each shown promise, there is a paucity of information regarding the comparative benefit. Additionally, there is little evidence of whether or not a treatment paradigm using a combination of ML therapy and technology-based therapy would be superior to ML alone, according to a homogeneous measure of the performance of actual complex functional tasks of everyday life; rather, a number of reported outcome measures contain a mixture of impairment items and functional task items. Finally, many studies have focused on mildly to moderately impaired stroke survivors, with significantly less attention paid to the severely impaired (<36 points on the upper-limb motor Fugl-Meyer [FM] score21). Therefore, in consideration of all these issues together, the purpose of this study was to investigate, for severely impaired, chronic stroke survivors, the comparative response to treatment using shoulder/elbow robotics plus ML versus wrist/hand FES plus ML versus ML alone according to a measure of actual complex functional tasks of everyday life.
Methods
Study design
This was a randomized controlled trial comparing response to treatment across 3 different treatment groups: robotics plus ML, FES plus ML, and ML. Subjects in all 3 groups received treatment for 5h/d for 5d/wk for 12 weeks (60 treatment visits). Measures were acquired at pre- and posttreatment.
Participants
There were 174 phone inquiries regarding the study. Of these, 135 did not meet criteria for an in-person screen (fig 1). Thirty-nine subjects participated in an in-person screen. Study inclusion criteria included persistent (>1y), upper-extremity impairment; at least a trace muscle contraction in the wrist extensors; single unilateral stroke; mobility and function sufficient for independent performance of activities (eg, toileting, eating lunch during the treatment days); stable medical condition; no other prior neurologic condition; and ability to follow 2-step commands. The study was conducted under the oversight of the institutional review board of the medical center. All subjects provided informed consent prior to study participation.
List of abbreviations:
AMAT Arm Motor Ability Test
AMAT-F AMAT Function scale
AMAT S/E Arm Motor Ability Test for shoulder/elbow
AMAT S/E-F AMAT S/E Function scale
AMAT W/H Arm Motor Ability Test for wrist/hand
AMAT W/H-F AMAT W/H Function scale
FES functional electrical stimulation
FM Fugl-Meyer
ML motor learning
Technology
Robotics training was implemented using the InMotion2 Shoulder-Elbow Robot.a This robotic device is a 2-degrees-of-freedom system that is back-drivable and impedance-controlled to allow for near-frictionless movement in a horizontal plane. The robot used the QNX real-time operating system,b which allowed for highperformance control and integrated graphics. Subjects were seated comfortably in a chair with their hemiplegic forearm and hand supported by a forearm cradle and cone-shaped hand support. Training movements were shoulder/elbow movements of flexion/ extension and horizontal shoulder movements from a center target to and from 8 points located on a circle around the center point.
FES was provided with the commercially available EMS+2 stimulator0 and surface gel electrodes (flexible PALS surface electrodesd). The EMS+2 is a portable, battery-operated, 2-channel surface electrical stimulator that delivers a biphasic, symmetrical, rectangular output for each of the 2 available channels. The stimulation parameters were as follows: 300-millisecond pulse width, 40Hz, and amplitude varied according to subject tolerance. The muscles stimulated included wrist and finger flexors/extensors and forearm supinators/pronators.
Interventions
The goal of training was recovery of the movement components composing functional tasks and recovery of performance of the whole complex task. Treatment was based on ML principles including the following: movement practice as close to normative as possible,22,23 high number of repetitions,24-27 attention to the motor task,28 and training specificity.29 Progression of training was based on the recovery of volitional capability and motor task difficulty, according to the motor task difficulty hierarchy shown in appendix 1. ML exercises were provided for training-isolated joint movement coordination of the scapula, shoulder, elbow, forearm, wrist, fingers, and thumb; task component movements; and whole arm/hand functional training (appendix 2).
Examples of practiced task components are reaching, grasp preparation, and grasp release. To encourage participation, functional tasks that were meaningful to the subject were used. We used a 1:3 group therapy paradigm, whereby 1 therapist treated a group of 3 subjects for 5h/d. There were 3 interventionists; each one was assigned to 1 of the 3 treatment groups. Standardization of treatment was addressed through weekly meetings that included identification of subject impairments and consensus of treatment addressing each given impairment.
Those in the robotics plus ML group used the robot for 1.5h/d. For the remainder of the day they were provided with ML without technologies (3.5h). Similarly, those in the FES plus ML group used FES for 1.5h/d. The ML group was provided with the ML intervention for 5h/d.
Primary outcome measure: Arm Motor Ability Test
All measures were acquired at pre- and posttreatment. There was 1 assessor, who was blinded to the group assignment of the subject. The primary outcome measure was the Arm Motor Ability Test (AMAT), which is a homogenous measure of functional tasks of everyday living.30 The AMAT consists of 13 complex functional tasks, which are videotaped and timed for performance. Examples of the AMAT tasks of everyday living are as follows: pick up and drink from a mug and pick up comb and comb hair.
174 Inquiries by Phone
Assessed for Eligibility
By in-person screen (n = 39)
Enrollment
(n = 39)
Randomized Allocation to Treatment
Allocated to FES ML (n = 14)
Received allocated intervention (n = 14)
Allocated to ML
(n = 13)
Received allocated intervention (n = 13)
Completed Treatment (n = 12)
Lost to follow-up (n = 2)
• Acute psychological problem
• Homesick (returned to home state)
Completed Treatment (n = 11)
Lost to follow-up (n = 2)
• Cardiovascular complications
• Dishonest about initial shoulder pain
Data Analysis (n = 12)
Data Analysis
(n = 11)
Excluded from In-person Screen (n = 135) 72 did not meet inclusion criteria 18 inquired for information only 1 declined in-person screen 1 lost caregiver 1 lost contact
23 inquiries from out of state; related issues
4 unable to commit time 10 transportation issue
5 medical issues developed for patient or caregiver, precluding their participation
Allocated to ROB ML (n = 12)
Received allocated intervention (n = 12)
Fig 1 Consolidated Standards of Reporting Trials diagram. Depiction of subject selection, group allocation, attrition and data analysis. Abbreviations: FES ML, FES plus ML group; ML, ML group; ROB ML, robotics plus ML group.
Secondary measures
Because the robotics and FES technologies were focused on shoulder/elbow or wrist/hand, respectively, we used 2 AMAT subscales: AMAT for shoulder/elbow (AMAT S/E) and AMAT for wrist/hand (AMAT W/H). The AMAT S/E and AMAT W/H subscales were scored by recording the time of the shoulder/elbow movements or the wrist/hand movements, respectively, performed during each AMAT task performance. These subscales have shown good validity and reliability (AMAT S/E intraclass correlation coefficient: .82; AMAT W/H intraclass correlation coefficient: .96).31
The FM coordination scale is a measure of limb joint movement coordination,32 with good validity and intra- and interrater reliability, used in most upper-limb rehabilitation studies of stroke.33 In addition to using the overall FM upper-limb motor score, we generated values for the shoulder/elbow movement items and wrist/hand items, summing them for each participant into the subscale scores of FM for shoulder/elbow and FM for wrist/hand, respectively.
In addition to the quantitative timed AMAT used for the primary measure, the AMAT can be scored using an ordinal,
observational scale (0—5 points; AMAT Function scale [AMAT-F]). To investigate clinical significance within groups, we used the AMAT-F for which others have reported that a gain of .21 points is indicative of clinical significance.34,35 The AMAT-F has shown correlation with the clinically meaningful FM score.36
Statistical analyses
Analyses were completed using the IBM SPSS version 19.0 statistical software package.6 Baseline measures were compared across the 3 treatment groups for the AMAT and FM coordination scale using the nonparametric Kruskal-Wallis test. For the primary study question of group treatment difference on the AMAT, the Kruskal-Wallis test was used on the improvement scores (pre-post). Additionally, 95% confidence intervals for mean differences for pairwise comparisons were completed. A similar group analysis was conducted on the secondary measure of the FM coordination scale. For the additional secondary within-group analyses, we made pre-/ posttreatment comparisons within each group using the Wilcoxon signed-rank test. A 95% Hodges-Lehman confidence interval was
Table 1 Subject characteristics
Group Stroke Type Years Poststroke Age Range (y) Sex Baseline FM Upper-Limb Score (SD)
Cortical Subcortical Both Brainstem 1-3 >4 21-49 50-81 Male Female
ML 6 1 2 2 8 3 2 9 6 5 23.58±5.86
FES plus ML 6 3 3 0 10 2 3 9 7 5 22.85±6.92
Robotics plus ML 3 4 4 1 9 3 2 10 10 2 22.62±5.66
NOTE. Values for stroke type, years poststroke, age range, and sex are n.
included for estimating the median change from pre- to posttreat-ment. To correct for multiple testing, sets of related hypotheses were grouped together,37 and then the Holm Bonferroni stepdown correction method was used to determine statistical significance.38 For the secondary measure, the ordinal AMAT-F measure and the subscales of AMAT S/E Function scale (AMAT S/E-F) and AMAT W/H Function scale (AMAT W/H-F), we calculated the following descriptive statistics: mean AMAT-F score for each individual across the task scores for both pre- and posttreatment, change score, and group means and change score. We inspected each group change score relative to the value of 0.21 point (clinically significant change for the AMAT-F).
Results
A total of 39 subjects enrolled in this study, with all but 1 subject in the severe range of impairment, according to the upper-extremity motor FM score <36 points21 (table 1). The attrition rate was 10% (4/39) (see fig 1). There were 4 subjects who enrolled (2 in the ML alone group, 2 in the FES plus ML group) but did not complete the study. Their characteristics did not alter the relative subject characteristics across groups, and
the characteristics are as follows: sex (FES plus ML group: 1 woman and 1 man; ML alone group: 2 men), stroke type (FES plus ML: cortical [n=1] and subcortical [n=1]; ML alone group: cortical [n = 2]), years poststroke (FES plus ML group: 1—3y [n = 2]; ML alone: 1-3y [n=1] and injury >4y [n = 1]), and age (FES plus ML group: 50—81y [n = 2]; ML alone group: 50—81y [n = 2]). The reasons for their withdrawing from the study were things such as transportation and family issues. A total of 35 subjects completed the study (see fig 1). The analyses subsequently reported were conducted on those who completed the study. No adverse events occurred as a result of participation in the study.
Prior to beginning treatment, there was no statistically significant difference among the 3 treatment groups based on baseline AMAT (P>.866) or baseline FM score (P>.966).
AMAT measure
Group comparison
For the primary measure (AMAT), there was no significant difference across groups regarding treatment response (P>.584). Similarly, for the secondary measures of the AMAT
Table 2 No significant difference between groups for AMAT measure of complex function
AMAT Measure Comparison Groups Pretreatment s) Posttreatment s) Mean Change Score s) Mean Difference 95% CI (s) P
AMAT ML vs FES+ML ML: 1794±479 ML: 1417+637 377 -124 -430 to 182) .584
FES+ML: 1868+501 FES+ML: 1367+566 501
ML vs ROB+ML ML: 1794+479 ML: 1417+637 377 -28 -334 to 278) .972
ROB+ML: 1868+597 ROB+ML: 1463+573 405
ROB+ML vs FES+ML ROB+ML: 1868+597 ROB+ML: 1463+573 405 -96 -395 to 206) .712
FES+ML: 1868+501 FES+ML: 1367+566 501
AMAT S/E ML vs FES+ML ML: 931+288 ML: 709+316 222 -27 -194 to 141) .917
FES+ML: 956+285 FES+ML: 707+263 249
ML vs ROB+ML ML: 931+288 ML: 709+316 222 -46 -213 to 122) .786
ROB+ML: 979+286 ROB+ML: 711+267 268
ROB+ML vs FES+ML ROB+ML: 979+286 ROB+ML: 711+267 268 18 -146 to 182) .960
FES+ML: 956+285 FES+ML: 707+263 249
AMAT W/H ML vs FES+ML ML: 864+250 ML: 682+326 182 -70 -257 to 117) .631
FES+ML: 912+245 FES+ML: 660+320 252
ML vs ROB+ML ML: 864+250 ML: 682+326 182 43 -143 to 231) .831
ROB+ML: 890+325 ROB+ML: 751+320 139
ROB+ML vs FES+ML ROB+ML: 890+325 ROB+ML: 751+320 139 -113 -297 to 69) .288
FES+ML: 912+245 FES+ML: 660+320 252
NOTE. Values are mean ± SD or as otherwise indicated.
Abbreviations: CI, confidence interval;FES+ML, FES plus ML group;ML, ML group;ROB+ML, robotics plus ML group.
Table 3 Within-group gains in functional task performance ( AMAT) for each of the 3 treatment groups
Treatment Group Functional Task Measure Pretreatment ( s) Posttreatment (s) Median Difference (95% CI) (s) P
ML AMAT 1794±479 1417±637 -277 -341 to -217) .003*
AMAT S/E 931±288 709±316 -209 -284 to -155) .003*
AMAT W/H 864±250 682±326 -144 -344 to -51) .009*
FES plus ML AMAT 1868±501 1367±566 -415 -655 to -290) .002*
AMAT S/E 956±285 707±263 -206 -387 to -115) .002*
AMAT W/H 912±245 660±320 -232 -374 to -133) .003*
Robotics plus ML AMAT 1868±597 1463±573 -402 -509 to -298) .002*
AMAT S/E 979±286 711±267 -262 -339 to -208) .002*
AMAT W/H 890±325 751±320 -119 -207 to -72) .003*
NOTE. Values are mean ± SD or as otherwise indicated. Abbreviation: CI, confidence interval. * Adjusted P value.
S/E and AMAT W/H subscales, there was no difference across groups in treatment response (P>.786 and P>.288, respectively) (table 2).
Within-group improvement
All 3 treatment groups demonstrated a statistically significant improvement according to the AMAT, AMAT S/E, and AMAT W/H, after adjusting for multiple tests (P<.009) (table 3).
Coordination impairment secondary measures
Group comparison
For the secondary measures of joint coordination (FM scale, FM scale for shoulder/elbow, FM scale for wrist/hand), there was no significant difference across groups regarding treatment response
(FM scale: P>.590; FM scale for shoulder/elbow: P>.979; FM scale for wrist/hand: P>.340) (table 4).
Within-group improvement
All 3 treatment groups demonstrated a statistically significant within-group improvement according to the FM scale, FM scale for shoulder/elbow, and FM scale for wrist/hand after adjusting for multiple tests (P<.007) (table 5).
Descriptive statistics for the AMAT-F scale
Table 6 provides descriptive statistics for the ordinal AMAT-F scale, AMAT S/E-F, and AMAT W/H-F for each of the 3 groups. Pre-/posttreatment change scores for all measures were >.21 point, which is considered the minimum value for clinically important improvement. All scores, except for 2 change
Table 4 No significant difference between groups according to gain in coordination (FM scale)
Mean Change Group Mean
Functional Task Score for Difference
Measure Groups Compared Pretreatment ( points) Posttreatment points) Each Group 95% CI) points) P
FM scale ML vs FES+ML ML: 23.6±5.8 ML: 33.5±8.3 9.9 1.1 -4.1 to 6.2) .867
FES+ML: 23.5±6.5 FES+ML: 32.3±7.9 8.8
ML vs ROB+ML ML: 23.6±5.8 ML: 33.5±8.3 9.9 2.2 -3.1 to 7.2) .590
ROB+ML: 23.6±5.9 ROB+ML: 31.3±6.2 7.7
ROB+ML vs FES+ML ROB+ML: 23.6±5.9 ROB+ML: 31.3±6.2 7.7 1.1 -4.0 to 6.0) .877
FES+ML: 23.5±6.5 FES+ML: 32.3±7.9 8.8
FM scale for ML vs FES+ML ML: 12.7±2.9 ML: 16.4±3.9 3.7 0.1 -2.6 to 2.2) .979
shoulders/elbows FES+ML: 12.7±3.5 FES+ML: 16.5±3.9 3.8
ML vs ROB+ML ML: 12.7±2.9 ML: 16.4±3.9 3.7 0 -2.5 to 2.4) .999
ROB+ML: 12.9±1.9 ROB+ML: 16.6±2.5 3.7
ROB+ML vs FES+ML ROB+ML: 12.9±1.9 ROB+ML: 16.6±2.5 3.7 0.1 -2.2 to 2.6) .984
FES+ML: 12.7±3.5 FES+ML: 16.5±3.9 3.8
FM scale for ML vs FES+ML ML: 9.1±2.6 ML: 14.7±4.7 5.6 1 -2.3 to 4.5) .728
wrists/hands FES+ML: 8.8±3.5 FES+ML: 13.4±4.2 4.6
ML vs ROB+ML ML: 9.1±2.6 ML: 14.7±4.7 5.6 1.9 -1.4 to 5.4) .340
ROB+ML: 8.3±4.3 ROB+ML: 12.0±4.1 3.7
ROB+ML vs FES+ML ROB+ML: 8.3±4.3 ROB+ML: 12.0±4.1 3.7 0.9 -2.4 to 4.2) .777
FES+ML: 8.8±3.5 FES+ML: 13.4±4.2 4.6
NOTE. Values are mean ± SD or as otherwise indicated.
Abbreviations: CI, confidence interval;FES+ML, FES plus ML group;ML, ML group;ROB+ML, robotics plus ML group.
Table 5 Within-group gains in impaired coordination (FM) for each of the 3 treatment groups
Median Gain
Treatment Pretreatment Posttreatment Score (95% CI) Mean Gain
Group Coordination Measure (points) (points) (points) P Score
ML FM 23.6±5.8 33.5±8.3 9 (7.5-12.5) .003* 11
FM scale for shoulders/elbows 12.7±2.9 16.4±3.9 3.5 (2.5-4.5) .003* 4
FM scale for wrists/hands 9.1±2.6 14.7±4.7 5 (4.0-7.5) .003* 6
FES+ML FM 23.5±6.5 32.3±7.9 8 (5.5-12) .002* 10
FM scale for shoulders/elbows 12.7±3.5 16.5±3.9 4 (2.0-6.0) .005* 4
FM scale for wrists/hands 8.8±3.5 13.4±4.2 5 (2.0-7.0) .003* 5
ROB+ML FM 23.6±5.9 31.3±6.2 7.8 (4.5-11) .003* 8
FM scale for shoulders/elbows 12.9±1.9 16.6±2.5 3.5 (2.5-5.0) .002* 3
FM scale for wrists/hands 8.3±4.3 12.0±4.1 4.0 (1.5-5.0) .007* 4
NOTE. Values are mean ± SD or as otherwise indicated.
Abbreviations: CI, confidence interval;FES+ML, FES plus ML group;ML, ML group;ROB+ML, robotics plus ML group.
* Significant according to adjusted P value.
scores, were less than or equal to twice the minimum value for clinically important improvement. The 2 individual participant change scores, which were the smallest, were in the robotics group plus ML group (AMAT-F, .37 point; AMAT W/H-F: .26 point).
Descriptive statistics for the FM coordination scale
Descriptive statistics for the FM coordination measure provide some additional insight into the level of clinically significant change for the subjects in each of the treatment groups. In the robotics plus ML and FES plus ML groups there were 75% and 92% of subjects, respectively, with a clinically significant gain in coordination impairment (>4.25 points on the FM coordination
scale). For the ML alone group, 100% of subjects were equal to or beyond a clinically significant gain. No subjects in the study worsened. Highest FM gain score for a participant in each group was: FES plus ML (25 points); robotics plus ML (15 points); and ML alone (18 points).
Discussion
Direct comparison of shoulder/elbow robotics, wrist/hand FES, and ML
To our knowledge, this is the first study of chronic stroke survivors making a comparison of robotics and FES and a direct comparison of either technology with intensive ML. We found no significant difference among the 3 groups in terms of treatment response, according to a measure of 13 complex functional tasks and an impairment measure of joint coordination. This could have been because all 3 groups received treatment that was based on ML principles (eg, as close to normative practice as is possible, focused attention on the task, high number of daily practice repetitions of motor task components, whole-task practice of functionally meaningful tasks, and generalization of movement component practice to >1 type of whole-task practice). In preliminary work, we reported that emphasis of shoulder/elbow robotics treatment resulted in significantly greater gains in AMAT S/E versus treatment with FES emphasis for the wrist/hand. We also found the converse; that is, emphasis of wrist/hand FES treatment resulted in significantly greater gain in AMAT W/H versus treatment with emphasis on shoulder/elbow robotics.31 However, that sample size was very small (n = 6 and n = 6, respectively).31 The current results did not bear out our findings from that preliminary work. Because all 3 groups had the benefit of comprehensive coordination training, any unique advantage of either robotics or FES could have been superseded by the importance of the general framework and principles of treatment. It could be that the hours of ML without the technologies served to consolidate newly learned joint coordination that was gained through the use of either robotics or FES. Alternatively, a larger sample size may show a significant group difference.
Table 6 AMAT-Function ordinal measure descriptive statistics
showing clinically significant change scores*
Change
Treatment Group Pretreatment Posttreatment Score*
a. AMAT function
measure
ML 1.82+0.48 2.30+0.77 0.48+0.34
FES+ML 1.78+0.53 2.22+0.62 0.44+0.24
ROB+ML 1.75+0.60 2.13+0.56 0.37+0.25
b. AMAT S/E function
measure
ML 2.12+0.53 2.55+0.67 0.43+0.23
FES+ML 2.04+0.52 2.47+0.56 0.42+0.35
ROB+ML 2.00+0.57 2.44+0.42 0.44+0.30
c. AMAT W/H function
measure
ML 1.37+0.57 1.89+0.93 0.53+0.61
FES+ML 1.42+0.67 1.92+0.71 0.50+0.27
ROB+ML 1.35+0.73 1.60+0.82 0.26+0.21
NOTE. Values are mean : + SD.
Abbreviations: FES+ML, FES plus ML group;ML, ML group;ROB+ML,
robotics plus ML group.
* Clinically significant improvement is >.21 points.
Considerations of extent of recovery, level of impairment, and treatment duration/intensity for severely impaired chronic stroke survivors
This study contributes to the literature in the extent of improvement that was shown in the FM joint coordination measure for all 3 groups of more severely involved participants in the chronic phase (>1y after stroke). In our consideration of the literature here, we are focusing on studies of others that enrolled stroke survivors at >6 months poststroke because some have reported spontaneous or endogenous recovery up to 3 to 6 months after stroke, which could confound a study of group difference.
Contrasting response to treatment for the less impaired subjects in other studies versus the more impaired subjects in the current work
Our study cohort was in the severely impaired category (baseline upper-limb motor FM score <36 points). Even still, compared with the work of others for mild to moderately impaired stroke survivors, the gain scores in our study of the severely impaired were either comparable (robotics plus ML group), higher (ML alone group), or almost twice as high (FES plus ML group) as that reported for the less impaired. For example, for mild to moderately impaired chronic stroke survivors, FM gains were reported in response to treatment as follows. In robotics therapy, gains reported ranged from 3.36 to 9 points.13'39'40 In FES therapy, a 5-point gain was reported.41 In ML or exercise, gains of 6 to 8 points were reported.5'7'10'42'43 In contrast with those studies of lesser impaired individuals, the current study focused on severely impaired stroke survivors and yielded the following results for FM mean gains: the robotics plus ML group yielded 8 points, the FES plus ML group yielded 9 points, and the ML alone group yielded 11 points. In terms of clinically important difference, other studies18 have suggested that the estimated clinically important difference for the upper-extremity FM coordination scale ranges from 4.25 to 7.25 points in scores. Our results for all 3 groups were beyond those values for the severely impaired. In addition, our gain scores for the AMAT-F scale (see table 6) were more than twice the clinically significant value of .21 point for the ML alone and FES plus ML groups and were greater than clinically significant for the robotics plus ML group.
Potential effect of treatment intensity (number of sessions, hours/session)
Emerging empirical evidence is supporting long-held clinical observation; that is, for recovery of persistent discoordination after stroke, many hours of specifically formulated practice are required.24'29'44'45 The current study included intensive practice of coordinated tasks (5h/ session, 60 sessions); this treatment intensity may help to explain the larger gains reported here in coordination and function for these severely involved stroke survivors. Although Kraft et al20 studied only 6 subjects in its FES group, they also provided 3 months of treatment, which may explain their high FM mean gain (8 points).
For our other 2 treatment groups (robotics plus ML and ML alone), treatment intensity may also explain gains that were greater for our severely impaired individuals than that reported by other researchers for severely impaired stroke survivors. For example, for severely impaired stroke survivors, others reported FM gains in response to robotics ranging only from 1.2 to 5 points,15,46-48 and ML alone was reported to have produced only a 4-point FM gain in 46 participants who were more severely impaired.15 In the current study, more hours of treatment were provided than for these cited studies. Given the results reported here (robotics plus ML group:
FM gain of 8; ML alone: FM gain of 11), it is reasonable to consider that greater treatment intensity is needed for the more severely impaired using those 2 types of interventions (ML alone or robotics plus ML groups) to achieve the greater FM score gains.
The current clinical practice milieu prevents the provision of long-duration, high-intensity treatment; therefore, this new information is an important contribution to the literature. One reason for the lack of provision of long-duration interventions in standard clinical care is the out-of-date belief that no more recovery can occur after 3 to 6 months poststroke. In contrast with these inaccurate beliefs, our results are consistent with others who have demonstrated the possibility of motor recovery beyond that time period, through the application of a variety of treatment methods.3'4'13-17'19'40'43'46-57
Functional task improvement
Although many research studies report significant gains in impairment, there is less information available regarding the recovery of actual functional tasks in response to experimental interventions, according to a homogenous measure of complex functional task performance (ie, everyday functional tasks). In the current work, the statistically significant improvement within each group for the AMAT (13 complex function tasks) can be explained in a number of ways. First, the high gain in the FM score in all 3 groups may have been sufficiently robust to produce a significant improvement in a measure of 13 actual complex function tasks. Second, the purposeful application of fundamental ML principles could explain both the relatively high gains in coordination (FM score) and AMAT gains. Third, the protocol was specific in practice of joint movement components within the context of actual task practice. (supplemental video S1 shows recovery of coordination and functional capability; available online only at http://www.archives-pmr.org/.)
Patient group delivery of intensive and long-duration intervention
We found that it was feasible to deliver the study protocol using a group treatment method (1:3 ratio of therapist to patients). The relatively high gain in FM scores for all 3 groups could serve as evidence to support the feasibility of the 1:3 therapist to patient ratio. According to the work of others58,59 and our study therapists' reports, the group treatment paradigm was more reasonably feasible with the use of the robotics or FES technologies because these practice-assist devices could be quickly set up in a manner enabling some independent practice, while the therapist could focus on other participants. This allowed a more calm therapeutic setting and a more satisfying work situation for the therapist.
Cost considerations
We calculated the cost of each of the 3 treatment protocols in this study. The following assumptions were used: therapist cost ($98,000, which is the annual salary for an experienced therapist in Ohio where the study was conducted; source: Department of Veterans Affairs and additional local hospital); shoulder/elbow clinical level robot cost ($89,000) and 5-year robot life; annual robot warranty and maintenance ($8000; source: robot distributing company); and FES cost for a 4-channel table top and 2-channel portable system ($4000), with a 5-year equipment life. We used the facts of our protocol (number of visits; duration of sessions for use of each piece of equipment and ML alone; and a ratio of 1:3, therapist to patient). Our calculations yielded the following costs per patient for the entire treatment protocol: ML alone ($4570), FES plus ML ($4604), and robotics plus
ML ($5686). These costs are in the ballpark in comparison with the calculations of others15; however, our treatment was considerably longer, the cost of our robot was significantly less because we used only 1 type of robot, and the robot cost is for a clinical robot. Our ML alone protocol was less expense than the robotics plus ML protocol by $1116 and less than the FES plus ML protocol by $34. Therefore, if a cost differential of approximately $1000 per patient is considered important, then the FES plus ML protocol and/or the ML alone protocol would be preferable.
Study limitations
There were a number of study limitations. First, the sample sizes per group were 11, 12, and 12, respectively. Although there was no significant difference and no indication of a trend in group difference, a larger sample size might have shown group differences. Second, this was a research trial. To determine whether the 1:3 ratio of therapist to patients is practical and beneficial in clinical practice, this treatment paradigm would require testing in a clinical environment. Third, in this study, FES and robotics were differentially targeted to either wrist/hand or shoulder/elbow, respectively; therefore, this study did not make a direct comparison of either robotics versus FES for shoulder/elbow or robotics versus FES for wrist/hand. Rather, this study design was selected and funded based on 2 assumptions. The first assumption was that FES may be preferable for wrist/hand intervention because it can be applied quickly and easily to wrist/hand flexors and extensors, and notably, it provides practice of an actual muscle contraction. In contrast, robotics can encourage and enable less therapeutic passive participation. The second assumption was that robotics may be preferable for shoulder/elbow intervention because it can be quickly and easily set up to support and guide movement of the complex shoulder/elbow movement components composing the reach task. In contrast, FES would have required time-consuming application of multiple electrodes for scapular and limb muscles and control of complex precision timing of multiple muscle activations for the greatest effectiveness.
Fourth, because of limitations in resources, it was not feasible to acquire follow-up data. However, others have documented good maintenance of gains after ML, robotics, and FES. For severely impaired individuals, robotics,15 ML (intensive therapy15), and FES60 produced gains in response to treatment that were maintained at follow-up. With these reported maintenance gains taken together, along with our high gains after treatment, it is reasonable to consider that gains may have been maintained in the current study. However, further study is required to quantitatively compare follow-up maintenance across groups.
Conclusions
Severely impaired chronic stroke subjects (>1y) with persistent upper-extremity dysfunction can make clinically significant gains in joint movement coordination and functional task performance in response to the 3 tested interventions (ML, combined robotics and ML, combined FES and ML) in an intensive and long-duration intervention. There was no difference in treatment response across the 3 intervention groups according to measures of joint movement coordination or complex functional task performance. It was feasible in the research laboratory to deliver effective group treatment for severely impaired stroke survivors in a 1:3 (therapist/patient) ratio.
Suppliers
b. QNX Software Systems.
c. Staodyn, Inc.
d. Axelgaard Manufacturing Co, Ltd.
e. IBM Corporation.
Keywords
Electric stimulation; FES; Randomized controlled trial, Rehabilitation; Robotics; Stroke; Upper extremity
Corresponding author
Janis J. Daly, PhD, MS, Director, Brain Rehabilitation Research Center, 151-A, Malcom Randall DVA Medical Center, 1601 SW Archer Rd, Gainesville, FL 32608. E-mail address: janis.daly@ neurology.ufl.edu.
Appendix 1 Upper-Limb Training Protocol: Treatment Progression Hierarchy for Coordinated Movement Practice
A. Muscle activation in synergy
B. Single joint movement in synergy
C. Single joint movement, out of synergy * Coordination training * Speed of movement training
D. Multiple joint movement, out of synergy * Coordination training * Speed of movement training
E. Alternating joint movement (flexion and extension) * Coordination training 'Speed of movement training
F. Task component practice * Coordination training 'Speed of movement training
G. Full functional task practice * Coordination training "Speed of movement training
a. Interactive Motion Technologies.
Appendix 2 Examples of Functional Tasks Practiced During Training Sessions
Stir food in a bowl.
Place objects in kitchen cupboard.
Carry objects (unilateral and bilateral).
Write with pen/pencil.
Type at computer.
Sweep with broom.
Throw ball.
Swing a golf club.
Sand wood.
References
1. Kollen BJ, Lennon S, Lyons B, et al. The effectiveness of the Bobath concept in stroke rehabilitation: what is the evidence? Stroke 2009; 40:e89-97.
2. van Vliet PM, Lincoln NB, Foxall A. Comparison of Bobath based and movement science based treatment for stroke: a randomized controlled trial. J Neurol Neurosurg Psychiatry 2005;76:503-8.
3. McCombe-Waller S, Liu W, Whitall J. Temporal and spatial control following bilateral versus unilateral training. Hum Mov Sci 2008;27: 749-58.
4. Wolf SL, Winstein CJ, Miller JP, et al. Effect of constraint-induced movement therapy on upper extremity function 3 to 9 months after stroke: the EXCITE randomized clinical trial. JAMA 2006;296: 2095-104.
5. Lin KC, Chung HY, Wu CY, et al. Constraint-induced therapy versus control intervention in patients with stroke: a functional magnetic resonance imaging study. Am J Phys Med Rehabil 2010; 89:177-85.
6. Wu CY, Chen CL, Tang SF, Lin KC, Huang YY. Kinematic and clinical analyses of upper-extremity movements after constraint-induced movement therapy in patients with stroke: a randomized controlled trial. Arch Phys Med Rehabil 2007;88:964-70.
7. Wu CY, Chen CL, Tsai WC, Lin KC, Chou SH. A randomized controlled trial of modified constraint-induced movement therapy for elderly stroke survivors: changes in motor impairment, daily functioning, and quality of life. Arch Phys Med Rehabil 2007;88:273-8.
8. Lin KC, Wu CY, Wei TH, Lee CY, Liu JS. Effects of modified constraint-induced movement therapy on reach-to-grasp movements and functional performance after chronic stroke: a randomized controlled study. Clin Rehabil 2007;21:1075-86.
9. Lin KC, Wu CY, Liu JS. A randomized controlled trial of constraint-induced movement therapy after stroke. Acta Neurochir Suppl 2008; 101:61-4.
10. Lin KC, Chang YF, Wu CY, Chen YA. Effects of constraint-induced therapy versus bilateral arm training on motor performance, daily functions, and quality of life in stroke survivors. Neurorehabil Neural Repair 2009;23:441-8.
11. Gauthier LV, Taub E, Perkins C, Ortmann M, Mark VW, Uswatte G. Remodeling the brain: plastic structural brain changes produced by different motor therapies after stroke. Stroke 2008;39:1520-5.
12. Fasoli SE, Krebs HI, Stein J, Frontera WR, Hogan N. Effects of robotic therapy on motor impairment and recovery in chronic stroke. Arch Phys Med Rehabil 2003;84:477-82.
13. Fasoli SE, Krebs HI, Stein J, Frontera WR, Hughes R, Hogan N. Robotic therapy for chronic motor impairments after stroke: follow-up results. Arch Phys Med Rehabil 2004;85:1106-11.
14. Volpe BT, Lynch D, Rykman-Berland A, et al. Intensive sensorimotor arm training mediated by therapist or robot improves hemiparesis in patients with chronic stroke. Neurorehabil Neural Repair 2008;22: 305-10.
15. Lo AC, Guarino PD, Richards LG, et al. Robot-assisted therapy for long-term upper-limb impairment after stroke. N Engl J Med 2010; 362:1772-83.
16. Ang KK, Guan C, Chua KS, et al. A clinical study of motor imagery-based brain-computer interface for upper limb robotic rehabilitation. Conf Proc IEEE Eng Med Biol Soc 2009;2009:5981-4.
17. Reinkensmeyer DJ, Maier MA, Guigon E, et al. Do robotic and non-robotic arm movement training drive motor recovery after stroke by a common neural mechanism? Experimental evidence and a computational model. Conf Proc IEEE Eng Med Biol Soc 2009;2009: 2439-41.
18. Page SJ, Fulk GD, Boyne P. Clinically importance differences for the upper-extremity Fugl-Meyer in minimally to moderately impaired, chronic stroke. Phys Ther 2012;92:791-8.
19. Chan MK, Tong RK, Chung KY. Bilateral upper limb training with functional electric stimulation in patients with chronic stroke. Neu-rorehabil Neural Repair 2009;23:357-65.
20. Kraft GH, Fitts SS, Hammond MC. Techniques to improve function of the arm and hand in chronic hemiplegia. Arch Phys Med Rehabil 1992;73:220-7.
21. Duncan PW, Goldstein LB, Matchar D, Divine GW, Feussner J. Measurement of motor recovery after stroke. Outcome assessment and sample size requirements. Stroke 1992;23:1083-9.
22. Nudo RJ, Wise BM, SiFuentes F, Milliken GW. Neural substrates for the effects of rehabilitative training on motor recovery after ischemic infarct. Science 1996;272:1791-4.
23. Nudo RJ, Milliken GW, Jenkins WM, Merzenich MM. Use-dependent alterations of movement representations in primary motor cortex of adult squirrel monkeys. J Neurosci 1996;16:785-807.
24. Butefisch C, Hummelsheim H, Denzler P, Mauritz KH. Repetitive training of isolated movements improves the outcome of motor rehabilitation of the centrally paretic hand. J Neurol Sci 1995;130:59-68.
25. Dean CM, Shepherd RB. Task-related training improves performance of seated reaching tasks after stroke. A randomized controlled trial. Stroke 1997;28:722-8.
26. Elbert T, Pantev C, Wienbruch C, Rockstroh B, Taub E. Increased cortical representation of the fingers of the left hand in string players. Science 1995;270:305-7.
27. Pascual-Leone A, Torres F. Plasticity of the sensorimotor cortex representation of the reading finger in Braille readers. Brain 1993; 116:39-52.
28. Singer R, Lidor R, Cauraugh JH. To be aware or not aware? What to think about while learning and performing a motor skill. Sport Psychol 1993;7:19-30.
29. Plautz EJ, Milliken GW, Nudo RJ. Effects of repetitive motor training on movement representations in adult squirrel monkeys: role of use versus learning. Neurobiol Learn Mem 2000;74:27-55.
30. Kopp B, Kunkel A, Flor H, et al. The Arm Motor Ability Test: reliability, validity, and sensitivity to change of an instrument for assessing disabilities in activities of daily living. Arch Phys Med Rehabil 1997;78:615-20.
31. Daly JJ, Hogan N, Perepezko EM, et al. Response to upper-limb robotics and functional neuromuscular stimulation following stroke. J Rehabil Res Dev 2005;42:723-36.
32. Fugl-Meyer AR, Jaasko L, Leyman I, Olsson S, Steglind S. The post-stroke hemiplegic patient. 1. A method for evaluation of physical performance. Scand J Rehabil Med 1975;7:13-31.
33. Duncan PW, Propst M, Nelson SG. Reliability of the Fugl-Meyer assessment of sensorimotor recovery following cerebrovascular accident. Phys Ther 1983;63:1606-10.
34. Harvey RL, Winstein CJ; Everest Trial Group. Design for the everest randomized trial of cortical stimulation and rehabilitation for arm function following stroke. Neurorehabil Neural Repair 2009;23: 32-44.
35. Page SJ, Levin L, Hermann V, Dunning K, Levine P. Longer versus shorter daily durations of electrical stimulation during task-specific practice in moderately impaired stroke. Arch Phys Med Rehabil 2012;93:200-6.
36. Chae J, Labatia I, Yang G. Upper limb motor function in hemi-paresis: concurrent validity of the Arm Motor Ability test. Am J Phys Med Rehabil 2003;82:1-8.
37. Miller RG Jr. Simultaneous statistical inference. New York: SpringerVerlag; 1981.
38. Holm S. A simple sequentially rejective Bonferroni test procedure. Scand J Stat 1979;6:65-70.
39. Takahashi CD, Der-Yeghiaian L, Le V, Motiwala RR, Cramer SC. Robot-based hand motor therapy after stroke. Brain 2008;131: 425-37.
40. Hu XL, Tong KY, Song R, Zheng XJ, Leung WW. A comparison between electromyography-driven robot and passive motion device on wrist rehabilitation for chronic stroke. Neurorehabil Neural Repair 2009;23:837-46.
41. de Kroon JR, Ijzerman MJ. Electrical stimulation of the upper extremity in stroke: cyclic versus EMG-triggered stimulation. Clin Rehabil 2008;22:690-7.
42. Lin KC, Wu CY, Liu JS, Chen YT, Hsu CJ. Constraint-induced therapy versus dose-matched control intervention to improve motor ability, basic/extended daily functions, and quality of life in stroke. Neurorehabil Neural Repair 2009;23:160-5.
43. Page SJ, Levine P, Khoury JC. Modified constraint-induced therapy combined with mental practice: thinking through better motor outcomes. Stroke 2009;40:551-4.
44. Lohse KR, Lang CE, Boyd LA. Is more better? Using metadata to explore dose-response relationships in stroke rehabilitation. Stroke 2014;45:2053-8.
45. Daly JJ, Ruff RL. Evidence-based construction and measurement of efficacious gait and upper limb functional interventions after stroke; a case for combination interventions. Scientific World 2007;7:2031-45.
46. Lum PS, Burgar CG, Shor PC, Majmundar M, Van der Loos M. Robot-assisted movement training compared with conventional therapy techniques for the rehabilitation of upper-limb motor function after stroke. Arch Phys Med Rehabil 2002;83:952-9.
47. Finley MA, Fasoli SE, Dipietro L, et al. Short-duration robotic therapy in stroke patients with severe upper-limb motor impairment. J Rehabil Res Dev 2005;42:683-92.
48. Fazekas G, Horvath M, Troznai T, Toth A. Robot-mediated upper limb physiotherapy for patients with spastic hemiparesis: a preliminary study. J Rehabil Med 2007;39:580-2.
49. Boyd LA, Winstein CJ. Impact of explicit information on implicit motor-sequence learning following middle cerebral artery stroke. Phys Ther 2003;83:976-89.
50. Piron L, Turolla A, Agostini M, et al. Motor learning principles for rehabilitation: a pilot randomized controlled study in post stroke patients. Neurorehabil Neural Repair 2010;24:501-8.
51. Hanlon RE. Motor learning following unilateral stroke. Arch Phys Med Rehabil 1996;77:811-5.
52. Yen JG, Wang RY, Chen HH, Hong CT. Effectiveness of modified constraint-induced movement therapy on upper limb function in stroke subjects. Acta Neurol Taiwan 2005;14:16-20.
53. Byl N, Roderick J, Mohamed O, et al. Effectiveness of sensory and motor rehabilitation of the upper limb following the principles of neuroplasticity: patients stable poststroke. Neurorehabil Neural Repair 2003;17:176-91.
54. Krebs HI, Hogan N, Volpe BT, Aisen ML, Edelstein L, Diels C. Overview of clinical trials with MIT-MANUS: a robot-aided neuro-rehabilitation facility. Technol Health Care 1999;7:419-23.
55. Lum PS, Burgar CG, Shor PC. Evidence for improved muscle activation patterns after retraining of reaching movements with the MIME robotic system in subjects with post-stroke hemiparesis. IEEE Trans Neural Syst Rehabil Eng 2004;12:186-94.
56. Ferraro M, Palazzolo JJ, Krol J, Krebs HI, Hogan N, Volpe BT. Robot-aided sensorimotor arm training improves outcome in patients with chronic stroke. Neurology 2003;61:1604-7.
57. Lo AC, Guarino P, Krebs HI, et al. Multicenter randomized trial of robot-assisted rehabilitation for chronic stroke: methods and entry characteristics for VA ROBOTICS. Neurorehabil Neural Repair 2009;23:775-83.
58. Kwakkel G, Kollen B, Krebs H. Effects of robot-assisted therapy on upper limb recovery after stroke: a systematic review. Neurorehabil Neural Repair 2009;22:111-21.
59. Norouzi-Gheidari N, Archambault PS, Fung J. Effects of robotassisted therapy on stroke rehabilitation in upper limbs: systematic review and meta-analysis of the literature. J Rehabil Res Dev 2012; 49:479-96.
60. Thorsen R, Cortesi M, Jonsdottir J, et al. Myoelectrically driven functional electrical stimulation may increase motor recovery of upper limb in poststroke subjects: a randomized controlled pilot study. J Rehabil Res Dev 2013;50:785-94.
