Constraint‐induced movement therapy versus bimanual intensive therapy in children with hemiplegia showing low/very low bimanual functional performance: A randomized clinical trial

Children with infantile hemiplegia with low or very low bimanual functional performance have great impediments to spontaneously use their affected upper limb, which affects their performance of day‐to‐day activities and their quality of life.


INTRODUCTION
Congenital hemiplegia is one of the most common forms of cerebral palsy, representing 1 in 1300 live births.These children develop motor and sensory disorders on one side of the body, which are more pronounced in the upper limb. 1,2The lack of information delivered to the affected arm at the cortical level causes a reduction in bimanual functional performance of the affected upper limb.This phenomenon influences autonomy and participation in bimanual activities of daily living, thereby decreasing the children's quality of life. 3,4The two most extensively used contemporary approaches for upper limb therapy 5 in children with hemiplegia are modified constraint-induced movement therapy (mCIMT) and bimanual intensive therapy (BIT), as well as hybrid-CIMT, which combines both approaches. 5Both therapies are intensive treatments, and they are based on motor skill learning, using self-generated active movements at high intensity and with positive reinforcement, where the practice is focused directly on attaining a goal set by the child (or by the parents, if necessary). 6CIMT combines repetitive uni-manual tasks practice, behavioral training techniques, and restraint of the unaffected upper limb (<3 hours per day). 7It has been used for improving impairments of the affected hand in children with hemiplegia. 7It is focused on increasing the spontaneous use of the affected upper limb, and two possible mechanisms may lead to greater use of the affected limb: overcoming the learned non-use of the more affected arm and stimulating use-dependent cortical reorganization. 8BIT 9 is focused on improving the coordination of the two hands using structured task practice embedded in bimanual play and functional activities.
Previous studies have demonstrated that BIT results in improved quality and quantity of bimanual upper limb use, and that it is as effective as mCIMT in children with moderate manual ability and whose affected hand acts as an assisting hand in the execution of bimanual activities.However, these interventions have been applied in summer camps or environments different from the children's usual contexts, [10][11][12][13][14][15] and they have not been studied in children within their home environment or with low/very low level of bimanual functional performance. 16Home therapy provides a wide variety of learning opportunities for the affected upper limb. 17Nevertheless, complying with the intervention and reducing the potential burden for the family requires that the parents to be trained in the therapy schedule. 17,18Parents can be more involved throughout the process, increasing opportunities for parent-child interaction.However, the involvement of parents in the treatment of children with low functional performance can make them feel frustrated when applying the therapy, and the level of anxiety can increase when observing the limitations of the affected upper limb. 191][22] The evidence shows that both intensive therapies have similar effects when the same dose is compared in children who do not have a significant non-use of their affected hand. 13,15,22However, what happens when the non-use of the affected hand is very pronounced or practically non-existent?Are the same results obtained in both interventions?Children with low functionality have shown greater improvements with combined protocols (high dose of mCIMT followed by low dose of BIT) and might benefit more from mCIMT than from the application of BIT alone. 23Nevertheless, this has not been studied in children with very low bimanual functional performance in comparison with other combined protocols (in the study of Deep et al., 23 combined therapy was compared only with a bimanual protocol).The influence of intensive therapies on quality of life in children with low and very low performance is unknown.Thus, the main objective of our research was to study whether the order of application and the dose of mCIMT within a combined protocol influences the results of bimanual functional performance of the affected upper limb and the quality of life of children with congenital hemiplegia (5 to 8 years old) with low/very low bimanual functional performance.
The hypothesis of this study was to accept that a combined protocol with an initial application of a high dose of mCIMT followed by a low dose of BIT allows for a greater increase in bimanual functional performance and quality of life in school-age children with congenital hemiplegia showing a low/very low functional performance.

METHODS
The study was approved (Reference N: 260) by the local ethics committee.The anonymity of the participants was preserved and complied with the precepts of Organic Law 3/2018, of 5 December on the Protection of Personal Data and Guarantee of Digital Rights (LOPDGDD).The principal investigator of this project was the only person with access to the data set.Informed consent was obtained from all the children's parents and an age-appropriate consent script was used to inform the children about their participation in the study.

Design
A single-blinded randomized controlled trial was performed in children with infantile hemiplegia with low/ very low bimanual functional performance (score ≤38 AHA units on the Assisting Hand Assessment scale). 24he randomization method was carried out through the Epidat software 4.2 version, which, following a process of simple randomization and consecutive sampling, divided the participants into two groups (experimental group and control group), with a minimum of 10 participants per group.A number sequence was obtained, assigning one number to each participant.These numbers were kept in individual, opaque, closed envelopes, which were opened by the blinded physiotherapist, who allocated the participants to either the experimental group or the control group, according to the randomization sequence.

Participants
Participants were recruited from two public hospitals and an infantile hemiplegia association in Spain (HEMIWEB).

Selection criteria
The children had to meet the following criteria to be included in the study: being diagnosed with congenital hemiplegia; being between 5 and 8 years of age; having low or very low bimanual functional performance [24][25][26] (0-38 AHA units); being classified as level I-III on the Manual Ability Classification System (MACS) 27 ; and being classified as level I-III on the Gross Motor Function Classification System (GMFCS). 28he exclusion criteria were disease not associated with congenital hemiplegia; low cognitive level compatible with attending a special education school; contractures in the affected upper limb that inhibited functional movement; surgery 6 months prior to the treatment; botulinum toxin 2 months prior to or during the intervention; and pharmacologically uncontrolled epilepsy.

Intervention dose
Two protocols of combined intensive therapy were designed to be delivered at home, each lasting 100 hours, applied throughout a 10-week period (structured activities for two discontinuous hours every day from Monday to Friday).The protocols were created based on the research on motor learning curve and long-term effectiveness of mCIMT by Geerdink et al. 29 In the experimental group, the 100 hours were divided into 80 hours of mCIMT followed by 20 hours of BIT, whereas, in the control group, the 100 hours were divided into 80 hours of BIT followed by 20 hours of mCIMT.

Characteristics of the intervention
Both interventions consisted of a family-centered program intended for home execution and included therapy based on principles of activity, motor learning, neuroplasticity, and overcoming non-use. 30dified Constraint-Induced Movement Therapy (mCIMT) For the mCIMT period, uni-manual activities were defined to include specific movements (flexion, abduction, and external rotation of the shoulder: throw balls in different directions above the head, grab objects from behind the child, draw on paper on the wall…; extension of the elbow: reach, touch, or pick up different objects, bowling, knock down a built tower; supination of the forearm: parents put stickers on the palm of the hand or on the forearm of the affected upper limb, play a trumpet or noisemaker, keep a paper cup as upright as possible and move it to another place; extension of the wrist: holding a flat object on the back of the hand, catching an object by raising the fingers, crushing plasticine with the heel of the hand…; and global-distal grasp: grasp, hold, and transfer of objects of different characteristics from one hand to another hand [big/small, rough/smooth, heavy/light, and large/spherical objects].Each week, 12 tasks were programmed to work a specific movement (6 tasks for each hour), each activity lasted 10 minutes.In the first hour, proximal work was enhanced with exercises for the shoulder, elbow, and forearm, and in the second hour, the activities were focused on working on wrist extension and different types of grasps.A glove glued on a cardboard base, to prevent grasp and possible mirror movements, was employed as a method of partial restraint for the unaffected hand, and it was manufactured by each family according to the child's interest (Figure 1).The child had to wear the glove for 2 hours per day, only during the execution of the therapy.

Bimanual Intensive Therapy (BIT)
For the BIT period, each hand was assigned a different role in the designed bimanual activities, beginning with the use of the affected hand as the assisting hand and changing later to act as the manipulative hand.The program comprised the following: complete tasks, including different actions, such as taking a marker out of the case, removing the lid, and drawing a picture on a paper (20 minutes); partial tasks, such as working a specific action, such as building a Lego tower (10 minutes); strength tasks with symmetrical or asymmetrical use of the hands, such as kneading play dough or tearing pieces of paper (10 minutes); and activities of daily living (20 minutes).The schedule for the first hour of daily therapy consisted of one complete task, three partial tasks, and one strength task, whereas the second hour included one complete task (different from the first hour), two partial tasks, and one activity of daily living.The protocol is available upon request to the authors.

Program supervision, parent training, and follow-up
A pediatric physiotherapist supervised the intervention, helping the parents to design an individualized program for each child.The activities were selected based on the functioning of the child's affected hand (basal MACS level and AHA score).The task demands were modified to increase their difficulty level depending on the child's progression.Parents were instructed to focus on making the intervention enjoyable and intrinsically motivating for their children.Age-appropriate reward systems and acknowledgement of results were provided (stickers, songs, games, choosing an activity…) as encouragement and to ensure that the activities were highly motivating and fun.
All parents received training from the same experienced supervisor over a series of three sessions, each lasting about 2 to 3 hours.Only the parents attended the first training session with the supervisor, where they were instructed on the general intervention procedures (experimental group and control group) for the execution of the activities.In the second session, both the parents and their children observed how the supervisor administered the different therapies.The parents were asked to model the activities with their children as administered by the supervisor and obtain feedback.The final training session took place at the family's home, where the supervisor provided feedback to the parents as they performed activities with the child, and also marked the beginning of the 100-hour intervention.The supervisor made a weekly follow-up of the therapy to ensure that the parents complied with the execution of activities via a registry of the performed tasks.If the family had any complication during any session, they could contact the supervisor and conduct a daily follow-up via webcam, thus avoiding complications and promoting motivation and the continuity of the intervention.

Primary Outcome Measure: Bimanual Functional Performance
Bimanual functional performance was measured on the AHA v. 5.0 scale, 31 which has been validated previously for children (ages between 18 months and 12 years) with infantile hemiplegia and obstetric brachial palsy. 31,322][33] The play session was recorded on video and later scored by a certified evaluator, who was blinded to the group allocation.The AHA has been Rasch analyzed, and the model provides measures of equal intervals in logits (log odds probability units) by converting ordinal rating scale observations to interval levels into a user-friendly 0 to 100 scale (namely, AHA units). 31condary Outcome Measure: Quality of Life Quality of life was assessed using the Pediatric Quality of Life Inventory, specifically via its module for cerebral palsy (PedsQL™ v. 3.0, CP module), 34 whose internal reliability coefficient is between 0.7 and 0.9.The CP module comprises 35 items assessing 7 dimensions: daily activities (9 items), school activities (4 items), movement and balance (5 items), pain (4 items), fatigue (4 items), eating activities (5 items), and speech and communication (4 items).A 5-point response scale is utilized (0 = never a problem; 1 = almost never a problem; 2 = sometimes a problem; 3 = often a problem; 4 = almost always a problem).The items are summed and linearly transformed into a 0 to 100 scale (0 = 100, 1 = 75, 2 = 50, 3 = 25, 4 = 0); thus, higher scores indicate better quality of life. 34The questionnaire was completed by the parents during each of the scheduled evaluations.To avoid complications and confusion regarding the questions on the questionnaires used to assess the quality of life of the children, the questionnaires were completed by the parents in the presence of a therapist and later scored by a blinded evaluator.
The assessments from the Assisting Hand Assessment scale and the quality-of-life questionnaires were scored by a blinded evaluator, who was certified in the use and interpretation of the measurement scales used in the study.

Sample Size
The sample size estimates were based on the primary outcome measure, that is, the AHA.The responsiveness to change has been shown previously by Eliasson et al. 35 Calculation by a statistician indicated a 1.16 effect size at a significant level of .05 and 80% power, requiring a minimum sample size of 10 participants per group.

Data Analysis
Statistical analyses were performed with SPSS v.24.0 (SPSS Inc., IBM).Statistical significance was set at p < .05.The Shapiro-Wilk test was employed to study the normality of the sample.
Data are expressed as mean and standard deviation (SD), unless otherwise stated.A descriptive analysis of demographic variables was performed.An independentsamples T-test was used to evaluate whether there were statistically significant intergroup differences in the means of any of the variables at baseline.A two-way (intervention Â time) repeated measures analysis of variance (ANOVA) was conducted to study the effects of the interventions (two protocols) on the dependent variables bimanual functional performance and quality of life at four different time points.The Bonferroni correction was employed for pairwise post hoc comparisons to further analyze significant interactions.The Greenhouse-Geisser adjustment was applied to correct for the lack of sphericity (Mauchly's sphericity test, p < .05)whenever necessary.

RESULTS
A total of 50 children were enrolled in the study, of whom 29 were excluded, and 21 were randomized in two intervention groups: 11 children were allocated in the experimental group and 10 in the control group, as is shown in the CONSORT flow diagram (Figure 3).The children in both groups completed the therapies and no dropouts or adverse events were reported.All of them were diagnosed with perinatal stroke (congenital hemiplegia).The descriptive characteristics of the participants are shown in Table 1.There were no statistically significant intergroup differences in age, baseline outcome of bimanual functional performance, or quality of life (p = .15;p = .68;p = .81,respectively).
Parents showed high compliance, with their children completing 77 of the 80 hours of mCIMT and 18.5 of the 20 hours of BIT in the experimental group, and 76 of the 80 hours of BIT and 19 of the 20 hours of mCIMT in the control group.

Bimanual Functional Performance
There were significant differences in the timepoints for the experimental and control groups (F = 61.3;p < .001)and the protocol-time intersection (F = 12.9; p < .001).Statistically significant intergroup differences were observed in the mean difference obtained from the assessments at week 4, week 8, and week 10, with significantly greater improvements in the experimental group compared to the control group at all time points (Table 2).The application of the protocol showed significant results for the experimental group (F = 6.7; p = .018).In the experimental group, all pairwise comparisons showed a progressive increase in the AHA score from week 0 to week 10 (p < .001).The greatest improvement was obtained at week 4, with an increase of 15.9 units, further reaching 22 units at week 8 (80 hour mCIMT) during the 20 hours of BIT (5.1 units).The functional gains in the experimental group included an increase in spontaneous use, faster initiation of use of the affected hand, and greater amplitude of forearm movement, as well as the capability of reaching and improving the quality of holding objects in the affected hand (measured through the AHA).
On the other hand, the results in the control group showed the greatest improvement (10.6 units) at week 10, when 20 hours of mCIMT were applied.The total increase in the combined protocols of intensive therapy was 27.1 units for the experimental group and 14.3 units for the control group (Figure 4).Significant clinical gains (>5 AHA units) 27 were found in all participants at each assessment in the case of the experimental group, and from week 8 to week 10 (20 hours of mCIMT) in the control group.

Quality of Life
Quality of life measurements showed improvements in both groups (F = 189; p < .001),as well as the protocol-time intersection (F = 39.6;p < .001).Statistically significant intergroup differences were observed in the mean difference obtained from the assessments at week 4, week 8, and week 10, with significantly greater improvements in the experimental group compared to the control group at all time points (Table 2).The protocol showed significant results for the experimental group (F = 8.5; p = .009).Changes in the quality of life in the experimental group were observed in all pairwise comparisons (p < .001):week 0 to 4: 4.6; week 4 to 8: 8.5; and week 8 to 10: 4.4.Moreover, the greatest improvement following the application of both protocols was obtained after the mCIMT period, with 13.1 points in the experimental group (week 0 to week 8: 80 hours), and 6.3 points in the control group from week 8 to week 10 (20 hours of mCIMT) (Figure 4).Significant clinical gains (>3.5 points) were found in all participants in all assessments in the case of the experimental group, and at week 10 in the control group.

DISCUSSION
The objective of this study was to determine whether the order of application and the dose of mCIMT within a combined (hybrid) protocol applied at home influences the results of bimanual functional performance of the affected upper limb and the quality of life of children with congenital hemiplegia.The application of 80 hours of mCIMT in the experimental group represented the highest increase in the bimanual functional performance and continued to increase during the 20 hours of BIT.In the control group, the improvement in the affected upper limb's functionality was obtained after 20 hours CIMT is most effective in improving hand function in children with low or very low bimanual function and should be included in the combined therapies even at a low dose.The findings reported by Geerdink et al., 29 Reidy et al., 36 and Deepe et al. 23 are similar to those observed in the experimental group.However, in the study by Geerdink et al., 29 children 5 to 8 years of age reached their maximum performance while receiving BIT, after receiving the 72-hour dose of combined CIMT and BIT.This contrasts the outcomes observed in our study for the experimental group, where the largest functional gains were observed during the application of mCIMT, which further improved during the application of BIT (20 hours).This difference could be the result of the higher doses of mCIMT (26 more hours) delivered in this study.The study by Reidy et al., 36 which combined an intensive therapy protocol of 114 hours of mCIMT followed by 12 hours of BIT, obtained similar results.However, the experimental group in our study scored 12.5 units higher than the children in the study by Reydi et al., 36 despite the use of a lower dose of mCIMT.Lesion etiology could potentially explain this difference, as in our study the etiology was congenital hemiplegia, whereas heterogeneous etiologies were present in the children included in the study by Reidy et al. 36 The study by Deepe et al. 23 applied intensive therapy combining 60 hours of mCIMT and 20 hours of BIT, achieving an increase of 9.1% in bimanual functional performance in contrast with the 34% improvement observed in this study in the experimental group.This difference could be due to the higher dose of applied mCIMT (80 hours).As in the case of our control group, the study by Deepe et al. 23 did not find a statistically significant relationship between the protocol and bimanual functional performance after the exclusive application of 60 hours of BIT.Most children with congenital hemiplegia manifest functional difficulties with deficiencies in the affected upper limb, including problems in activities of daily living, which limit their quality of life. 37To date, we have not found any studies that assess quality of life in children with low bimanual functional performance who participated in intensive therapy programs.When there is significant disuse and lack of integration of the affected upper limb, the quality of life of children with hemiplegia decreases, thereby affecting the execution of activities of daily living, 38 such as eating, school, and hygiene activities, etc., which improved after mCIMT application for both groups, with the experimental group also acquiring increases with the BIT.Children with low/very low bimanual functional performance have no experience using their affected arm, 39 and they develop strategies with the dominant hand to avoid the use of the affected hand.Therefore, performing bimanual activities at the beginning of a combined protocol may be less beneficial due to the lack of spontaneous use of the affected upper limb.This means that continuous verbal and physical feedback is needed from the adult  (parents, therapist) to involve the hand in activities with a bimanual component and, therefore, a reduction in learning, since there is more interaction of the unaffected hand.Thus our study demonstrates that the mCIMT should be applied at the beginning of combined protocols to increase the one-handed use of the affected upper limb, allowing the child to learn how to use their arm through trial and error and to acquire functional strategies for reaching, grasping, and increasing manipulative dexterity. 40,41This results in an increased use of the affected hand that subsequently allows for greater participation in bimanual activities without verbal or physical reference for its use and, therefore, greater learning based on a "hands-off" approach to allow the child to learn using intrinsic feedback or information provided by sensory systems during repetitive movements. 42The article by Gordon, 43 entitled "To constrain or not to constrain, and other stories of intensive upper extremity training for children with unilateral cerebral palsy," reported that constraining the unaffected upper limb has limitations when applied in the pediatric population and is insufficient for the attainment of bimanual functional gains, concluding that the BIT program is a better treatment.Our results from the control group suggest that this hypothesis may not be valid for children with low/very low bimanual functional performance, since no improvements in the affected hand functionality were observed during the application of 80 hours of BIT, prior to receiving mCIMT.
The implementation of both combined intensive therapies at home was possible due to the great involvement of the families and children in the treatment, with no follow-up drop-outs.Parents and health professionals should learn from each other and share each other's perspectives on the rehabilitation of the child.Home-based programs may be the preferred or even the only feasible option in specific contexts, for example, in cases of long commutes from the child's home to the institution, 44 reducing the hours of direct intervention (clinical environment) of the health professional due to follow-up care for the family and the child, since the family is empowered 45 and capable of redirecting strategies within the home through the support of the professional.The execution of previous training for therapy at home, deciding on the activities and times of the day to do them, as well as the weekly follow-up, favor the family's adherence to the intervention. 46eaknesses of our study include the use of a small sample and the lack of long-term follow-up of results.Future research should include long-term follow-up and extend the age of the participants beyond 8 years.
With respect to strengths, this study allows us to consider mCIMT as the main therapy and, as such, to apply it prior to the BIT in children with hemiplegia who have significant disuse in their affected upper limb.It shows the importance, for the clinical practice, of determining the baseline bimanual functional performance level before starting an intensive therapy, since decision-making and the success of the intensive therapies would be related to the bimanual functional performance at baseline situation.

CONCLUSIONS
In this study, mCIMT was more beneficial than BIT for the improvement of upper limb functionality and quality of life in children with congenital hemiplegia showing low/very low bimanual functional performance.
The order and the dose of mCIMT in combined protocols influence the increase of the bimanual functional performance and quality of life in children (5-8 years) with congenital hemiplegia showing low/very low bimanual performance.It would be appropriate to start the combined protocol with a high dose of mCIMT followed by a low dose of BIT.

F
I G U R E 1 Partial restraint of unaffected hand.Freedom of movement at the wrist joint was maintained to avoid complications in case of losing balance.

F I G U R E 2
Evaluation time in experimental and control groups.mCIMT: modified Constraint-Induced Movement Therapy, BIT: Bimanual Intensive Therapy.

F I G U R E 4
Evolution of bimanual functional performance and quality of life from baseline to week 10 in the experimental and control groups.(a) Increases in bimanual functional performance are observed at all evaluation time points (week 0-10) in the experimental group, whereas the highest increase in the control group is observed between weeks 8 and 10 (20 hours of modified constraint-induced movement therapy).(b) Similar results are found for quality of life in both groups.A repeated measures analysis of variance (ANOVA) was conducted for bimanual functional performance and quality of life at four time points.The Bonferroni correction was employed for pairwise post hoc comparisons (n = 6) to further analyze significant interactions.
T A B L E 1 Baseline descriptive characteristics of participants.
T A B L E 2 Mean inter-group difference (95% CI) at each assessment time-point (AHA scale and ICP module of the PedsQL3.0Questionnaire).: BFP, bimanual functional performance; AHA, assisting hand assessment; SD, standard deviation; QoL, quality of life; PedsQL 3.0 questionnaire, ICP, pediatric quality of life questionnaire in infantile cerebral palsy; Experimental Group: 80 h mCIMT+20 h BIT.Control Group: 80 h BIT+ 20 h mCIMT.