Scholarly article on topic 'Constraint-induced movement therapy improves upper limb activity and participation in hemiplegic cerebral palsy: a systematic review'

Constraint-induced movement therapy improves upper limb activity and participation in hemiplegic cerebral palsy: a systematic review Academic research paper on "Psychology"

CC BY-NC-ND
0
0
Share paper
Academic journal
Journal of Physiotherapy
OECD Field of science
Keywords
{"Cerebral palsy" / "Systematic review" / Meta-analysis / "Randomised controlled trials" / "Constraint-induced movement therapy"}

Abstract of research paper on Psychology, author of scientific article — Hsiu-Ching Chiu, Louise Ada

Abstract Questions: Does constraint-induced movement therapy improve activity and participation in children with hemiplegic cerebral palsy? Does it improve activity and participation more than the same dose of upper limb therapy without restraint? Is the effect of constraint-induced movement therapy related to the duration of intervention or the age of the children? Design: Systematic review of randomised trials with meta-analysis. Participants: Children with hemiplegic cerebral palsy with any level of motor disability. Intervention: The experimental group received constraint-induced movement therapy (defined as restraint of the less affected upper limb during supervised activity practice of the more affected upper limb). The control group received no intervention, sham intervention, or the same dose of upper limb therapy. Outcome measures: Measures of upper limb activity and participation were used in the analysis. Results: Constraint-induced movement therapy was more effective than no/sham intervention in terms of upper limb activity (SMD 0.63, 95% CI 0.20 to 1.06) and participation (SMD 1.21, 95% CI 0.41 to 2.02). However, constraint-induced movement therapy was no better than the same dose of upper limb therapy without restraint either in terms of upper limb activity (SMD 0.05, 95% CI –0.21 to 0.32) or participation (SMD –0.02, 95% CI –0.34 to 0.31). The effect of constraint-induced movement therapy was not related to the duration of intervention or the age of the children. Conclusions: This review suggests that constraint-induced movement therapy is more effective than no intervention, but no more effective than the same dose of upper limb practice without restraint. Registration: PROSPERO CRD42015024665. [Chiu H-C, Ada L (2016) Constraint-induced movement therapy improves upper limb activity and participation in hemiplegic cerebral palsy: a systematic review. Journal of Physiotherapy 62: 130–137]

Academic research paper on topic "Constraint-induced movement therapy improves upper limb activity and participation in hemiplegic cerebral palsy: a systematic review"

Journal of

PHYSIOTHERAPY

journal homepage: www.elsevier.com/locate/jphys

Research

Constraint-induced movement therapy improves upper limb activity and participation in hemiplegic cerebral palsy: a systematic review

Hsiu-Ching Chiua, Louise Adab

a Department of Physical Therapy, I-Shou University, Kaohsiung, Taiwan (ROC); b Discipline of Physiotherapy,The University of Sydney, Sydney, Australia

KEY WORDS

ABSTRACT

Cerebral palsy Systematic review Meta-analysis

Randomised controlled trials Constraint-induced movement therapy

Questions: Does constraint-induced movement therapy improve activity and participation in children with hemiplegic cerebral palsy? Does it improve activity and participation more than the same dose of upper limb therapy without restraint? Is the effect of constraint-induced movement therapy related to the duration of intervention or the age of the children? Design: Systematic review of randomised trials with meta-analysis. Participants: Children with hemiplegic cerebral palsy with any level of motor disability. Intervention: The experimental group received constraint-induced movement therapy (defined as restraint of the less affected upper limb during supervised activity practice of the more affected upper limb). The control group received no intervention, sham intervention, or the same dose of upper limb therapy. Outcome measures: Measures of upper limb activity and participation were used in the analysis. Results: Constraint-induced movement therapy was more effective than no/sham intervention in terms of upper limb activity (SMD 0.63,95% CI 0.20 to 1.06) and participation (SMD 1.21, 95% CI 0.41 to 2.02). However, constraint-induced movement therapy was no better than the same dose of upper limb therapy without restraint either in terms of upper limb activity (SMD 0.05,95% CI -0.21 to 0.32) or participation (SMD -0.02, 95% CI -0.34 to 0.31). The effect of constraint-induced movement therapy was not related to the duration of intervention or the age of the children. Conclusions: This review suggests that constraint-induced movement therapy is more effective than no intervention, but no more effective than the same dose of upper limb practice without restraint. Registration: PROSPERO CRD42015024665. [Chiu H-C, Ada L (2016) Constraint-induced movement therapy improves upper limb activity and participation in hemiplegic cerebral palsy: a systematic review. Journal of Physiotherapy 62: 130-137]

© 2016 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Introduction

Cerebral palsy is a non-progressive neurological condition resulting in motor impairments that can change over time.1 The impairments may originate directly from damage to an immature brain, or indirectly from compensatory movements or disuse during development.1 Such impairments may result in activity limitations that require rehabilitation throughout life.1 Among children with cerebral palsy, 29% have hemiplegia, that is, one side of the body is affected much more than the other, and the upper limb is typically more involved than the lower limb.2 They may develop 'learned non-use' in their affected upper limb, because they tend to learn alternative strategies to manage daily tasks using the less affected limb.3-5 Performance of tasks is often more efficient using the less affected upper limb, even if there is only mild impairment in the more affected limb.3 Children with hemiplegic cerebral palsy usually have the intellectual capacity to attend regular schools, yet impaired upper limb function tends to restrict their participation in education and leisure, and impact their social image.

Therapists working with children with hemiplegic cerebral palsy encourage movement of the affected limb by repetitive practice of unilateral and bimanual activities. Constraint-induced movement therapy (CIMT) aims to overcome 'learned non-use' by

intensive, targeted practice with the more affected limb during restraint of the less affected limb.5 While restrained, only the affected upper limb can be used to carry out activities, forcing children to find solutions to their movement problems.4,5

There are four systematic reviews specifically examining the effect of CIMT in children with cerebral palsy or hemiplegia from other causes.5-8 Two of the reviews included all published studies, regardless of design, and included low levels of evidence such as case studies.5,8 The Cochrane review on this topic has not been updated since 2007 and includes only three randomised trials. These three trials were not pooled into a meta-analysis but the authors concluded that there was a trend towards a beneficial effect of CIMT.7 The most recent review6 to focus on CIMT reported a standardised effect size of 0.55 from the pooled estimate of 27 randomised trials of CIMT versus conventional therapy. One of the post-hoc analyses carried out was to divide the trials on the basis of the equivalence of dose of intervention. When CIMT was compared with a dose-equivalent intervention, the effect was much smaller (SMD 0.37) than the effect among trials without a dose-equivalent comparison group (SMD 0.84). These results give insight into the mechanism of CIMT. The effect of CIMT may be due to nothing more than the large amounts of practice that restraint of the less affected upper limb produces.

http://dx.doi.org/10.1016/jjphys.2016.05.013

1836-9553/© 2016 Australian Physiotherapy Association. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http:// creativecommons.org/licenses/by-nc-nd/4.0/).

In order to fully investigate the effect of CIMT on children with hemiplegic cerebral palsy, trials where CIMT is compared with no intervention need to be pooled separately from trials where CIMT is compared with the same dose of practice without restraint of the unaffected limb. The present systematic review therefore took this approach. In addition, this review examined outcomes at the level of activity and participation, because not only is the effect of CIMT on upper limb activity of interest, but also how improved activity might translate into the broader context of using the upper limb to participate at home, at school and in the community. This review also sought to determine whether the amount of benefit obtained from CIMT is associated with certain characteristics of the children or the CIMT.

Therefore, the research questions for this systematic review were:

1. Does CIMT improve activity and participation in children with hemiplegic cerebral palsy?

2. Does CIMT improve activity and participation more than the same dose of upper limb therapy without restraint?

3. Is the effect of CIMT related to the duration of intervention or the age of the child?

Method

Identification and selection of studies

Searches were conducted of Medline (1966 to June 2015), CINAHL (1982 to June 2015), PubMed (1966 to June 2015), Embase (1974 to June 2015), the Cochrane Library (1966 to June 2015), Web of Science (1945 to June 2015) and the Physiotherapy Evidence Database (PEDro) (to June 2015), without language restrictions using words related to cerebral palsy and randomised controlled trials and words related to constraint-induced movement therapy (such as constraint-induced movement therapy, forced and massed practice) (see Appendix 1 for full search strategy). Titles and abstracts were displayed and screened by one reviewer to identify relevant studies. Full-text copies of relevant studies were retrieved and their reference lists were screened. The methods of the retrieved papers were screened independently by two reviewers against the inclusion criteria: randomised or quasi-randomised trials; children or adolescents (< 18 years of age) with hemiplegic cerebral palsy; experimental intervention of CIMT; control intervention of no/sham intervention or same dose of upper limb therapy; and outcome measure(s) of activity or participation (Box 1).

Assessment of characteristics of studies

Quality

The quality of included studies was assessed by extracting PEDro scores from the PEDro website. Each score on the PEDro website is generated by two accredited raters scoring the trial, with any discrepancies in rating resolved by a third accredited rater.

Participants

Studies involving participants of either gender, regardless of the level of initial disability, were included. The Manual Ability Classification System was used to quantify the severity of upper limb disability. The Manual Ability Classification System classifies how children with cerebral palsy use their hands to handle objects in daily activities, with I = minor limitations and V = severe limitations.9 Age and Manual Ability Classification System level were recorded so that the similarity of participants between studies could be examined. If the Manual Ability Classification System level was not reported, reviewers classified the participants based on the available information.

Intervention

The experimental group had to have received CIMT (defined as restraint of the less affected upper limb during task practice of the more affected upper limb). To be eligible to answer the first study question, the control group had to receive no/sham intervention, defined as usual therapy < 20% of the time that the experimental group spent restrained. To be eligible to answer the second study question, the control group received the same dose of upper limb therapy (unilateral or bilateral or both), defined as equal to or greater than the time that the experimental group spent restrained. Participants could be receiving other therapy as long as both groups received it. The frequency and duration of the intervention was recorded so that the similarity of intervention between studies could be examined.

Outcome measures

Measures that reflected upper limb activity and participation were used in the analysis. Upper limb activity was measured as what the child could do with their more affected limb. Therefore, measures using direct observation of unimanual performance of standardised upper limb tasks, such as Jebsen-Taylor Test of Hand Function, Nine-Hole Peg Test or Bruininks-Oseretsky Test of Motor Proficiency, Quality of Upper Extremity Skills Test or Melbourne Assessment of Unilateral Upper Limb Function, were used and reported as either level of difficulty or time taken. Upper limb participation was measured as what the child did in real life. Therefore, measures using direct observation or parent perception of bimanual real-life play, such as the Assisting Hand Assessment or Pediatric Motor Activity Log, were used and reported as level of difficulty.

Data analysis

Data were extracted from the included studies by one reviewer and cross-checked by a second reviewer. Information about the method (ie, design, participants, intervention and measures) and outcome data (ie, number of participants, mean (SD) activity and participation) were extracted. Authors of papers with missing data were contacted.

Most studies reported post-intervention scores immediately after intervention; therefore, these scores were used to obtain the pooled estimate of the effect of intervention. Since different measurement tools were used, the effect size was reported as Cohen's standardised mean difference (SMD, 95% CI). A random-effects model was used. The analyses were performed using MIX 2.0, which is a statistical add-in for performing meta-analysis in Excel.10'11

Simple linear regression was used to determine the association between the duration of CIMT and the effect of CIMT (on activity

Box 1. Inclusion criteria.

Design

• randomised or quasi-randomised trial

Participants

• children (ie, < 18 years old)

• hemiplegic cerebral palsy

• any level of disability

Intervention

• constraint-induced movement therapy (ie, restraint of the less affected limb) applied during supervised activity practice of the more affected upper limb

Outcome measures

• measures of activity or participation

Comparisons

• constraint-induced movement therapy vs no/sham intervention (sham defined as usual therapy < 20% of time restrained)

• constraint-induced movement therapy vs same dose of upper limb therapy (defined as > time restrained)

Figure 1. Flow of studies through the review.

RCT = randomised controlled trial, CIMT = constraint-induced movement therapy, mCIMT = modified constraint-induced movement therapy.

and participation) and as well as between age and the effect of CIMT (on activity and participation). Only the no/sham treatment-controlled studies were used in this analysis.

Results

Flow of studies through the review

The search strategy identified 597 studies. After screening titles and abstracts, 46 full papers were retrieved. After being assessed against the inclusion criteria, 31 papers12-42 of 21 studies12'14-18'21-24'27'28'30'31'36-42 were included in the review (Figure 1; see Appendix 2 for excluded papers).

Characteristics of included studies

The 21 studies provided 22 comparisons that were relevant to this review because one study had three arms (reported in Fedrizzi et al 201 324 and Facchin et al 201125). Among the 22 comparisons, 15 investigated CIMT versus no/sham intervention12'15-18'22-24'30'36-40'42 and seven investigated CIMT

versus the same dose of upper limb therapy.14,21,24,27,28,31,41

A summary of the studies is presented in Table 1. Quality

The mean PEDro score of the papers was 5.8 (range 3 to 8) (Table 2). The majority of the papers: were randomised (100%), analysed the between-group difference (97%), reported point estimates and variability (87%), had similar groups at baseline (84%), reported < 15% loss to follow-up (74%) and had blinded assessors (65%). The majority of studies did not conceal the allocation list (58%), carry out an intention-to-treat analysis (65%), nor blind participants or therapists (100%).

Participants

Participants were children and adolescents who were classified as having hemiplegic cerebral palsy, with the mean age across studies ranging from 2.4 to 10.2 years. Fourteen studies (63%) involved participants aged < 4 years. Most of the studies

investigated children classified as Manual Ability Classification System level I/II/III in the more affected limb (ie, able to handle objects without or with adaptive behaviour).

Intervention

The experimental group received CIMT with supervised upper limb practice (22 comparisons). The types of restraint included slings (seven studies), splints (four studies), gloves (eight studies), casts (two studies), and bandage (one study). Restraint was worn for 35 hours/week (range 2 to 84) for a duration of 5 weeks (range 2 to 10) in the comparisons with no/ sham intervention and 23 hours/week (range 12 to 30) for a duration of 5 weeks (range 2 to 10) in the comparisons with the same dose of upper limb therapy. Supervised practice was undertaken for 50% of the time that the restraint was worn in the comparisons with no/sham intervention and 100% of the time that restraint was worn in the comparisons with the same dose of upper limb therapy. The control group received no/sham intervention (15 comparisons) or the same dose of upper limb therapy (seven comparisons). Sham intervention was a small amount (range 0.3 to 2.2 hours/week) of usual therapy, which was not necessarily specific to the upper limb and often not specified. The same dose of upper limb therapy usually involved bilateral training. Both groups received usual therapy in six comparisons.

Outcome measures

Measures of upper limb activity were reported in 19 studies. The measures chosen for the analysis of upper limb activity were: Jebsen-Taylor Test of Hand Function (four studies), Bruininks-Oseretsky Test of Motor Proficiency (one study), Quality of Upper Extremity Skills Test (six studies), Nine-Hole Peg Test (one study), Melbourne Assessment of Unilateral Upper Limb Function (three studies), Box and Block Test (two studies) and Pediatric Arm Function Test (two studies).

Measures of upper limb participation were reported in 12 studies. The measures chosen for the analysis of upper limb participation were: Assisting Hand Assessment (eight studies), Pediatric Motor Activity Log (three studies) and Caregiver Functional Use Survey (one study).

Effect of constraint-induced movement therapy versus no/sham intervention

Activity

The immediate effect of CIMT compared with no/sham intervention on activity was examined by pooling post-intervention data from 11 comparisons with a PEDro score of 5.5 and 302 participants, using a random-effects model. Four studies were unable to be included in the analysis because they had no appropriate measure of activity13'22'40 or because of missing data.38 CIMT increased activity (SMD 0.63, 95% CI 0.20 to 1.06) compared with no/sham intervention (Figure 2). See Figure 3 on the eAddenda for the detailed forest plot. There was substantial statistical heterogeneity (I2 = 65%), indicating that the variation between the results of the trials was above the variation expected by chance. Sensitivity analyses revealed that the heterogeneity was not explained by the quality of the trials, assessor blinding, number or severity of participants.

Participation

The immediate effect of CIMT compared with no/sham intervention on participation was examined by pooling postintervention data from eight comparisons with a PEDro score of 5.5 and 215 participants, using a random-effects model. Seven studies were unable to be included in the analysis because they had no appropriate measure of participation.1718'23'24'36'37'42 CIMT increased participation (SMD 1.21, 95% CI 0.41 to 2.02) compared with no/sham intervention (Figure 4). See Figure 5 on the eAddenda for the detailed forest plot. There was

Table 1

Characteristics of included studies (n = 21 studies across 31 papers, with 22 comparisons).

Study Design Participants Intervention Outcome measures

Experimental Control Both

CIMT versus no/sham intervention

Aarts et al 201012 RCT n = 50 Restraint» sling Usual therapy • Activity: MAUULF

Aarts et al 201113 Mean age (range) = 2.9 yr 3 h/d x 3/wk x 6 wk 1.5 h/wk x 8 wk (0 to 100)

Geerdink (2.5 to 8) (Total: 54 h) (Total: 12 h) • Participation: AHA

et al 201326 Classification = hemiplegia, Super=3 h/d x 3/wk (0 to 100)

MACS I to III x 6 wk+ Bimanual • Timing = 0, 9, 26, 52 wk

practice

3 h/d x 3/wk x 2 wk

(Total: 72 h)

Al-Oraibi RCT n= 14 Restraint» glove Usual therapy • Participation: AHA

et al 201115 Mean age (range)=4.8 yr 2 h/d x 6 or 7/wk x 2 h/wk x 8 wk (0 to 100)

(1.8 to 9) 8 wk (Total: 92 h) (Total: 16 h) • Timing = 0, 8 wk

Classification = hemiplegia, Super=2 h/d x 6/wk

MACSa I to V x 8 wk (Total: 92 h)

Charles RCT n = 22 Restraint» sling No intervention • Activity: JTTHFb (s)

et al 200616 Mean age (range) = 6.7 yr 6 h/d x 5/wk x 2 wk • Participation: CFUS-freq

(4 to 8) (Total: 60 h) (0 to 5)

Classification = spastic Super= 7 h/d x 5/wk • Timing = 0, 3 wk

hemiplegia, MACSa I to III x 2 wk (Total: 70 h)

Choudhary RCT n = 31 Restraint» sling No intervention Usual therapy • Activity: QUEST-grasp

et al 201317 Mean age (range) = 5.1 yr 3 h/d x 2-3/wk x 0.3 h/d x 7/ (0 to 100)

(3 to 8) 4 wk + 2 h/d x 4-5/ wk x 4 wk • Timing = 0, 4, 12 wk

Classification = hemiplegia, wk x 4 wk (Total: (Total: 8.5 h)

MACSa I to III 66 h)

Super=2 h/d x 2-3/

wk x 4 wk (Total:

De Brito Brandao RCT n= 16 Restraint» sling Usual therapy • Activity: JTTHFb (s)

et al 201018 Mean age (range) = 6 yr 10 h/d x 5/wk x 2 wk 0.75 h/wk x 3 wk • Timing = -1, 3, 7 wk

(4 to 8) (Total »100 h) (Total »2 h)

Classification = spastic Super»3 h/d x 5/wk

hemiplegia, MACS I to III x 2 wk+ bimanual

practice

0.75 h/d x 3/wk x

1 wk (Total: 32 h)

Eliasson et al 201122 Cross-over n = 25 Restraint »glove No intervention Usual therapy • Participation: AHA

RCT Mean age (range) = 2.4 yr 2 h/d x 7/wk x 8 wk (0 to 100)

(1.5 to 5) (Total: 102 h) • Timing = 0, 8 wk

Classification = hemiplegia, Super»2 h/d x 7/wk

MACSa I to V x 8 wk (Total: 102 h)

Eugster-Buesch RCT n = 23 Restraint» splint No intervention Usual therapy • Activity: MAUULF

et al 201223 Mean age (range) = 10.7 yr 6 h/d x 7/wk x 2 wk (0 to 100)

(6 to 16) (Total: 84 h) • Timing = 0, 2, 12, 52 wk

Classification = hemiplegia, Super»2 h/d x 7/wk

MACSa I to III x 2 wk (Total: 28 h)

Fedrizzi RCT n = 48 Restraint »glove Usual therapy • Activity: QUEST-grasp (%)

et al 201324 Mean age (range)=4.3 yr 3 h/d x 7/wkx 10 wk 1-2/wk x 10 wk • Timing = 0, 10, 12, 26 wk

Facchin (2 to 8) (Total: 210 h) (Total: 15 h)

et al 201125 Classification = hemiplegia, Super»3 h/d x 7/wk

MACSa I to V x 10 wk (Total:

210 h)

Rostami RCT n= 16 Restraint» splint No intervention Usual therapy • Activity: BOTMP-subtest

et al 201230 Mean age (range) = 8.2 yr 5 h/d x 7/wk x 4wk 0.5 h/d x 2/ 8 (0 to 9)

(6 to 12) (Total: 140 h) wk x 4 wk • Participation: PMAL-quality

Classification = spastic Super » 1.5 h/d x 3/wk (Total: 4 h) (0 to 5)

hemiplegia, MACSa I to III x 4 wk (Total: 18 h) • Timing = 0, 4, 12 wk

Smania Cross-over n=10 Restraint »glove Usual therapy • Activity: PAFT (0 to 120)

et al 200936 RCT Mean age (range) = 3.3 yr 8 h/d x 7/wk x 5 wk 1 h/d x 2/wk x 5 wk • Timing = 0, 5 wk

(1 to 9) (Total: 280 h) (Total: 10 h)

Classification = hemiplegia, Super» 1 h/d x 2/wk

MACSa I to III x 5 wk (Total: 10 h)

Sung et al 200537 RCT n=31 Restraint» cast Usual therapy • Activity: BBT (blocks)

Mean age (range) = 3.1 yr 12 h/d x 7/wk x 6 wk 0.5 h/d x 2/wk x 6 wk • Timing = 0, 6 wk

(< 8) (Total: 500 h) (Total: 6 h)

Classification = hemiplegia, Super»0.5 h/d x 2/

MACSa I to III wk x 6 wk (Total:

Taub RCT n= 18 Restraint= cast Usual therapy • Activity: QUEST (%)

et al 200438 Mean age (range) = 3.5 yr 12 h/dx 7/wkx3 wk 2.2 h/wk x 3 wk • Participation: PMAL-quality

DeLuca (0.5 to 8) (Total: 250 h) (Total = 7 h) (0 to 5)

et al 200620 Classification = hemiplegia, Super=6 h/d x 7/wk • Timing = 0, 3 wk

MACSa I to V x3wk (Total: 125 h)

Table 1 (Continued )

Design

Participants

Intervention

Outcome measures

Experimental

Control

Taub et al 201139 RCT n = 20

Mean age (range) = 3.7 yr (2 to 6)

Classification = hemiplegia, MACSa I to V

Restraint» splint 12 h/d x 5/wkx 3 wk (Total: 90 h) Super = 6 h/d x 5/wk x 3 wk (Total: 90 h)

No intervention

Usual therapy 1.5 hr/wk x 3 wk (Total: 4.5 h)

. Activity: PAFT (0 to 64) I Participation: PMAL-quality

(0 to 5) • Timing = 0, 3 wk

Wallen et al 201140

RCT n = 50

Mean age (range) = 3.1 yr (1.5 to 8)

Classification = spastic hemiplegia, MACS I to IV

Restraint = glove 2 h/d x 7/wk x 8 wk (Total: 112 h) Super = 2 h/d x 7/wk x 8 wk (Total: 112 h)

Usual therapy 0.3 h x 7/wk x 8 wk (Total: 17 h)

• Participation: AHA (0 to 100)

• Timing = 0, 10, 26 wk

Yu et al 20124

RCT n = 20

Mean age (range) = 9.4 yr (9 to 10)

Classification = hemiplegia, MACSa I to III

Restraint» sling 1 h/d x 2/wk x 10 wk (Total: 20 h) Super = 1 h/d x 2/wk x 10 wk (Total: 20 h)

No intervention

Usual therapy 0.5 h/d x 2/wk x 10 wk (Total: 10 h)

. Activity: BBT (blocks) Timing = 0, 10 wk

CIMT versus same dose of upper limb therapy

Abd el-Kafy et al 201414

RCT n=27

Mean age (range) = 6.1 yr (4 to 8)

Classification = spastic hemiplegia, MACS II to IV

Restraint = sling Bimanual practice

6 h/d x 5/wk x 4 wk 6 h/d x 5/wk x 4 wk

(Total: 120 h) (Total: 120 h) Super = 6 h/d x 5/wk x4 wk (Total: 120 h)

. Activity: QUEST (%) Timing = 0, 4, 12 wk

Deppe et al 201321

RCT n = 29

Mean age (range) = 6.3 yr (3.3 to 12)

Classification = hemiplegia, MACS I to III

Restraint = bandage 4 h/d x 5/wk x 3 wk (Total: 60 h) Super = 4 h/d x 5/wk x 3 wk

+ bimanual practice 4 h/d x 5/wk x 1 wk (Total: 80 h)

Bimanual practice 4 h/d x 5/wk x 4 wk (Total: 80 h)

. Activity: MAUULF (0 to 122) Participation: AHA (22 to 88) Timing = 0, 4 wk

Fedrizzi et al 201324 Facchin et al 201125

RCT n = 48

Mean age (range) = 4 yr (2 to 8)

Classification = hemiplegia, MACSa I to V

Restraint = glove 3 h/d x 7/wkx 10 wk (Total: 210 h) Super = 3 h/d x 3/wk x 10 wk (Total: 210 h)

Bimanual practice 3 h/d x 7/wk x 10 wk (Total: 210 h)

. Activity: QUEST-grasp (%) 1 Timing = 0, 10, 26 wk

Gelkop et al 20 1 527 Cross-over n=12

RCT Mean age (range) = 4.3 yr

(1.5 to 7)

Classification = spastic hemiplegia, MACS I to III

Restraint = glove 2 h/d x 6/wk x 8 wk (Total: 96 h) Super = 2 h/d x 6/wk x 8 wk (Total: 96 h)

Bimanual practice 2 hx 6/wk x 8 wk (Total: 96 h)

Activity: QUEST-grasp (%) Participation: AHA (0 to 100) Timing = 0, 8 wk

Gordon et al 201128 De Brito Brandao et al 201219 Hunget al 201129

RCT n = 42

Mean age (range) = 6.3 yr (3.5 to 10)

Classification = hemiplegia, MACS I to III

Restraint = sling 6 h/d x 5/wk x 3 wk (Total: 90 h) Super = 6 h/d x 5/wk x 3 wk (Total: 90 h)

Bimanual practice 6 h/d x 5/wk x 3 wk (Total: 90 h)

. Activity: JTTHFb (s) Participation: AHA (logits) Timing = 0, 3 wk

Sakzewski et al 2011a31 Sakzewski et al 2011b32 Sakzewski et al 2011c33 Sakzewski et al 2011d34 Sakzewski et al 201235

RCT n = 63

Mean age (range) = 10.2 yr (5 to 16)

Classification = spastic hemiplegia, MACS I to III

Restraint = glove 6 h/d x 5/wk x 2 wk (Total: 60 h) Super = 6 h/d 5/wk x 2 wk (Total: 60 h)

Bimanual practice 6 h/d x 10 days (Total: 60 h)

Activity: JTTHFb (s) Participation: AHA (0 to 100) Timing = 0, 3 wk

Xu et al 201241 RCT n = 45 Restraint = splint Bimanual practice

Mean age (range) = 4.6 yr 3 h/d x 5/wk x 2 wk 3 h/d x 5/wk x 2 wk

(2 to 14) (Total: 30 h) (Total: 30 h)

Classification = hemiplegia, Super = 4 h/d x 5/wk

MACSa I to III x 2 wk (Total: 40 h)

Activity: 9-HPTb (s) Timing = 0, 2 wk

a MACS level estimated by reviewers.

b Experimental and control scores reversed for analysis because a smaller score denotes better performance. AHA=Assisting Hand Assessment, BBT = Box and Block Test, BOTMP = Bruininks-Oseretsky Test of Motor Proficiency, CFUS freq = Caregiver Functional Use Survey, JTFHT=Jebsen-Taylor Test of Hand Function, MACS = Manual Ability Classification System, MAUULF = Melbourne Assessment of Unilateral Upper Limb Function, PAFT = Pediatric Arm Function Test, PMAL=Pediatric Motor Activity Log, QUEST = Quality of Upper Extremity Skills Test, RCT = randomised controlled trial, Super = supervised practice, 9-HPT = Nine Hole Peg Test.

substantial statistical heterogeneity (I2 = 84%), indicating that the variation between the results of the trials was above the variation expected by chance. Sensitivity analyses revealed that the heterogeneity was not explained by the quality of the trials, assessor blinding or the number or severity of participants.

Effect of constraint-induced movement therapy versus same dose of upper limb therapy

Activity

The immediate effect of CIMT compared with the same dose of upper limb therapy on activity was examined by pooling data after

Table 2

PEDro scores for included papers (n = 31).

Study Random Concealed Groups Participant Therapist Assessor < 15% Intention- Between-group Point Total

allocation allocation similar at baseline blinding blinding blinding dropouts to-treat analysis difference reported estimate and variability reported (0 to 10)

Aarts et al 201012 Y N Y N N Y Y N Y Y 6

Aarts et al 201113 Y N Y N N Y Y N Y Y 6

Abd el-Kafy et al 201414 Y N Y N N Y Y N Y N 5

Al-Oraibi et al 201115 Y N N N N Y N N Y Y 4

Charles et al 200616 Y N Y N N Y N N Y Y 5

Choudhary et al 201317 Y Y Y N N Y Y Y Y Y 8

De Brito Brandao et al 201018 Y Y Y N N Y Y Y Y Y 8

De Brito Brandao et al 201219 Y Y Y N N N Y N Y Y 6

De Luca et al 200620 Y N N N N Y Y N Y N 4

Deppe et al 201321 Y Y Y N N Y Y N Y Y 7

Eliasson et al 201122 Y N N N N Y N N Y Y 4

Eugster-Buesch et al 201223 Y Y Y N N Y Y N Y Y 7

Fedrizzi et al 201324 Y N Y N N N Y N Y N 4

Facchin et al 201125 Y N Y N N N Y N Y N 4

Geerdink et al 201326 Y N Y N N N Y Y Y Y 6

Gelkop et al 201527 Y Y Y N N Y Y Y Y Y 8

Gordon et al 201128 Y Y Y N N Y Y N Y Y 7

Hung et al 201129 Y N N N N Y Y N Y Y 5

Rostami et al 201230 Y N Y N N Y Y Y Y Y 7

Sakzewski et al 2011a31 Y Y Y N N N Y Y Y Y 7

Sakzewski et al 2011b32 Y Y Y N N Y Y Y Y Y 8

Sakzewski et al 2011c33 Y Y Y N N N Y Y Y Y 7

Sakzewski et al 2011d34 Y Y Y N N N Y Y Y Y 7

Sakzewski et al 201235 Y Y Y N N N Y Y Y Y 7

Smania et al 200936 Y N N N N Y N N Y Y 4

Sung et al 200537 Y N Y N N N N N Y Y 4

Taub et al 200438 Y N Y N N Y N N Y Y 5

Taub et al 201139 Y N Y N N N Y N Y Y 5

Wallen et al 201140 Y Y Y N N Y Y Y Y Y 8

Xu et al 201241 Y N Y N N Y N N Y Y 5

Yu et al 201242 Y N Y N N N N N N Y 3

Y=yes, N = no, PEDro = Physiotherapy Evidence Database. PEDro scores extracted from website www.pedro.org.au

intervention from five comparisons with a PEDro score of 6.8 and 218 participants, using a random-effects model. Two studies were unable to be included in the analysis because they had no postintervention data41 or missing data.14 CIMT did not increase activity (SMD 0.05,95% CI -0.21 to 0.32,12 = 0%) compared with the same dose of upper limb therapy (Figure 6). See Figure 7 on the eAddenda for the detailed forest plot.

SMD (95% CI) Study Random

Aarts 2010 - -

Charles 2006 —

Choudhary 2013 --

De Brito 2010 --

Eugster 2012 —

Fedrizzi 2013 -

Rostami 2012

Smania 2009 —

Sung 2005 — —

Taub 2011 --

Yu 2012 --

Pooled i i i i i i i i

-2 -1 0 1 2 3 4 5 6 Favours con Favours exp

Figure 2. SMD (95% CI) of effect of constraint-induced movement therapy compared with no/sham intervention on upper limb activity immediately after intervention by pooling data from 11 trials (n = 302) using a random-effects model (I2 = 65%).

Participation

The immediate effect of CIMT compared with the same dose of upper limb therapy on participation was examined by pooling data after intervention from four comparisons with a PEDro score of 7.5 and 146 participants, using a random-effect model. Three studies were unable to be included in the analysis because they had no appropriate participation measure.14,24,41 CIMT did not increase participation (SMD -0.02, 95% CI -0.34 to 0.31, I2 = 0%) compared with the same dose of upper limb therapy (Figure 8). See Figure 9 on the eAddenda for the detailed forest plot.

SMD (95% CI) Study Random

Aarts 2010 --

Al-Oraibi 2011--

Charles 2006 --

Eliasson 2011 --

Rostami 2012 --

Taub 2004 --

Taub 2011

Wallen 2011 --

Pooled i i i i i i i i

i-1--1-1-1-1-1-1

-2 -1 0 1 2 3 4 5 6 Favours con Favours exp

Figure 4. SMD (95% CI) of effect of constraint-induced movement therapy with no intervention on upper limb participation immediately after intervention by pooling data from eight trials (n = 215) using a random-effects model (I2 = 84%).

SMD (95% CI) Random

Deppe 2013--

Fedrizzi 2013 Gelkop 2015 — Gordon 2011

Sakzewski 2011a — —

Pooled

~i-1-1

-2-1 0 12 3 Favours con Favours exp

Figure 6. SMD (95% CI) of effect of constraint-induced movement therapy with same dose of upper limb therapy on upper limb activity immediately after intervention by pooling data from five trials (n = 218) using a random-effect model (I2 = 0%).

SMD (95% CI) Random

Deppe 2013 Gelkop 2015 Gordon 2011 Sakzewski 2011a Pooled

-2 -1 0 1 2 3 Favours con Favours exp

Figure 8. SMD (95% CI) of effect of constraint-induced movement therapy with same dose of upper limb therapy on upper limb participation immediately after intervention by pooling data from four trials (n = 146) using a random-effects model (I2 = 0%).

Relation between the effect and duration of constraint-induced movement therapy and age for activity and participation

There was no significant relation between duration of CIMT (total duration of CIMT) and effect of CIMT on activity (r = -0.25, p = 0.46) or participation (r = -0.10, p = 0.81). Neither was there a significant relation between age and effect of CIMT on activity (r = 0.37, p = 0.26) or participation (r = 0.58, p = 0.13).

Discussion

This systematic review found that CIMT had a beneficial effect compared with no/sham intervention for children with hemiplegic cerebral palsy. Furthermore, the effect was beneficial in terms of both activity and participation, suggesting that the improved upper limb activity carried over into what the children actually did in real life with their upper limb. On the other hand, when CIMT was compared with the same dose of upper limb therapy, there was little effect on activity or participation. Neither duration of CIMT nor age influenced the size of the effect of CIMT.

This review was based on randomised trials of reasonable to good quality. Given that 8 was the likely maximum PEDro score achievable, because it is not possible to blind the therapists or participants during complex interventions such as CIMT, the mean PEDro score of 5.8 for the papers included in this review suggested that the findings were reasonably credible. Although CIMT was

effective, it was no more effective than the same dose of upper limb therapy without restraint. This suggests that the mechanism of the effect is the dose of practice undertaken, rather than the type of practice (ie, CIMT). In the studies where CIMT was compared with no/sham intervention, children in the CIMT group were restrained for an average of 5 hours/day and they spent just over 50% of this period engaged in supervised practice. In the studies where CIMT was compared with the same dose of upper limb practice without restraint, children in the CIMT group were restrained for an average of 4 hours/day and they spent 100% of this period engaged in supervised practice, sometimes carrying out extra unrestrained practice.

The findings from this review are supported by the findings of the only other systematic review to specifically examine CIMT where a meta-analysis was performed.5 In this previous review, when all studies were pooled, CIMT provided a moderate beneficial effect of 0.55. When only the studies of CIMT against a non-equivalent dose intervention were analysed, the estimated effect size was 0.84, which was similar to our estimated effect size of 0.63 for activity and 1.21 for participation. On the other hand, when only the studies of CIMT against an equivalent dose of practice were analysed, the estimated effect size was 0.37, which is larger than our estimated effect size of 0.05 for activity and -0.02 for participation. This may be because Chen et al5 included seven studies in their analysis that were not considered dose equivalent by our definition.1213'26'36'40'43'44 Since the control groups in these studies received less practice than the CIMT groups, this may explain why the effect that Chen et al found was larger than in our review. In a general review, Sazewski et al45 also came to the conclusion that the mechanism of the effect was the dose of practice undertaken, rather than the type of practice.

There were some limitations to this review. First, there were some missing data, so not all the included studies are represented in the final pooled estimate, although this only amounts to 15% of the total data. Second, although a large number of studies were represented, most were of a small sample size. On average, there were 33 participants per study included in the meta-analyses, leaving the review vulnerable to small trial bias. Third, there were high levels of statistical heterogeneity (I2 > 60%) in the analyses of CIMT against no/sham intervention, and the source of this heterogeneity was not obvious.

This review generates several implications for clinical practice with children with hemiplegic cerebral palsy. CIMT is an effective way to improve upper limb function at the activity level and this can be expected to carry over into participation in real life. Given that the same dose of practice without restraint is likely to result in the same outcome, it seems that as long as large amounts of practice are carried out, regardless of whether that is with restraint (unimanual) or without restraint (bimanual and unimanual), improvement will occur. In the studies of CIMT against no/sham intervention, children were supervised to practise using their upper limb for an average of 2.5 hours a day, with a further 2.5 hours of restraint forcing more practice. Ultimately, the way in which practice is achieved may be best chosen by a combination of the child and the parents, as well as the therapists. For example, it may be easier to 'force' practice over long periods of time at home using CIMT than practising without restraint under the supervision of parents.

What is already known on this topic: Children with hemiplegia due to cerebral palsy may have impairments due to damage to an immature brain, indirectly from compensatory movements or from learned disuse. Such impairments may result in limitations in activity requiring rehabilitation throughout life.

What this study adds: Constraint-induced movement therapy is an effective way to improve upper limb function, but as long as large amounts of practice are carried out, regardless of whether that is achieved with or without restraint, this benefit can be expected.

eAddenda: Figures 3, 5, 7 and 9; Appendices 1 and 2, can found at: doi:10.1016/j.jphys.2016.05.013 Ethics approval: Not applicable Competing interests: Nil

Source of support: This research was, in part, supported by I-Shou University, under Grant no. ISU 103-04-06A. Acknowledgement: Nil Provenance: Not invited. Peer reviewed. Correspondence: Hsiu-Ching Chiu, Department of Physical Therapy, I-Shou University, Taiwan (ROC). Email: hsiuchingchiu@ isu.edu.tw

References

1. Wilson Howle J. Cerebral palsy. In: Campbell SK, ed. Decision making in pediatric neurologic physical therapy. New York: Churchill Livingstone; 1999:23-83.

2. Beaman J, Kalisperis FR, Miller-Skomorucha K. The infant and child with cerebral palsy. In: Tecklin J, ed. Pediatric physical therapy 5 ed. Sydney: Lippincott Williams & Wilkins; 2015:187-246.

3. Kuhtz-Buschbeck JP, Sundholm LK, Eliasson AC, Forssberg H. Quantitative assessment of mirror movements in children and adolescents with hemiplegic cerebral palsy. Dev Med Child Neurol. 2000;42:728-736.

4. Taub E, Uswatte G, Pidikiti R. Constraint-Induced Movement Therapy: a new family of techniques with broad application to physical rehabilitation-a clinical review. J Rehabil Res Dev. 1999;36:237-251.

5. Charles J, Gordon AM. A critical review of constraint-induced movement therapy and forced use in children with hemiplegia. Neural Plast. 2005;12:245-261.

6. Chen Y-P, Pope S, Tyler D, Warren GL. Effectiveness of constraint-induced movement therapy on upper-extremity function in children with cerebral palsy: a systematic review and meta-analysis of randomized controlled trials. Clin Rehabil. 2014;28:939-953.

7. Hoare B, Imms C, Carey L, Wasiak J. Constraint-induced movement therapy in the treatment of the upper limb in children with hemiplegic cerebral palsy: a Cochrane systematic review. Clin Rehabil. 2007;21:675-685.

8. Huang HH, Fetters L, Hale J, McBride A. Bound for success: a systematic review of constraint-induced movement therapy in children with cerebral palsy supports improved arm and hand use. Phys Ther. 2009;89:1126-1141.

9. Eliasson AC, Krumlinde-Sundholm L, Rosblad B, Beckung E, Arner M, Ohrvall AM, et al. The Manual Ability Classification System (MACS) for children with cerebral palsy: scale development and evidence of validity and reliability. Dev Med Child Neurol. 2006;48:549-554.

10. Bax L, Yu LM, Ikeda N, Tsuruta H, Moons KG. Development and validation of MIX: comprehensive free software for meta-analysis of causal research data. BMC Med Res Methodol. 2006;6:50.

11. Bax L, Yu LM, Ikeda N, Tsuruta H, Moons KG. MIX: comprehensive free software for meta-analysis of causal research data. Version 2.0. 2014.

12. Aarts PB, Jongerius PH, Geerdink YA, van Limbeek J, Geurts AC. Effectiveness of modified constraint-induced movement therapy in children with unilateral spastic cerebral palsy: a randomized controlled trial. Neurorehabil Neural Repair. 2010;24:509-518.

13. Aarts PB, Jongerius PH, Geerdink YA, van Limbeek J, Geurts AC. Modified Constraint-Induced Movement Therapy combined with Bimanual Training (mCIMT-BiT) in children with unilateral spastic cerebral palsy: how are improvements in arm-hand use established? Res Dev Disabil. 2011;32:271-279.

14. Abd El-Kafy EM, Elshemy SA, Alghamdi MS. Effect of constraint-induced therapy on upper limb functions: a randomized control trial. Scand J Occup Ther. 2014;21: 11-23.

15. Al-Oraibi S, Eliasson AC. Implementation of constraint-induced movement therapy for young children with unilateral cerebral palsy in Jordan: a home-based model. Disabil Rehabil. 2011;33:2006-2012.

16. Charles JR, Wolf SL, Schneider JA Gordon AM, Charles JR, Wolf SL, et al. Efficacy of a child-friendly form of constraint-induced movement therapy in hemiplegic cerebral palsy: a randomized control trial. Dev Med Child Neurol. 2006;48:635-642.

17. Choudhary A, Gulati S, Kabra M, Singh UP, Sankhyan N, Pandey RM, et al. Efficacy of modified constraint induced movement therapy in improving upper limb function in children with hemiplegic cerebral palsy: a randomized controlled trial. Brain Dev. 2013;35:870-876.

18. De Brito Brandao M, Mancini MC, Vaz DV, Pereira de Melo AP, Fonseca ST. Adapted version of constraint-induced movement therapy promotes functioning in children with cerebral palsy: a randomized controlled trial. Clin Rehabil. 2010;24:639-647.

19. De Brito Brandao M, Gordon AM, Mancini MC. Functional impact of constraint therapy and bimanual training in children with cerebral palsy: a randomized controlled trial. Am J Occup Ther. 2012;66:672-681.

20. Deluca SC, Echols K, Law CR, Ramey SL. Intensive pediatric constraint-induced therapy for children with cerebral palsy: randomized, controlled, crossover trial.J Child Neurol. 2006;21:931-938.

21. Deppe W, Thuemmler K, Fleischer J, Berger C, Meyer S, Wiedemann B. Modified constraint-induced movement therapy versus intensive bimanual training for children with hemiplegia - a randomized controlled trial. Clin Rehabil. 2013;27:909-920.

22. Eliasson AC, Shaw K, Berg E, Krumlinde-Sundholm L. An ecological approach of Constraint Induced Movement Therapy for 2-3-year-old children: a randomized control trial. Res Dev Disabil. 2011;32:2820-2828.

23. Eugster-Buesch F, de Bruin ED, Boltshauser E, Steinlin M, Kuenzle C, Muller E, et al. Forced-use therapy for children with cerebral palsy in the community setting: a single-blinded randomized controlled pilot trial. J Pediatr Rehabil Med. 2012;5:65-74.

24. Fedrizzi E, Rosa-Rizzotto M, Turconi AC, Pagliano E, Fazzi E, Pozza LV, et al. Unimanual and bimanual intensive training in children with hemiplegic cerebral palsy and persistence in time of hand function improvement: 6-month follow-up results of a multisite clinical trial. J Child Neurol. 2013;28:161-175.

25. Facchin P, Rosa-Rizzotto M, Visona Dalla Pozza L, Turconi AC, Pagliano E, Signorini S, et al. Multisite trial comparing the efficacy of constraint-induced movement therapy with that of bimanual intensive training in children with hemiplegic cerebral palsy: postintervention results. Am J Phys Med Rehabil. 2011;90:539-553.

26. Geerdink Y, Aarts P, Geurts AC. Motor learning curve and long-term effectiveness of modified constraint-induced movement therapy in children with unilateral cerebral palsy: a randomized controlled trial. Res Dev Disabil. 2013;34:923-931.

27. Gelkop N, Burshtein DG, Lahav A, Brezner A, Al-Oraibi S, Ferre CL, et al. Efficacy of constraint-induced movement therapy and bimanual training in children with hemiplegic cerebral palsy in an educational setting. Phys Occup Ther Pediatr. 2015;35:24-39.

28. Gordon AM, Hung YC, Brandao M, Ferre CL, Kuo HC, Friel K, et al. Bimanual training and constraint-induced movement therapy in children with hemiplegic cerebral palsy: a randomized trial. Neurorehabil Neural Repair. 2011;25:692-702.

29. Hung YC, Casertano L, Hillman A, Gordon AM. The effect of intensive bimanual training on coordination of the hands in children with congenital hemiplegia. Res Dev Disabil. 2011;32:2724-2731.

30. Rostami HR, Arastoo AA, Nejad SJ, Mahany MK, Malamiri RA, Goharpey S. Effects of modified constraint-induced movement therapy in virtual environment on upper-limb function in children with spastic hemiparetic cerebral palsy: a randomised controlled trial. NeuroRehabil. 2012;31:357-365.

31. Sakzewski L, Ziviani J, Abbott DF, Macdonell RA, Jackson GD, Boyd RN. Equivalent retention of gains at 1 year after training with constraint-induced or bimanual therapy in children with unilateral cerebral palsy. Neurorehabil Neural Repair. 2011;25:664-671.

32. Sakzewski L, Ziviani J, Abbott DF, Macdonell RA, Jackson GD, Boyd RN. Randomized trial of constraint-induced movement therapy and bimanual training on activity outcomes for children with congenital hemiplegia. Dev Med Child Neurol. 2011;53:313-320.

33. Sakzewski L, Ziviani J, Boyd RN. Best responders after intensive upper-limb training for children with unilateral cerebral palsy. Arch Phys Med Rehabil. 2011;92:578-584.

34. Sakzewski L, Ziviani J, Abbott DF, Macdonell RA, Jackson GD, Boyd RN. Participation outcomes in a randomized trial of 2 models of upper-limb rehabilitation for children with congenital hemiplegia. Arch Phys Med Rehabil. 2011;92:531-539.

35. Sakzewski L, Carlon S, Shields N, Ziviani J, Ware RS, Boyd RN. Impact of intensive upper limb rehabilitation on quality of life: a randomized trial in children with unilateral cerebral palsy. Dev Med Child Neurol. 2012;54:415-423.

36. Smania N, Aglioti SM, Cosentino A, Camin M, Gandolfi M, Tinazzi M, et al. A modified constraint-induced movement therapy (CIT) program improves paretic arm use and function in children with cerebral palsy. Eur J Phys Rehabil Med. 2009;45:493-500.

37. Sung I, Ryu J, Pyun S, Yoo S, Song W, Park M. Efficacy of forced-use therapy in hemiplegic cerebral palsy. Arch Phys Med Rehabil. 2005;86:2195-2198.

38. Taub E, Ramey SL, DeLuca S, Echols K. Efficacy of constraint-induced movement therapy for children with cerebral palsy with asymmetric motor impairment. Pediatr. 2004;113:305-312.

39. Taub E, Griffin A, Uswatte G, Gammons K, Nick J, Law CR. Treatment of congenital hemiparesis with pediatric constraint-induced movement therapy. J Child Neurol. 2011;26:1163-1173.

40. Wallen M, Ziviani J, Naylor O, Evans R, Novak I, Herbert RD. Modified constraint-induced therapy for children with hemiplegic cerebral palsy: a randomized trial. Dev Med Child Neurol. 2011;53:1091-1099.

41. Xu K, Wang L, Mai J, He L. Efficacy of constraint-induced movement therapy and electrical stimulation on hand function of children with hemiplegic cerebral palsy: a controlled clinical trial. Disabil Rehabil. 2012;34:337-346.

42. Yu J, Kang H, Jung J. Effects of modified constraint induced movement therapy on hand dexterity, grip strength and activities of daily living of children with cerebral palsy: a randomized control trial. J Phys Ther Sci. 2012;24:1029-1031.

43. Chen CL, Kang LJ, Hong WH, Chen FC, Chen HC, Wu CY. Effect of therapist-based constraint-induced therapy at home on motor control, motor performance and daily function in children with cerebral palsy: a randomized controlled study. Clin Rehabil. 2013;27:236-245.

44. Hsin YJ, Chen FC, Lin KC, Kang LJ, Chen CL, Chen CY. Efficacy of constraint-induced therapy on functional performance and health-related quality of life for children with cerebral palsy: a randomized controlled trial. J Child Neurol. 2012;27:992-999.

45. Sakzewski L, Ziviani J, Boyd RN. Efficacy of upper limb therapies for unilateral cerebral palsy: a meta-analysis. Pediatr. 2014;133:e175-e204.

Websites

PEDro www.pedro.org.au