Abstract Background Congenital hemiplegia is the most common form of cerebral palsy accounting for 1 in 1300 live births. Children usually present with greater upper limb than lower limb involvement. Impaired unimanual capacity of the involved upper limb and deficits in bimanual performance contribute to difficulties with day to day activities and participation in home, school and community life. Interventions to address these deficits in upper limb unimanual capacity and bimanual performance have recently shifted focus to address limitations in activity performance rather than underlying impairments. One intensive intervention approach is constraint induced movement therapy, which entails placing a constraint on the unimpaired upper limb to focus intense and repetitive training of the impaired upper limb. To date, it is unclear whether constraint induced movement therapy is superior to a more traditional bimanual therapy to improve activity performance and participation outcomes for children with congenital hemiplegia, as there has been no direct comparison of the two approaches. Aim The primary aim of this research was to determine whether constraint induced movement therapy was more effective than bimanual training to improve activity performance and participation for children with congenital hemiplegia. The specific aims were to: i) determine the efficacy of therapeutic upper limb interventions on activity and participation outcomes for children with congenital hemiplegia, ii) systematically review the clinimetric properties (psychometric properties and clinical utility) of participation assessment tools for children with congenital hemiplegia, iii) examine the relationship between impairments, unimanual capacity and bimanual performance in children with congenital hemiplegia and, iv) determine whether constraint induced movement therapy is more effective than bimanual training to improve activity and participation outcomes for children with congenital hemiplegia. Research Design A matched pairs randomised design was chosen with children matched for age, gender, side of hemiplegia and upper limb function. Children were randomised within pairs to receive either constraint induced movement therapy or bimanual training in equal dosages. Both interventions used a day camp model, with groups receiving the same dosage and content of intervention delivered in the same environment. A novel circus theme was used in the camps to enhance children’s engagement and motivation. Children in the constraint induced movement therapy group wore a tailor made glove on their unimpaired hand during the intervention camp. Outcomes were measured across all domains of the International Classification of Functioning, Disability and Health at baseline, 3 and 26 weeks post intervention. The primary outcome measure for unimanual capacity of the impaired upper limb was the Melbourne Assessment of Unilateral Upper Limb Function, and bimanual performance was the Assisting Hand Assessment. A secondary outcome measure for unimanual capacity was the Jebsen Taylor Test of Hand Function. The Canadian Occupational Performance Measure was used as the primary outcome for participation and three measures, the Assessment of Life Habits, Children’s Assessment of Participation and Enjoyment and the School Function Assessment were included to explore their research utility and responsiveness to change. Results Two systematic reviews were performed prior to the commencement of the randomised trial. The first systematic review and meta-analysis of all upper limb interventions for children with congenital hemiplegia identified four treatment approaches with varying evidence to support their efficacy. Interventions included the use of intramuscular Botulinum toxin A injections to the upper limb augmenting upper limb training, neurodevelopmental treatment, constraint induced movement therapy and hand arm intensive bimanual training. Data were pooled for upper limb, self care and individualised outcomes. Results indicated a small to medium treatment effect favouring all four interventions on upper limb outcomes. Large treatment effects favoured intramuscular Botulinum toxin A injections combined with upper limb training for individualised outcomes. Overall, the systematic review and meta-analysis found no upper limb training approach to be superior although Botulinum toxin A injections appeared to provide a consistent supplementary benefit to a variety of upper limb training approaches. However it was unclear which type of upper limb training was optimal. Findings suggested that the two intensive intervention approaches that are the focus of this randomised controlled trial, constraint induced movement therapy and bimanual intensive training, required further research to support their efficacy. The second systematic review was performed to inform choice of participation measures for the randomised comparison trial. The review identified five specific measures of participation suitable for school aged children with congenital hemiplegia (Assessment of Life Habits, Children’s Assessment of Participation and Enjoyment, School Function Assessment (participation domain), Children Helping Out: Responsibilities and Expectations, School Outcome Measure) and two measures of individualised outcomes that could include specific participation goals (Goal Attainment Scaling and Canadian Occupational Performance Measure). Results suggested that no one measure adequately captured all aspects of participation as outlined in the International Classification of Functioning, Disability and Health, and a combination of assessments would be required to broadly assess children’s participation in home, school and community life. The Canadian Occupational Performance Measure was selected as the primary outcome measure in the randomised trial as it had strong evidence for validity and reliability, had been used in paediatric clinical trials and was responsive to change. Three measures of participation, the Assessment of Life Habits which was completed by the parent/caregiver, the Children’s Assessment of Participation and Enjoyment which was completed by the child, and the School Function Assessment, which was completed by the child’s teacher, were selected to explore the research utility of the measures and their responsiveness to change. Analysis of cross-sectional data collected during screening and baseline assessments for the randomised trial found a strong relationship between bimanual performance and unimanual capacity. Scores on the Melbourne Assessment of Unilateral Upper Limb Function and stereognosis accounted for a significant amount of variance in scores on the Assisting Hand Assessment. There were only moderate associations between impairments (eg. sensory deficits and reduced grip strength) and bimanual performance and unimanual capacity. Age, gender, grip strength and two-point discrimination did not significantly influence bimanual performance. Results of the randomised controlled trial found no differences between groups on any baseline measure. A significant difference between groups favouring the constraint induced movement therapy group was found at 26 weeks on the Melbourne Assessment of Unilateral Upper Limb Function. There were no differences between groups on any other measure at either immediately post intervention at 3 weeks or in the medium term at 26 weeks. The constraint induced movement therapy group made significant gains in unimanual capacity (Melbourne Assessment of Unilateral Upper Limb Function and Jebsen Taylor Test of Hand Function) from baseline to 3 and 26 weeks. The bimanual group demonstrated significant improvement in movement efficiency (Jebsen Taylor Test of Hand Function) by 26 weeks. Significant gains in bimanual performance (Assisting Hand Assessment) were evidenced for both groups from baseline to 3 weeks. These gains were maintained at 26 weeks by the bimanual group only. There were no differences between groups on any participation measures. Both constraint induced movement therapy and bimanual training groups made statistically and clinically significant changes in perceived performance and satisfaction of identified functional goals from baseline to 3 and 26 weeks. Significant gains were made by both groups in personal care on the Assessment of Life Habits from baseline to 26 weeks. There were no changes for either group on the School Function Assessment and Children’s Assessment of Participation and Enjoyment. Conclusions This study found minimal differences between the two training approaches. Outcomes achieved by children reflected the mode of upper limb training, that is, improved and sustained gains in unimanual capacity were achieved with a unimanual approach (constraint induced movement therapy), and significant change in bimanual performance was achieved following bimanual training. The constraint induced movement therapy group made initial improvements in bimanual performance that were not sustained at 26 weeks, suggesting that intensive unimanual training may need to be followed by bimanual training in order to retain effects. Both interventions resulted in significant improvements in the achievement of individualised outcomes. Small gains in participation appeared to correspond with specific goal areas identified by children and their caregivers and highlighted the importance of goal directed training and measuring individualised outcomes. Regardless of the type of approach, intervention needs to be goal-directed, focusing on areas of central importance for children and their families.