‘Both constraint‐induced movement therapy and bimanual training lead to improved performance of upper extremity function in children with hemiplegia’

SIR–In recent years increasing attention has been afforded to the potential benefit of constraint-induced movement therapy (CIMT) as a means of improving upper extremity (UE) function in children with hemiplegia. Indeed, there is mounting evidence of treatment efficacy for some children with hemiplegia across a wide array of delivery methods, including various restraint types, treatment models, intensities and durations, as well as age.1–3 Despite the success across the diversity of these studies, recently it has been suggested that treatment outcomes are compromised if the delivery method deviates from the methods typically used in adults, including use of a cast worn continuously for 3 weeks on the less-affected UE.4 However, these claims are unsubstantiated as comparisons are based on caregiver surveys without established validity or reliability. Furthermore, treatment efficacy has been demonstrated with standardized measures using far less restrictive restraints (mitts) during just 2 hours per day.5 In addition, we have demonstrated that improvements in UE function can be achieved during intensive bimanual training (i.e. without any restraint whatsoever).6 However, currently it is not known whether distributing practice across both UEs is as beneficial or whether there may be short-term compromises in treatment outcome. In the present study we compared the efficacy of CIMT and bimanual training using a quasi-randomized design (i.e. alternation assignment in groups of four children). Sixteen children (eight children in each treatment; age 3y 8mo–13y 7mo) with mild to moderate hemiplegic cerebral palsy were provided either CIMT or Hand-Arm Bimanual Intensive Therapy (HABIT).7 All children met inclusion criteria and underwent procedures established in our earlier studies.8 Informed consent was obtained from all children and their caregivers. The study was approved by Teachers College, Columbia University Institutional Review Board. Briefly, the CIMT and HABIT interventions were provided one-on-one by trained interventionists 6 hours per day on 10 out of 12 consecutive days in a day-camp environment at our laboratory. Details of the intervention procedures are reported elsewhere.7,9 Two standardized measures, the Jebsen-Taylor Test of Hand Function and the Assisting Hand Assessment (AHA), were used to assess hand function immediately before and after treatment. We also determined the percent time of affected UE use during performance of the AHA as measured by accelerometers. The assessments were administered by a blind evaluator. Table I shows that similar improvements were demonstrated for each group from the pretest to the post-test in all three measures (p<0.05 in all cases). Specifically, there was a 16% and 13% decrease in time to complete the Jebsen-Taylor Test for the CIMT and HABIT groups respectively. Furthermore the AHA scores increased approximately 8% and the accelerometry scores increased approximately 16% for both treatment groups. These findings are in agreement with our earlier studies that demonstrated efficacy for both treatments separately. This represents the first attempt to compare efficacy of constraint therapy and bimanual training. Here we show that the amount of improvement is not dependent on use of a restraint. These results give further credence to our argument that using an adult CIMT model is invasive and unnecessary to achieve UE gains. Generally we espouse the belief that the goal of UE rehabilitation should be to increase functional independence by improving use of both hands in cooperation. Our results do not support the notion that this requires specificity of practice since both groups demonstrate similar improvement. One reason for this finding could be that that neither group practiced items in which they were later tested. Thus, both treatment groups were asked to generalize what they learn during testing, and the tests may not be sensitive to outcome differences in this regard. Several limitations should be noted. First, although the results were remarkably similar for both groups with medium to large effect sizes (Table I, eta2), individuals with hemiplegia are a heterogeneous population, and thus a larger study is warranted. Second, the study used a quasi-randomized design, and thus a randomized study with stratification based on initial severity is merited. Finally, long-term retention of the reported gains are not known. Future stratified randomized trials are needed to home in on effects of severity, dose response, specificity of training, and retention in order to truly begin to unravel the key ingredients that lead to optimal UE treatment outcomes. This work was supported by the Thrasher Research Fund.