Sakzewski etal. reply.

Abstract

SIR–We would like to thank Dr Basu and Professor Eyre for their comments regarding our randomized clinical trial comparing equal doses of modified constraint-induced movement therapy (mCIMT) and bimanual training in school-aged children with unilateral cerebral palsy. The primary concern proposed is the potential that constraint of the unimpaired hand used in mCIMT may have a deleterious impact on functioning of that hand, and that for many children with unilateral cerebral palsy, the unimpaired hand may also have subtle difficulties. We agree with Basu and Eyre that the unimpaired upper limb needs to be considered and may not be truly ‘unimpaired’. Our findings in a previous cross-sectional study which included all baseline data of children in the randomized trial confirmed a modest reduction in stereognosis ability for the unimpaired upper limb, although this may reflect developmental progression as all of the children who had difficulties were under the age of 8 years. We also found that movement efficiency of the unimpaired upper limb, (measured on the Jebsen Taylor Test of Hand Function [JTTHF]), was more than two standard deviations below published norms for 68% of the group. It is important to note in a small group of typically developing children, 47% also scored greater than one standard deviation below published norms. This certainly highlights the need to re-evaluate normative data for the JTTHF which is now nearly 40 years old. The question of whether constraining the unimpaired hand would have a deleterious effect on that limb was a consideration in our study. We sought to measure movement efficiency (JTTHF), grip strength, and sensation (stereognosis and moving two-point discrimination) on the unimpaired upper limb at baseline and each follow-up. In the process of preparing this reply, we found an error in Table I in the original paper regarding the 3-week follow-up score on the JTTHF for the mCIMT group. We would like to clarify these results. Data for the JTTHF, stereognosis and moving two-point discrimination were highly skewed so that non-parametric Mann–Whitney U tests were used to determine whether follow-up scores were significantly different from baseline. Generalized estimating equations were used to determine within group changes for grip strength between baseline and each follow-up. Our results for the unimpaired limb before and after each intervention are summarized in Table I in this letter. Our results demonstrate that movement efficiency on the unimpaired upper limb was not compromised following