Constraint-Induced Movement Therapy (CIMT)/Constraint-Induced Therapy (CIT)

According to the Children’s Hemiplegia and Stroke Association ( constraint-induced movement therapy (CIMT), sometimes called “forced use therapy”, has been used in the adult stroke population for years. Recently, this type of therapy has gained the attention of therapists who work with children who have hemiplegia (weakness on one side of the body due to an injury to the brain on the opposite side).

CIMT focuses on regaining movement on the affected side of the body by restraining the non-affected arm, thus forcing the child to learn to move the affected arm more efficiently and effectively. There is increasing evidence that this therapy may result in positive structural changes in the brain, prompting Brady and Garcia, in an excellent review of CIMT (Dev Disabil Res Rev 2009;15:102-111), to comment that CIMT is an example of an emerging “paradigm shift” in rehabilitation of CNS injury, from an emphasis on compensatory skills to a hope for partial restoration.

Accumulating research reports have generally shown a favorable response to CIMT, although questions remain about what is the critical level of intensity of therapy necessary for a positive effect (how much? how frequently?).  As with any new therapy, another important question is whether it is superior to what is already available and being implemented, perhaps at less expense.

A recent article by Wallen and colleagues from Sydney, Australia compared a modified form of CIMT with intensive occupational therapy on activities of daily living and upper limb outcomes in children with hemiplegic cerebral palsy. They concluded from their study that modified constraint-induced therapy is no more effective than intensive occupational therapy.(Dev Med Child Neurol 2011;53:1091-1099)

In a Letter to the Editor (accepted but not yet published by Developmental Medicine and Child Neurology)  Dr. Stephanie DeLuca (University of Alabama at Birmingham) and colleagues long involved in CIMT research raise some interesting questions about the Wallen study. An excerpt follows:

We raise many serious issues about the [Wallen et al.] study as well as present directly comparative data from a multisite trial of CIMT that we recently completed (and is forthcoming as a manuscript in the  Am J Occup Ther, January, 2012). The purpose of the comparative data is to help readers better interpret the magnitude of changes reported among children in the two Wallen et al treatment groups – for an objective outcome (the Assisting Hand Assessment) and a subjective one (parental ratings on thePediatric Motor Activity Log).

What concerns us most is that when clinical trials are conducted in a way that fails to clearly specify the intervention treatment and  to document its fidelity of implementation, then readers are at a loss as to how to use the findings. Rigorous clinical trials have clearly agreed upon standards about what constitutes adequate, objective outcome data. Based on the published article, the Wallen et al. study did not meet criteria of a rigorous clinical trial with appropriate primary outcomes.

The field is eager to resolve critical questions about whether Constraint-Induced Movement Therapy (CIMT) works, and for whom it works best, and what format (dosage, constraint) yields the best results. The Wallen et al study is described as though it answers some of these questions. In fact, we judge the form of administration (parent delivered almost exclusively) and the dosage (below 1.5 hr/day) and constraint (a mitt (worn less than 1.5 hr/day) of the so-called “modified” Constraint-Induced Therapy to be insufficient to know if it really WAS CIT.

We think the field needs to develop clear and agreed upon definitions for different therapy approaches, with operational definitions and measures of the delivery of the components of a specified form of therapy. Otherwise, we fear that CIMT – which thus far is one of the most promising evidence-based therapies available for children with unilateral cerebral palsy  –  may go the way of earlier “popular” therapies that became so ill-defined (such as Neurodevelopmental Therapy – NDT) that it becomes a “discounted” or disrespected therapy, because no one can describe exactly what it is. In our view, use of a short-term form of constraint and only slightly more than a one hour therapy session per week cannot qualify as CIMT!

Children need evidence-based treatments. The field needs a solid, trustworthy database to inform treatment recommendations and the training for therapists who deliver treatments. Wallen et al, unfortunately, failed to clarify or advance the role of CIMT per se. It did, however, perhaps show low dosages of CIMT fail to produce large and statistically significant improvements in function (despite parents liking the intervention and being satisfied with their children’s progress).

This kind of interplay between researchers is very healthy and will lead to a better understanding of the most cost-effective approaches to therapy for cerebral palsy. I encourage parents, therapists, physicians and all other interested in cerebral palsy treatment and research to read the Wallen et al. article and the more complete Letter to the Editor by DiLuca et al that is to follow in Dev Med Child Neurol.

We asked Dr. Wallen to respond to this critique and she kindly furnished the following:
Our trial evaluated a modified form of CIMT (modCIT), devised in response to families requesting CIMT which was less intensive than pre-existing models, and therapists who proposed that these models were not clinically feasible within the Australian health services context. We compared modCIT with an intensive block of occupational therapy, arguing that a constraint-based intervention needed to be substantially more effective than the best available service currently offered, in order to justify its additional intensity and intrusiveness.

The statement that modCIT “cannot qualify as CIMT” is disingenuous. What is CIMT?  Case-Smith, DeLuca and colleagues1 employed a cast worn 24 hours per day for 18 days during which time children participated in an intervention protocol of either 3 or 6 hours per day. This was followed by a period in which children participated in bimanual intervention.   How do we delineate the effects of CIMT from those of the bimanual therapy or an interaction between the two in this protocol?  Diverse CIMT protocols are reported in the literature variously using casts, splints, slings, mitts and even holding to achieve constraint for 1 to 24 hours per day over periods from 9 days to 8 weeks. Which of these options is CIMT?  Case-Smith, DeLuca and colleagues very accurately stated that “consensus has not been reached on the differential effects of dosage (or intensity) of therapy or a minimum threshold to produce significant effects” (p.16).  Each study adds its own unique contribution to the ever-increasing and complex knowledge base.  

Dr DeLuca stated that our trial “did not meet criteria of a rigorous clinical trial with appropriate primary outcomes.”  In our response to the Letter to the Editor of DMCN we provide evidence that the Canadian Occupational Performance Measure is valid for use with young children with cerebral palsy and reiterate that we specifically chose to use this measure as it individualizes and prioritizes outcomes.  Furthermore, it is consistent with family-centred care, a fundamental philosophy of contemporary practice. We take this opportunity to further highlight the aspects of our study which demonstrate methodological rigor:  randomisation with allocation concealment, blinding of raters and data analysis, adequate sample size determined by sample size calculation, a priori selection of primary and secondary outcome measures and endpoints, data analysis completed according to principles of intention to treat and so on.

In our response to the Letter to the Editor of DMCN we expressed strong concerns about the integrity of comparing results from Case-Smith’s trial with our trial.  There was no consideration of the heterogeneity of the trials (e.g., age group of participants), and information from Case-Smith’s report (e.g., variability of data, severity of disability) which would facilitate an informed and responsible comparison, was not provided. Furthermore it is erroneous to compare data from different versions of one of the measures, the Pediatric Motor Activity Log.

We concur with Professor Blackman’s observation that scholarly dialogue is healthy and contributes to the evidence-base informing stakeholders about intervention for children with cerebral palsy.  We also urge stakeholders to read our article, the aforementioned Letter to the Editor of DMCN and our response to this letter in DMCN.


1. Case-Smith J, deLuca S, Stevenson R, Landesman Ramey S. Multicenter randomized controlled trial of pediatric constraint-induced movement therapy: 6-month follow up. The American Journal of Occupational Therapy. 2012;66:15-23.