Constraint-induced movement therapy (CIMT) is recommended for patients with upper limb dysfunction after stroke, yet evidence to support the implementation of CIMT in ambulatory care is insufficient. We assessed the efficacy of home CIMT, a modified form of CIMT that trains arm use in daily activities within the home environment.

In this parallel, cluster-randomised controlled trial, we selected 71 therapy practices in northern Germany that treat adult patients with upper limb dysfunction after stroke. Practices were stratified by region and randomly allocated by an external biometrician (1:1, block size of four) using a computer-generated sequence. 37 practices were randomly assigned to provide 4 weeks of home CIMT and 34 practices to provide 4 weeks of standard therapy. Eligible patients had mild to moderate impairment of arm function at least 6 months after stroke and a friend or family member willing to participate as a non-professional coach. Patients of both groups received 5 h of professional therapist contact in 4 weeks. In the home CIMT group, therapists used the contact time to instruct and supervise patients and coaches in home CIMT. Patients in the standard therapy group received conventional physical or occupational therapy, but additional home training was not obligatory. All assessments were done by masked outcome assessors at baseline, after 4 weeks of intervention, and at 6 month follow-up. The primary outcomes were quality of movement, assessed by the Motor Activity Log (MAL-QOM, assessor-assisted self-reported), and performance time, assessed by the Wolf Motor Function Test (WMFT-PT, assessor-reported). Primary outcomes were tested hierarchically after 4 weeks of intervention and analysed by intention to treat, using mixed linear models. This trial is registered with ClinicalTrials.gov, NCT01343602.

Between July 11, 2011, and June 4, 2013, 85 of 156 enrolled patients were assigned home CIMT and 71 patients were assigned standard therapy. 82 (96%) patients in the home CIMT group and 71 (100%) patients in the standard therapy group completed treatment and were assessed at 4 weeks. Patients in both groups improved in quality of movement (MAL-QOM; change from baseline 0·56, 95% CI 0·41–0·71, p<0·0001 for home CIMT vs 0·31, 0·15–0·46, p=0·0003 for standard therapy). Patients in the home CIMT group improved more than patients in the standard therapy group (between-group difference 0·26, 95% CI 0·05–0·46; p=0·0156). Both groups also improved in motor function performance time (WMFT-PT; change from baseline −25·60%, 95% CI −36·75 to −12·49, p=0·0006 for home CIMT vs −27·52%, −38·94 to −13·94, p=0·0004 for standard therapy), but the extent of improvement did not differ between groups (2·65%, −17·94 to 28·40; p=0·8152). Nine adverse events (of which six were serious) were reported in the home CIMT group and ten (of which seven were serious) in the standard therapy group; however, none was deemed related to the study intervention.

Home-based CIMT can enhance the perceived use of the stroke-affected arm in daily activities more effectively than conventional therapy, but was not superior with respect to motor function. Further research is needed to confirm whether home CIMT leads to clinically significant improvements and if so to identify patients that are most likely to benefit.

German Federal Ministry of Education and Research.