Nine RCTs (452 patients) were included. Follow-up ranged from one day to one year after intervention. Overall quality scores ranged from 5 to 8 out of 10. All studies used randomisation but only three reported allocation concealment methods. Intention-to-treat analysis was used in only four trials. Eight studies blinded assessors. The nature of the interventions precluded blinding of participants and therapists.
In chronic patients with a mild upper limb paresis, marginally significant differences favouring unilateral training were found for upper limb activity performance (SMD 0.34, 95% CI 0.04 to 0.63; four RCTs), perceived upper limb activity performance (amount of use MD 0.42, 95% CI 0.09 to 0.76; three RCTs) and quality of movement (MD 0.45, 95% CI 0.12 to 0.78; three RCTs).
No statistically significant differences were found with another upper limb activity scale (motor assessment scale). No statistically significant results were found for patient in the acute phase post-stroke. There was no evidence of significant heterogeneity in any of the analyses.