Bimanual therapy: the evidence for training both hands together

For children with upper-limb motor impairments, daily life can be a struggle. Simple tasks such as dressing or eating can become the most difficult parts of the day. Any improvement in upper limb function can make a significant difference to their independence and quality of life.

For two decades the dominant approach to upper limb therapy has been constraint-induced movement therapy (CIMT), which restrains the less affected hand to force use of the more affected one. The literature of the past few years suggests that bimanual therapy, training movements with both hands together, achieves the same outcomes without restraining a child’s hand, while training the coordination that daily life actually requires. It also suggests the choice of technique is largely settled, and that what decides the outcome is how much practice a child actually does.

Daily life needs two hands

• Coordination of both hands is needed for most activities of daily living (ADLs): dressing, eating, playing, opening bottles, folding clothes. Training one hand in isolation may not transfer to them. (Ouyang et al., 2020)
• CIMT was designed for adults who lost the use of one hand after a stroke, where restraint forces a return to a movement pattern they already knew. Children with upper limb motor-impairments often never learned that pattern, so their difficulty is the coordination between the two hands rather than disuse of one. (Ouyang et al., 2020)

Bimanual training builds upper limb function

• A systematic review of 15 studies, 11 of them randomised controlled trials, found bimanual training significantly improved upper limb function immediately after intervention (d = 0.36, p = 0.017), and the improvement was maintained at follow-up. (Ouyang et al., 2020)
• The effect on self-care function was moderate (d = 0.52, p = 0.003), and the effect on the goals children set for themselves was large (d = 1.78 to 2.28, p < 0.001). (Ouyang et al., 2020)
• Bimanual training is one of a minority of interventions to reach a “green light” rating, in the traffic-light review that grades every cerebral palsy intervention against its evidence base. (Novak et al., 2020)

Constraint therapy adds nothing but frustration

• CIMT is no better than bimanual therapy. Where the two were compared at the same dose there was no evidence of any difference in bimanual performance, nor in manual ability or unimanual capacity on almost every measure used, across 36 trials and 1,264 participants. (Hoare et al., 2019)
• This is supported by the largest comparison to date, a network meta-analysis of 48 randomised controlled trials and 1,629 children, which ranked bimanual therapy at least as effective as modified CIMT on all three outcomes it measured: hand function, individual goal attainment and self-care. (Burgess et al., 2025)
• The key difference: children often will not tolerate the constraint in CIMT. The same review found adverse events reported only for CIMT: frustration, refusal of the restraint, and reversible skin irritation from casting. (Hoare et al., 2019)

Frequency matters more than technique

• A review of 74 trials found that any measurable improvement to general upper limb function takes around 40 hours of practice in total, across a course of therapy. Below that, no reliable functional gain was seen. (Jackman et al., 2020)
• Supporting this, at 30 hours neither modified CIMT nor bimanual therapy altered upper limb function. (Sakzewski et al., 2015)
• The prescribed therapeutic dose is rarely delivered. Compliance with home-based upper limb programmes ranges from 56% to 99%, and the studies attributed it to children being too tired or upset to complete the training. (Beckers et al., 2020) Across home programmes generally, family dropout has been reported as high as 66%. (Medina-Valera et al., 2025)
• Parents are asked to prompt the same movement over and over, which is itself a source of stress and a barrier to adherence. Implicit motor learning, where the child practises through play rather than instruction, has been proposed as the way to keep compliance high and parental stress low. (Beckers et al., 2020)

Conclusion

Much like the physical tools used for upper limb therapy, the CIMT approach used in paediatric practice was intended for adults, then adapted for children. Restraining the limb of an adult recovering from a stroke is one thing; restraining the arm of a child is going to be met with resistance more often than not.

Moreover, where the two approaches have been compared at equal dose the results are indistinguishable, but only the bimanual approach trains the coordination that daily life actually requires.

A theme running through the literature is that as long as a child is doing some form of therapy, the choice of approach matters less than the frequency and intensity of practice. That reframes the problem slightly. The constraint is not which technique to prescribe, it is how many hours actually get done. The best approaches are therefore the ones that meet the least resistance, and the best tools are those engaging enough that children want to repeat the movements willingly.