Twenty-five articles that included 22 trials (n=808 participants, range 10 to 100) were included in the review. All studies achieved scores of between 1 and 8 out of 11 within the quality criteria, with four studies considered to be high quality. None of the trials was blinded. All trials had similar outcome assessment timing. Four trials fulfilled the criteria of adequate randomisation method, allocation concealment, prognostic similarity and acceptable drop-out rate. Follow-up ranged from one to 18 months. Further details were provided in online appendices.
The reviewers formed eight intervention categories.
Comprehensive physical therapy approaches (six trials, one high quality): Significant differences between groups were observed in four of the six trials.
Upper extremity treatments (four trials, two high quality): Significant differences between groups were observed in three of the four trials for some outcomes. Two trials showed moderate evidence for improvement in active supination and goals achieved using occupational therapy and improved developmental status with neurodevelopmental therapy (NDT) twice a week compared to NDT once a week. One trial showed limited evidence in improvement quality of hand movement.
Strength training (four trials): Two of the four trials found limited evidence of effectiveness, but all studies were of low quality.
Cardiovascular fitness/aerobic programmes (two trials): Both studies found limited evidence of effectiveness, but were of low quality.
Constraint-induced therapy (two trials): Both studies reported positive effects; one study was of lower quality and one of high quality and so these results were considered moderate evidence of effectiveness.
Therapy with animals (two trials): One of the two lower quality trials found evidence of effectiveness for one sub-item of grasping.
Sensorimotor therapy (one trial): This low-quality trial found positive short-term effectiveness that suggested limited evidence of effectiveness.
Balance training (one trial): This low-quality trial found limited evidence of effectiveness.