Study	Reason for exclusion
Caçola 2015	Study design: not a randomised controlled trial.
De Milander 2015	Study design: not a cluster‐randomised controlled trial, but a quasi‐experimental study, which lacked an element of random assignment.
Ferguson 2013	Study design: not a randomised controlled trial; participants not randomised but allocated according to schools that were not sufficiently similar.
Hammond 2014	Intervention: Wii Fit and active control 'Jump Ahead' may or may not be task‐oriented interventions.
Peens 2008	Intervention: motor‐based intervention programme involved both task‐specific and process‐oriented (kinaesthetic and sensory integration) treatment.
Tsai 2014	Intervention: exercise intervention not considered to be a task‐specific intervention.
Yu 2016	Study design: two clusters only and intra‐cluster correlation coefficient not clear.