To summarise tne main results, very small to small but n non-significant (with the exception of self-efficacy for v UCTs) intervention effects for all outcomes variables were b apparent (Fig. 2). Subgroup analyses showed that training volume, age and study design typically have a small 3 influence on the overall intervention effects (for PedsQL, physical health) with high training volumes, younger E children/adolescents and RCTs yielding the largest effects. o However, results for each subgroup were still in favour of s the control group. As a result of the small number of studies