Looking at that meta study, the conclusions seems to be that this kind of studies on children take a lot of time and generally lack any control group, thus conclusions are going to be weak.