Participation (permission to measure BMI) was around 60%. That seems like a pretty low rate, and it is not too hard to imagine that parents who know their child has a BMI problem might be more likely to not want it measured, for fear of stigmatizing them or their child. So, while I don't doubt that a program of this sort might not be the answer, I don't think their data is sufficient to prove it one way or the other, unfortunately. I would expect you need well over 90% compliance/permission to measure before you could say that. Disclosure: I am not a statistician (and I would love to hear from one about this issue).