I don't think those numbers really capture the full story since the majority of children rarely or never cycle on the road. I couldn't find a resource for the number of children who regularly cycle on roads, but one of the main reasons children might be doing this is going to and from school. Only 1% of children commute this way. There are certainly other reasons children might cycle on roads, for the sake of argument let's add another 1% for children who regularly ride on roads but do not ride to school. Using your estimate of 40 million children between the age of 5 and 14 that gives us just 800k regular cyclists and a 30.6 child pedestrian deaths per 100k.

https://www.saferoutespartnership.org/media/pressreleases/US...