I wonder if tumbling practice is still done in PE? It is a good thing to have a little experience with IMO. Anyway, if you were to try something like a diving forward roll with a bicycle helmet on, you would find that it gets in the way. In a tumble that would result in no head impact otherwise, the tail (which sticks out quite far) can hit the ground hard. I recall a friend mentioned a crash where he fell straight backwards onto his back. He was able to avoid hitting his helmetless head because his instinctive reaction was to keep it tucked to his chest. If he had been wearing his helmet, the tail would likely have struck the ground and he would have gotten some bad neck trauma, and maybe a concussion.

I'm certainly no acrobat, but I've noticed that in the 20+ bad cycling accidents I've had, my instinctive tendency is to tuck my head down to my chest and take the brunt of the impact on the shoulder and back. If I fall sideways or backward I will also tuck my head and try to keep it from hitting anything. Only once did I hit my head (helmet) and in that case I'm pretty sure that I wouldn't have hit my head if I hadn't been wearing a helmet. It was a forward somersault in which the tail of the helmet struck the pavement.

Anyway, I'm only presenting these as possible explanations on why the fatality rates don't go down when helmet use goes up.