I imagine they can use the same process from the wiki article where they studied the mechanical efficiency of Pistorius's running form and compared to other runners. They'd need to broaden the study to include turns and starts, and possibly more. That should show whether prosthetic runners have a net advantage.