The problem with the statistic that you cite for probability of death due to asteroid impact is that it is computed from the very low probability of a large (dinosaur killer class ) asteroid strike on the planet (on the order I believe of one every 100,000,000 years) time the death of everyone on the planet.

Since the scale of the event is on the order of everyone dying if the asteroid hits, the defence is to not allow that to happen. This is much like thermonuclear war, nobody ever thought that thermonuclear war was a good idea, the arguements were how to keep it from happneing.

The way to prevent a large asteroid strike is to detect and classify all Earth orbit crossing asteroids and develop technology to modify their orbits (if necessary) to insure that they miss. There are programs currently in place to do the former, but none that I am aware of to do the latter. Of course, the more lead time you have, the smaller the orbital change (in degrees) needs to be made.

Seth