I would say that there's increasing emphasis on using Finite Element Method.
It has always been used for structural analysis with varying precision in the old days to be fast enough to calculate.
It can be approximately done in real time now, but the problem of optimization still has more room to go.
I really want to try it out, but I have yet to find many usable implementations that are opensource and general purpose.
I think xissburg is working on this.
FEM is not really collision detection though.
I agree that it is a very interesting area of research when it comes to physics simulation though since you want more realistic deformation and breaking of different materials.
Traditional techniques for destruction usually give a very concrete-like behaviour.
CCD still has a lot of unresolved problems.
Especially rotational movement is very hard to handle in a correct fashion and unless you allow it to drop time you can sometimes end up with serious oscillation issues.