The convex hull are special for collision detection because of a mathematical theorem called the Minkowski sum of two set of points. you can look that up in Wikipedia.
it just happen that the Minkowski sum of a set of points that happen to be form a convex hull is also another convex hull.
the collision detection is converted from a problem of calculation interstation between to shapes to the problem of calculations the closest distance from a convex hull to the origin.
if the origin is outside the hull the shapes do no intersect, If the origin is inside or touches the shape they the intersect.
here he where another algorithm called JGK is use to calculation the closest point form a convex hull to the origin iteratively. This algorithm does no need to build the complete convex Hull of the two input set of vertices,
instead is build them incrementally as it need them.
No Newton does not use quickhull to build convex hulls, newton use my own algorithm which has being extracted as a stand alone algorithm by some nvidea people
juliohull is much faster that quickull, more robust and much cleaner implementation. and it si a truly N dimension convex hull generator. I use also for making 4 dimensional convex hulls
http://code.google.com/p/juliohull/