I do remember that while I was still employed (in the UK) as a hands-on basher of instruments, the QFE that you got from the tower seldom, if ever, bore any relation to the altimeter reading obtained. The surface pressure changes constantly and may vary by a millibar or two from one part of the field to another. If QFE only gives a more-or-less figure for altitude at touch-down why bother?
As to procedures, isn't it better to have everyone working to the same standard? An altimeter is calibrated to a 'standard' atmosphere and is supposed to indicate your barometric height above a standard pressure for MSL. Not your real height, mark you. A notional standard. Terrain clearance using the standard is then obtained by nominating a minimum safe clearance to allow for instrument errors and the difference between the standard atmosphere and the current version of the real world. So, why override these wise precautions when you descend and fly closer to the lumpy stuff?
I know its an extreme example, but when I worked in Kathmandu I knew that the lumpy bit on the regular approach from the south was around nine thousand feet above MSL and the threshold was around five thousand above MSL. If the altimeter were set to QFE and your minimum clearance is, say, two thousand feet, what should be on the clock five miles from touch down? OK, now we've worked it out, to avoid bumping into the nine thousand foot mountain that the radio altimeter cannot see until its too late, we'll fly at six thousand feet. Logical eh? Or would you prefer QNH and have eleven thousand on the clock until you can see KTM?
Human Factors. Why ask for trouble?