On the point about hull compression raising the internal air pressure, the normal volume change hence air-pressure rise by shell compression should be minute, provided the material's movement is within its elastic limit.
It it is not, the shell is collapsing anyway.
I don't know how long an emergency ascent from 3000m would take but even then the consequent reduction to ambient air pressure will be as small as its increase was, and relatively slow. So I would not think "the bends" any likely hazard.
'
The Bends cited by Hopper, or Decompression-sickness, was originally called "Caisson Disease" as it was first identified among civil-engineering labourers working in sub-estuarine caissons for building bridge piers*. It is not directly an an effect of water at high-pressure on the body.
Instead it is an effect of breathing compressed-air at pressures equivalent to ambient even at quite modest pressures, depending on time. The bridge-builders were working hard, surrounded by high-pressure air, for some hours at a time.
In these conditions the blood absorbs nitrogen, and if the pressure is released too rapidly it cannot come out of solution in a gentle manner, but effervesces slightly, creating bubbles that typically lodge in the joints, causing great pain.
Treatment is by putting the patient in a decompression-chamber in which the air pressure is raised to the diving depth then released at a very slow, controlled rate over as many hours as calculated for the case.
SCUBA Divers usually stay within fairly shallow depths but if venturing deeper, especially to 30+m with some time at depth, take enormous care over gas-mixtures, the potentially toxic or narcotic effects of atmospheric gases at high pressure, times at depth, ascent decompression-stops, short decompression-oxygen breathing, etc. Even then things can go wrong. The complications are multiplied greatly if the dive profile (depth, distance and time) is not a single dive to a steady depth but rises and falls like a switchback, even alternating from open air to considerable depths a few times – but this is normally only met in cave-diving, not the sea.
So air-pressure change by a submarine's hull compression is not going to lead to The Bends. I would expect the shell to contract slightly, but if enough to raise the internal pressure significantly, it is passing beyond its elastic limit….
… and that is a design consideration. As are the mind-bogglingly complicated stress and strain calculations for external pressures on a shell that although basically cylindrical, is not of single material and contains shape-changers like interpenetrations and mountings.
I have seen a video of a plain cylinder collapsing, in a hydrostatic test-tank. It was a fairly small container intended for marine electronic equipment, but failed well within its intended working-pressure for some reason, while undergoing the test as specified. It gave no visible warning such as slow buckling, but ruptured suddenly and implosively.
.
*I believe the first cases questioned, against considerable medical and commercial opposition, were among the men building the Brooklyn Bridge, in New York.