Testing is costly in terms of time and money, particularly with a submersible. It's one thing to test a piece of code, but testing a submersible under realistic conditions means basically arranging a full mission for every test run (which is hugely expensive, requiring a support ship, etc.). There's no such thing as perfect safety, which always has to be balanced with cost. I'm not saying Rush's approach was sound (he apparently eschewed more thorough testing) but testing that submersible a hundred times surely would have been cost-prohibitive.
Also the Titan did successfully complete ~20 deep dives from what I've read, so the design was apparently sufficient to stand up to the pressure more than once, meaning even a non-trivial testing program might not have detected the problem. Unfortunately the problem (as many have stated) seems to be in the choice of carbon fiber, which (particularly when taking into account the interfaces with other materials) will weaken over time even if the initial flaws are very small. All materials experience fatigue but the problem is much worse with carbon particularly under those kind of pressures.