[A4ET8a8uTh0]

I think the issue is that some people ( and this may be HN-specific ) think that medicine in general, and maybe even diagnostics in particular, is almost exactly like software testing. In a lot of ways, it really isn't. Some reasons get a little esoteric, but the more important one is rather simple: until more recently, software did not have a direct impact on life and as such was mostly given a pass on some otherwise heavy blunders. That is slowly changing, but missing something during quality control is not likely to have the same impact.

But this brings me to the other important reason, statistical check can only get you so far and that is assuming we can now trust it was even done. Some people do rely on being able to say, 'this was false positive once, you know what are the odds of it being false positive twice'? Now, we add variable of uncertainty into the system in the form of 'well, it was outsourced so there is a non-zero chance it is bs anyway'.

Not exactly a recipe for success.