> All of this notwithstanding the growing knowledge that no, you can't just give mice cancer/MS/autism and check if a drug works there, expecting the same result in humans.
...to state the obvious, scientists aren't giving drugs to mice and then shipping them off to pharmacies. They're testing the drugs in mice first to see if it's worth testing in humans. Then, the human trials are used to determine efficacy.
It's not a perfect system to be sure, but what is the alternative? Bypass animal testing and go straight to human trials? I'd have a lot of obvious ethical issues with that.