It's understandable that the regulatory process errs way on the side of restrictiveness: if someone dies because it's not restrictive enough, that's easy to identify, and it can become a lawsuit and a political campaign, Senators can get involved, agency heads can resign, civil servants can get fired. Meanwhile, the people who die because it's too restrictive are much more difficult to identify, and their deaths can't be blamed on a single decision. (Maybe the unapproved treatment wouldn't have worked anyway, and maybe nobody even knows it could have existed.)

So, given that whatever decision they make will kill masses of people, regulators have very strong incentives to kill them by being too restrictive, rather than kill a different mass of people by not being restrictive enough.

The solution to this problem is clear: the person with the strongest incentive to keep you alive is you. Any regulator or drug-company executive will reliably put other considerations above your survival under some circumstances. So, except in cases of people who really cannot be trusted with life-and-death decisions — children and some severely mentally impaired people — you should have the final say on what medical treatments you try or don't try.

In short, the whole regulatory structure should be scrapped. We still need clinical trials, but not before new treatments become available; after.

This sounds like a radical proposal, but in fact it's how the process worked in the US during the time that we owe most of our revolutionary medical treatments to, the 1920s through the 1960s. That's when most (all?) of our currently marketed antibiotic families were discovered, when one dude spent a couple of years prototyping the pacemaker in his barn and then convinced his friend at a hospital to start trying it in patients, when artificial joints were developed. We're still coasting on the legacy from that era.