In terms of outcome, it maybe a tolerable risk, but anything that flat out doubles the risk ratio is fairly concerning; in particular, where the mechanisms are not well understood.
You are right in that the mechanisms may not be well-understood. This does warrant further study, as the paper concludes.
Your chances of being hit by a meteorite are _at most_ 1 in 1,000,000.
My happiness would not extend across several orders of magnitude like that, but perhaps, some people just need the $10k more than I would.
Some are. Some are simply prescribed because the patient asks for them.
> even if little is known about their mechanism.
I wouldn't take it unless my condition was immediately life threatening and there were no other medications available. There are very few classes of treatment that fall within these parameters.