> We have problem A, which is alleviated by crutch B.
It's not a crutch. It's "we have textbook market failure A, which is alleviated by textbook government intervention B; we can remove the intervention, but we have to substitute one of the other textbook interventions suitable for addressing market failures."
> If there is a problem with the FDA approval system, let's fix that.
First, the FDA approval system is a major but not the biggest cost in getting a new drug to market. Making it dramatically cheaper wouldn't eliminate the underlying problem.
Second, the FDA process is thorough and therefore expensive for good reason. The drug market is problematic because its exceedingly difficult to trace a condition to a particular drug, and because conditions sometimes show up years after treatment with a particular drug.
And the cost of letting a bad drug through is astronomical. A popular drug like Vioxx might be used by 50 million people worldwide. David Graham, and FDA scientist, estimated that Vioxx caused 90-140k preventable heart attacks, and estimated 25-55k preventable deaths from those heart attacks. What's the cost of this damage? How much would you have to be paid to have a heart attack? Remember the target demographic of Vioxx, middle aged men with decent jobs that provide health insurance to be able to afford an expensive medication, is one of the most expensive demographics in terms of lost income from premature death. Between premature deaths and heart attacks, the actual cost of the Vioxx mistake could be $50-$100 billion in the US. Merck had spectacular lawyers and ended up settling for a fraction of that, but that's the nature of class-action settlements--the tortfeasor only ever pays pennies on each dollar of damage their products cause. The public bears the rest of those costs.