It's just the good old problem of incentives misalignment. And the article doesn't seem to have any solution that is practical to implement.

Suppose we did form an independent research committee. That committee now has the incentive to test medicines in such a rigorous and harsh ways that nothing would get approved. In a situation of ambiguity, this committee would rather reject a newer medicine based on some vague or ill-founded doubts, rather than approve a medicine that's reasonably effective. They are not paid to get out an effective medicine in public.