Earlier this month, Senators Ted Cruz and Mike Lee introduced a bill that would allow for automatic FDA approval of drugs already approved in other developed country, specifically, the provisions are:
Amending the Food, Drug and Cosmetic Act to allow for reciprocal approval of drugs, devices and biologics from foreign sponsors in certain trusted, developed countries including EU member countries, Israel, Australia, Canada and Japan.
Encouraging the FDA to expeditiously review life-saving drug and device applications, this legislation would provide the FDA with a 30-day window to approve or deny a sponsor’s application….
The HHS Secretary is instructed to approve a drug, device or biologic if the FDA confirms the product is:
- Lawfully approved for sale in one of the listed countries;
- Not a banned device by current FDA standards;
- There is a public health or unmet medical need for the product.
If a promising application for a life-saving drug is declined Congress is granted the authority to disapprove of a denied application and override an FDA decision with a majority vote via a joint resolution.
This is a pretty interesting proposal, which some (like Tabarrok) are very in favor of it. Julia Belloz, at Vox, argues instead that the FDA is already quick enough, and doing this would abdicate governmental responsibility to protect its citizens from dangerous drugs. Personally, I think it’s probably a good idea (Scott Alexander, my favorite blogger, is a big fan and knows a lot more about drug approval than I do, as a doctor and medical researcher.)
However, I just want to look at the bill in the context of our class, and specifically significance testing. Basically, what this bill proposes to do is lower the significance threshold for drugs to be legalized in the United States. This isn’t a 100% precise analogy, as all the other states may have the same explicit significance levels, but having many more places to potentially get approval in would presumably make getting approval easier. If not, why pass the bill at all?
By raising the effective alpha, what this bill does is trade off type one for type two errors. The current FDA system is designed around having very few type 1 errors: Drugs have to really prove their (safety and) efficacy to be legalized, because the risks of ineffective drugs with really bad side effects, or even just high costs, are so high. That is, the null hypothesis for all drugs is that they don’t work, and if you don’t show them working to a really low alpha, they don’t get approved. The null hypothesis, that they don’t work, is only rejected if we’re really sure they work.
The other side is more concerned about Type 2 errors. Cruz wrote, “We continue to lose far too many of our loved ones to the “invisible graveyard,” as economist Alex Tabarrok has described: lives that could have been saved but for a bureaucratic barrier that rejects medical cures and innovation.” This reflects a concern about type 2 errors, not type 1 errors: we are rejecting too many drugs for insufficiently proving their efficacy, and doing so costs lives.
Both positions are reasonable. Personally, I think we can trust the countries included in the reciprocity clause of the bill enough to continue to avoid too many type 1 errors, while shaving off a few type 2s. However, I’m not so sure about the idea of letting congress override the FDA to approve specific drugs, as the last bullet point above proposes.