Lockdowns would never receive FDA approval

A mediocre vaccine is still better than the alternatives

At this point, it seems unlikely that the FDA will approve the AstraZeneca vaccine before data from the US trial is available in April. Although the data from the international trial was good enough to earn European approval, the FDA has signaled that it would reject an EUA from AstraZeneca supported by that data. That’s a shame; adding AstraZeneca’s vaccine to the mix would accelerate progress towards herd immunity and the end of the pandemic. While some thinkers (like Alex Tabbarok) continue to forcefully advocate for its approval, they’re very much the exception. The response to this delay feels like a collective yawn.

I suspect that the apathy follows from the sense that AstraZeneca’s vaccine isn’t very good. And to be clear, it is mediocre compared to its mRNA competitors: roughly 70% effective in its clinical trial compared to 95% for Moderna and Pfizer. There’s also the uncomfortable reality that AstraZeneca screwed up their trial by incorrectly dosing a large portion of the participants. That makes it easy to dismiss the AstraZeneca vaccine as unproven or ineffective; I’ve heard people who’ve spent the last year diligently social distancing say that they wouldn’t take the AZ vaccine even if they were offered it.

That’s a mistake. We hold vaccines and drugs to very high standards both for efficacy and for evidence of efficacy. As a result, even a mediocre vaccine works better than the nonpharmaceutical interventions (NPIs) that we’ve subjected ourselves to over the last year. And the evidence we have about the AstraZeneca vaccine is far stronger than the evidence supporting measures like lockdowns.

We can’t precisely quantify the costs and benefits of voluntary social distancing, voluntary mask wearing, mandated mask wearing, business closures, school closures, and stay-at-home orders. That doesn’t mean they don’t work; for instance, there really is evidence that mask wearing reduces spread (albeit only moderately). On the other hand, the evidence supporting those measures is largely observational or even anecdotal — which is to say far less compelling than AstraZeneca’s trial with all its flaws.

We can also put an upper limit on the effectiveness of NPIs based on Covid’s progression over the last year. Simply put, if the mix of NPIs endured by the average American over the last year matched AstraZeneca’s 70% effectiveness [1], the virus would have stopped spreading long ago. That’s for the same reason that the herd immunity threshold in a simple SIR model is quoted at 60-70%; if you take Covid’s R0 value (which is quoted at 2.5-3) and reduce spread by 70%, the virus will start to fizzle out. This doesn’t mean that social distancing hasn’t been worth it, but it strongly implies that it’s less effective than the AstraZeneca vaccine.

The bottom line is that lockdowns could never gain FDA approval. It’s a dumb analogy to make; for starters, we can’t study policies like stay-at-home orders through a randomized controlled trial. The high standard of proof required for FDA approval isn’t the right bar for every decision, and we had to make decisions in an emergency.

But it’s also not an analogy. The AstraZeneca vaccine apparently falls just short of the FDA’s standard for approval, so we’re directly substituting measures are less effective, far more painful, and and supported by far more ambiguous evidence. The vaccine is 70% effective, costs a few dollars, and has only minor side effects. Those claims are backed by an RCT, albeit an imperfect one. None of the anti-Covid measures we’ve had in place for the last year have such an attractive cost-benefit ratio and none of them are supported by such clear evidence.

I firmly believe in appropriately testing drugs and have written as much in the past. I also think that many of the social distancing measures taken over the last year have been necessary. At the same time, I’m also increasingly troubled by the toll lockdowns have taken on people, whether that shows up as drug overdoses, suicidal ideation, or just the general unhappiness that has characterized the last year for so many. We’ve sacrificed so much to achieve protection that doesn’t even match what AstraZeneca’s vaccine offers. Why is the FDA still dragging its feet on approval?

Notes:

[1]: I use the headline figure of 70% reduction in symptomatic cases throughout this piece. In the context of calculating herd immunity, the relevant number is the reduction in transmission which is estimated at a very similar 67%