Read on Twitter
I’m really surprised not to have seen more quantitative analysis of the one/two/delayed dose vaccine options. I’ve seen loads of qualitative takes from qualified people, but don’t we need to plug some numbers in to work out what to do next? Has anyone and I’ve missed it?
This thread is good on trying to make decisions under uncertainty, and I think this could be the key point—but it really doesn’t seem like there’s much expected value calculation going on. https://twitter.com/robertwiblin/status/1345800499019264003
Which course of action will result in the fewest infections/deaths? That’s the key here, and the only way to make a decent best-guess is surely to plug some numbers in.
Relatedly, it absolutely staggers me that the clinical trials didn’t try one dose, and different dose intervals. Sure, 40,000 people is a big clinical trial, but couldn’t we have made it 100,000, or 200,000 and tried a few more options dose-wise?
Perhaps I’m being entirely naïve about clinical trial logistics and there’s some insurmountable bottleneck even with infinite money, but the value of that data to potentially effectively double vaccination rates (or tell us we shouldn’t) seems absolutely huge to me.
Surely smart vaccine people and pharma companies knew that numbers of doses could be varied in March, when the seriousness of covid was obvious, and could have made the case to do bigger trials so we could have all the options on the table? Was there a good reason not to?
This vaccine escape stuff also seems ripe for a bit of numerical modelling. Here’s a virology expert arguing specifically for two doses, basically to make sure the immune response totally overwhelms the virus and doesn’t give it time to mutate. https://twitter.com/florian_krammer/status/1345791187236646912
This is the counterpoint—and just to highlight that most of this excellent thread is about the risk of escape generally—but if single doses would reduce spread that could be more effective at ensuring vaccines continue to work. https://twitter.com/Williams_T_C/status/1345708328643682307
Say single doses double numbers vaccinated so have twice the effect reducing transmission, but triple the risk of vaccine-escaping mutations by weakening the immune response a bit…that would make single doses a bad idea in terms of avoiding vaccine escape.
Then what matters is the absolute risk of a vaccine-escaping mutation, how bad that would be, how quickly we could pivot (good news about mRNA vaccines is they can be redesigned in days and hopefully reapproved quickly without full-scale trials)… https://twitter.com/ATabarrok/status/1345829856538464263
…and how that compares to the number of deaths avoided with the best transmission/death-avoiding vaccine schedule in the event mutations don’t occur.
Any quantitative model for this would involve a significant number of best-guess numbers, but it would at least tell us which of these parameters are most significant, which it’s really hard to do based on narrative descriptions, even from qualified people.
And I think we need to have a long, hard look at ourselves for the next pandemic and do some proper contingency planning, so these entirely predictable policy debates aren’t being had on Twitter, on-the-fly, based on incomplete data, while people are dying.
