TLDR: The value of a vaccine, & thus your distribution priorities, depends on the counterfactual. If CF is no/delayed vaccine & *suppression* (eg lockdown), main benefit of vaccine is economic activity. If CF is no/delayed vaccine & *no suppression*, main benefit is health. 2/20
In the former CF, you prioritize the most productive people. In the latter CF, you would prioritize people at greatest risk of health harm from COVID. 3/20
Background: I'm thinking about how to value a vaccine w/ efficacy Z & price p? What's the value to difft groups, defined by observables like location, demographics? We use that to advise LMICs on distribution priority & ultimately, how many doses to purchase. 4/20
The value of a vaccine depends on the counterfactual (CF). To most simply illustrate the core issue, assume the CF involves no vaccine. But that's not enuf to describe outcomes in the counterfactual. We need to know what would happen with suppression policy w/o a vaccine. 5/20
Please indulge me -- for pedagogical reasons -- in some bad generalizations, which I'll revisit later. Assume, incorrectly, that there is no social distancing. Also, assume economic outcomes include on wages or profits, not health. Then we might think... 6/20
In the CF of vaccine & no suppression, health outcomes = bad, but economic outcomes = good.
In CF of no vaccine & suppression, then health = good, but economy = bad.
Now consider outcomes w/ vaccine: health = good & economy = good. (I assumed vaccine ends suppression.) 7/20
This implies that if the CF includes no suppression, the main benefit of vaccines is improvement in health.
If CF includes suppression, the main benefit is economic.
This has big implications for vaccine distribution priorities. 8/20
If the counterfactual includes no suppression, we should prioritize people facing the greatest health risk from COVID. (I'll call this the conventional approach to distribution.) 9/20
If the CF includes suppression, we should prioritize those who are the most productive.
So your distribution plan depends on yr belief about continued suppression w/o vaccine. This result was surprising to me cuz I assumed suppression in the CF but started valuing vaccines on health benefit.
Or: the key to valuing a vaccine is predicting how long suppression lasts.
Some caveats. 1/ If there is a lot of voluntary distancing, the no vaccine + no suppression CF is incorrect. Even if the govt doesn't suppress activity, ppl do so on their own. But then vaccine priorities should focus on productivity. 11/20
2/ Delay complicates things. With short delay, ppl may be willing to suppress a bit longer. With long delay, suppression fatigue erodes distancing. So short delay suggests prioritizing productive ppl, long delay prioritizing ppl w health risk. 12/20
The key is predicting what ppl think the delay til the vaccine will be. (Are there any surveys on this?) 13/20
3/ I assumed a vaccine eases suppression. That's true if we vaccine everyone we need to. But if there is a shortage of doses, & we can only vaccine 1/2 the needed ppl in a location, then suppression may have to continue. 14/20
(This suggests value in vaccinating 1 area then another rather than spreading doses out, but I am guessing prioritizing certain areas over others may be politically difficult.) 15/20
If partial vaccination includes continued suppression, then the benefit of partial vaccination is different than full vaccination. If the CF is no vaccine + suppression, then benefit becomes smaller: either way, economy is bad & health is medium. 16/20
Implication: Countries should be willing to pay less for initial doses unless they can store them up and vaccinate everyone at once -- though I doubt that is politically feasible. 17/20
If CF (to partial vaccination) is no/delayed vaccine + no suppression, then the benefit of vaccination is the same lives v. livelihoods tradeoff: partial vaccination increases health a bit but reduces economic activity. 18/20
But there is a hitch: even if you think the health benefits of suppression outweigh the economic costs, you may not think that a partial vaccination that offers only medium health is enough to outweigh the economic costs of continued suppression. 19/20
I know there is a lot I'm ignoring. This is thinking-in-progress tweet. Eg, time discounting is critical. That may outweigh everything -- though much more so in the partial vaccination analysis than the full vaccination analysis. END OF THREAD 20/20
• • •
Missing some Tweet in this thread? You can try to
force a refresh
where CI is confidence. But economically this is telling you the cost you must incur (N) for a given level of quality (inputs into f, including MDE).
That has economic value, but is not really how studies are done or items are purchased. Consumers don't say, I have to have a car of this quality, just tell me what to spend. They see a vector of (quality, price) pairs & choose pair that maximizes U s.t. their budget constraint.
I want to highlight this MR post by @ATabarrok. It has AFAIK a novel approach to trials that warrants serious discussion: release a drug & randomize access via lottery. Then compare outcome across ppl that won lottery and those who lost. Like the Oregon Health Ins Experiment.
If we do this, we may want to subsidize the price, so that demand for the lottery is high. This may also address ethical issues.
But there are some important issues with this approach. Not all insurmountable. Here is a first cut.
1/Selection. Ppl who are randomized to no access will get other treatment. Some will take nothing, others will take another conventional treatment if available.
But this is a problem with existing trials. I noted an old article on selection into RCTs.
CAN YOU HELP SLUMS WITH HANDWASHING?
This is an urgent request form NGOs working to stop COVID in Indian slums. Details of problem in THREAD. Summary: we need solution that slums can implement themselves w/ reused water. Please retweet to increase eyeballs. 1/ @vandanagoyal01
Handwashing requires water & soap. Both are in short supply in slums cuz they usually don't have water pipes. Govt sometimes sent water tankers, but those have slowed. 2/ @adamschilton@tariqthachil@adam_m_auerbach
We can't create new handwashing stations. This takes time to scale w/o a lockdown. With India's 21 day lockdown, definitely cannot ship them there. Plus, they require water. 3/ @pritika13@nebuer42
A key issue in the cost-benefit analysis on #COVID19 is how ending suppression would affect the economy*. I assume main mechanism is through *voluntary* social distancing. What is the best estimate of that? 1/4 @ATabarrok
*NOTE: Release is terrible for health.
Is there an estimate of prevalence response elasticity (using econ epi terms) for COVID? If the policy switched, should we assume that there would be a comms strategy to reduce fear? 2/4
A 2nd reason release may hurt the economy is death. But those effects are mitigated because most of those who pass are above 70 (sadly). There will be absenteeism among young. 3/4
Thinking about suppression v. mitigation for India. One issue that came up is breadth of immunity & future corona flare-ups. Suppression relies on vaccine acquired immunity (VAI); mitigation on natural immunity (NI). This may have implications for comparing loss of life. 1/6
By immune breadth I mean the⬇️in probability of infection/harm as you ⬆️ distance between current & next strain in genetic/antigenic space. 2/6
For old time gamers, think of it as the radius of the explosion when you fire at incoming missiles in Missile Command: bigger breadth is better cuz you knock out more missiles. The bigger the immune breadth, the more protected you are against future strains.* 3/6