The main reason vaccine trials normally take years is money. Normally you wouldn't start manufacturing a vaccine before you know it at least works in animals. You also wouldn't _simultaneously_ vaccinate the 30000 people needed for phase 3 data. The Moderna study has over 100 study locations [1]. A normal vaccine trial might have 10-20. With that number, it takes much longer to get the same number of people in the study. Given the virtually unlimited financial resources coming from the government, they were able to conduct the study, as well as begin manufacturing, in record time.

All the vaccine trials are following participants for over 2 years, at which point many of the 1-in-10000 effects should be known. If you look at previous problematic vaccines [2], most rare side effects (> 1-in-100000) were only seen after MILLIONS of innoculations, meaning you only see them in Phase 4 (post-approval) studies as they are extremely rare. Given the high case fatality rate of COVID-19 [3], the risk of an unseen extremely rare reaction is much lower than the risk of dying from COVID-19, so general speaking, you should get vaccinated.

[1] https://clinicaltrials.gov/ct2/show/NCT04470427#contactlocat...

[2] https://www.cdc.gov/vaccinesafety/concerns/concerns-history....

[3] https://ourworldindata.org/mortality-risk-covid

Honest question: would you wait if there was a pact to the effect that either everyone takes the vaccine, or no one does?

Because obviously not getting vaccinated when everybody else does is almost always the best choice game-theoretically speaking; it doesn't really say anything on the vaccine (it may say more on the civic sense of the individual).

Let's see... it's a calculus of death vs potential long term adverse effect:

P(death) < 0.05% for the vast majority of the population (Age < 65) vs P(unknown) long term harm in life quality, my wild estimate: P(0.1%). It's within the same order of magnitude.

I'd take the vaccine in a heartbeat, long term effects and all, if P(death) was 10% or higher.

If you get covid, there's a P(unknown) of long term harm as well.

It's an unknown vs an unknown. Novel virus, novel immunization (first mRNA or maybe first widespread modified adenovirus)

But from what I understand the vaccine does not prevent you from getting the Virus and it's unknown side effects

You're correct about long term harm, I lumped them together. A better formula would be:

  P(covid death or long term harm) < 0.05%
  vs.
  P(vaccine long term harm) = ?

since my estimate is they're about the same order of magnitude, I prefer to take the devil-I-know vs. the small-scale-tested vaccine.

I model the potential harm based on a similarly rushed vaccine of 1976 which caused a large increase in cases of Guillain-Barré Syndrome [0]. Longer trials and phase IV monitoring was instituted in response to this rushed vaccine.

[0] https://en.wikipedia.org/wiki/1976_swine_flu_outbreak

What is your reason for assuming P < 0.05% for long-term harm due to COVID-19? You quite clearly assumed 0.05% for death only, therefore attributing even less to long-term health effects. But so far, there appear to be many indications that COVID-19 does have significant medium (and possibly long) term consequences even in asymptomatic cases.

I fully admit that I am not up to speed in proven COVID-19 medium- or long-term effects beyond having seen headlines, but to essentially assume none when there have been very obvious and multi-faceted indications of them implies you must have a very good basis of research to base your estimate on. Care to share?

I take the IFR (infection fatality rate) for age < 65, and multiply by 10 to get a worst case of IFR+long term harm. Note that we're talking about severe long term harm - such as autoimmune disease etc. Not "it took me 3 months to recover" stories.

For example, IFR(age<65) in geneva[0] is 0.0045%. multiply by 10 to include also long term harm, you get 0.045%.

[0] https://www.thelancet.com/pdfs/journals/laninf/PIIS1473-3099...

Ok, I can get behind that thought process.

I will note that the cited 0.0045% infection fatality rate number is quite old/outdated (June 1st). It's not easy to match different areas under the age curve (which seems to be exponential) and that it can vary a lot by region, but here is a review paper which has your number for ages 0-34 (Table 3) and (substantially) higher ones for ages above that:

https://www.medrxiv.org/content/10.1101/2020.07.23.20160895v...

How about P(get covid, infect others, they die)? Does that factor in to your calculations?

What about the known long term effects of Covid?