Vaccines

Viewpoints about vaccines are an interesting case study in epistemology.

This page is a stub, created on 2021-10-02 (last updated on 2020-10-03). Its contents are notes on the issues and angles I want to address about this topic.


So much to cover...

A note I sent to Andrew about what data I want to make an informed decision:

So as I was thinking about it, here's basically what I want to know:

Within as granular a dataset as possible (ie, a cohort as close to my characteristics as possible; eg, age, height/weight, other health factors, etc), of all COVID-positive cases, what is the percentage of fatalities? What is the percentage of some kind of long-term effect (eg, indefinite loss of smell/taste, cognitive impairment, etc)? The statements I'd like to see would be like "Within the cohort, x% of COVID-positive cases resulted in death.".

Same dataset, but now looking at adverse effects from J&J and mRNA vaccines (as separate figures)? The statements I'd like to see would be like "Within the cohort, (J&J | Pfizer | Moderna) had x% rate of adverse reactions.", and "adverse reactions" would need to be qualified somehow.

Same dataset, but now looking independently at J&J and mRNA effectiveness rate each at (1) preventing COVID infection in the first place, and (2) reducing death / adverse long-term effect rates of contracting COVID by some amount. So these statements would be something like (1) "Within the cohort, those vaccinated with (J&J | Pfizer | Moderna) had their risk of contracting COVID reduced from x% to y%." and (2) "Within the cohort, those vaccinated with (J&J | Pfizer | Moderna) who had a breakthrough case of COVID had their risk of death reduced from x% to y%." and "Within the cohort, those vaccinated with (J&J | Pfizer | Moderna) who had a breakthrough case of COVID had their risk of adverse long-term effects reduced from x% to y%." for some qualification of "adverse long-term effects".

I know that might be a tall order, but that strikes me as a reasonable desire for information to make an informed decision.

Data for an Informed Decision

Baseline cohort (or as close this as possible):

Frankly, with how well-studied COVID-19 and its vaccines are (especially given how vociferously its proponents claim that point), as well as how huge our sample size is and how much data I know firsthand is being collected (from participating in lots of surveys), I don't think it's unrealistic to think that this data exists for something very close to the exact cohort I'm looking for. Whether it's readily made available is a different question entirely, but that's an interesting philosophical/cultural discussion.

"long-term adverse effects" may be any of the known phenomena, including

And the same principle applies for adverse effects from the vaccines (for which we inherently can't have long-term data), but I think it's fair to lump in hospitalization and death into a monolithic "adverse effects" there.

Ideally, we'd have a breakdown by each different adverse effect.

risk for sub-cohort →
sub-cohort ↓
COVID-19 infection
hospitalization
from COVID-19
long-term adverse effects
from COVID-19
death
from COVID-19
adverse effects
from vaccine
unvaccinated         n/a
Johnson & Johnson          
Moderna          
Pfizer