4 ms·

> This waters down case rates in the unvaccinated group and ends up lowering the vaccine efficacy estimates. Okay. Let's be clear on something. Telling the tr

> This waters down case rates in the unvaccinated group and ends up lowering the vaccine efficacy estimates. Okay. Let's be clear on something. Telling the truth is never a downside. It is not a negative that the U.S. doesn't pretend that immunity due to prior infection is due to the vaccine. Suggesting that these cases should be counted in favor of vaccine efficacy is brazenly dishonest. This article is flamebait.

FooBarBizBazz 5y ago
The title is unfortunate flamebait -- that's for sure -- but the logic later on is ok. Like, when she writes, > This waters down case rates in the unvaccinated group and ends up lowering the vaccine efficacy estimates. the implication is that she'd prefer for three groups to be tracked: 1. never infected, never vaccinated 2. previously infected but not vaccinated 3. vaccinated. The idea isn't to lump 2 and 3 together, but merely to stop lumping 1 and 2 together. Fair enough.
- loonster 5y ago
  The logic is bad. You need to compare 5 groups. Your group 3 lumps together people that were previously infected with those that were not previously infected. By doing that, it would artificially show the vaccine was more effective than it should be. The 5 groups: 1a. Never vaccinated, No prior infection 1b. Never vaccinated, prior infection 2a. Vaccinated, No prior infection 2b. Vaccinated, prior infection (Infected then Vaccinated) 2c. Vaccinated, prior infection (Vaccinated then Infected)
  - SideburnsOfDoom 5y ago
    Finer grained is better, yes. But I don't think that this supports the idea that anyone "pretends that immunity due to prior infection is due to the vaccine."
    - loonster 5y ago
      Yes it does. Just replace the vaccine with snake oil and follow it to its logical conclusions. I'll use some BS numbers to make the point. Let's assume: study population of 100, 20% of the population had a prior infection, half are given the snake oil. Starting conditions: 1. Never infected, Never Snaked: 40 2. Prior infected, Never Snaked:10 3. Never infected, Snaked: 40 4. Prior infected, Snaked: 10 Now let's assume wave 2 happens. 20% of the never infected gets infected, and 0% of the previously infected get reinfected. Both groups 1+2 combined and 3+4 combined have an average infection rate of 16% for wave 2. The groups are equal because snake oil is bullshit. What the Ars author was suggesting is to compare group 1 with group 3+4 combined. Now you have a 20% infection rate vs a 16% infection rate. The snake oil appears to have an (20-16)/20 = 25% effectiveness just by including previously acquired immunity.
  - nojokes 5y ago
    You forget boosting :) and infected after boosting and ...
SideburnsOfDoom 5y ago
> Suggesting that these cases should be counted in favor of vaccine efficacy is brazenly dishonest. This article is flamebait. This seems to be a complete misreading of what the article is saying. Nowhere does it claim that "immunity due to prior infection should be counted as due to the vaccine" merely that it makes the "unvaccinated" group look better than it actually is, as it should be further subdivided to "account for past infection" in a subgroup. Drawing the conclusion that we should therefore ignore the article entirely is unwarranted. Let us not be brazenly dishonest.
- DannyBee 5y ago
  1. This would require testing tons of people for antibody response all the time. That seems hard. We do at least have antibody tests that will test for infection vs vaccination. 2. If you are going to go this far, you'd also want the "vaccinated but had no effect" group. The vaccine will not generate an immune response (or an effective one) in everyone, that's just how it works. So the vaccinated group (and "breakthrough cases") include some percent of people who were vaccinated, but it didn't really do anything to them. You'd want to know this number, because you want to know things like "how many breakthrough cases are people where the vaccine had no real effect vs people where the vaccine generated an effective antibody response, but not against this variant" Especially because the "breakthrough" rate is small.
  - SideburnsOfDoom 5y ago
    OK, but if doing this subdivision well is prohibitively hard, that does not not make the article dishonest about these groups. It does not "pretend that immunity due to prior infection is due to the vaccine".
  - nojokes 5y ago
    Not necessarily. You can also sample the subset of people.

FooBarBizBazz 5y ago

The title is unfortunate flamebait -- that's for sure -- but the logic later on is ok. Like, when she writes, > This waters down case rates in the unvaccinated group and ends up lowering the vaccine efficacy estimates. the implication is that she'd prefer for three groups to be tracked: 1. never infected, never vaccinated 2. previously infected but not vaccinated 3. vaccinated. The idea isn't to lump 2 and 3 together, but merely to stop lumping 1 and 2 together. Fair enough.

loonster 5y ago
The logic is bad. You need to compare 5 groups. Your group 3 lumps together people that were previously infected with those that were not previously infected. By doing that, it would artificially show the vaccine was more effective than it should be. The 5 groups: 1a. Never vaccinated, No prior infection 1b. Never vaccinated, prior infection 2a. Vaccinated, No prior infection 2b. Vaccinated, prior infection (Infected then Vaccinated) 2c. Vaccinated, prior infection (Vaccinated then Infected)
- SideburnsOfDoom 5y ago
  Finer grained is better, yes. But I don't think that this supports the idea that anyone "pretends that immunity due to prior infection is due to the vaccine."
  - loonster 5y ago
    Yes it does. Just replace the vaccine with snake oil and follow it to its logical conclusions. I'll use some BS numbers to make the point. Let's assume: study population of 100, 20% of the population had a prior infection, half are given the snake oil. Starting conditions: 1. Never infected, Never Snaked: 40 2. Prior infected, Never Snaked:10 3. Never infected, Snaked: 40 4. Prior infected, Snaked: 10 Now let's assume wave 2 happens. 20% of the never infected gets infected, and 0% of the previously infected get reinfected. Both groups 1+2 combined and 3+4 combined have an average infection rate of 16% for wave 2. The groups are equal because snake oil is bullshit. What the Ars author was suggesting is to compare group 1 with group 3+4 combined. Now you have a 20% infection rate vs a 16% infection rate. The snake oil appears to have an (20-16)/20 = 25% effectiveness just by including previously acquired immunity.
- nojokes 5y ago
  You forget boosting :) and infected after boosting and ...

SideburnsOfDoom 5y ago

> Suggesting that these cases should be counted in favor of vaccine efficacy is brazenly dishonest. This article is flamebait. This seems to be a complete misreading of what the article is saying. Nowhere does it claim that "immunity due to prior infection should be counted as due to the vaccine" merely that it makes the "unvaccinated" group look better than it actually is, as it should be further subdivided to "account for past infection" in a subgroup. Drawing the conclusion that we should therefore ignore the article entirely is unwarranted. Let us not be brazenly dishonest.

DannyBee 5y ago
1. This would require testing tons of people for antibody response all the time. That seems hard. We do at least have antibody tests that will test for infection vs vaccination. 2. If you are going to go this far, you'd also want the "vaccinated but had no effect" group. The vaccine will not generate an immune response (or an effective one) in everyone, that's just how it works. So the vaccinated group (and "breakthrough cases") include some percent of people who were vaccinated, but it didn't really do anything to them. You'd want to know this number, because you want to know things like "how many breakthrough cases are people where the vaccine had no real effect vs people where the vaccine generated an effective antibody response, but not against this variant" Especially because the "breakthrough" rate is small.
- SideburnsOfDoom 5y ago
  OK, but if doing this subdivision well is prohibitively hard, that does not not make the article dishonest about these groups. It does not "pretend that immunity due to prior infection is due to the vaccine".
- nojokes 5y ago
  Not necessarily. You can also sample the subset of people.