4 ms·

CDC says vaccines are only 66 percent effective at preventing Delta infection. As far as I can understand, this means 34% of vaccinated people develop a covid i

CDC says vaccines are only 66 percent effective at preventing Delta infection. As far as I can understand, this means 34% of vaccinated people develop a covid infection. I'm having a hard time squaring this with 'vaccines prevent infection' thesis. Perhaps I'm missing something obvious? (Yes, I'm aware the fraction is better against severe Delta infection. Yes, I agree 3x reduction is infection rate is a good thing.) https://www.webmd.com/vaccines/covid-19-vaccine/news/20210825/cdc-covid-19-vaccines-66-percent-effective-against-delta-variant https://www.webmd.com/vaccines/covid-19-vaccine/news/2021082...

lamontcg 5y ago
> CDC says vaccines are only 66 percent effective at preventing Delta infection WITH A 95% CONFIDENCE INTERVAL OF 26% TO 84% They only statistically modeled how much their unvaccinated population was already recovered from coronavirus based on estimates of community spread, among other issues with that number. https://www.cdc.gov/mmwr/volumes/70/wr/mm7034e4.htm https://www.cdc.gov/mmwr/volumes/70/wr/mm7034e4.htm And when you have a CI of something like that size, of course you wind up in the middle of it -- the actual value could very well be closer to 80% due to systematic uncertainties in the modeling.
- disgruntledphd2 5y ago
  Or 30%, to be fair.
  - lamontcg 5y ago
    published on the CDC site today: 77% (74%-80%) VE for pfizer and 92% (89%-93%) VE for moderna. With only some evidence of reduced VE against delta and/or waning in those > 75 years old: https://www.cdc.gov/mmwr/volumes/70/wr/mm7037e2.htm https://www.cdc.gov/mmwr/volumes/70/wr/mm7037e2.htm
    - disgruntledphd2 5y ago
      That's a very reassuring result, thanks for sharing. Interesting that moderna continues to do better, and looking forward to seeing the studies VE versus time since vaccination.
MatteoFrigo 5y ago
> this means 34% of vaccinated people develop a covid infection That's not what "effective" means. It means that if your probability of getting infected without the vaccine is p, your probability of getting infected with the vaccine is .34*p. One way to think about it is that the "ineffectiveness" is the relative risk P[infected|vaccinated]/P[infected|unvaccinated], and effectiveness is 1-ineffectiveness. A vaccine can be 100% effective even if P[infected] is negligible. P[infected] is per unit time and does not account for the dynamics of the situation. For example, if the vaccine is not powerful enough to make R<1, then everybody becomes infected eventually regardless of the vaccine.
- angelzen 5y ago
  Thanks!

lamontcg 5y ago

> CDC says vaccines are only 66 percent effective at preventing Delta infection WITH A 95% CONFIDENCE INTERVAL OF 26% TO 84% They only statistically modeled how much their unvaccinated population was already recovered from coronavirus based on estimates of community spread, among other issues with that number. https://www.cdc.gov/mmwr/volumes/70/wr/mm7034e4.htm https://www.cdc.gov/mmwr/volumes/70/wr/mm7034e4.htm And when you have a CI of something like that size, of course you wind up in the middle of it -- the actual value could very well be closer to 80% due to systematic uncertainties in the modeling.

disgruntledphd2 5y ago
Or 30%, to be fair.
- lamontcg 5y ago
  published on the CDC site today: 77% (74%-80%) VE for pfizer and 92% (89%-93%) VE for moderna. With only some evidence of reduced VE against delta and/or waning in those > 75 years old: https://www.cdc.gov/mmwr/volumes/70/wr/mm7037e2.htm https://www.cdc.gov/mmwr/volumes/70/wr/mm7037e2.htm
  - disgruntledphd2 5y ago
    That's a very reassuring result, thanks for sharing. Interesting that moderna continues to do better, and looking forward to seeing the studies VE versus time since vaccination.

MatteoFrigo 5y ago

> this means 34% of vaccinated people develop a covid infection That's not what "effective" means. It means that if your probability of getting infected without the vaccine is p, your probability of getting infected with the vaccine is .34*p. One way to think about it is that the "ineffectiveness" is the relative risk P[infected|vaccinated]/P[infected|unvaccinated], and effectiveness is 1-ineffectiveness. A vaccine can be 100% effective even if P[infected] is negligible. P[infected] is per unit time and does not account for the dynamics of the situation. For example, if the vaccine is not powerful enough to make R<1, then everybody becomes infected eventually regardless of the vaccine.

angelzen 5y ago
Thanks!