6 ms·

The fact that the confidence intervals for the excess risks aren't roughly symmetric around zero, but rather biased to mostly positive, is worrying. Most of th

by ponow 4y ago

The fact that the confidence intervals for the excess risks aren't roughly symmetric around zero, but rather biased to mostly positive, is worrying. Most of the probability is for positive excess risk. If it were just noise, wouldn't we expect the excess risk CI's to be roughly centered around zero?

benjaminwootton 4y ago
The fact that benefits are not an absolute slam dunk over risks is also worrying given the coercion that was used to get people to take these vaccines.
- hprotagonist 4y ago
  that metric is not quantified here.
- cortesoft 4y ago
  This study doesn't talk about benefits at all.
  - josephcsible 4y ago
    Isn't "the risk reduction for COVID-19 hospitalization" a benefit?
    - Arnt 4y ago
      "A" benefit? I'd say no. It's a wide range of benefits ranging from "not spending a night in hospital" to "surviving". If you know that the benefits and risks result in similar distributions of the risks, then you can just compare the numbers. If not, you have to check the distribution first.
checkyoursudo 4y ago
It's been a couple years since I had to report risk ratio (RR), and it's not something that I have used much in my research, but hear me out. The RR CI must be >= 0. The calculation is based on the incidence in a population and the incidence in a sub-population, which are themselves fractions of positive integers (e.g., 1/1000 and 1/100). I don't think RR can ever be negative, or at least I'm not sure how you would get that without messing up your data or calculations? RR of >1 means there is excess risk, and <1 means anti-excess risk which is a word I cannot think of right now (just, less risk than the baseline anyway). But RR is also just a measure of risk over baseline to begin with, so a 1.5 RR on something that happens 1/10,000 times is ... not that concerning? I wouldn't be concerned, anyway. Depending on what the risk is, I guess. So, no, I don't think we would expect RR CI to be centered around zero at all. The fact that the main effect is so small and crosses from neg to pos is by far more concerning. I would not submit results like that for publication.
- ponow 4y ago
  Then if we throw a logarithm in there we should be seeking zero-centered intervals, I'm thinking, as opposed to non-negative intervals containing 1 (no effect). Any why normal distributions are even part of such discussions of ratios seems crazy, given that ratios can't be negative, but log-ratios can. Not my area; could it be that a log or similar non-linearity was already effectively included? If the confidence interval method was only using ratios, then the lower bound should have been zero, and negatives impossible. Sanity checking results shouldn't be so difficult.
LorenPechtel 4y ago
Actually, I would not expect them to be. There's inherently a risk from a vaccine because there is inherently a risk from riling up the immune system. Note, however, that almost all the effects blamed on the vaccine are also effects we see with the virus. Does this not strongly suggest that what we are actually seeing is how the body reacts to the spike protein? If so, that means that if the patients got the real thing they would have the same or worse outcome? Realistically, the comparison shouldn't be vaccine vs nothing, but vaccine vs infection--and by that yardstick they aren't even in the same ballpark.
gus_massa 4y ago
If there is no effect, you expect that half of the times the confidence interval are biased to mostly positive values and you expect that half of the times the confidence interval are biased to mostly negative values. Moreover, if there is no effect you expect that some of the intervals are almost centered but you also expect that many of them are very biased. Moreover^2, if there is no effect, You even expect that in a 5% of them the 95% confidence interval does not include 0. This is a nice trick to detect fake data in blab reports from students. The intervals are too centered.

by ponow 4y ago

benjaminwootton 4y ago

The fact that benefits are not an absolute slam dunk over risks is also worrying given the coercion that was used to get people to take these vaccines.

hprotagonist 4y ago
that metric is not quantified here.
cortesoft 4y ago
This study doesn't talk about benefits at all.
- josephcsible 4y ago
  Isn't "the risk reduction for COVID-19 hospitalization" a benefit?
  - Arnt 4y ago
    "A" benefit? I'd say no. It's a wide range of benefits ranging from "not spending a night in hospital" to "surviving". If you know that the benefits and risks result in similar distributions of the risks, then you can just compare the numbers. If not, you have to check the distribution first.

checkyoursudo 4y ago

It's been a couple years since I had to report risk ratio (RR), and it's not something that I have used much in my research, but hear me out. The RR CI must be >= 0. The calculation is based on the incidence in a population and the incidence in a sub-population, which are themselves fractions of positive integers (e.g., 1/1000 and 1/100). I don't think RR can ever be negative, or at least I'm not sure how you would get that without messing up your data or calculations? RR of >1 means there is excess risk, and <1 means anti-excess risk which is a word I cannot think of right now (just, less risk than the baseline anyway). But RR is also just a measure of risk over baseline to begin with, so a 1.5 RR on something that happens 1/10,000 times is ... not that concerning? I wouldn't be concerned, anyway. Depending on what the risk is, I guess. So, no, I don't think we would expect RR CI to be centered around zero at all. The fact that the main effect is so small and crosses from neg to pos is by far more concerning. I would not submit results like that for publication.

ponow 4y ago
Then if we throw a logarithm in there we should be seeking zero-centered intervals, I'm thinking, as opposed to non-negative intervals containing 1 (no effect). Any why normal distributions are even part of such discussions of ratios seems crazy, given that ratios can't be negative, but log-ratios can. Not my area; could it be that a log or similar non-linearity was already effectively included? If the confidence interval method was only using ratios, then the lower bound should have been zero, and negatives impossible. Sanity checking results shouldn't be so difficult.

LorenPechtel 4y ago

Actually, I would not expect them to be. There's inherently a risk from a vaccine because there is inherently a risk from riling up the immune system. Note, however, that almost all the effects blamed on the vaccine are also effects we see with the virus. Does this not strongly suggest that what we are actually seeing is how the body reacts to the spike protein? If so, that means that if the patients got the real thing they would have the same or worse outcome? Realistically, the comparison shouldn't be vaccine vs nothing, but vaccine vs infection--and by that yardstick they aren't even in the same ballpark.

gus_massa 4y ago

If there is no effect, you expect that half of the times the confidence interval are biased to mostly positive values and you expect that half of the times the confidence interval are biased to mostly negative values. Moreover, if there is no effect you expect that some of the intervals are almost centered but you also expect that many of them are very biased. Moreover^2, if there is no effect, You even expect that in a 5% of them the 95% confidence interval does not include 0. This is a nice trick to detect fake data in blab reports from students. The intervals are too centered.