4 ms·

This data needs to be framed against the % of the population that was vaccinated at the time they contracted the virus, right? Otherwise, for all we know, the

This data needs to be framed against the % of the population that was vaccinated at the time they contracted the virus, right? Otherwise, for all we know, the vaccine could actually increase mortality if the % vaccinated was smaller than the % of deaths that group represents. I'm assuming that isn't the case but hard to make meaningful conclusions about efficacy without that context. The apparent efficacy of vaccination is at the very least somewhat inflated because between Jan and July is when everyone got vaccinated, I'm assuming in England, like here in the US, basically no one was vaccinated in Jan, Feb. Also, bucketing people who were vaccinated with previously infected makes sense as an additional lens to look at things through, but it seems odd that they then omitted reporting those group separately. I'm very curious how one did compared to the other.

lmilcin 5y ago
> This data needs to be framed against the % of the population that was vaccinated at the time they contracted the virus, right? No, it does not. It compares chance of you dying to covid compared to all other possible causes. This dimensionless number does not depend on relative population sizes. So if in one population you get 458 deaths from Covid compared to 57,263 from all other causes and in the other population you get 38,964 from covid compared to 65,170 from all other causes, we can immediately tell that one population fared much better at least when it comes to your chance to die of covid compared to chance of dying to all other causes. Just by the fact of being in one population than in the other you get (38964/(38964+65170)) / (458/(458+57263)) ~= 47 times less chance of dying to covid assuming rates of deaths to all other causes were the same between both populations. So what are these population? To be in the better-faring population you had to be fully vaccinated for at least 21 days and the worse population was completely unvaccinated. Now, I am pretty sure there are other factors that factor in that number 47. It is likely that vaccinated people treat Covid more seriously and are more likely to practice social distancing, washing hands and be generally mindful of various things you do to minimize chance of getting infected. They might also possibly wait shorter amount of time before calling for help if they see signs. While this is pure conjecture on my part I think it is reasonable to assume this happens to some extent and reduces rates of infection within vaccinated population.
- TechBro8615 5y ago
  Note this does assume that you are not immortal and will eventually die. If you’re holding out hope for Mr Bezos to discover an anti-aging elixir, you might interpret these stats more pessimistically.
- eagsalazar2 5y ago
  >>> So if in one population you get 458 deaths from Covid compared to 57,263 from all other causes and in the other population you get 38,964 from covid compared to 65,170 from all other causes, we can immediately tell that one population fared much better at least when it comes to your chance to die of covid compared to chance of dying to all other causes. I think your math is off. What if the first population only has 458 people in it? 100% mortality for that population would obviously suggest you want to be part of the other population even if their total deaths were higher. The issue I'm calling out is that we don't know the size of each population. I certainly don't think vaccines increase mortality, I actually believe it has a large positive impact. The point is this data doesn't help us understand how impactful it really is and, given all the debate over this, if they'd framed the data with better context, this could actually be a very useful study in ending that debate.
  - lmilcin 5y ago
    > I think your math is off. What if the first population only has 458 people in it? I think your thinking is off. How do you get 57,263 other deaths in a population of 458? This data is not comparing population sizes. It avoids it by comparing your chance of dying to covid to something else that is supposed stable between populations -- chance of dying to all other causes. We just need to know that population sizes being compared are large enough (as evidenced from the number of deaths included) to be statistically significant.
- jollybean 5y ago
  The problem with this generalization is that the vaxxed/unvaxxed populations are ultimately very different. Vaxxed group is older, unvaxxed are younger. i.e. groups are not randomly distributed. They are going to have different risk factors and die of different things. This showed up in British Columbia's health data that showed infections and hospitalizations way, way down for the vaxxed relative to unvaxxed, but the ratio between the two for actual deaths was not as extreme, because vaccinated population skewed older, more vulnerable, more likely to die even when fully vaxxed. Should note that age-standardized mortality rates are included in the study.

lmilcin 5y ago

> This data needs to be framed against the % of the population that was vaccinated at the time they contracted the virus, right? No, it does not. It compares chance of you dying to covid compared to all other possible causes. This dimensionless number does not depend on relative population sizes. So if in one population you get 458 deaths from Covid compared to 57,263 from all other causes and in the other population you get 38,964 from covid compared to 65,170 from all other causes, we can immediately tell that one population fared much better at least when it comes to your chance to die of covid compared to chance of dying to all other causes. Just by the fact of being in one population than in the other you get (38964/(38964+65170)) / (458/(458+57263)) ~= 47 times less chance of dying to covid assuming rates of deaths to all other causes were the same between both populations. So what are these population? To be in the better-faring population you had to be fully vaccinated for at least 21 days and the worse population was completely unvaccinated. Now, I am pretty sure there are other factors that factor in that number 47. It is likely that vaccinated people treat Covid more seriously and are more likely to practice social distancing, washing hands and be generally mindful of various things you do to minimize chance of getting infected. They might also possibly wait shorter amount of time before calling for help if they see signs. While this is pure conjecture on my part I think it is reasonable to assume this happens to some extent and reduces rates of infection within vaccinated population.

TechBro8615 5y ago
Note this does assume that you are not immortal and will eventually die. If you’re holding out hope for Mr Bezos to discover an anti-aging elixir, you might interpret these stats more pessimistically.
eagsalazar2 5y ago
>>> So if in one population you get 458 deaths from Covid compared to 57,263 from all other causes and in the other population you get 38,964 from covid compared to 65,170 from all other causes, we can immediately tell that one population fared much better at least when it comes to your chance to die of covid compared to chance of dying to all other causes. I think your math is off. What if the first population only has 458 people in it? 100% mortality for that population would obviously suggest you want to be part of the other population even if their total deaths were higher. The issue I'm calling out is that we don't know the size of each population. I certainly don't think vaccines increase mortality, I actually believe it has a large positive impact. The point is this data doesn't help us understand how impactful it really is and, given all the debate over this, if they'd framed the data with better context, this could actually be a very useful study in ending that debate.
- lmilcin 5y ago
  > I think your math is off. What if the first population only has 458 people in it? I think your thinking is off. How do you get 57,263 other deaths in a population of 458? This data is not comparing population sizes. It avoids it by comparing your chance of dying to covid to something else that is supposed stable between populations -- chance of dying to all other causes. We just need to know that population sizes being compared are large enough (as evidenced from the number of deaths included) to be statistically significant.
jollybean 5y ago
The problem with this generalization is that the vaxxed/unvaxxed populations are ultimately very different. Vaxxed group is older, unvaxxed are younger. i.e. groups are not randomly distributed. They are going to have different risk factors and die of different things. This showed up in British Columbia's health data that showed infections and hospitalizations way, way down for the vaxxed relative to unvaxxed, but the ratio between the two for actual deaths was not as extreme, because vaccinated population skewed older, more vulnerable, more likely to die even when fully vaxxed. Should note that age-standardized mortality rates are included in the study.