8 ms·

Germany stats all COVID cases wrt vaccination? Ah, in the Netherlands most if not all stats I've seen involve hospitalisation and especially IC. Those are often

by icoder 5y ago

Germany stats all COVID cases wrt vaccination? Ah, in the Netherlands most if not all stats I've seen involve hospitalisation and especially IC. Those are often misattributed the other way around, 50/50 sounds equal, until you consider 85% of the population is vaccinated (and most of the 50% vaccinated IC patients are old / sick / immunodeprived).

native_samples 5y ago
Yes, the German stat I'm talking about is all cases. Hospitalization and IC numbers have other problems, especially with confounding. I don't think Germany provides enough data to check for problems and anyway their stats are meaningless for as long as they are playing such obviously dishonest games. But the UK data seems to be more honest. There a problem can be seen: vaccines don't get administered to the very sickest people who are dying anyway, which makes vaccines look artificially good / non-vaccination look artificially bad. This shows up in the UK data as a spike in deaths in unvaccinated people for non COVID causes that correlates with vaccine rollouts. Obviously, that's not biologically caused. Rather it's because the sickest people aren't given the vaccines because there's no point, and end up concentrated into the unvaccinated cohort. Then they pick up COVID in hospital anyway as so many do, and this is used as evidence for why everyone should take the shot without de-confounding. Paper explaining the problems with the UK data that appear to show this confounding: https://www.researchgate.net/publication/356756711_Latest_statistics_on_England_mortality_data_suggest_systematic_mis-categorisation_of_vaccine_status_and_uncertain_effectiveness_of_Covid-19_vaccination https://www.researchgate.net/publication/356756711_Latest_st... They also try to control for this problem and when they do, effectiveness goes away, which is very distressing. VE against death is supposedly very strong so it shouldn't disappear when the rigor of the analysis is increased. There are other problems, like the standard delay period where people have been given a vaccine but count as unvaccinated. In one study, buried in the appendix, they show data that when this period is included the incidence between those who were vaccinated and those who weren't was equal, i.e. no effectiveness. This article summarizes some of these statistical issues: https://dailysceptic.org/2021/12/12/is-vaccine-effectiveness-against-death-mostly-a-statistical-illusion/ https://dailysceptic.org/2021/12/12/is-vaccine-effectiveness... It's very problematic that these data issues can occur but inevitable. Normally you want RCT evidence exactly to avoid these confounders, but COVID kills so few people that nobody was able to mount a trial big enough to show impact on mortality. Actually in the Pfizer trial more people died in the vaccine arm than the control arm, but there were so few deaths in total (from anything) they said this wasn't significant / meaningful. Hence you end up trying to find evidence in real-world observational data but those are hard to control.
- icoder 5y ago
  That's interesting, such confounding is an example for how difficult such research is. It's also two articles hammering on a single piece of data and making conclusions using Occam's razor. The comments on the second article don't help either. Most importantly it's not just death I'm worried about, it's also not getting hospitalised (or worse, staying in coma for days on the IC). Not only for the individual, but also for society. At least here in the Netherlands, that was spot on the point we had to go back to lockdown like measures. There's a lot of numbers that suggest vaccins do wonders for that, for all ages (a bit less for elderly actually). An obvious one is the much bigger amount of daily infections we could now support with less beds compared to before vaccinations. These numbers are dropping, possibly because vaccinations being longer ago. Hopefully boosters will mitigated this.
  - native_samples 5y ago
    The second article also discusses the problem of the 2 week delay period and links to the study appendix - different data and argument. Hospitalization is unfortunately also hopelessly confounded by the same factors - if you don't vaccinate people who are already dying then they are much more likely to be or end up in hospital - but then additional problems too. Again, RCT data didn't prove any effects on hospitalization, that just wasn't an end point they cared about. Why not? Well, partly because it's extremely subjective. To be hospitalized for a disease like COVID you must either catch it in hospital, or choose to go. COVID isn't a sudden collapse like a heart attack or stroke where bystanders may rush you to hospital in an ambulance. It's a steady degradation and at some point someone has to call a doctor or emergency number and say, I think I need to go to the hospital. Hospitals can turn you away if your COVID isn't serious enough but in many places they have monetary incentives not to do so, and that's ignoring all the psychological issues - if you have some spare beds giving someone who's terrified some attention may be better than turning them away. In normal/sane times this wouldn't really matter, as we can hope that people are roughly OK at judging whether they need this or not, and hospitals wouldn't spend resources on cases that aren't really needing to be there. But with COVID, years of relentless misinformation from newspapers and authority figures have totally destroyed people's understanding of the severity of the disease. US adults over-estimate the risk of hospitalization if you're unvaccinated by orders of magnitude: https://news.gallup.com/opinion/gallup/354938/adults-estimates-covid-hospitalization-risk.aspx https://news.gallup.com/opinion/gallup/354938/adults-estimat... "The results show that most Americans overstate the risk of hospitalization for both groups: 92% overstate the risk that unvaccinated people will be hospitalized, and 62% overstate the risk for vaccinated people .... For unvaccinated hospitalization risk, 2% of Democrats responded correctly, compared with 16% of Republicans. In fact, 41% of Democrats replied that at least 50% of unvaccinated people have been hospitalized due to COVID-19." 41% of left leaning voters think HALF of unvaccinated people have been hospitalized for COVID! The true number is less than 1%. That's an astonishing stat. I suspect these polls aren't representative and the true numbers are better than this but even so, they are indicative of the problem. What happens if you trust the media and government so much that you're totally convinced testing positive for COVID = required hospital visit, and then those same authority figures tell you that vaccines will "protect" you from hospitalization? Well, you're much less likely to go, even in situations where previously you would have gone straight to the ER and demanded attention. And that's fine, they'll recover on their own as they almost certainly would have done anyway, and they'll say "thank god I was vaccinated otherwise it would have been much worse". This is the sort of confounder a rigorous trial would seek to control for, but we don't have any useful trials. The vaccine trials were complete failures: for the only end point they targeted, they reported 95% effectiveness, which started dwindling to near or even below zero in the months after the trial period ended. No usefully accurate information was provided by them, which is shocking.

by icoder 5y ago

native_samples 5y ago

Yes, the German stat I'm talking about is all cases. Hospitalization and IC numbers have other problems, especially with confounding. I don't think Germany provides enough data to check for problems and anyway their stats are meaningless for as long as they are playing such obviously dishonest games. But the UK data seems to be more honest. There a problem can be seen: vaccines don't get administered to the very sickest people who are dying anyway, which makes vaccines look artificially good / non-vaccination look artificially bad. This shows up in the UK data as a spike in deaths in unvaccinated people for non COVID causes that correlates with vaccine rollouts. Obviously, that's not biologically caused. Rather it's because the sickest people aren't given the vaccines because there's no point, and end up concentrated into the unvaccinated cohort. Then they pick up COVID in hospital anyway as so many do, and this is used as evidence for why everyone should take the shot without de-confounding. Paper explaining the problems with the UK data that appear to show this confounding: https://www.researchgate.net/publication/356756711_Latest_statistics_on_England_mortality_data_suggest_systematic_mis-categorisation_of_vaccine_status_and_uncertain_effectiveness_of_Covid-19_vaccination https://www.researchgate.net/publication/356756711_Latest_st... They also try to control for this problem and when they do, effectiveness goes away, which is very distressing. VE against death is supposedly very strong so it shouldn't disappear when the rigor of the analysis is increased. There are other problems, like the standard delay period where people have been given a vaccine but count as unvaccinated. In one study, buried in the appendix, they show data that when this period is included the incidence between those who were vaccinated and those who weren't was equal, i.e. no effectiveness. This article summarizes some of these statistical issues: https://dailysceptic.org/2021/12/12/is-vaccine-effectiveness-against-death-mostly-a-statistical-illusion/ https://dailysceptic.org/2021/12/12/is-vaccine-effectiveness... It's very problematic that these data issues can occur but inevitable. Normally you want RCT evidence exactly to avoid these confounders, but COVID kills so few people that nobody was able to mount a trial big enough to show impact on mortality. Actually in the Pfizer trial more people died in the vaccine arm than the control arm, but there were so few deaths in total (from anything) they said this wasn't significant / meaningful. Hence you end up trying to find evidence in real-world observational data but those are hard to control.

icoder 5y ago
That's interesting, such confounding is an example for how difficult such research is. It's also two articles hammering on a single piece of data and making conclusions using Occam's razor. The comments on the second article don't help either. Most importantly it's not just death I'm worried about, it's also not getting hospitalised (or worse, staying in coma for days on the IC). Not only for the individual, but also for society. At least here in the Netherlands, that was spot on the point we had to go back to lockdown like measures. There's a lot of numbers that suggest vaccins do wonders for that, for all ages (a bit less for elderly actually). An obvious one is the much bigger amount of daily infections we could now support with less beds compared to before vaccinations. These numbers are dropping, possibly because vaccinations being longer ago. Hopefully boosters will mitigated this.
- native_samples 5y ago
  The second article also discusses the problem of the 2 week delay period and links to the study appendix - different data and argument. Hospitalization is unfortunately also hopelessly confounded by the same factors - if you don't vaccinate people who are already dying then they are much more likely to be or end up in hospital - but then additional problems too. Again, RCT data didn't prove any effects on hospitalization, that just wasn't an end point they cared about. Why not? Well, partly because it's extremely subjective. To be hospitalized for a disease like COVID you must either catch it in hospital, or choose to go. COVID isn't a sudden collapse like a heart attack or stroke where bystanders may rush you to hospital in an ambulance. It's a steady degradation and at some point someone has to call a doctor or emergency number and say, I think I need to go to the hospital. Hospitals can turn you away if your COVID isn't serious enough but in many places they have monetary incentives not to do so, and that's ignoring all the psychological issues - if you have some spare beds giving someone who's terrified some attention may be better than turning them away. In normal/sane times this wouldn't really matter, as we can hope that people are roughly OK at judging whether they need this or not, and hospitals wouldn't spend resources on cases that aren't really needing to be there. But with COVID, years of relentless misinformation from newspapers and authority figures have totally destroyed people's understanding of the severity of the disease. US adults over-estimate the risk of hospitalization if you're unvaccinated by orders of magnitude: https://news.gallup.com/opinion/gallup/354938/adults-estimates-covid-hospitalization-risk.aspx https://news.gallup.com/opinion/gallup/354938/adults-estimat... "The results show that most Americans overstate the risk of hospitalization for both groups: 92% overstate the risk that unvaccinated people will be hospitalized, and 62% overstate the risk for vaccinated people .... For unvaccinated hospitalization risk, 2% of Democrats responded correctly, compared with 16% of Republicans. In fact, 41% of Democrats replied that at least 50% of unvaccinated people have been hospitalized due to COVID-19." 41% of left leaning voters think HALF of unvaccinated people have been hospitalized for COVID! The true number is less than 1%. That's an astonishing stat. I suspect these polls aren't representative and the true numbers are better than this but even so, they are indicative of the problem. What happens if you trust the media and government so much that you're totally convinced testing positive for COVID = required hospital visit, and then those same authority figures tell you that vaccines will "protect" you from hospitalization? Well, you're much less likely to go, even in situations where previously you would have gone straight to the ER and demanded attention. And that's fine, they'll recover on their own as they almost certainly would have done anyway, and they'll say "thank god I was vaccinated otherwise it would have been much worse". This is the sort of confounder a rigorous trial would seek to control for, but we don't have any useful trials. The vaccine trials were complete failures: for the only end point they targeted, they reported 95% effectiveness, which started dwindling to near or even below zero in the months after the trial period ended. No usefully accurate information was provided by them, which is shocking.