| The Dutch data has the same problem as the ONS data that I describe in my other comment: > For the first mRNA vaccination, almost all [hazard ratios], for the different weeks after vaccination and different subgroups, are significantly lower than 1, which indicates a reduced risk of non-COVID-19 death shortly after vaccination. Obviously vaccination is not supposed to reduce your risk from things unrelated to COVID-19 yet in their data it does. This isn't because it's some magical cure-all. It's because their data is biased. The correct hazard ratio is found only in the 12-49 age group. As they say in their discussion, this is because "in the event of a very short remaining life expectancy, (a subsequent dose of) COVID-19 vaccination may be decided against ... Vaccination is also postponed in the event of a fever ... This is also called the 'healthy vaccine effect' and it is difficult to correct for this in observational research ... the results should be interpreted with caution" They also have the problem that "people who have been vaccinated, but have not given permission for registration in CIMS, have been incorrectly classified as unvaccinated (5 to 7 percent of the people)" which means that "This will have influenced the analysis of the risk of death after the first vaccination, because in this analysis the unvaccinated person time served as a reference" So this dataset is not usable for detecting vaccine side effects or induced deaths. It is a hopelessly corrupted dataset in which large numbers of people are misclassified and it's not a randomized study. But this is public health research, so after saying results should be interpreted with caution they go ahead and make the totally incautious claim that "Based on these results there are no indications at a population level that COVID-19 vaccination increases the risk of death due to an adverse event." which is a false statement. Their data doesn't allow them to draw such conclusions. The reason this topic is neuralgic is because if you want to be scientifically correct then you can't actually know whether vaccines reduced deaths or not, but people really want to believe they did. The trials that were supposed to unambiguously measure this failed, in the sense that they showed no effect on deaths and yielded incorrect conclusions about reductions in infection rates (the number started at 95% effective and then was regularly revised, this is not supposed to happen and was due to their time windowing problems). Then attempts to measure it using mass datasets collected outside of RT context all yield incorrect conclusions due to dataset bias, but are presented as definitive anyway for political reasons. Science is truly in a bad shape :( |
Is there any official data of this kind you would trust right now?
To be fair, I have seen enough by now; the main open question for me is what role endemic covid in all those believing the cure-all narrative and going out partying before milder variants came out played.