| Topic List | Page List: 1 |
|---|---|
| Topic | Explain why they would fire nurses in the middle of a pandemic. |
| bladegash 10/11/21 1:42:53 AM #85: | -GalacticBass- posted... That's a website where people post their opinions. Here's some actual data... A worthy comment at bottom of the article: To the study authors: As an epidemiologist, if I was reviewing this paper for publication I would send it back to you for major revisions or reject it outright. I would not even bother looking at the results. The major issue is that you have conditioned study group entry by an event that happens at the end of the study. I.e. you have created a cohort of unvaccinated persons who must remain unvaccinated throughout the study. This is guaranteed to introduce selection bias, more specifically immortal-time bias. This further guarantees a biased estimate. This topic has been written about many times. Cf any of many articles by Sammie Souza at McGill. Imagine someone in your unvaccinated cohort. Soon after the initial study date they develop an infection. 5 weeks later they have recovered and decided they should have had the vaccine, so they get one. Because you have insisted this group remain vaccine free you throw them out of the group and you lose their data. You have just thrown out an infection. Do this just a few times and it is guaranteed that your unvaccinated group is not reporting as many infections as it actually experienced. This easily accounts for the effect you report. Note that this does NOT happen with the fully vaccinated/boosted group who must receive all vaccinations prior to study entry. You capture each and every infection with no drop out. Thus youve created a situation where you have non-random drop-out between the groups. That is selection bias. To get around this problem you MUST use methods such as Cox proportional hazards modeling with time-varying exposure variables so that persons can move between cohorts based on exposure to the vaccine during the study period --- Snappin necks & Cashin checks ... Copied to Clipboard! |
| Topic List | Page List: 1 |