Fact: There is no immunity or protection against The Law of Scoreboards.
Did you know: What the media does NOT want you to read is at https://market-ticker.org/nad.
You are not signed on; if you are a visitor please register for a free account!
|The Market Ticker Single Post Display (Show in context)||
|User Info||Covid-19 -- A White Paper - To @RealDonaldTrump and @CDC; entered at 2020-12-02 21:54:37|
@Zappafan - That's directly from the study linked in the footnotes, and yes -- it's simply a function of the power of the study, which is a function of events, time and participants -- and how you define the endpoint.|
In short you can define an endpoint that is supposed to show a given thing but if you get that endpoint you can only infer that anything ELSE will happen from what you showed; that is, you can only generate a HYPOTHESIS, not a conclusion.
The issue is that with a disease that causes relatively little severe disease, and often presents as an asymptomatic infection it is not possible to get statistical significance on actual infections and outcomes, especially on sub-groups, from a short trial and thus you cannot obtain anything more than a hypothesis.
Further, since the endpoint is overall infections in controls, at which point we stop and call it good, there is no effective control because there are too many confounders and worse, when it comes to subgroups in the population, such as (for example) elderly people where we REALLY care if works, we simply don't get the data to KNOW if it does. There's not enough of them and not enough time to generate statistical significance.
A vaccine that prevents people from getting a clinically-detected flu but does not stop someone from being killed by a severe case is functionally worthless. Nobody cares if you run a fever or cough for a day or two yet distinguishing between that and severe or fatal cases simply cannot happen given these trials because there is not enough data generated. So the best you can do is generate a hypothesis, that if the reduction in "cases" is X% then they should distribute ratably across outcomes -- but that is nothing more than a hypothesis.
Consider this -- a healthy young person has an alleged risk of 3/100,000 of death from infection of Covid-19. So you would need to have roughly 33,000 healthy people to get infected before ONE person would be expected to die, statistically-speaking. So when you have a study that terminates when 150 people get infected as regards severe and fatal outcomes you learned exactly nothing. All you got is a hypothesis, but no evidence beyond that because there are simply not enough events to generate even one expected death. It will be years before there are enough events to generate enough statistical power to know if the vaccine works to prevent those events and there is no possible way to short-circuit that process.
It took 20 years to license the chicken pox vaccine in the US because, as with Covid-19, severe and fatal cases are VERY RARE, especially among kids. In fact among kids chicken pox is about as serious and fatal as Covid-19 (roughly 3/100,000 death rate.) It simply took THAT LONG to get the data to know that severe and fatal cases - as opposed to nuisances - were prevented.
(There is also a serious problem with the data from one of these candidates, in that it appears there was co-mingling of results. This is the one where the "half-dose" thing happened -- I've seen some very disturbing data from that which strongly implies that they have literally nothing that would survive any sort of peer review at all at this point.)
Last modified: 2020-12-02 22:11:20 by tickerguy