@freemo Have you seen the USC antibody study results? Interesting, and somewhat expected https://news.usc.edu/168987/antibody-testing-results-covid-19-infections-los-angeles-county/
🇳🇱
@obi indeed, not surprising
@realcaseyrollins thats what it has always seemed to me, but what do i know @freemo
Nothing about this changes the fact that it was as serious as they said. We always knew, and accounted for, the fact that many more were infected than tested. We also always knew that a large portion were asymptomatic though this doesnt change the fact that a large portion still died, these facts arent at odds either.
@freemo @realcaseyrollins I know its a lot to base off just one study, and to scale it towards the entire population, but if it were to that scale in that range (2.8-5.6%) wouldn't that bring the mortality rate down to somewhere between .02 and .06%? That added with the fact that COVID19 attributed deaths dont require a positive test, I'm just saying its not as dire as the media makes it out to be. Of course a lot of people died. A lot of people are always dying. Never good, but we don't do shit about the rest.
No there are sooooo many things wrong with that assumption.. putting aside anything to do with scaling that figure to the whole population you are forgetting one very fundamental fallacy in your thinking.. its called the False Positive Paradox..
In any disease where the number of people who have the disease is a minority of the population, even if the test for the disease has a very low false-positive rate then when you randomly sample and test the population the **overwhelming** majority of positive results will be false-positives.
this is a more specific form of the Base Rate Fallacy logic: https://en.wikipedia.org/wiki/Base_rate_fallacy
No I never said that. No studies show a high infection rate in terms of percentage of the population infected. New york city, for example, one of the worst hit is around a percentage point, most areas much lower.. it has a high R0 but the total number infected in terms of percentages are relatively low thankfully.
As far as I know that too was the same test as we are discussing here so subject to the same False Positive Paradox... point is we need other types of information, like what I discussed above, to really draw any sort of conclusion either way.
I never said the study is worthless, its very valuable, it just doesnt draw the conclusion you (or the media) seems to think it draws.
For the moment it only provides data, it doesn't draw a conclusion. However over time as we acquire more data it will certainly help draw conclusions. If we do some of the things mentioned above than the data from that study, when used with other data of the nature I mentioned, can eventually be used to determine the actual percentage of people infected. But so far we dont have the data to do that.
Data and studies aside if we just look at the high profile people being infected and dieing or otherwise having severe complications should tell you something.
Actor Nick Cordero just had his leg amputated due to getting COVID-19, when the last time you heard of any celebrity getting a leg amputated due to the flu? Or all the high profile people who have died from COVID-19, again, when was the last time you heard of any famous person even dying of the flu?
Any studies or numbers aside its pretty clear even if this is over or under hyped, its a pretty serious disease.
@freemo Truthfully i don't know of anyone famous who has had it bad. Never heard of this guy you referenced lol @realcaseyrollins
Thankfully they havent been huge actors, but famous enough I suppose:
Well that wasnt really the conclusion in the first place.. thought by whom?
All the experts up until then and still now at the moment would say the same thing.. we dont know the number infected very well.
The link i provided explained why the False Positive Paradox means that we cant draw conclusions from the test results without other types of data needed to calculate the true-positive rate.
We would need to first know the actual incidence of the virus in the population along with the false-positive rate of the test, with those two pieces of data then we can conclude meaningful results from the test.
Basically your working backwards, your trying to use a test to determine incident rate when you need to know the incident rate first in order to interprit the results of the test. which is exactly why we need other types of data which we dont have before this particular data comes useful.
@freemo @obi I thought you have to have had the #Coronavirus before you can get the antibodies. Would that count as an "incident"? Perhaps some of these people are naturally immune and were born with antibodies.
False positives happen for numerous reasons. technical inaccuracies of a test (such as detecting antibodies from another source or similar) can be one, but it can also include things as simple as human error, cross contamination, and countless other reasons.
The reason it is called the false positive paradox is exactly because of the reaction you are exhibiting now.. it is counter intuitive to what your instinct tells you when interpreting such a test.
I usually find when dealing with counter intuitive logic it can be helpful to think about it in the most extream case to help you see why it is the way it is... imagine some imaginary disease that is very rare, no one knows how rare it is though (just as we have no clue with coronavirus how common it really is).. so we develop some test to test for the disease so we can figure out how common it is.
Now lets say, for simplicity, we know due to human error or whatever other error inherent int he testing process that 1% of the time the test will say someone has the disease when they really do not (false positive). Lets also assume that in reality (though unknown to the scientists) only one person in the entire world actually has the disease, or for simplicity sake, no one does.
So armed with our 99% certain test we test the entire population of the world in an effort to determine how wide spread the disease is. What result would we get?
Well according to our 99% reliable rest 78 million people in the world have the disease (we tested 7.8 billion and 1% of them got a false positive).. Since we dont know the actual number of people in the world who had the disease, or even the the false positive rate of the test going in, we will happily conclude the disease has reached 78 million people even though in reality the disease is non-existant.
In other words.. the test is only actually helpful to us if we know the incident rate in the population and the false-positive rate of the test.. without these additional data points the test doesnt tell us anything of value.
Depends.. it has long been the case that laypeople reading studies, or even news media outlets reporting on them, have done more harm than good. Its pretty common for the media and laypeople to draw incorrect conclusions that professionals might know better not to do.
The study **is** valuable, it just doesnt draw any conclusions on its own and will simply be part of a bigger puzzle as time goes on. But the lay person is likely to draw premature conclusions from it anyway.
Among scientists its a huge gripe about how people and the news tend to report on studies. Its almost always wrong.
@freemo @realcaseyrollins I didn't even know the media was even covering this. I'm gonna go search it, wondering what they said now
@freemo That's interesting. Never knew false positives could go that high. Does this apply to this test tho? Because these people tested negitive to COVID19, but were had antibodies for it. Are antibody tests also susceptible to false positives? @realcaseyrollins
Every test has some degree of false-positive and false-negatives, yes.
@realcaseyrollins
@freemo So what do you do to account for that? Why are the studies even done if they don't mean anything? That's not meant to be snarky, honest question lol @realcaseyrollins
It isnt meaningless, you as the math on that link shows if you know the false positive rate, the size of the population, and a few other factors you can calculate the actual number of true-positive tests...
Right now we just dont have enough information to really caclulate that out is all. But this is one piece of the puzzle in getting there.
@freemo So more testing and larger scale testing and different region testing? Seems like that would help @realcaseyrollins
More tests, larger scale, specifically random testing, and case studies (identifying people actively infected and performing the tests on them even though we already know the results).. all of which should give us enough data to get a better idea of things
@freemo
So...just ignore the studies showing high infection rates? 🤔
@obi