I was wondering why it took so long for a simple random sample study to be done, but finally it has been done, and the result is pretty much as I expected - millions are carrying the virus, making the true mortality rate tiny.
This looks like a reliable study since its testing for antibodies, rather than the more common swab tests.
If the true mortality rate is tiny, how did this virus manage to kill 153k people in 3 months? Can someone ELI5 me that? People have been saying "Flu kills 30k Americans every year" and now 30k Americans died of this disease in 3 months AFTER locking up the entire country. Something doesn't add up.
If the most vulnerable did not have flu shots, the mortality of the flu would be shocking. You may or may not need one (or notice if you didn't) but every high-viral-load-facing healthcare worker, cystic fibrosis patient, etc. gets a flu shot like clockwork. And while they don't always work, they mostly work, and that hugely reduces mortality. The problem with this one is the novelty. Human immune systems get better trained (not necessarily better) with age. Cytokine storms in the eldery with covid-19 is the biological equivalent of an overfit AI with memory leaks that runaway. If the R0 turns out to be 5.6, basically everyone has been exposed at this point (not necessarily infected).
The argument, as I understand it, is that they typically have near 100% vaccination. Akin to herd immunity in that cohort. If they did not have that, it would be much worse.
yea as some people have stated, if even 30% of the world gets infected and we have a 0.5% mortality rate, then that is 11.5 million people... not a trivial amount....
Because it's novel and incredibly contagious. Unlike the flu, where many get seasonal flu shots, no one has antibodies for the virus so there's no herd immunity. Meaning it's just spreading like wild fire. And it's all happening at once so healthcare systems are overwhelmed, which affects patient outcomes. I suspect the number of deaths are quite a bit higher because of those dying without getting tested.
> If the true mortality rate is tiny, how did this virus manage to kill 153k people in 3 months?
Because there's 7 billion people on earth.
> People have been saying "Flu kills 30k Americans every year" and now 30k Americans died of this disease in 3 months AFTER locking up the entire country.
This wasn’t totally random:
“ This study had several limitations. First, our sampling strategy selected for members of Santa Clara County with access to Facebook and a car to attend drive-through testing sites. This resulted in an over- representation of white women between the ages of 19 and 64, and an under-representation of Hispanic and Asian populations, relative to our community. Those imbalances were partly addressed by weighting our sample population by zip code, race, and sex to match the county. We did not account for age imbalance in our sample, and could not ascertain representativeness of SARS-CoV-2 antibodies in homeless populations. Other biases, such as bias favoring individuals in good health capable of attending our testing sites, or bias favoring those with prior COVID-like illnesses seeking antibody confirmation are also possible. The overall effect of such biases is hard to ascertain”
I know at least one person was inclined to get tested because he had a bad flu earlier.
You should read what they have to say about the study's limitations before concluding too much. It's not a random sample (though they tried to correct for this), and the test is new, so their estimates of the test's accuracy are uncertain.
It will be really interesting once large random samples of people are tested -- quite possibly we will find out that the virus was spreading super widely in certain areas for Jan / Feb / March before lockdowns started happening.. the model might end up being something like given X person has the virus, what is the likelihood that they will interact with somebody over the age of 60..
I don't know why this was downvoted, it seems reasonable to me. A lot of people would have traveled for christmas, new years and chinese new year. Testing was almost non existent. I would bet that everyone by now at least knows someone who suspects they have already had it.
> Why would it be spread in certain areas and not others?
I would guess because of different social distancing practices.
> Why would knowing who has antibodies hurt the economy when everyone is already secluded?
Knowing who has antibodies is going to be conflated with who has covid-19, because the general public is not going to get the nuance. And everyone is secluded right now but that will change.
However, I don't understand why you are asking me to prove the premise of the parent poster. I thought I was answering the question why he might have been downvoted, not whether what he was proposing was feasible. Maybe you could explain why you are asking me this. It seems you agree that a map targeting certain groups of people as likely of having covid-19 would be a bad outcome, isn't it? If so, you could probably understand why he might be downvoted.
To be clear, you think that enough people will be paying attention to this study on testing for antibodies that it will hurt the local economy of the place being tested, because the people paying attention don't understand antibodies?
What “study”? Are we not talking about what the parent poster wrote anymore?
And to answer your question, you see, there are these people called reporters, and they get a lot of information from very different sources. They filter out what is interesting and broadcast it to their audience. Perhaps you've heard of them.
It's unlikely. I don't think we know the exact progression enough to say for sure, but typically the specific immune response won't start that quickly. (It couldn't account for anything close to the study results in either case - there's no realistic way that there are 50x-85x cases that are about to become symptomatic soon.)
1. There were 2 false positives in 401 pre-covid samples (known negative). Assuming a normal distribution, 2 sigma confidence can account for up to 80% of the positive results reported in this study.
2. For the virus to have already infected this many people, it should’ve spread so fast that it doesn’t match what we know about similar epidemics.
3. The researchers recruited volunteers via facebook ads. There could be sampling bias here.
I encourage everyone to read the article: my summary < The original article.
My summary is when betting on any test with a low positive rate take under not the over.
Saying at least this many people have been infected, wrong.
Better. At most this many people have been infected.
Reminds me of a survey 15 years ago that said something like 2-3% of black people approved of the job the president was doing. That 2-3% was more likely 'rectified noise'.
In normal times, where there aren't critically urgent decisions to be made, that's the right approach. We don't have that luxury right now; we're continually charting the path forward, and if we don't have clean data then we'll need to chart it with unclean data.
And I think it's hard to argue that this study is worse than our existing case data, which is simply a daily aggregate of the number of new positive PCR tests labs report they've run. That's not even attempting to be a statistically valid estimate.
for 3, sampling bias would not account for much as the coronavirus will catch you regardless of someone's socio-economical status. The difference is mostly just survival rate.
Surely richer people can more effectively self isolate? Less likely to be a key worker, can easily get items delivered to house, more likely to be able to exercise at home etc
31 comments
[ 3.9 ms ] story [ 77.3 ms ] threadThis looks like a reliable study since its testing for antibodies, rather than the more common swab tests.
SARS-1 (2002/2003) wiped out ICU staff in both Toronto and S. Korea.
"High viral load" is an undefined term, so I won't comment on that.
Because there's 7 billion people on earth.
> People have been saying "Flu kills 30k Americans every year" and now 30k Americans died of this disease in 3 months AFTER locking up the entire country.
It's been spreading world-wide since Dec.
> Something doesn't add up.
GIGO.
I know at least one person was inclined to get tested because he had a bad flu earlier.
You should read what they have to say about the study's limitations before concluding too much. It's not a random sample (though they tried to correct for this), and the test is new, so their estimates of the test's accuracy are uncertain.
Main discussion is here: https://news.ycombinator.com/item?id=22899272
Emphasis mine
I would guess because of different social distancing practices.
> Why would knowing who has antibodies hurt the economy when everyone is already secluded?
Knowing who has antibodies is going to be conflated with who has covid-19, because the general public is not going to get the nuance. And everyone is secluded right now but that will change.
However, I don't understand why you are asking me to prove the premise of the parent poster. I thought I was answering the question why he might have been downvoted, not whether what he was proposing was feasible. Maybe you could explain why you are asking me this. It seems you agree that a map targeting certain groups of people as likely of having covid-19 would be a bad outcome, isn't it? If so, you could probably understand why he might be downvoted.
And to answer your question, you see, there are these people called reporters, and they get a lot of information from very different sources. They filter out what is interesting and broadcast it to their audience. Perhaps you've heard of them.
https://medium.com/@balajis/peer-review-of-covid-19-antibody...
A summary of his post:
1. There were 2 false positives in 401 pre-covid samples (known negative). Assuming a normal distribution, 2 sigma confidence can account for up to 80% of the positive results reported in this study.
2. For the virus to have already infected this many people, it should’ve spread so fast that it doesn’t match what we know about similar epidemics.
3. The researchers recruited volunteers via facebook ads. There could be sampling bias here.
I encourage everyone to read the article: my summary < The original article.
Saying at least this many people have been infected, wrong. Better. At most this many people have been infected.
Reminds me of a survey 15 years ago that said something like 2-3% of black people approved of the job the president was doing. That 2-3% was more likely 'rectified noise'.
And I think it's hard to argue that this study is worse than our existing case data, which is simply a daily aggregate of the number of new positive PCR tests labs report they've run. That's not even attempting to be a statistically valid estimate.