Thanks Doc

Good analysis of the study. Some points have been brought up here and some haven't.

Also worth noting that the Stanford authors wrote several Op Eds promoting their conclusion
#1 before the study even began
#2 without disclosing that their involvement with the study

k2aggie07 said:

---

Unfortunately none of those are papers for serological studies. There's some media reports but the papers are for swab / PCR tests, which only tell you active infections. The media reports for blood tests don't give key information like sensitivity of the test, sample size, etc.

---

This is a big issue. In too many cases, we don't understand what we are reading. Active virus/antibody. Asymptomatic at time of test vs never showed symptoms. We are mixing and matching different tests with different goals.

The Stanford and this LA study seem consistent, just need to have more of these antibody tests, in different states/population centers, and hopefully with a non-Chinese test kit. I'm not saying the Chinese kit is for sure bad, but I just don't like the question mark.

Jeez the guy that wrote the article was on the study???? For ****s sake.

Stanford and LA are basically the same study. The lead of the USC study is third author on the Stanford paper. They did the same test kit. With the same small sample size specificity, the USC lead author mentioned the Stanford testing.

And if you look at his CV he's coauthored papers with Bendavid and Bhattcharya before.

If the point of this discussion is reopening, arguing serum vs antibody vs PCR is useless. PCR has a lot of false negatives, antibodies has a lot of false positives. Bam, you're in the ballpark anyway. If this is an exercise to pinpoint a specific number, that's useless. Have fun with the academic masturbation.

lol at Peak Prosperity.

The point is I want to read papers with error bars, not media blurbs. Without the paper I'd have no idea that the Stanford study was using basically a made up specificity number and a demographic adjusted raw prevalence at 1.86 times the raw prevalence. It's hard enough filtering through the academic stuff, when you add the media filter it becomes impossible. At least researchers who publish are held to some standard of scientific rigor.

https://www.medrxiv.org/content/10.1101/2020.04.14.20062463v1.full.pdf

Yes, thank you, that's the Stanford article being discussed. It's the first out there I think. And I think it's going to get torn up in peer review.

What is the point of that comment? Do you think that's being helpful, or contributing to the discussion?

Can we please be respectful of one another on here? k2 has been respectful of you.

Healthy skepticism of any new information coming is appropriate. Too many apply skepticism only to information that refutes their prior beliefs.

Because there hasn't been anything constructive out of his comments. Skepticism is fine but when everything is poo-pooed for this or that, there's no point in discussing these things with him.

Snap E Tom said:

---

Because there hasn't been anything constructive out of his comments. Skepticism is fine but when everything is poo-pooed for this or that, there's no point in discussing these things with him.

---

The Dr that was pretty much the impetus for this board has serious doubts about this "study" as well...

Feel free to ignore all the other studies I linked to.

He's been having a good faith discussion about the merits of the study and hasn't been stooping to vulgarity or name calling to do it. You can disagree with him without being disrespectful.

I think it's reasonable to be skeptical about this study, considering what the authors have published prior to the study, considering some of the issues with the study itself. Revellie has echoed some of those sentiments.

If someone posted an article tomorrow saying that 5 mm people would die in the US before the end of the year due to a new study on the deadliness of the virus in certain populations, wouldn't that raise skepticism in yourself. Wouldn't you want to dig through the assumptions and check for potential flaws in the study? Why does this rigor only apply to information that doesn't reinforce our priors. It's the pinnacle of confirmation bias.

HotardAg07 said:

---

He's been having a good faith discussion about the merits of the study and hasn't been stooping to vulgarity or name calling to do it. You can disagree with him without being disrespectful.

I think it's reasonable to be skeptical about this study, considering what the authors have published prior to the study, considering some of the issues with the study itself. Revellie has echoed some of those sentiments.

If someone posted an article tomorrow saying that 5 mm people would die in the US before the end of the year due to a new study on the deadliness of the virus in certain populations, wouldn't that raise skepticism in yourself. Wouldn't you want to dig through the assumptions and check for potential flaws in the study? Why does this rigor only apply to information that doesn't reinforce our priors. It's the pinnacle of confirmation bias.

---

My point is, as I have said before, in context with other studies with different methodologies, different countries, different institutions, different researchers, the conclusions are aligning. You can't see that. Talk about confirmation bias. Am I technically derailing this thread? Fine. However, I don't know why you are so hell bent on me focusing focusing on him.

The issue is PCR studies (without diligent follow-ups) don't really answer the the question we urgently need to know - what is the prevalence rate in the population? We know that we have asymptomatics, we know it's probably a lot. But there is a huge difference in the implication of "a lot" and "a whole lot". Unfortunately that difference might be masked by a 92% vs 99% specificity in a test kit when you're looking at the research.

Consider in the Stanford paper, they note "if new estimates indicate test specificity to be less than 97.9%, our SARS-CoV-2 prevalence estimate would change from 2.8% to less than 1%, and the lower uncertainty bound of our estimate would include zero."

And that's the population adjusted 2.8% - that takes their non-population adjusted 1.5% down to 0.5%.

Another study in Europe independently tested this same kit from China:
https://www.medrxiv.org/content/10.1101/2020.04.09.20056325v1.full.pdf

They found a specificity of 87%. If we use that number with the 63.7% or 91.8% sensitivity from the Stanford study the point estimate for prevalence with a 50/3300 test result is below zero.

What on earth are we supposed to make of this??

When I said - hey, lets find some more antibody studies, there are none to read. Only media, which doesn't give us any of the critical information necessary to understand what researchers are actually finding.

At some point you just have to give up. They want a perfect study, across many different cities, across a huge random population, without any hint of bias or collaboration, with high confidence for accurate results, peer reviewed, and published in Nature or Journal of Medicine. All within a month so that we can make sound decisions.

Follow up studies corroborating results are required to well.

Until then it's trash and we need to continue to stay on lockdown because the virus has only infected .1% of the world and we need to keep it that way or fear of dead bodies in the streets will cause the earth to completely shut down and we'll be in even worse shape.

Logic and connecting dots don't matter because k2 has way more experience than all these Stanford medical doctors and PhDs, and he sees through their lies.

NASAg03 said:

---

At some point you just have to give up. They want a perfect study, across many different cities, across a huge random population, without any hint of bias or collaboration, with high confidence for accurate results, peer reviewed, and published in Nature or Journal of Medicine. All within a month so that we can make sound decisions.

Follow up studies corroborating results are required to well.

Until then it's trash and we need to continue to stay on lockdown because the virus has only infected .1% of the world and we need to keep it that way or fear of dead bodies in the streets will cause the earth to completely shut down and we'll be in even worse shape.

Logic and connecting dots don't matter because k2 has way more experience than all these Stanford medical doctors and PhDs, and he sees through their lies.

---

If only they required the same perfection to implement the lockdowns in the first place. There are 500 deaths in Texas. 500!

Look, we have exactly one serological paper to look at right now. I've probably read 50 studies for modeling and PCR testing, there's not much more we can learn from those. The next phase is antibody testing, and so we need papers. In the meantime, we have this one, and it has issues.

I'm curious, what do you think about the impact of the specificity of the test kit on the study?

Forget the ad hominem against me. The numbers are there, this is why we publish papers. They're there for people to do exactly what I'm doing.

So where is my mistake? I mean, if you want to talk about logic, and connecting dots, let's go.

This is a ridiculous false equivalence. I criticized early studies coming out China just as harshly. I'm an equal opportunity skeptic. Go read the original politics board threads. The only difference is here guys like NASAg03 bristle because I had the audacity to call the authors skeptics and note their bias.

And yet, he offers no defense, no insight, no rebuttal. Just accuses me of, I don't know what, some kind of illogical and wanting perfection. Forget perfection - I'd settle for "reasonable" and "not having a negative prevalence within the a likely range of point estimates for my study".

PS one of the silent co-authors of the study referred to them / including himself as skeptics in this op ed where he touts their paper without disclosing that he was part of it. And they scoffed about bias and said skeptic was a bad word or something.

k2aggie07 said:

---

Look, we have exactly one serological paper to look at right now. I've probably read 50 studies for modeling and PCR testing, there's not much more we can learn from those. The next phase is antibody testing, and so we need papers. In the meantime, we have this one, and it has issues.

I'm curious, what do you think about the impact of the specificity of the test kit on the study?

Forget the ad hominem against me. The numbers are there, this is why we publish papers. They're there for people to do exactly what I'm doing.

So where is my mistake? I mean, if you want to talk about logic, and connecting dots, let's go.

---

Any guesses on how far off the recorded numbers are?

This site (https://www.worldometers.info/coronavirus/country/us/) says

Coronavirus Cases:

788,172

I think it could really be 2-3 times higher. I don't think it is 10x higher.

In heavy hit areas, 20x is possible. 10x wouldn't surprise me one bit. I would be very surprised if it is less than 5x. But 50 or 80? I don't know.

This is an interesting look:

https://www.imperial.ac.uk/media/imperial-college/medicine/mrc-gida/2020-03-30-COVID19-Report-13.pdf

You can back into the possible under-reporting by looking at total cases/population on March 28.

k2aggie07 said:

---

This is a ridiculous false equivalence. I criticized early studies coming out China just as harshly. I'm an equal opportunity skeptic. Go read the original politics board threads. The only difference is here guys like NASAg03 bristle because I had the audacity to call the authors skeptics and note their bias.

And yet, he offers no defense, no insight, no rebuttal. Just accuses me of, I don't know what, some kind of illogical and wanting perfection. Forget perfection - I'd settle for "reasonable" and "not having a negative prevalence within the a likely range of point estimates for my study".

PS one of the silent co-authors of the study referred to them / including himself as skeptics in this op ed where he touts their paper without disclosing that he was part of it. And they scoffed about bias and said skeptic was a bad word or something.

---

Fair enough. I do not need to see another single study or model. There are 500 deaths in Texas with projected deaths around 1000.

Chicago gang bangers have shot more people than that this year. Open this MF'er up!! May the odds ever be in your favor.

k2aggie07 said:

---

Stanford and LA are basically the same study. The lead of the USC study is third author on the Stanford paper. They did the same test kit. With the same small sample size specificity, the USC lead author mentioned the Stanford testing.

And if you look at his CV he's coauthored papers with Bendavid and Bhattcharya before.

---

Yea it's this kind of crap that gives science a bad name.
The **** studies where this BS goes on, gets picked up by the media and we are left to explain why it's not valid but it reflects bad on a lot of groups and good scientists that are playing by the rules to put out good info.

NASAg03 said:

---

At some point you just have to give up. They want a perfect study, across many different cities, across a huge random population, without any hint of bias or collaboration, with high confidence for accurate results, peer reviewed, and published in Nature or Journal of Medicine. All within a month so that we can make sound decisions.

Follow up studies corroborating results are required to well.

Until then it's trash and we need to continue to stay on lockdown because the virus has only infected .1% of the world and we need to keep it that way or fear of dead bodies in the streets will cause the earth to completely shut down and we'll be in even worse shape.

Logic and connecting dots don't matter because k2 has way more experience than all these Stanford medical doctors and PhDs, and he sees through their lies.

---

Have their been any studies by medical doctors and PhDs, predicting a large amount of deaths with no mitigation steps, or hospitals being overrun, that you disagreed with?

I say this not to suggest one study is wrong, or another right. I say this because when medical doctors and PhDs were putting out models showing what might happen with no mitigation steps, some have retroactively said "all the models were wrong". We can pick and choose which doctors and PhDs to listen to, but if its just to support something you "feel in your gut" then I think we can admit what we are choosing to believe is due to personal bias.

I find them all valuable - the predictions in February, the predictions now, the studies of actual antibodies now. All will inform leaders on what actions to take. If your criteria is we should listen to medical doctors and PhDs, then don't pick and choose which data produced by people meeting your criteria supports your personal belief. Listen to it all, and be open to changing your opinion as data is added.

I can't control when it opens up. Totally along for the ride. I didn't have an opinion on what the right thing to do was, and I don't know now.

All I'm trying to do is understand the situation so I can anticipate - for my job, for my family. For my own sanity and "need to know."

Knowing the severity - and by extension, the asymptomatic rate and critically how that is distributed in the population has been basically the only anyone wants to know the answer to since January when we realized the R0 of this thing is over 2. So a study like this is hugely important. A bigtime time claim like 50-80x prevalence... which is like 4-20 times what everyone else is saying...is huge. If true.

Quote:

---

In heavy hit areas, 20x is possible. 10x wouldn't surprise me one bit. I would be very surprised if it is less than 5x. But 50 or 80? I don't know.

---

50x is, or is nearing, mathematically impossible in a growing number of locations, yet alone 80x. That's a huge point of confusion, bordering on "skepticism", for me.

Absolutely there have been models that I've disagreed with - and certainly with how people are using them. The IHME model has limitations, and people are badly abusing it. It's using death curves from this epidemic to back into hospitalizations. To do that, they're assuming an infection rate in the population at the end of May. That's a huge assumption. They're also looking at limited death curves, with all the baggage and confounding implications. The minute the IHME model says "you're not going to bust hospital capacity" you can basically ignore it. That's its only purpose - not to predict deaths, not to predict spread, whatever. It basically never said we had issues in Texas, so people keep looking at it, and I don't get why.

Same with that horrible model that came out that was used by Dallas County to justify their shelter in place. It was egregious, really bad assumptions, really bad data on hospital beds / ICU capacity. Also - not made by epidemiologists, which was part of the issue.

Now everyone is like 'hurr hurr models r dum' and you get a bunch of luddites coming out of the woodwork to talk about how stupid all these academics are.

k2aggie07 said:

---

PS one of the silent co-authors of the study referred to them / including himself as skeptics in this op ed where he touts their paper without disclosing that he was part of it. And they scoffed about bias and said skeptic was a bad word or something.

---

Side note:

Andrew Bogan who wrote that op ed and helped fund the study isn't really a researcher. He's a venture capitalist with a vested interest in re-opening the economy.

Bobcat06 said:

---

k2aggie07 said:

---

PS one of the silent co-authors of the study referred to them / including himself as skeptics in this op ed where he touts their paper without disclosing that he was part of it. And they scoffed about bias and said skeptic was a bad word or something.

---

Side note:

Andrew Bogan who wrote that op ed and helped fund the study isn't really a researcher. He's a venture capitalist with a vested interest in re-opening the economy.

---

A vested interest in reopening the economy?? That is some bias you've identified. Wouldn't that be basically everybody?

Not researchers who publish an "unbiased" study

Bobcat06 said:

---

Not researchers who publish an "unbiased" study

---

Be careful, your bias is showing.

Speak for yourself.

When scientists are paid to find a certain conclusion, they magically find ways to reach that conclusion (using tainted sample pools, tests that generate false positives, etc). The same problem exists with research grants that sponsor a predetermined conclusion (e.g. climate change).

When reading research, I've learned to ask who's paying for it. There's a reason I left academia. [/end rant]