[burger]

and, in fact, an argument can be made that over time the IFR has been a bit less scary).
BC

Do you think the IFRs cited in this document are accurate?
https://www.medrxiv.org/content/10.1101/...20101253v3

This Twitter thread suggests no. I haven't read the paper myself (and I'm not going to. One episode of near fraud is enough for an author to get a permanent ban from me).

[tweet]https://twitter.com/AVG_Joseph96/status/1283234509639618560[/tweet]

I believe cdc is saying that ifr is 0.65% now. I'd bet that's closer to correct than 0.3.

Here's another Twitter thread with more criticisms.

[tweet]https://twitter.com/GidMK/status/1283232023402868737?s=19[/tweet]

[burger]

and, in fact, an argument can be made that over time the IFR has been a bit less scary).
BC

Do you think the IFRs cited in this document are accurate?
https://www.medrxiv.org/content/10.1101/...20101253v3

This Twitter thread suggests no. I haven't read the paper myself (and I'm not going to. One episode of near fraud is enough for an author to get a permanent ban from me).

[tweet]https://twitter.com/AVG_Joseph96/status/1283234509639618560[/tweet]

I believe cdc is saying that ifr is 0.65% now. I'd bet that's closer to correct than 0.3.

Here's another Twitter thread with more criticisms.

[tweet]https://twitter.com/GidMK/status/1283232023402868737?s=19[/tweet]

[lex24]

and, in fact, an argument can be made that over time the IFR has been a bit less scary).
BC

Do you think the IFRs cited in this document are accurate?
https://www.medrxiv.org/content/10.1101/...20101253v3

This Twitter thread suggests no.  I haven't read the paper myself (and I'm not going to.  One episode of near fraud is enough for an author to get a permanent ban from me).

[tweet]https://twitter.com/AVG_Joseph96/status/1283234509639618560[/tweet]

I believe cdc is saying that ifr is 0.65% now.  I'd bet that's closer to correct than 0.3.

Here's another Twitter thread with more criticisms. 

[tweet]https://twitter.com/GidMK/status/1283232023402868737?s=19[/tweet]

What would we do without twitter threads.......These twitter folks certainly have better credentials than Ioannidis : https://profiles.stanford.edu/john-ioannidis

Particularly Gideon M-K. One day he may just get that PHD from the Univ of Wollongong. Until then he can write about how Red Bull’s are actually good for us......

[lex24]

and, in fact, an argument can be made that over time the IFR has been a bit less scary).
BC

Do you think the IFRs cited in this document are accurate?
https://www.medrxiv.org/content/10.1101/...20101253v3

This Twitter thread suggests no.  I haven't read the paper myself (and I'm not going to.  One episode of near fraud is enough for an author to get a permanent ban from me).

[tweet]https://twitter.com/AVG_Joseph96/status/1283234509639618560[/tweet]

I believe cdc is saying that ifr is 0.65% now.  I'd bet that's closer to correct than 0.3.

Here's another Twitter thread with more criticisms. 

[tweet]https://twitter.com/GidMK/status/1283232023402868737?s=19[/tweet]

What would we do without twitter threads.......These twitter folks certainly have better credentials than Ioannidis : https://profiles.stanford.edu/john-ioannidis

Particularly Gideon M-K. One day he may just get that PHD from the Univ of Wollongong. Until then he can write about how Red Bull’s are actually good for us......

[BostonCard]

Best estimates I've seen are between 0.5% (if you use confirmed COVID-19 deaths) to about 1% if you use excess all-cause mortality in areas where rigorous sampling has been used to calculate the seroprevalence using a random sampling method (Spain is the best study).  Because of the strong age gradient, to a large degree IFR will be very dependent on the age distribution of the infected population.

BC

---

@Burger, if you have never been the victim of confirmation bias, you are a better man than I am.  Rather than attribute malice to Ioannidis' motives, maybe consider the possibility that a contrarian by nature got caught up a bit in his contrarianism, and accept that he was correct directionally; the ultimate IFR is substantially lower than the initial estimates from the WHO, though probably not as low as his estimates.

@Lex, even guys with the credentials of Ioannidis make mistakes.  HEre's a Nobel Prize winner who retracted a Science paper because the results were not reproducible.  I still admire and respect Ioannidis a ton, and in my response to Burger, I note that he was directionally correct, though (probably) wrong on the magnitude.

BC

---

Lastly, Gideon Meyerowitz-Katz and his co-author, Lea Merone have their own pre-print paper:

https://www.medrxiv.org/content/10.1101/...20089854v4

Their conclusions:

Based on a systematic review and meta-analysis of published evidence on COVID-19 until May, 2020, the IFR of the disease across populations is 0.68% (0.53-0.82%). However, due to very high heterogeneity in the meta-analysis, it is difficult to know if this represents the true point estimate. It is likely that, due to age and perhaps underlying comorbidities in the population, different places will experience different IFRs due to the disease. Given issues with mortality recording, it is also likely that this represents an underestimate of the true IFR figure. More research looking at age-stratified IFR is urgently needed to inform policy-making on this front.

Would love to see a mixed-effects meta analysis adjusting for age of the seropositive population to see if that would get rid of the heterogeneity.

BC

[BostonCard]

Best estimates I've seen are between 0.5% (if you use confirmed COVID-19 deaths) to about 1% if you use excess all-cause mortality in areas where rigorous sampling has been used to calculate the seroprevalence using a random sampling method (Spain is the best study).  Because of the strong age gradient, to a large degree IFR will be very dependent on the age distribution of the infected population.

BC

---

@Burger, if you have never been the victim of confirmation bias, you are a better man than I am.  Rather than attribute malice to Ioannidis' motives, maybe consider the possibility that a contrarian by nature got caught up a bit in his contrarianism, and accept that he was correct directionally; the ultimate IFR is substantially lower than the initial estimates from the WHO, though probably not as low as his estimates.

@Lex, even guys with the credentials of Ioannidis make mistakes.  HEre's a Nobel Prize winner who retracted a Science paper because the results were not reproducible.  I still admire and respect Ioannidis a ton, and in my response to Burger, I note that he was directionally correct, though (probably) wrong on the magnitude.

BC

---

Lastly, Gideon Meyerowitz-Katz and his co-author, Lea Merone have their own pre-print paper:

https://www.medrxiv.org/content/10.1101/...20089854v4

Their conclusions:

Based on a systematic review and meta-analysis of published evidence on COVID-19 until May, 2020, the IFR of the disease across populations is 0.68% (0.53-0.82%). However, due to very high heterogeneity in the meta-analysis, it is difficult to know if this represents the true point estimate. It is likely that, due to age and perhaps underlying comorbidities in the population, different places will experience different IFRs due to the disease. Given issues with mortality recording, it is also likely that this represents an underestimate of the true IFR figure. More research looking at age-stratified IFR is urgently needed to inform policy-making on this front.

Would love to see a mixed-effects meta analysis adjusting for age of the seropositive population to see if that would get rid of the heterogeneity.

BC

[burger]

@Burger, if you have never been the victim of confirmation bias, you are a better man than I am.  Rather than attribute malice to Ioannidis' motives, maybe consider the possibility that a contrarian by nature got caught up a bit in his contrarianism, and accept that he was correct directionally; the ultimate IFR is substantially lower than the initial estimates from the WHO, though probably not as low as his estimates.

Come on, really?  It's a truism that all people suffer from confirmation bias.  But there's a chasm between that and publishing papers where one's biases influence the results.  The methods that Ioannidis is using have been roundly criticized by epidemiologists as biased and designed to produce a low estimate of IFR.

I don't attibute malice to Ioannidis (though see below).  I just think he's being a shitty scientist.  A good scientist does not let their personal bias influence their results.  Good scientists don't get "caught up" in anything.  They analyze the data in an objective manner and let it speak from themselves.  And really good scientists change their beliefs when the data shows that they were wrong.  I've had this experience a number of times.  

Your line about "directionally correct" is laughable.  There is no such thing.  Getting an accurate estimate of IFR is the goal here, not a direction.

I don't know what's in Ioannidis' heart, but he did go on TV multiple times to hype his studies and argue that the virus was not as big of a threat as others were suggesting.  Maybe he's just out for the self-promotion.  Maybe he really thinks the IFR is low.  Either way, he is massaging the numbers to artificially lower his IFR estimates, and that should not stand.

---

What would we do without twitter threads.......These twitter folks certainly have better credentials than  Ioannidis :  https://profiles.stanford.edu/john-ioannidis

Particularly Gideon M-K. One day he may just get that PHD from the Univ of Wollongong.  Until then he can write about how Red Bull’s are actually good for us......

Ad hominem attacks are the last refuge of those who have no substantive arguments to make.  If you think those arguments in the twitter feeds are wrong, I'd like to hear why.  Otherwise, you're just deflecting.  You don't need a nobel prize to how when someone else has been fudging the numbers.

[burger]

@Burger, if you have never been the victim of confirmation bias, you are a better man than I am.  Rather than attribute malice to Ioannidis' motives, maybe consider the possibility that a contrarian by nature got caught up a bit in his contrarianism, and accept that he was correct directionally; the ultimate IFR is substantially lower than the initial estimates from the WHO, though probably not as low as his estimates.

Come on, really?  It's a truism that all people suffer from confirmation bias.  But there's a chasm between that and publishing papers where one's biases influence the results.  The methods that Ioannidis is using have been roundly criticized by epidemiologists as biased and designed to produce a low estimate of IFR.

I don't attibute malice to Ioannidis (though see below).  I just think he's being a shitty scientist.  A good scientist does not let their personal bias influence their results.  Good scientists don't get "caught up" in anything.  They analyze the data in an objective manner and let it speak from themselves.  And really good scientists change their beliefs when the data shows that they were wrong.  I've had this experience a number of times.  

Your line about "directionally correct" is laughable.  There is no such thing.  Getting an accurate estimate of IFR is the goal here, not a direction.

I don't know what's in Ioannidis' heart, but he did go on TV multiple times to hype his studies and argue that the virus was not as big of a threat as others were suggesting.  Maybe he's just out for the self-promotion.  Maybe he really thinks the IFR is low.  Either way, he is massaging the numbers to artificially lower his IFR estimates, and that should not stand.

---

What would we do without twitter threads.......These twitter folks certainly have better credentials than  Ioannidis :  https://profiles.stanford.edu/john-ioannidis

Particularly Gideon M-K. One day he may just get that PHD from the Univ of Wollongong.  Until then he can write about how Red Bull’s are actually good for us......

Ad hominem attacks are the last refuge of those who have no substantive arguments to make.  If you think those arguments in the twitter feeds are wrong, I'd like to hear why.  Otherwise, you're just deflecting.  You don't need a nobel prize to how when someone else has been fudging the numbers.

[lex24]

Best estimates I've seen are between 0.5% (if you use confirmed COVID-19 deaths) to about 1% if you use excess all-cause mortality in areas where rigorous sampling has been used to calculate the seroprevalence using a random sampling method (Spain is the best study).  Because of the strong age gradient, to a large degree IFR will be very dependent on the age distribution of the infected population.

BC

---

@Burger, if you have never been the victim of confirmation bias, you are a better man than I am.  Rather than attribute malice to Ioannidis' motives, maybe consider the possibility that a contrarian by nature got caught up a bit in his contrarianism, and accept that he was correct directionally; the ultimate IFR is substantially lower than the initial estimates from the WHO, though probably not as low as his estimates.

@Lex, even guys with the credentials of Ioannidis make mistakes.  HEre's a Nobel Prize winner who retracted a Science paper because the results were not reproducible.  I still admire and respect Ioannidis a ton, and in my response to Burger, I note that he was directionally correct, though (probably) wrong on the magnitude.

BC

---

Lastly, Gideon Meyerowitz-Katz and his co-author, Lea Merone have their own pre-print paper:

https://www.medrxiv.org/content/10.1101/...20089854v4

Their conclusions:

Based on a systematic review and meta-analysis of published evidence on COVID-19 until May, 2020, the IFR of the disease across populations is 0.68% (0.53-0.82%). However, due to very high heterogeneity in the meta-analysis, it is difficult to know if this represents the true point estimate. It is likely that, due to age and perhaps underlying comorbidities in the population, different places will experience different IFRs due to the disease. Given issues with mortality recording, it is also likely that this represents an underestimate of the true IFR figure. More research looking at age-stratified IFR is urgently needed to inform policy-making on this front.

Would love to see a mixed-effects meta analysis adjusting for age of the seropositive population to see if that would get rid of the heterogeneity.

BC

BC,  My point was not to suggest that he’s correct. I really don’t know. And of course he makes mistakes. This is a trial and error process is it not?

---

@Burger, if you have never been the victim of confirmation bias, you are a better man than I am.  Rather than attribute malice to Ioannidis' motives, maybe consider the possibility that a contrarian by nature got caught up a bit in his contrarianism, and accept that he was correct directionally; the ultimate IFR is substantially lower than the initial estimates from the WHO, though probably not as low as his estimates.

Come on, really?  It's a truism that all people suffer from confirmation bias.  But there's a chasm between that and publishing papers where one's biases influence the results.  The methods that Ioannidis is using have been roundly criticized by epidemiologists as biased and designed to produce a low estimate of IFR.

I don't attibute malice to Ioannidis (though see below).  I just think he's being a shitty scientist.  A good scientist does not let their personal bias influence their results.  Good scientists don't get "caught up" in anything.  They analyze the data in an objective manner and let it speak from themselves.  And really good scientists change their beliefs when the data shows that they were wrong.  I've had this experience a number of times.  

Your line about "directionally correct" is laughable.  There is no such thing.  Getting an accurate estimate of IFR is the goal here, not a direction.

I don't know what's in Ioannidis' heart, but he did go on TV multiple times to hype his studies and argue that the virus was not as big of a threat as others were suggesting.  Maybe he's just out for the self-promotion.  Maybe he really thinks the IFR is low.  Either way, he is massaging the numbers to artificially lower his IFR estimates, and that should not stand.

---

What would we do without twitter threads.......These twitter folks certainly have better credentials than  Ioannidis :  https://profiles.stanford.edu/john-ioannidis

Particularly Gideon M-K. One day he may just get that PHD from the Univ of Wollongong.  Until then he can write about how Red Bull’s are actually good for us......

Ad hominem attacks are the last refuge of those who have no substantive arguments to make.  If you think those arguments in the twitter feeds are wrong, I'd like to hear why.  Otherwise, you're just deflecting.  You don't need a nobel prize to how when someone else has been fudging the numbers.

I have no idea who’s right and who’s wrong Burger Which I freely admit.  My argument is that Twitter allows essentially everyone to go out there and create their own universe for lack of a better way to put it. So you’re going to get all sorts of criticisms from all sorts of people. I picked on Gideon because I spent some time researching him and he doesn’t have much in the way of credentials number one and number two a lot of his stuff as far as I can tell is based upon fancy ways for people to lose weight. I don’t think he’s anywhere nearly as credentialed as the one he attacked. Who of course may be wrong. I  fully admit that. It was never my point to suggest otherwise. 

I think there are a lot of “unknown unknowns” In this whole thing and one of the “unprecedented” about this is the amount of information that is being thrown out there by way of Internet 24 seven news stations social media etc. Which makes it very very difficult to separate the wheat from the chaff – even for those people who have a scientific ability to do so.

As for personal bias – you don’t think Gideon has personal bias? You don’t think those going after Ioannadis have personal bias? Please.

[lex24]

Best estimates I've seen are between 0.5% (if you use confirmed COVID-19 deaths) to about 1% if you use excess all-cause mortality in areas where rigorous sampling has been used to calculate the seroprevalence using a random sampling method (Spain is the best study).  Because of the strong age gradient, to a large degree IFR will be very dependent on the age distribution of the infected population.

BC

---

@Burger, if you have never been the victim of confirmation bias, you are a better man than I am.  Rather than attribute malice to Ioannidis' motives, maybe consider the possibility that a contrarian by nature got caught up a bit in his contrarianism, and accept that he was correct directionally; the ultimate IFR is substantially lower than the initial estimates from the WHO, though probably not as low as his estimates.

@Lex, even guys with the credentials of Ioannidis make mistakes.  HEre's a Nobel Prize winner who retracted a Science paper because the results were not reproducible.  I still admire and respect Ioannidis a ton, and in my response to Burger, I note that he was directionally correct, though (probably) wrong on the magnitude.

BC

---

Lastly, Gideon Meyerowitz-Katz and his co-author, Lea Merone have their own pre-print paper:

https://www.medrxiv.org/content/10.1101/...20089854v4

Their conclusions:

Based on a systematic review and meta-analysis of published evidence on COVID-19 until May, 2020, the IFR of the disease across populations is 0.68% (0.53-0.82%). However, due to very high heterogeneity in the meta-analysis, it is difficult to know if this represents the true point estimate. It is likely that, due to age and perhaps underlying comorbidities in the population, different places will experience different IFRs due to the disease. Given issues with mortality recording, it is also likely that this represents an underestimate of the true IFR figure. More research looking at age-stratified IFR is urgently needed to inform policy-making on this front.

Would love to see a mixed-effects meta analysis adjusting for age of the seropositive population to see if that would get rid of the heterogeneity.

BC

BC,  My point was not to suggest that he’s correct. I really don’t know. And of course he makes mistakes. This is a trial and error process is it not?

---

@Burger, if you have never been the victim of confirmation bias, you are a better man than I am.  Rather than attribute malice to Ioannidis' motives, maybe consider the possibility that a contrarian by nature got caught up a bit in his contrarianism, and accept that he was correct directionally; the ultimate IFR is substantially lower than the initial estimates from the WHO, though probably not as low as his estimates.

Come on, really?  It's a truism that all people suffer from confirmation bias.  But there's a chasm between that and publishing papers where one's biases influence the results.  The methods that Ioannidis is using have been roundly criticized by epidemiologists as biased and designed to produce a low estimate of IFR.

I don't attibute malice to Ioannidis (though see below).  I just think he's being a shitty scientist.  A good scientist does not let their personal bias influence their results.  Good scientists don't get "caught up" in anything.  They analyze the data in an objective manner and let it speak from themselves.  And really good scientists change their beliefs when the data shows that they were wrong.  I've had this experience a number of times.  

Your line about "directionally correct" is laughable.  There is no such thing.  Getting an accurate estimate of IFR is the goal here, not a direction.

I don't know what's in Ioannidis' heart, but he did go on TV multiple times to hype his studies and argue that the virus was not as big of a threat as others were suggesting.  Maybe he's just out for the self-promotion.  Maybe he really thinks the IFR is low.  Either way, he is massaging the numbers to artificially lower his IFR estimates, and that should not stand.

---

What would we do without twitter threads.......These twitter folks certainly have better credentials than  Ioannidis :  https://profiles.stanford.edu/john-ioannidis

Particularly Gideon M-K. One day he may just get that PHD from the Univ of Wollongong.  Until then he can write about how Red Bull’s are actually good for us......

Ad hominem attacks are the last refuge of those who have no substantive arguments to make.  If you think those arguments in the twitter feeds are wrong, I'd like to hear why.  Otherwise, you're just deflecting.  You don't need a nobel prize to how when someone else has been fudging the numbers.

I have no idea who’s right and who’s wrong Burger Which I freely admit.  My argument is that Twitter allows essentially everyone to go out there and create their own universe for lack of a better way to put it. So you’re going to get all sorts of criticisms from all sorts of people. I picked on Gideon because I spent some time researching him and he doesn’t have much in the way of credentials number one and number two a lot of his stuff as far as I can tell is based upon fancy ways for people to lose weight. I don’t think he’s anywhere nearly as credentialed as the one he attacked. Who of course may be wrong. I  fully admit that. It was never my point to suggest otherwise. 

I think there are a lot of “unknown unknowns” In this whole thing and one of the “unprecedented” about this is the amount of information that is being thrown out there by way of Internet 24 seven news stations social media etc. Which makes it very very difficult to separate the wheat from the chaff – even for those people who have a scientific ability to do so.

As for personal bias – you don’t think Gideon has personal bias? You don’t think those going after Ioannadis have personal bias? Please.

[chrisk]

The immediate crisis is the record-setting number of new cases every day. The number of hospitalizations is ready to surge past the April high. The trend of deaths has now reversed and is heading back up. Texas is scrambling to find refrigerator trucks.

Other countries took the virus seriously and contained it, while too many Americans listened to leaders who said wearing masks and socially distancing were optional, or even worse discouraged. Other countries were able to get even lesser educated citizens to comply, but in the US we have people of all social strata flouting common sense,

[chrisk]

The immediate crisis is the record-setting number of new cases every day. The number of hospitalizations is ready to surge past the April high. The trend of deaths has now reversed and is heading back up. Texas is scrambling to find refrigerator trucks.

Other countries took the virus seriously and contained it, while too many Americans listened to leaders who said wearing masks and socially distancing were optional, or even worse discouraged. Other countries were able to get even lesser educated citizens to comply, but in the US we have people of all social strata flouting common sense,

[M T]

Georgia hospitals are having to search out of state for beds for COVID patients, according to this article.

Houston (or, at least, Texas Medical Center) has managed to delay running out of beds so far.   Forecasts are OK for the next 2 weeks.  There was a dip in cases when masks were required again, but the numbers are climbing again.

---
On a separate note, this preprint (out of Univ of Col.-Boulder & Harvard, discussed on TWIV this past week) suggests using cheap & rapid (but lower sensitivity) tests for (daily) surveillance to knock the disease down.  I don't know how anybody could produce those kind of numbers in the near future, but the concept seemed reasonable to me.

On a similar note, there are at least a couple of breathalyzer style tests being researched.  I'm not sure I'd want to do a breathalyzer test knowing that someone had used it before me.  Maybe blow into a balloon and use the balloon to feed the breathalyzer.

-----

In regard to Chinese medical reports, there's a new WSJ article about paper mills producing papers by Chinese authors that appear bogus, but still got peer-reviewed and published.  This originally was about papers that were pre-COVID (I don't have access to WSJ to know if the subject matter has moved into COVID-19).
A January posting on this.  The main investigator's blog (former Stanford Med School lab person)
The claim is that clinicians are required to publish in order to be promoted but they don't have the time or the tools, so they buy papers.
I haven't waded through this but, if it is right, I worry about COVID-19 publications being polluted by bogus papers.

[M T]

Georgia hospitals are having to search out of state for beds for COVID patients, according to this article.

Houston (or, at least, Texas Medical Center) has managed to delay running out of beds so far.   Forecasts are OK for the next 2 weeks.  There was a dip in cases when masks were required again, but the numbers are climbing again.

---
On a separate note, this preprint (out of Univ of Col.-Boulder & Harvard, discussed on TWIV this past week) suggests using cheap & rapid (but lower sensitivity) tests for (daily) surveillance to knock the disease down.  I don't know how anybody could produce those kind of numbers in the near future, but the concept seemed reasonable to me.

On a similar note, there are at least a couple of breathalyzer style tests being researched.  I'm not sure I'd want to do a breathalyzer test knowing that someone had used it before me.  Maybe blow into a balloon and use the balloon to feed the breathalyzer.

-----

In regard to Chinese medical reports, there's a new WSJ article about paper mills producing papers by Chinese authors that appear bogus, but still got peer-reviewed and published.  This originally was about papers that were pre-COVID (I don't have access to WSJ to know if the subject matter has moved into COVID-19).
A January posting on this.  The main investigator's blog (former Stanford Med School lab person)
The claim is that clinicians are required to publish in order to be promoted but they don't have the time or the tools, so they buy papers.
I haven't waded through this but, if it is right, I worry about COVID-19 publications being polluted by bogus papers.

[Mick]

Best estimates I've seen are between 0.5% (if you use confirmed COVID-19 deaths) to about 1% if you use excess all-cause mortality in areas where rigorous sampling has been used to calculate the seroprevalence using a random sampling method (Spain is the best study).  Because of the strong age gradient, to a large degree IFR will be very dependent on the age distribution of the infected population.

BC

Same numbers in this WSJ article, 0.5% to 1.0% IFR.
https://www.wsj.com/articles/how-deadly-...1595323801

[Mick]

Best estimates I've seen are between 0.5% (if you use confirmed COVID-19 deaths) to about 1% if you use excess all-cause mortality in areas where rigorous sampling has been used to calculate the seroprevalence using a random sampling method (Spain is the best study).  Because of the strong age gradient, to a large degree IFR will be very dependent on the age distribution of the infected population.

BC

Same numbers in this WSJ article, 0.5% to 1.0% IFR.
https://www.wsj.com/articles/how-deadly-...1595323801