[BostonCard]

New paper in Nature Medicine estimates mortality at about 1.4% of symptomatic patients; the estimate is that 50% of patients will be symptomatic so the overall CFR of all infected people is 0.7%, which is awful but not quite as bad as the estimate based on confirmed cases (3.4%).

https://www.nature.com/articles/s41591-020-0822-7

They did some clever stuff looking at the proportion of expats who tested positive after evacuating to get a sense of the disease prevalence.

BC

[BostonCard]

New paper in Nature Medicine estimates mortality at about 1.4% of symptomatic patients; the estimate is that 50% of patients will be symptomatic so the overall CFR of all infected people is 0.7%, which is awful but not quite as bad as the estimate based on confirmed cases (3.4%).

https://www.nature.com/articles/s41591-020-0822-7

They did some clever stuff looking at the proportion of expats who tested positive after evacuating to get a sense of the disease prevalence.

BC

[burger]

New paper in Nature Medicine estimates mortality at about 1.4% of symptomatic patients; the estimate is that 50% of patients will be symptomatic so the overall CFR of all infected people is 0.7%, which is awful but not quite as bad as the estimate based on confirmed cases (3.4%).

https://www.nature.com/articles/s41591-020-0822-7

They did some clever stuff looking at the proportion of expats who tested positive after evacuating to get a sense of the disease prevalence.

BC

There are a lot of different CFR estimates coming out, almost daily now.  Each uses different methods or data from a different country.  As an interested person, I'd say the best thing is to just take a meta-analytic approach and average the CFRs instead of worrying about which used the best methodology*.  I suspect that would put it in the 1-1.5% range.

*The only exception would be to totally ignore the new preprint, which I won't even link to, claiming a CFR of something like 0.05%.  They also claim a staggering number of asymptomatic cases.  That beggars the imagination.

[burger]

New paper in Nature Medicine estimates mortality at about 1.4% of symptomatic patients; the estimate is that 50% of patients will be symptomatic so the overall CFR of all infected people is 0.7%, which is awful but not quite as bad as the estimate based on confirmed cases (3.4%).

https://www.nature.com/articles/s41591-020-0822-7

They did some clever stuff looking at the proportion of expats who tested positive after evacuating to get a sense of the disease prevalence.

BC

There are a lot of different CFR estimates coming out, almost daily now.  Each uses different methods or data from a different country.  As an interested person, I'd say the best thing is to just take a meta-analytic approach and average the CFRs instead of worrying about which used the best methodology*.  I suspect that would put it in the 1-1.5% range.

*The only exception would be to totally ignore the new preprint, which I won't even link to, claiming a CFR of something like 0.05%.  They also claim a staggering number of asymptomatic cases.  That beggars the imagination.

[2006alum]

Thanks for sharing - they made some savvy (and transparent) assumptions and it was written simply enough for those with stats knowledge but next to no medical knowledge to more or less follow. I wish this was what people reported more, rather than who's giving the virus what name. I feel like so many people are in the dark because the topline news coverage is so political rather than epidemiological.

[2006alum]

Thanks for sharing - they made some savvy (and transparent) assumptions and it was written simply enough for those with stats knowledge but next to no medical knowledge to more or less follow. I wish this was what people reported more, rather than who's giving the virus what name. I feel like so many people are in the dark because the topline news coverage is so political rather than epidemiological.

[burger]

Thanks for sharing - they made some savvy (and transparent) assumptions and it was written simply enough for those with stats knowledge but next to no medical knowledge to more or less follow. I wish this was what people reported more, rather than who's giving the virus what name. I feel like so many people are in the dark because the topline news coverage is so political rather than epidemiological.

I don't think that the problem is bad reporting.  The reporting has been excellent from what I've seen (admittedly, I get all my news from newspapers and scientific sources, so I'm probably an outlier).

The problem is more things like this:
[tweet]https://twitter.com/LisPower1/status/1241042357228916737[/tweet]

[burger]

Thanks for sharing - they made some savvy (and transparent) assumptions and it was written simply enough for those with stats knowledge but next to no medical knowledge to more or less follow. I wish this was what people reported more, rather than who's giving the virus what name. I feel like so many people are in the dark because the topline news coverage is so political rather than epidemiological.

I don't think that the problem is bad reporting.  The reporting has been excellent from what I've seen (admittedly, I get all my news from newspapers and scientific sources, so I'm probably an outlier).

The problem is more things like this:
[tweet]https://twitter.com/LisPower1/status/1241042357228916737[/tweet]

[2006alum]

But your post is precisely what I'm saying is the problem - topline news should be:

-How many states are asking for testing kits that haven't gotten them. X number were promised Y days ago - why haven't they been delivered?
-How many hospitals are out of critical care beds? What is the plan to get those hospitals the beds they need?
-When will the average American with average ILI symptoms be able to get a test? Will it be free in all circumstances?

Etc. That Trump lashes out at reporters is a tale as old as time, and distracts from the federal government's response (or lack thereof) to the illness. The more the topline coverage is stuff like this and whether or not he scribbled out Corona and replaced it with China, the less coverage and attention there is to what is critical right now. Given that we know Trump is not a credible source of information, stop covering what he says and just focus on Fauci et al. Instead the coverage is which things Trump got wrong, which helps literally no one in a time when there is limited attention span and a desire for useful information.

[2006alum]

But your post is precisely what I'm saying is the problem - topline news should be:

-How many states are asking for testing kits that haven't gotten them. X number were promised Y days ago - why haven't they been delivered?
-How many hospitals are out of critical care beds? What is the plan to get those hospitals the beds they need?
-When will the average American with average ILI symptoms be able to get a test? Will it be free in all circumstances?

Etc. That Trump lashes out at reporters is a tale as old as time, and distracts from the federal government's response (or lack thereof) to the illness. The more the topline coverage is stuff like this and whether or not he scribbled out Corona and replaced it with China, the less coverage and attention there is to what is critical right now. Given that we know Trump is not a credible source of information, stop covering what he says and just focus on Fauci et al. Instead the coverage is which things Trump got wrong, which helps literally no one in a time when there is limited attention span and a desire for useful information.

[burger]

But your post is precisely what I'm saying is the problem - topline news should be:

-How many states are asking for testing kits that haven't gotten them. X number were promised Y days ago - why haven't they been delivered?
-How many hospitals are out of critical care beds? What is the plan to get those hospitals the beds they need?
-When will the average American with average ILI symptoms be able to get a test? Will it be free in all circumstances?

Etc. That Trump lashes out at reporters is a tale as old as time, and distracts from the federal government's response (or lack thereof) to the illness. The more the topline coverage is stuff like this and whether or not he scribbled out Corona and replaced it with China, the less coverage and attention there is to what is critical right now. Given that we know Trump is not a credible source of information, stop covering what he says and just focus on Fauci et al. Instead the coverage is which things Trump got wrong, which helps literally no one in a time when there is limited attention span and a desire for useful information.

Yeah, I agree on all those points.  Softball questions are not helpful (even when they are revealing on the mindset of leaders).

I listened to Gov. Newsom's press conference last night, and the some of the questions (and answers) were actually quite intelligent, asking for details on the hospital situation, etc.  I have no idea how much of that made it into the news reports--probably not much.

[burger]

But your post is precisely what I'm saying is the problem - topline news should be:

-How many states are asking for testing kits that haven't gotten them. X number were promised Y days ago - why haven't they been delivered?
-How many hospitals are out of critical care beds? What is the plan to get those hospitals the beds they need?
-When will the average American with average ILI symptoms be able to get a test? Will it be free in all circumstances?

Etc. That Trump lashes out at reporters is a tale as old as time, and distracts from the federal government's response (or lack thereof) to the illness. The more the topline coverage is stuff like this and whether or not he scribbled out Corona and replaced it with China, the less coverage and attention there is to what is critical right now. Given that we know Trump is not a credible source of information, stop covering what he says and just focus on Fauci et al. Instead the coverage is which things Trump got wrong, which helps literally no one in a time when there is limited attention span and a desire for useful information.

Yeah, I agree on all those points.  Softball questions are not helpful (even when they are revealing on the mindset of leaders).

I listened to Gov. Newsom's press conference last night, and the some of the questions (and answers) were actually quite intelligent, asking for details on the hospital situation, etc.  I have no idea how much of that made it into the news reports--probably not much.

[BostonCard]

New paper in Nature Medicine estimates mortality at about 1.4% of symptomatic patients; the estimate is that 50% of patients will be symptomatic so the overall CFR of all infected people is 0.7%, which is awful but not quite as bad as the estimate based on confirmed cases (3.4%).

https://www.nature.com/articles/s41591-020-0822-7

They did some clever stuff looking at the proportion of expats who tested positive after evacuating to get a sense of the disease prevalence.

BC

There are a lot of different CFR estimates coming out, almost daily now.  Each uses different methods or data from a different country.  As an interested person, I'd say the best thing is to just take a meta-analytic approach and average the CFRs instead of worrying about which used the best methodology*.  I suspect that would put it in the 1-1.5% range.

*The only exception would be to totally ignore the new preprint, which I won't even link to, claiming a CFR of something like 0.05%.  They also claim a staggering number of asymptomatic cases.  That beggars the imagination.

We won't know until we have a serological test and we figure out who had asymptomatic (or minimally symptomatic) disease on a population level.  I think there are some ways of estimating, if you can identify subsets of people who are thoroughly screened (for example, isolated contacts) and can calculate the proportion of asymptomatic positives.  I do think the 3.4% number the WHO calculated is artificially inflated and I suspect it's probably around 1%.

The two cruise ships that had widespread infection might be a good place to test that hypothesis; I wish I knew what the breakdown of symptoms was.

ANother interesting place to look: Luxemburg.  They reported 149 cases overnight, out of a population of 600,000.  One could actually test the entire population.

BC

[BostonCard]

New paper in Nature Medicine estimates mortality at about 1.4% of symptomatic patients; the estimate is that 50% of patients will be symptomatic so the overall CFR of all infected people is 0.7%, which is awful but not quite as bad as the estimate based on confirmed cases (3.4%).

https://www.nature.com/articles/s41591-020-0822-7

They did some clever stuff looking at the proportion of expats who tested positive after evacuating to get a sense of the disease prevalence.

BC

There are a lot of different CFR estimates coming out, almost daily now.  Each uses different methods or data from a different country.  As an interested person, I'd say the best thing is to just take a meta-analytic approach and average the CFRs instead of worrying about which used the best methodology*.  I suspect that would put it in the 1-1.5% range.

*The only exception would be to totally ignore the new preprint, which I won't even link to, claiming a CFR of something like 0.05%.  They also claim a staggering number of asymptomatic cases.  That beggars the imagination.

We won't know until we have a serological test and we figure out who had asymptomatic (or minimally symptomatic) disease on a population level.  I think there are some ways of estimating, if you can identify subsets of people who are thoroughly screened (for example, isolated contacts) and can calculate the proportion of asymptomatic positives.  I do think the 3.4% number the WHO calculated is artificially inflated and I suspect it's probably around 1%.

The two cruise ships that had widespread infection might be a good place to test that hypothesis; I wish I knew what the breakdown of symptoms was.

ANother interesting place to look: Luxemburg.  They reported 149 cases overnight, out of a population of 600,000.  One could actually test the entire population.

BC

[oldalum]

According to Rachel Maddow last night, Iceland is going to test everyone

[oldalum]

According to Rachel Maddow last night, Iceland is going to test everyone

[burger]

The CDC monitors influenza by calculating the % of outpatient visits for flu-like illnesses.  That upturn in the last few weeks is probably the virus--flu tends to drop off by this time (and is dropping off per some lab reports).



ETA: New York state is at 7,800 cases now. Oy.

[burger]

The CDC monitors influenza by calculating the % of outpatient visits for flu-like illnesses.  That upturn in the last few weeks is probably the virus--flu tends to drop off by this time (and is dropping off per some lab reports).



ETA: New York state is at 7,800 cases now. Oy.

[Goose]

I listened to Gov. Newsom's press conference last night, and the some of the questions (and answers) were actually quite intelligent, asking for details on the hospital situation, etc.  I have no idea how much of that made it into the news reports--probably not much.

YES! This is the problem. The government at all levels doesn't necessarily want to give out all the information, and when they do, the media doesn't report it because it isn't a headline or a sound bite. The media also doesn't attempt to ask for details generally, because it won't be published anyway. And the devil is often in the details. The politicians can blow smoke and nobody pins them down on exactly what they mean and where the data came from.

[Goose]

I listened to Gov. Newsom's press conference last night, and the some of the questions (and answers) were actually quite intelligent, asking for details on the hospital situation, etc.  I have no idea how much of that made it into the news reports--probably not much.

YES! This is the problem. The government at all levels doesn't necessarily want to give out all the information, and when they do, the media doesn't report it because it isn't a headline or a sound bite. The media also doesn't attempt to ask for details generally, because it won't be published anyway. And the devil is often in the details. The politicians can blow smoke and nobody pins them down on exactly what they mean and where the data came from.

[cardcrimson]

The CDC monitors influenza by calculating the % of outpatient visits for flu-like illnesses.  That upturn in the last few weeks is probably the virus--flu tends to drop off by this time (and is dropping off per some lab reports).



ETA: New York state is at 7,800 cases now.  Oy.

Wondering if phone and/or online consultations are included. For example, in late January, I had a 102 fever, dry cough, scratchy throat. All contact with Kaiser was online or with the advice nurse on the phone.

[cardcrimson]

The CDC monitors influenza by calculating the % of outpatient visits for flu-like illnesses.  That upturn in the last few weeks is probably the virus--flu tends to drop off by this time (and is dropping off per some lab reports).



ETA: New York state is at 7,800 cases now.  Oy.

Wondering if phone and/or online consultations are included. For example, in late January, I had a 102 fever, dry cough, scratchy throat. All contact with Kaiser was online or with the advice nurse on the phone.

[M T]

The CDC monitors influenza by calculating the % of outpatient visits for flu-like illnesses.  That upturn in the last few weeks is probably the virus--flu tends to drop off by this time (and is dropping off per some lab reports).

This report is of doctor visits.  It would be hard to distinguish cases that would not have gone to the doctor without the worry that it is COVID-19.  I can't remember how many years it has been since I might have made a visit for a flu-like illness.   I would say there's likely to much confounding causes for these visits to put ANY trust in this as indicating a higher than normal number of flu-like illnesses.

More visits? yes. 
More flu like illness?  No evidence.  If you recognize the public is scared of COVID-19, this isn't evidence of anything but increased fear.

[M T]

The CDC monitors influenza by calculating the % of outpatient visits for flu-like illnesses.  That upturn in the last few weeks is probably the virus--flu tends to drop off by this time (and is dropping off per some lab reports).

This report is of doctor visits.  It would be hard to distinguish cases that would not have gone to the doctor without the worry that it is COVID-19.  I can't remember how many years it has been since I might have made a visit for a flu-like illness.   I would say there's likely to much confounding causes for these visits to put ANY trust in this as indicating a higher than normal number of flu-like illnesses.

More visits? yes. 
More flu like illness?  No evidence.  If you recognize the public is scared of COVID-19, this isn't evidence of anything but increased fear.

[M T]

There are a lot of different CFR estimates coming out, almost daily now.  Each uses different methods or data from a different country.  As an interested person, I'd say the best thing is to just take a meta-analytic approach and average the CFRs instead of worrying about which used the best methodology*.  I suspect that would put it in the 1-1.5% range.

*The only exception would be to totally ignore the new preprint, which I won't even link to, claiming a CFR of something like 0.05%.  They also claim a staggering number of asymptomatic cases.  That beggars the imagination.

I agree. Here are two papers analyzing China's numbers and making various assumptions, and coming to two numbers differing by a factor of 2 to 3 for China's CFR.

For the purposes here, and as used in one of the papers, I will use Case Fatality Rate for fatality rate/diagnosed cases  and Infection Fatality Rate for fatality rate/infections (which includes some estimate of asymptomatic patients).  In both papers, IFR was calculated from CFR based on the percentage of asymptomatic patients on the cruise ship.

Study 1        Study 2

The overall CFR they report are  1: 1.38%   2: 3.3%.  IFR:  1: 0.657%  2: 1.6%    Big difference.

They each go through a step of trying to deal with the fact that there were active cases in China when they took their numbers and adjusting the raw/crude/naive CFR (fatalities to date / diagnosed cases to date) to account for that.  That number for the two studies were  1: 3.67%  2: 6%
As the number of unresolved cases in China decreases, that number can be better estimated to be between 4.0% and 4.4%.
As of March 19, the crude CFR in China is 4.0% (probably an underestimate).     fatalities to date / (fatalities + recoveries) is 4.4% (probably an overestimate)

---

These two studies also gave their numbers for the 10-year age groups. 

The overall CFR for a country is dependent on its demographics, as the CFR is much higher for older citizens.   I multiplied their CFRs for age groups by the population percentage in each age group  and added those up.   This gives a rough estimate for the overall CFR that can be expected.  (Note this is mostly independent of how a country fights the pandemic, or how many get sick (modulo overwhelming the health system).)

Country	Study 1 CFR	Study 2 CFR
China	1.22%	3.28%
India	0.73%	1.96%
Italy	2.47%	6.47%
Spain	2.16%	5.66%
UK	1.96%	5.11%
US	1.22%	3.32%

[M T]

There are a lot of different CFR estimates coming out, almost daily now.  Each uses different methods or data from a different country.  As an interested person, I'd say the best thing is to just take a meta-analytic approach and average the CFRs instead of worrying about which used the best methodology*.  I suspect that would put it in the 1-1.5% range.

*The only exception would be to totally ignore the new preprint, which I won't even link to, claiming a CFR of something like 0.05%.  They also claim a staggering number of asymptomatic cases.  That beggars the imagination.

I agree. Here are two papers analyzing China's numbers and making various assumptions, and coming to two numbers differing by a factor of 2 to 3 for China's CFR.

For the purposes here, and as used in one of the papers, I will use Case Fatality Rate for fatality rate/diagnosed cases  and Infection Fatality Rate for fatality rate/infections (which includes some estimate of asymptomatic patients).  In both papers, IFR was calculated from CFR based on the percentage of asymptomatic patients on the cruise ship.

Study 1        Study 2

The overall CFR they report are  1: 1.38%   2: 3.3%.  IFR:  1: 0.657%  2: 1.6%    Big difference.

They each go through a step of trying to deal with the fact that there were active cases in China when they took their numbers and adjusting the raw/crude/naive CFR (fatalities to date / diagnosed cases to date) to account for that.  That number for the two studies were  1: 3.67%  2: 6%
As the number of unresolved cases in China decreases, that number can be better estimated to be between 4.0% and 4.4%.
As of March 19, the crude CFR in China is 4.0% (probably an underestimate).     fatalities to date / (fatalities + recoveries) is 4.4% (probably an overestimate)

---

These two studies also gave their numbers for the 10-year age groups. 

The overall CFR for a country is dependent on its demographics, as the CFR is much higher for older citizens.   I multiplied their CFRs for age groups by the population percentage in each age group  and added those up.   This gives a rough estimate for the overall CFR that can be expected.  (Note this is mostly independent of how a country fights the pandemic, or how many get sick (modulo overwhelming the health system).)

Country	Study 1 CFR	Study 2 CFR
China	1.22%	3.28%
India	0.73%	1.96%
Italy	2.47%	6.47%
Spain	2.16%	5.66%
UK	1.96%	5.11%
US	1.22%	3.32%