What's new
Fantasy Football - Footballguys Forums

Welcome to Our Forums. Once you've registered and logged in, you're primed to talk football, among other topics, with the sharpest and most experienced fantasy players on the internet.

Covid IFR - Wisdom of THIS Crowd (1 Viewer)

How has YOUR estimate of Covid' 19's IFR changed from 2 months ago?

  • I now think Covid is MUCH LESS deadly than I believed it was 2 months ago

    Votes: 12 14.6%
  • I now think Covid is SOMEWHAT LESS deadly than I believed it was 2 months ago

    Votes: 25 30.5%
  • I now think Covid is ABOUT AS deadly as I thought it was 2 months ago

    Votes: 38 46.3%
  • I now think Covid is SOMEWHAT MORE deadly than I believed it was 2 months ago

    Votes: 5 6.1%
  • I now think Covid is MUCH MORE deadly than I believed it was 2 months ago

    Votes: 2 2.4%

  • Total voters
    82
  • Poll closed .

Morton Muffley

Footballguy
There's been alot of discussion about how deadly Covid-19 is. And with more data coming out daily (including estimates from epidemiologists everywhere), it's seems we can all pick and choose which study to believe to back up whatever opinion we had/have.  So, thought it made sense to get a feel for what you all thinks is YOUR best guess at the IFR and if your beliefs have changed much over the past 2 months.

Note: IFR is the % of people who die from Covid-19 divided by the percentage of people who contracted it.  As such, the denominator INCLUDES those who are asymptomatic and/or never got tested.  In this way it is different from the Case Fatality Rate (CFR).  So, please answer the questions for Covid's IFR.

Thanks.

 
There's been alot of discussion about how deadly Covid-19 is. And with more data coming out daily (including estimates from epidemiologists everywhere), it's seems we can all pick and choose which study to believe to back up whatever opinion we had/have.  So, thought it made sense to get a feel for what you all thinks is YOUR best guess at the IFR and if your beliefs have changed much over the past 2 months.

Note: IFR is the % of people who die from Covid-19 divided by the percentage of people who contracted it.  As such, the denominator INCLUDES those who are asymptomatic and/or never got tested.  In this way it is different from the Case Fatality Rate (CFR).  So, please answer the questions for Covid's IFR.

Thanks.
i am not very familiar with these terms or how they come up with them but the bolded seems wrong to me.  How do they include people that don't have symptoms and never got tested as being people that contracted the virus?  Wouldn't this just be everyone?  Is it just a complete guess assuming how many people they think have the virus?  If that is the case it seems like it is very unreliable. 

 
There's been alot of discussion about how deadly Covid-19 is. And with more data coming out daily (including estimates from epidemiologists everywhere), it's seems we can all pick and choose which study to believe to back up whatever opinion we had/have.  So, thought it made sense to get a feel for what you all thinks is YOUR best guess at the IFR and if your beliefs have changed much over the past 2 months.

Note: IFR is the % of people who die from Covid-19 divided by the percentage of people who contracted it.  As such, the denominator INCLUDES those who are asymptomatic and/or never got tested.  In this way it is different from the Case Fatality Rate (CFR).  So, please answer the questions for Covid's IFR.

Thanks.


i am not very familiar with these terms or how they come up with them but the bolded seems wrong to me.  How do they include people that don't have symptoms and never got tested as being people that contracted the virus?  Wouldn't this just be everyone?  Is it just a complete guess assuming how many people they think have the virus?  If that is the case it seems like it is very unreliable. 
Am fairly certain that my definition is correct, but certainly would ask others to chime in if they feel differently.  And fwiw, the terminology is INCREDIBLY important and often misunderstood.  The WHO cited a CFR of 3.4% early on that has been continuously (and I think intentionally) mis-used to demonstrate how greatly the over-estmated the deadliness of Covid - when as you can see CFR will nearly always be multitudes higher than IFR.  As for IFR containing asymptomatic people, that's kinda what makes this so hard and also why I think any calculation (am looking at you CDC) that reports the IFR only among the symptomatic population is unhelpful at best and misleading at worst.

As always, just my opinion.  

 
Am fairly certain that my definition is correct, but certainly would ask others to chime in if they feel differently.  And fwiw, the terminology is INCREDIBLY important and often misunderstood.  The WHO cited a CFR of 3.4% early on that has been continuously (and I think intentionally) mis-used to demonstrate how greatly the over-estmated the deadliness of Covid - when as you can see CFR will nearly always be multitudes higher than IFR.  As for IFR containing asymptomatic people, that's kinda what makes this so hard and also why I think any calculation (am looking at you CDC) that reports the IFR only among the symptomatic population is unhelpful at best and misleading at worst.

As always, just my opinion.  
What exactly does IFR stand for?

 
Am fairly certain that my definition is correct, but certainly would ask others to chime in if they feel differently.  And fwiw, the terminology is INCREDIBLY important and often misunderstood.  The WHO cited a CFR of 3.4% early on that has been continuously (and I think intentionally) mis-used to demonstrate how greatly the over-estmated the deadliness of Covid - when as you can see CFR will nearly always be multitudes higher than IFR.  As for IFR containing asymptomatic people, that's kinda what makes this so hard and also why I think any calculation (am looking at you CDC) that reports the IFR only among the symptomatic population is unhelpful at best and misleading at worst.

As always, just my opinion.  


What exactly does IFR stand for?
Let me translate:

The Who: rock band from way back then

CDC: Rap song. "I'm down with CDC, yeah you know me."

CFR: Corn fed Republicans

IFR: I Fart Responsibly

 
What exactly does IFR stand for?
Infection fatality rate, i.e. what percent of people who are infected with the virus end up dying.

vs. Case fatality rate, i.e. what percent of people who are infected, and suffer symptoms badly enough that they seek out medical treatment, end up dying. 

 
Last edited by a moderator:
Currently 5.8% of all confirmed covid-19 cases die.

But we know there are a lot of missed tests and we know there are a lot of asymptomatic cases.  

Efforts to figure this out have been made with antibody testing.  Based on the tests done in New York, Spain, ITaly and France, I'd say between .8% and 1.2%.  All four of those areas had numbers in this range.  How accurate were the tests and how "random" were their processes?  Who knows.  But it seems like a good floor.

I haven't seen any good studies or any solid data that would suggest anything even remotely close to .5% or below (if it exists, someone please link it).

The Diamond Princess was a great example of a closed system, and it was 1.8%.

I think way too much has been made of the ifr, especially as it relates to asymptomatic cases.  Yes, the ultimate ifr will be much lower than we can determine now, but it doesn't change the fact that if you have a confirmed covid-19 case, and if it's serious enough that you go get tested, you currently are in a group with a 5.8% death rate.  It's not a good virus to catch, at all.  If a significant portion of the world ends up getting mild cases or show no symptoms, in the long run that's great, but it still doesn't change the fact that many are dying and getting full-blown cases, which cause serious issues.

 
Last edited by a moderator:
Infection fatality rate, i.e. what percent of people who are infected with the virus end up dying.

vs. Case fatality rate, i.e. what percent of people who are infected, and suffer symptoms badly enough that they seek out medical treatment, end up dying. 
So how do they come up with the number of people infected?  Guess?  If it includes people not tested and showing no symptoms how do they know they are infected?  This is where my confusion comes in.  This seems like a mythical number that there is no way to accurately calculate.

 
I don't like the bands - I'm not ready to think about this in terms of +/- 0.2% yet...I'm thinking in terms of +/- 0.5% or more.  

 
So how do they come up with the number of people infected?  Guess?  If it includes people not tested and showing no symptoms how do they know they are infected?  This is where my confusion comes in.  This seems like a mythical number that there is no way to accurately calculate.
We can derive reasonably accurate estimates with statistical methods but yeah, obviously if there are a lot of asymptomatic cases you'll have a hard time nailing down the exact number.  

 
So how do they come up with the number of people infected?  Guess?  If it includes people not tested and showing no symptoms how do they know they are infected?  This is where my confusion comes in.  This seems like a mythical number that there is no way to accurately calculate.
Exactly.  Asymptomatic people who never get tested DON'T know they have it.  It's like asking me if I enjoy cheeseburgers if I've never had one.

 
For the record, CDC published 5 scenarios.  you can multiply their CFR by % symptomatic (or, 1- asymptomatic) and come up with IFR.  Their 5 scenarios range from 0.1% -> 0.8% - that's a pretty big spread.

 
So how do they come up with the number of people infected?  Guess?  If it includes people not tested and showing no symptoms how do they know they are infected?  This is where my confusion comes in.  This seems like a mythical number that there is no way to accurately calculate.
They use antibody testing on a segment of the population. 

From what I've seen, there are two potential issues with this antibody testing:

1. specificity of the antibody tests.  Many of the tests used in some areas are essentially worthless, as they give out a false report half the time.  But there are a number of tests that are performing excellently and have a specificity well over 99%

2. How random are the samples?  If they advertise free antibody testing, they will get anyone and everyone who has been sick over the past few months and wants to know if they had covid.  That could lead to inaccurate numbers very easily.  They need the pool to be as random as possible.

 
We can derive reasonably accurate estimates with statistical methods but yeah, obviously if there are a lot of asymptomatic cases you'll have a hard time nailing down the exact number.  
How can we possibly derive reasonably accurate estimates?  I live in a family of 3.  None of us are sick, and none of us have been tested.  Maybe we all are asymptomatic, maybe we aren't.  Do we count as 0, 1, 2, or 3?

 
For the record, CDC published 5 scenarios.  you can multiply their CFR by % symptomatic (or, 1- asymptomatic) and come up with IFR.  Their 5 scenarios range from 0.1% -> 0.8% - that's a pretty big spread.
And they provided very little data behind it.  Since they didn't provide the data, we really can't use their numbers, unless we're just taking their word for it.

 
They use antibody testing on a segment of the population. 

From what I've seen, there are two potential issues with this antibody testing:

1. specificity of the antibody tests.  Many of the tests used in some areas are essentially worthless, as they give out a false report half the time.  But there are a number of tests that are performing excellently and have a specificity well over 99%

2. How random are the samples?  If they advertise free antibody testing, they will get anyone and everyone who has been sick over the past few months and wants to know if they had covid.  That could lead to inaccurate numbers very easily.  They need the pool to be as random as possible.
Neither of those scenarios seems like it is trustworthy at all to get what is needed.  Way too many factors to get anything near a reliable representation of data.

 
How can we possibly derive reasonably accurate estimates?  I live in a family of 3.  None of us are sick, and none of us have been tested.  Maybe we all are asymptomatic, maybe we aren't.  Do we count as 0, 1, 2, or 3?
The same way I have no idea what your family thinks of President Trump but we have statistical methods to determine reasonably accurate estimates of his popularity.  We don't need to test everyone, just a large enough sample of people (which is something we probably haven't done yet for COVID-19, which is why we probably don't yet have a reasonably accurate estimate of the IFR).  

 
between .8% and 1.2%.
I believe this is about right, discerning through the information with admittedly no expertise.

You don't want to get this virus. It's universally transmissible, and people get extremely sick from it. Like go to the hospital and potentially die from it sick.

Percentages and likelihoods matter until they don't. Avoid becoming a statistic. 

 
Neither of those scenarios seems like it is trustworthy at all to get what is needed.  Way too many factors to get anything near a reliable representation of data.
In many cases, this is true.  As an example, two states published info on their antibody testing.  Florida did a great job of random testing.  But the test they used had a specificity of less than 95%.  Which means that their data was basically rubbish.

New York State used really good tests, and they did a decent job of trying to get random data, but there were still some major flaws in that methodology.  For instance, they would suddently show up at random stores in the New York State area with testing tables.  Pretty good so far...but the issue is that people quickly used social media to announce where they were doing free antibody testing, which skews things.

The Stanford study used a mix of bad tests AND did a terrible job of getting a random sample, as they advertised it on social media in advance, which just negates the whole point of the study.

Spain, France and Italy have recently released some studies all showing .8% to 1.2%, but I haven't investigated their methods all that much.

 
I believe this is about right, discerning through the information with admittedly no expertise.

You don't want to get this virus. It's universally transmissible, and people get extremely sick from it. Like go to the hospital and potentially die from it sick.

Percentages and likelihoods matter until they don't. Avoid becoming a statistic. 
Yeah, I think the focus on the IFR has casued people to forget that getting a bad case of this virus is not a good thing.  If the IFR ultimately ends up being .8%, that doesn't mean your chances of dying if you get a serious case are .8%, because the asymptomatic are included in the IFR.

 
Yeah, I think the focus on the IFR has casued people to forget that getting a bad case of this virus is not a good thing.  If the IFR ultimately ends up being .8%, that doesn't mean your chances of dying if you get a serious case are .8%, because the asymptomatic are included in the IFR.
Sure but it's not like you can go out and protect yourself only from a "serious case," you're either going to get infected or not.  Some percentage of people who get infected will develop a serious case, and some percentage of those people will die.  It's useful to know both metrics. 

 
So how do they come up with the number of people infected?  Guess?  If it includes people not tested and showing no symptoms how do they know they are infected?  This is where my confusion comes in.  This seems like a mythical number that there is no way to accurately calculate.
Some places, like South Korea, have tested so many people it's likely they're picking up on most of the asymptomatic cases.

Also, there are a handful of decent antibody-based tests at this point.  Where a group of people are randomly tested to see how prevalent the disease is.  Since asymptomatic people show up in those counts you get a pretty good measure of the true infection rate.

There are some problems with both methods, but the good news is that all of them that I'm aware of have yielded IFRs in the .68%-.89% range (IIRC).  Depending on what you believe about false positives and false negatives, and the degree of the undercount in fatalities, that range moves some.  I think .7% to maybe 1.1% makes a lot of sense.

I don't believe the initial 3%+ (this is only true for places where asymptomatics were being missed) or <.3% (generally only being promoted by allies of the Trump Admin) are reasonable estimates.  There's too much evidence out there all pointing to the ranges above.

 
Last edited by a moderator:
Sure but it's not like you can go out and protect yourself only from a "serious case," you're either going to get infected or not.  Some percentage of people who get infected will develop a serious case, and some percentage of those people will die.  It's useful to know both metrics. 
I agree with that...but there is just so much bad info out there.  The reality is that currently, if you get diagnosed with covid, you're in a group with a 5.8% fatality rate.  So many are saying, "I only have a 1 percent chance of dying if I get it".  While this may be true, it's not really a true representation of the risk once you've gotten a confirmed case.

 
Some places, like South Korea, have tested so many people it's likely they're picking up on most of the asymptomatic people.

Also, there are a handful of decent antibody tests at this point.  Where a group of people are randomly tested to see how prevalent the disease is -- since asymptomatic people show up in those counts you get a pretty good measure of the true infection rate.

There are some problems with both methods, but the good news is that ALL of them that I'm aware of have yielded IFRs in the .68-.89 range.  Depending on what you believe about false positives and false negatives and the degree of the undercount that's happening with fatalities that range moves some.  I think .7% to 1.0% makes a lot of sense.

I don't believe the initial 3+% (this is only true for places where asymptomatics were being missed) or .3% (generally only being promoted by allies of the Trump Admin) are reasonable estimates.  There's too much evidence out there all pointing to the ranges above.
:goodposting:

 
I agree with that...but there is just so much bad info out there.  The reality is that currently, if you get diagnosed with covid, you're in a group with a 5.8% fatality rate.  So many are saying, "I only have a 1 percent chance of dying if I get it".  While this may be true, it's not really a true representation of the risk once you've gotten a confirmed case.
I think the IFR is a truer representation of the risk.  Getting a confirmed case is an arbitrary threshold.  Similarly, given my age and health history, etc., the reality is not at all true that I'm in a group with a 5.8% fatality rate.  But that's why all of this requires context. 

When I'm trying to decide what level(s) of risk I'm willing to take when I leave my house, for example, I'm essentially interested in three things: what are the odds that I catch the disease; if I catch the disease, what are the odds I develop severe symptoms; if I develop severe symptoms, what are the odds that I die.  I think it's kind of pointless to argue that one of those is more important than another, they're all part of the calculus.  

 
I hear you and don't disagree, but this is where the "wisdom of the crowd" works its magic.  Make your best guess
ok.  i'm not ready to put down my vote yet until I get a chance to play with some numbers now that I have a better feeling for how CFR is calculated (weighted for demographics).  I am comfortable saying I believe it to be far less than I did 2 months ago, but that's mostly because of my new understanding (per yesterday).  In fact, I'd be comfortable saying I think it's much less than I thought 2 days ago.  Big thanks to @Dr_Zaius and this post.

 
ok.  i'm not ready to put down my vote yet until I get a chance to play with some numbers now that I have a better feeling for how CFR is calculated (weighted for demographics).  I am comfortable saying I believe it to be far less than I did 2 months ago, but that's mostly because of my new understanding (per yesterday).  In fact, I'd be comfortable saying I think it's much less than I thought 2 days ago.  Big thanks to @Dr_Zaius and this post.
I think it more likely that young people are being disproportionately exposed than the other way around. Most older people I know are being much more strict about staying at home. 

 
I think it more likely that young people are being disproportionately exposed than the other way around. Most older people I know are being much more strict about staying at home. 
As it should be.  If you have more risk if you get it then you should take more precautions to avoid it.  It's not worth it to you to go to the store for a non-essential so don't go.  Where if your a healthy, young individual your risk tolerance is more so you are more willing to risk it.

 
As it should be.  If you have more risk if you get it then you should take more precautions to avoid it.  It's not worth it to you to go to the store for a non-essential so don't go.  Where if your a healthy, young individual your risk tolerance is more so you are more willing to risk it.
Yes, but to the context of this discussion, it means the true IFR is higher than it appears. 

 
Some places, like South Korea, have tested so many people it's likely they're picking up on most of the asymptomatic cases.

Also, there are a handful of decent antibody-based tests at this point.  Where a group of people are randomly tested to see how prevalent the disease is.  Since asymptomatic people show up in those counts you get a pretty good measure of the true infection rate.

There are some problems with both methods, but the good news is that all of them that I'm aware of have yielded IFRs in the .68%-.89% range (IIRC).  Depending on what you believe about false positives and false negatives, and the degree of the undercount in fatalities, that range moves some.  I think .7% to maybe 1.1% makes a lot of sense.

I don't believe the initial 3%+ (this is only true for places where asymptomatics were being missed) or <.3% (generally only being promoted by allies of the Trump Admin) are reasonable estimates.  There's too much evidence out there all pointing to the ranges above.
Of interest

Officially, I don’t know, and I think the true IFR will remain unknown for some time. Gun to my head I think  0.60%-0.80% is a reasonable range of outcomes. But it may be higher. Unless it’s lower.

Math is easy, epidemiology is hard af.

 
Last edited by a moderator:
And they provided very little data behind it.  Since they didn't provide the data, we really can't use their numbers, unless we're just taking their word for it.
the CDC ***official*** death data is here.  I don't know where they get the case data from.

 
Of interest

Officially, I don’t know, and I think the true IFR will remain unknown for some time. Gun to my head I think  0.60%-0.80% is a reasonable range of outcomes. But it may be higher. Unless it’s lower.

Math is easy, epidemiology is hard af.
Right.  I think there are still enough unknowns that the current estimates could be wrong.

Having said -- I don't see much evidence at all for anything much outside that range.  If you think we're missing like 50% of all deaths (which seems not likely) you can get to higher numbers but it's pretty hard to get to much above 1.0% with the info we've got today.

 
I think way too much has been made of the ifr, especially as it relates to asymptomatic cases.  Yes, the ultimate ifr will be much lower than we can determine now, but it doesn't change the fact that if you have a confirmed covid-19 case, and if it's serious enough that you go get tested, you currently are in a group with a 5.8% death rate.  It's not a good virus to catch, at all.  If a significant portion of the world ends up getting mild cases or show no symptoms, in the long run that's great, but it still doesn't change the fact that many are dying and getting full-blown cases, which cause serious issues.
Yep. And if you’re symptomatic, you’ll end up in the hospital a fifth of the time, with no guarantee all your organs still work when you go home.

I’ve said it before, the pandemic would be a healthcare crisis even if the CFR/IFR were 0.

 
Last edited by a moderator:
I think the IFR is a truer representation of the risk.  Getting a confirmed case is an arbitrary threshold.  Similarly, given my age and health history, etc., the reality is not at all true that I'm in a group with a 5.8% fatality rate.  But that's why all of this requires context. 

When I'm trying to decide what level(s) of risk I'm willing to take when I leave my house, for example, I'm essentially interested in three things: what are the odds that I catch the disease; if I catch the disease, what are the odds I develop severe symptoms; if I develop severe symptoms, what are the odds that I die.  I think it's kind of pointless to argue that one of those is more important than another, they're all part of the calculus.  
Probably should add: what are the odds I catch it and unknowingly spread it to another person who gets sick? Or more accurately, what are the odds I promote transmission which ultimately harms someone else?

 
Question for the OP and the Audience: What were the total number of deaths in America...say Jan of 2019 thru April of 2019 vs Same time period in 2020?

I believe 100k people have died and had CV-19 attributed to their deaths. What I am curious about is simply the total death count in this country year over year vs right now.

-The idea here is that it's possible a few less people have died of Diabetes, Cancer and some of the other top life takers we are used to reading about or seeing first hand in our families. 

I'm not a huge doubter of the number reported, I just question if there are a few less dying of the usual culprits. 

I also believe more people have committed suicide over the same time vs last year as this pandemic has pushed people and also PITTED people against each other which has a major impact on mental health in this country which was already out of control and especially with drug companies on the TV every night begging people to come forward with illness. 

Cheers

 
Last edited by a moderator:
So you gotta wait until the end of the year or several months because that neat little graph does show a spike but the median doesn't seem to rock that much. 
Not sure what you mean by that, but to more directly answer your question, there are tens of thousands more deaths over the last two months than would normally be expected at this time of year.  This has been observed in countries all over the world.  It's being tracked in a few different places, just google "excess deaths" or "excess mortality" for more resources.  It's a useful way to estimate the impact of the pandemic, especially while we're still in the thick of it, because it avoids the complications of "what's being classified as a COVID death?" etc.  

 
Not sure what you mean by that, but to more directly answer your question, there are tens of thousands more deaths over the last two months than would normally be expected at this time of year.  This has been observed in countries all over the world.  It's being tracked in a few different places, just google "excess deaths" or "excess mortality" for more resources.  It's a useful way to estimate the impact of the pandemic, especially while we're still in the thick of it, because it avoids the complications of "what's being classified as a COVID death?" etc.  
🤙 

Thank You IE, it's been hard for me to balance watching any news at all, I've mostly just been working on my health and immune system if folks have been following in the OFD thread and not chiming in with much. 

You probably are more involved in the back and forth of all this but I'll share that my wife works in a hospital where there's been a spike in the last 2 weeks here in the Palm Beaches and there are 15 NEW open CV-19 cases that are not being reported at her hospital, 2 on ventilators. 

I can't really get into the CV-19 thread and I wasn't received well even though I felt I had a lot of info to share and wanted to share but this place gets dicey and it was just better for me to take some quiet solace watching things unfold. 

But I'm glad we can see the actual surge on those graphs you are linking to, thanks again.  

 
🤙

Thank You IE, it's been hard for me to balance watching any news at all, I've mostly just been working on my health and immune system if folks have been following in the OFD thread and not chiming in with much. 

You probably are more involved in the back and forth of all this but I'll share that my wife works in a hospital where there's been a spike in the last 2 weeks here in the Palm Beaches and there are 15 NEW open CV-19 cases that are not being reported at her hospital, 2 on ventilators. 

I can't really get into the CV-19 thread and I wasn't received well even though I felt I had a lot of info to share and wanted to share but this place gets dicey and it was just better for me to take some quiet solace watching things unfold. 

But I'm glad we can see the actual surge on those graphs you are linking to, thanks again.  
The second poster to mention a spike in South Florida hospitals...That definitely doesn't jive with the official numbers from the state.  Quite concerning.

 
Last edited by a moderator:
The second poster to mention a spike in South Florida hospitals...That definitely doesn't jive with the official numbers from the state.  Quite concerning.
It makes you wonder what else we don't know or not having the best information to make decisions for ourselves just as citizens trying to move ahead, slowly I might add. 

I find myself questioning the entire machine top to bottom, it's not a comfortable feeling is it?

 

Users who are viewing this thread

Top