5 ms·

The CDC revised their risk estimates towards the end of last month. "The fifth scenario is the CDC's "current best estimate about viral transmission and diseas

The CDC revised their risk estimates towards the end of last month. "The fifth scenario is the CDC's "current best estimate about viral transmission and disease severity in the United States." In that scenario, the agency described its estimate that 0.4% of people who feel sick with Covid-19 will die. For people age 65 and older, the CDC puts that number at 1.3%. For people 49 and under, the agency estimated that 0.05% of symptomatic people will die." https://www.cnn.com/2020/05/22/health/cdc-coronavirus-estimates-symptoms-deaths/index.html https://www.cnn.com/2020/05/22/health/cdc-coronavirus-estima...

gridspy 6y ago
Even if you do not die, you might have a very unpleasant experience with lasting consequences. Perhaps a better way to frame it would be - 2% (?) of patients have serious life-changing consequences. Those under 49 tend to survive them.
- TheSpiceIsLife 6y ago
  > serious life-changing consequences This is a much more interesting data point. Death tends to have very few ongoing life-altering consequences, and in and of itself has a very low R0.
- GiorgioG 6y ago
  The problem for me is we don't have any way at this time of knowing at an individual level what the consequences are likely to be. I might have no symptoms if I get sick but I may transmit it my son (type 1 diabetic) who might die. Or maybe he'd be fine and I might die, etc. There's just no way to gauge what one's reaction might be to getting the coronavirus. A local, healthy 8 year old girl started having symptoms last week Thursday and died the following Tuesday. I don't expect my family's lockdown to end before well into next year at the earliest. The odds are against dying or long term complications for younger people, but is going out to dinner, a bar, sporting or music event worth the risk? Not for me. Stay safe folks.
  - ashtonkem 6y ago
    There’s the risk of death and injury, which is an interesting discussion, and then there’s the fear. With cases beginning to rise rapidly in multiple states again, I expect the latter to have a more dominant effect on our behavior
  - blub 6y ago
    That's really unfortunate with that young girl. So far it looks like these are the absolute exception, but it does sadly happen that young healthy people and kids too get seriously sick and sometimes even die. The German pediatric association publishes some (incomplete) stats about the disease in children (https://dgpi.de/covid-19-survey-update-kw22/ https://dgpi.de/covid-19-survey-update-kw22/) and there were so far 159 hospitalizations, with 21 of those in intensive care. 140 kids were released, 1 died. ~9600 <19yos registered as sick with COVID.
    - hef19898 6y ago
      Key word: Incomplete. As was Drostens study about the virus concentration in children. The German Fox News wannabe BILD tried to kill him for it. Incomplete studies and data is the norm right now. I am happy to not have to make decissions about life and death based on this. Someone has, so. And erring on the side of caution isn't necessarily a bad thing.
      - blub 6y ago
        The DGPI data is not a study, but rather they collect data from children's clinics and it is incomplete in the sense that not all clinics deliver information to them. If anything there could be more hospitalizations and deaths, but that seems rather unlikely. In any case I agree that the topic COVID and children is particularly sensitive and still blurry right now.
acqq 6y ago
In the very article you link to the case of New York is mentioned by Bergstrom, specifically to critique that estimate, and point that New York disproves it. I verified that claim: We know that the antibody tests in NYC gave the estimate of "all infected" of 20% of all. But there were around 20K deaths, and NYC has 8.3M population. Also even if we expect that the spread will stop once the 70% of population is infected the result is: 100e3 * 0.7 / 8e6 = 0.87% Still much more than 0.4%. That it's closer to 1% matches all the statistics of the countries of the world that did a lot of testing compared to the number of cases and deaths. Back to the article: Bergstrom in the article: "Given that these parameter sets underestimate fatality by a substantial margin compared to current scientific consensus, this is deeply problematic." In the same article, even CDC disclaims that their numbers are predictions: "The scenarios are intended to advance public health preparedness and planning. They are not predictions or estimates of the expected impact of COVID-19" the CDC says. So it seems you intentionally misinterpret CDC, given that it's all in the very article you linked.
JamesBarney 6y ago
Those CDC numbers are really weird btw. 0.4% of symptomatic cases is like a 0.2-0.3% IFR. 0.3% of the New York City population has already died of corona-virus. So assuming everyone in New York was infected that gives us an IFR of .3%. But if a more realistic 20% of the population was infected then the IFR could be as high as 1.5%. I tried going through all the most recent research and almost all of it came up with an IFR in the 0.5-1.2% range. And the few that had a lower rate were looking at anti-body tests in places with very low base rates.
- in_cahoots 6y ago
  Maybe the 20% of the population was more susceptible? It seems that the outbreak was concentrated in long-term care facilities, healthcare workers (who received a larger viral load than average), and the poor (who tend to have more pre-existing conditions). If the 20% is not randomly distributed the math can still work out.
  - JamesBarney 6y ago
    The math could work out, but if that was true then that means if NYC got 5 times as many infections the number of fatalities would barely change. I would find that hard to believe and I also doubt they would have enough data to validate that type of hypothesis. We're still guessing at the number of people infected, not to mention how old are they, what pre-existing conditions did they have, how do those pre-existing condition affect Covid-19 mortality in different age groups, what viral load did they receive, etc..
    - in_cahoots 6y ago
      Fair points! You could probably make an estimate with a stratified approach but if 0.3% have already died there’s no wiggle room for additional deaths.
- jakeogh 6y ago
  Influenza and pneumonia deaths by influenza season and age United States, 2008–2015: https://www.cdc.gov/nchs/data/health_policy/influenza-and-pneumonia-deaths-2008-2015.pdf https://www.cdc.gov/nchs/data/health_policy/influenza-and-pn...
mrb 6y ago
The CDC estimate (IFR 0.26%, S-IFR 0.4%) is heavily criticized because it ignores, thus predates, the 3 largest and most reliable serosurveys which all indicate an IFR of ~1% (S-IFR ~1.54%, assuming 35% of cases are asymptomatic): 1.0% in Brazil based on 25,025 samples https://twitter.com/GidMK/status/1267670267624476672 https://twitter.com/GidMK/status/1267670267624476672 1.14% in Spain based on 60,897 samples https://twitter.com/zorinaq/status/1265405628853305345 https://twitter.com/zorinaq/status/1265405628853305345 0.92% in NYC and 0.84% in New York State based on 15,101 samples https://twitter.com/GidMK/status/1267669314657611777 https://twitter.com/GidMK/status/1267669314657611777 Also YYG, the guy who runs https://covid19-projections.com/ https://covid19-projections.com/ (a model that has consistently produced the most accurate forecasts) reported his model found that a ~1% IFR gave the best fit: https://mobile.twitter.com/youyanggu/status/1256051255253757953 https://mobile.twitter.com/youyanggu/status/1256051255253757... Note: S-IFR is the fatality of symptomatic cases only. IFR is the fatality of all cases (asymptomatic and symptomatic)
- cm2187 6y ago
  The wider NY state has an IFR of 0.5%. The IFR should also be very sensitive to the demographics of the infection. Both the UK and NY State had a policy of sending recovering (but potentially contagious) older covid patients back to nursing homes and it would have exacerbated the mortality rate vs regions that did a better job at protecting the elderly. I am also starting to hear some noise that there might be cross immunity between this virus and previous coronavirus. In a recent video Prof Raoult was suggesting it might explain the low infection rate of people under 20 [1]. I am not qualified to tell whether the theory has any leg but it may be that we underestimate the spread of the virus when looking at a specific antibody. [1] https://www.youtube.com/watch?time_continue=734&v=zUbiYhknaK0&feature=emb_logo https://www.youtube.com/watch?time_continue=734&v=zUbiYhknaK... (in french)
  - acqq 6y ago
    > Both the UK and NY State had a policy of sending recovering (but potentially contagious) older covid patients back to nursing homes and it would have exacerbated the mortality rate vs regions that did a better job at protecting the elderly. Isn't the death rate the percentage of population that dies when all are infected? Then nothing can be "exacerbated" by that alone? It then just reflects the death rate which results from the people being infected, no matter the speed?

gridspy 6y ago

Even if you do not die, you might have a very unpleasant experience with lasting consequences. Perhaps a better way to frame it would be - 2% (?) of patients have serious life-changing consequences. Those under 49 tend to survive them.

TheSpiceIsLife 6y ago
> serious life-changing consequences This is a much more interesting data point. Death tends to have very few ongoing life-altering consequences, and in and of itself has a very low R0.
GiorgioG 6y ago
The problem for me is we don't have any way at this time of knowing at an individual level what the consequences are likely to be. I might have no symptoms if I get sick but I may transmit it my son (type 1 diabetic) who might die. Or maybe he'd be fine and I might die, etc. There's just no way to gauge what one's reaction might be to getting the coronavirus. A local, healthy 8 year old girl started having symptoms last week Thursday and died the following Tuesday. I don't expect my family's lockdown to end before well into next year at the earliest. The odds are against dying or long term complications for younger people, but is going out to dinner, a bar, sporting or music event worth the risk? Not for me. Stay safe folks.
- ashtonkem 6y ago
  There’s the risk of death and injury, which is an interesting discussion, and then there’s the fear. With cases beginning to rise rapidly in multiple states again, I expect the latter to have a more dominant effect on our behavior
- blub 6y ago
  That's really unfortunate with that young girl. So far it looks like these are the absolute exception, but it does sadly happen that young healthy people and kids too get seriously sick and sometimes even die. The German pediatric association publishes some (incomplete) stats about the disease in children (https://dgpi.de/covid-19-survey-update-kw22/ https://dgpi.de/covid-19-survey-update-kw22/) and there were so far 159 hospitalizations, with 21 of those in intensive care. 140 kids were released, 1 died. ~9600 <19yos registered as sick with COVID.
  - hef19898 6y ago
    Key word: Incomplete. As was Drostens study about the virus concentration in children. The German Fox News wannabe BILD tried to kill him for it. Incomplete studies and data is the norm right now. I am happy to not have to make decissions about life and death based on this. Someone has, so. And erring on the side of caution isn't necessarily a bad thing.
    - blub 6y ago
      The DGPI data is not a study, but rather they collect data from children's clinics and it is incomplete in the sense that not all clinics deliver information to them. If anything there could be more hospitalizations and deaths, but that seems rather unlikely. In any case I agree that the topic COVID and children is particularly sensitive and still blurry right now.

acqq 6y ago

In the very article you link to the case of New York is mentioned by Bergstrom, specifically to critique that estimate, and point that New York disproves it. I verified that claim: We know that the antibody tests in NYC gave the estimate of "all infected" of 20% of all. But there were around 20K deaths, and NYC has 8.3M population. Also even if we expect that the spread will stop once the 70% of population is infected the result is: 100e3 * 0.7 / 8e6 = 0.87% Still much more than 0.4%. That it's closer to 1% matches all the statistics of the countries of the world that did a lot of testing compared to the number of cases and deaths. Back to the article: Bergstrom in the article: "Given that these parameter sets underestimate fatality by a substantial margin compared to current scientific consensus, this is deeply problematic." In the same article, even CDC disclaims that their numbers are predictions: "The scenarios are intended to advance public health preparedness and planning. They are not predictions or estimates of the expected impact of COVID-19" the CDC says. So it seems you intentionally misinterpret CDC, given that it's all in the very article you linked.

JamesBarney 6y ago

Those CDC numbers are really weird btw. 0.4% of symptomatic cases is like a 0.2-0.3% IFR. 0.3% of the New York City population has already died of corona-virus. So assuming everyone in New York was infected that gives us an IFR of .3%. But if a more realistic 20% of the population was infected then the IFR could be as high as 1.5%. I tried going through all the most recent research and almost all of it came up with an IFR in the 0.5-1.2% range. And the few that had a lower rate were looking at anti-body tests in places with very low base rates.

in_cahoots 6y ago
Maybe the 20% of the population was more susceptible? It seems that the outbreak was concentrated in long-term care facilities, healthcare workers (who received a larger viral load than average), and the poor (who tend to have more pre-existing conditions). If the 20% is not randomly distributed the math can still work out.
- JamesBarney 6y ago
  The math could work out, but if that was true then that means if NYC got 5 times as many infections the number of fatalities would barely change. I would find that hard to believe and I also doubt they would have enough data to validate that type of hypothesis. We're still guessing at the number of people infected, not to mention how old are they, what pre-existing conditions did they have, how do those pre-existing condition affect Covid-19 mortality in different age groups, what viral load did they receive, etc..
  - in_cahoots 6y ago
    Fair points! You could probably make an estimate with a stratified approach but if 0.3% have already died there’s no wiggle room for additional deaths.
jakeogh 6y ago
Influenza and pneumonia deaths by influenza season and age United States, 2008–2015: https://www.cdc.gov/nchs/data/health_policy/influenza-and-pneumonia-deaths-2008-2015.pdf https://www.cdc.gov/nchs/data/health_policy/influenza-and-pn...

mrb 6y ago

The CDC estimate (IFR 0.26%, S-IFR 0.4%) is heavily criticized because it ignores, thus predates, the 3 largest and most reliable serosurveys which all indicate an IFR of ~1% (S-IFR ~1.54%, assuming 35% of cases are asymptomatic): 1.0% in Brazil based on 25,025 samples https://twitter.com/GidMK/status/1267670267624476672 https://twitter.com/GidMK/status/1267670267624476672 1.14% in Spain based on 60,897 samples https://twitter.com/zorinaq/status/1265405628853305345 https://twitter.com/zorinaq/status/1265405628853305345 0.92% in NYC and 0.84% in New York State based on 15,101 samples https://twitter.com/GidMK/status/1267669314657611777 https://twitter.com/GidMK/status/1267669314657611777 Also YYG, the guy who runs https://covid19-projections.com/ https://covid19-projections.com/ (a model that has consistently produced the most accurate forecasts) reported his model found that a ~1% IFR gave the best fit: https://mobile.twitter.com/youyanggu/status/1256051255253757953 https://mobile.twitter.com/youyanggu/status/1256051255253757... Note: S-IFR is the fatality of symptomatic cases only. IFR is the fatality of all cases (asymptomatic and symptomatic)

cm2187 6y ago
The wider NY state has an IFR of 0.5%. The IFR should also be very sensitive to the demographics of the infection. Both the UK and NY State had a policy of sending recovering (but potentially contagious) older covid patients back to nursing homes and it would have exacerbated the mortality rate vs regions that did a better job at protecting the elderly. I am also starting to hear some noise that there might be cross immunity between this virus and previous coronavirus. In a recent video Prof Raoult was suggesting it might explain the low infection rate of people under 20 [1]. I am not qualified to tell whether the theory has any leg but it may be that we underestimate the spread of the virus when looking at a specific antibody. [1] https://www.youtube.com/watch?time_continue=734&v=zUbiYhknaK0&feature=emb_logo https://www.youtube.com/watch?time_continue=734&v=zUbiYhknaK... (in french)
- acqq 6y ago
  > Both the UK and NY State had a policy of sending recovering (but potentially contagious) older covid patients back to nursing homes and it would have exacerbated the mortality rate vs regions that did a better job at protecting the elderly. Isn't the death rate the percentage of population that dies when all are infected? Then nothing can be "exacerbated" by that alone? It then just reflects the death rate which results from the people being infected, no matter the speed?