6 ms·

CDC’s New ‘Best Estimate’ Implies a Covid-19 Infection Fatality Rate Below 0.3%

37 points by mrfusion 6y ago

ajsnigrutin 6y ago
This is all great, but what is also important here is, how many of those 99.7% of people need hospital care to survive? As far as I understood all the "flatten-the-curve" motions, the plan was to keep hospitals in a state, where they're able to process the patients with complications, so they don't become the 0.3%++. My country (Slovenia) has pretty-much zero new cases per day currently, with 2-4% of people (there was a study) with antibodies.... in a potential second wave with exponential growth, our hospitals are basically fucked.
- rpiguy 6y ago
  Flattening the curve to preserve hospital beds, ,at least in the US, only really applied to NYC and northern New Jersey. Nowhere else in the country ever even came close to overwhelming the hospitals. If NJ and NYC had shut down public transportation the curve wouldn’t have been so bad - hotspots map 1:1 with subway and train stops. 40% of US counties haven’t had a single COVID death. Most hospitals are so under utilized they are losing money.
  - ajsnigrutin 6y ago
    Slovenia has pretty-much opened up everything (masks still required indoors, nightclubs still closed, public events limited to 50 people, but everything else, from stores to gyms open). We did lose like 2 months of business, so if we knew what to expect, it might have gone better (but italy is our neighbor, so we went 'better safe than sorry' way). Closings should basically be linked to number of cases in that area.
  - downrightmike 6y ago
    There was a point in Arizona that of the 1019 ICU beds, only 9 weren't being used.
  - dopylitty 6y ago
    Birmingham Alabama's hospitals are currently overwhelmed[0]. This thing is going to get even uglier quickly as it tears through rural areas that have little to no nearby healthcare options. And as the people who are out mingling this weekend start to get ill in the next week and start to feel bad enough to need hospitalization in the next two to three weeks we're going to see some scenes just like in NYC with overwhelmed hospitals and morgues being replaced by air conditioned semi trucks (or worse, bodies piling up in the summer heat). It's already happening in some rural areas[1]. The only counties that have no deaths are those with very few people and/or no testing [2]. [0] https://www.businessinsider.com/dire-coronavirus-alabama-icu-no-beds-2020-5 https://www.businessinsider.com/dire-coronavirus-alabama-icu... [1] https://www.sfgate.com/news/article/A-deadly-checkerboard-Covid-19-s-new-surge-15292024.php https://www.sfgate.com/news/article/A-deadly-checkerboard-Co... [2] https://www.texasmonthly.com/news/coronavirus-spread-rural-counties-university-texas-researchers/ https://www.texasmonthly.com/news/coronavirus-spread-rural-c...
    - odkamkfn 6y ago
      I like the slight, vague implication in your post that no testing means no deaths.
    - rpiguy 6y ago
      According to the Alabama department of public health as of 5 hours ago there are only 571 people in ICU. https://www.alabamapublichealth.gov/covid19/assets/cov-al-cases-052420.pdf https://www.alabamapublichealth.gov/covid19/assets/cov-al-ca... According to CDC data, 81% of deaths from COVID-19 in the United States are people over 65 years old, most with preexisting conditions. If you add in 55-64-year-olds that number jumps to 93%. For those below age 55, preexisting conditions play a significant role, but the death rate is currently around 0.0022%, or one death per 45,000 people in this age range. Below 25 years old the fatality rate of COVID-19 is 0.00008%, or roughly one in 1.25 million. I am sure some hospitals will come close to capacity in some areas due to some case clusters, but Florida, Texas, and Georgia have been largely fine after opening up even in rural areas. I expect over the next two to three weeks we will continue to see drops in deaths.
    - masonic 6y ago
  - 0d9eooo 6y ago
    The hospitals losing money part is complicated because most of them are losing money because of lost elective procedures. In many hospitals, especially private ones, those elective procedures are the main source of profit and underwrite the acute/ICU cases. So you could have a hospital that's at occupancy with intensive/acute care cases, and still losing a ton of money if electives are being canceled either to quarantine orders, or due to patients canceling them out of fear. There's also essentially a U-shaped curve between intensive care occupancy and profitability. As ICU/acute care cases decrease below some point, profits decrease, and as they increase they increase, but at some point the cases become overwhelming and the hospital loses money again. Beds are not all equal when it comes to profit/financial solvency at hospitals.
  - duckMuppet 6y ago
    Yes but it wasn't about a virus, this was about control. https://mobile.twitter.com/EthicalSkeptic/status/1263662455219240962 https://mobile.twitter.com/EthicalSkeptic/status/12636624552...
    - epakai 6y ago
      I tried to understand where these numbers were coming from. The 32,768,000 cases led me to seroprevalence studies (total infection estimates by sampling the population for antibodies). I think this wired article [1] really covers the signifigance of these results. Some of the issues are the results are not nationwide (Santa Clara County only), the studies weren't peer reviewed before the twitter analysis happened, there may be issues with testing method accuracy, etc. In any case these are fairly new results (April 11). Our decisions about lockdown were based on data available at the time which had a worse outlook in part due to lack of early action. I think it's also important to point out the CDC's planning scenarios state "sCFR: ... reflects existing standard of care". I presume this means our current ability, not our ability a month ago when fatalities peaked. [1] https://www.wired.com/story/new-covid-19-antibody-study-results-are-in-are-they-right/ https://www.wired.com/story/new-covid-19-antibody-study-resu...
      - elcritch 6y ago
        > I think this wired article [1] really covers the signifigance of these results. Some of the issues are the results are not nationwide (Santa Clara County only), the studies weren't peer reviewed before the twitter analysis happened, there may be issues with testing method accuracy, etc. For future consideration by others, while the Stanford paper was pre-peer review the deluge of Twitter and media articles denouncing the results based on those limitations is as bad or worse than the drawbacks of the original research. Given that multiple papers from groups around the world have produced similar or more drastic results from serological testing, it seems most likely the Stanford researchers were more correct than wrong and the Twitter crowd (included many scientists) were more incorrect. The wisdom of the crowds is not always that wise. It was largely people disagreeing with the Stanford results for various non-scientific reasons using the limitations as a way to invalidate the entirety of the results rather than understand that the issues would most probably only affect the range of the results but even accounting for that still indicated order of magnitude lower infection fatality rate than assumed from the known case rate at the time. The same reasoning should apply to the early research from Imperial College that had (IMHO) even more serious limitations, but even their results weren't completely invalidated and the current numbers are still within their estimated ranges. Similar logical failures happened when the WHO stated that there wasn't evidence that getting infected would result in immunity. While technically true and worth investigating, taking it as a justification to extend lockdown or denounce states planning to reopen was unfounded given the paucity of data showing people don't develop some degree of resistance to sars-cov-2 virus after infection. Given there are few viruses that can elude the immune system and prevent immunity, the prior knowledge by itself strongly indicates some level of immunity will be formed. None of the traits of a lingering viral infections were there, aside from a few anomalous results from Wuhan or South Korea. Luckily, later research showed those possible cases of relapse were due to false positives, but it was also a highly likely outcome based on the known of prior information about viruses and the immune system.
jimrandomh 6y ago
I'm pretty sure the CDC is just wrong about this, and the true IFR is about 1.0%. The CDC page in question doesn't have much of any information about their methodology, and the claimed 0.3% is not consistent with what other research has found. Given the CDC's extremely poor track record, I don't think it's right to assign this estimate much weight.
- pnako 6y ago
  I'm pretty sure you are just wrong about this, and the true IFR is about 17.282%. "Believe Science(R). Wait, no, not this one!"
- boomboomsubban 6y ago
  The CDC is basically testing different methods of creating models based on the available research. They note that these results >Are not predictions of the expected effects of COVID-19 Nothing wrong with their results, just some grunt work that other researchers may use. It isn't worth this media reaction though.
- nradov 6y ago
  What is your basis for being pretty sure about that?
- ericb 6y ago
  Could this channel be politically influenced, or do we think this is direct from scientists? I have no idea how to evaluate this possibility.
  - boomboomsubban 6y ago
    This is not trying to evaluate the actual fatality rate, they're generating scenarios for use with mathematic models elsewhere in the federal government.
- svaha1728 6y ago
  Agree. New York fatalities are already at .3% of the population (estimating population at 8.39 million and deaths at 23k). Antibody estimates are that only ~20% were infected.
boomboomsubban 6y ago
By "best estimate" they mean >Parameter values for disease severity, viral transmissibility, and pre-symptomatic and asymptomatic disease transmission that represent the best estimate, based on the latest surveillance data and scientific knowledge. Also, the "best estimate" predicts a .4% fatality rate, below .3% comes from the best performing model using the lower bound for disease severity...
- nradov 6y ago
  The 0.4% fatality rate is for symptomatic patients. At least a third of infections are asymptomatic, so the actual infection fatality rate is lower.
woodandsteel 6y ago
Interesting, but the question is what it implies for public policy. In particular, a lot of people on the right argue this ifr is not much worse than the flu, so there should be no government dictated restrictions. Well to address that we need to calculate the numbers. And to start, the flu death numbers are as low as they are because we have a vaccine, and every year there is an aggressive campaign aimed at the people most likely to die for the disease. We don't yet have a covid vaccine, and it is highly transmissible, so if we did nothing to prevent its spread it would likely reach a majority of people. In the US that means at least 200 million. Now multiply that by 0.2% and you get at least 400,000 deaths, 8 times the number for a bad flu season. I say that is high enough to merit aggressive governmental action to halt transmission until we get a vaccine. For those who say no, I would ask what, if anything, would be a high enough number?
- SpicyLemonZest 6y ago
  It's not a binary decision. Every reasonable person agrees the government should take action; the questions are how much is needed and how severe it can be.
  - woodandsteel 6y ago
    I agree, it's not a binary solution. But there is a significant portion of the American public, including Trump for a good while, and off and on today, that thinks the government should be doing little or nothing.
- vikramkr 6y ago
  The flu is also really bad. Making sure you are vaccinated before working in a hospital is a requirement is taken seriously, and the 1918 flu outbreak was one of the deadliest pandemics in human history. Don't underestimate the flu. Sometimes, I wonder if we would respect it more if we had truly or nearly vanquished it like Smallpox and polio. The way we have it now, just barely under control with the constant specter of a new pandemic, such as h5n1, seems to lead to less respect for how dangerous it really is, since we just see it as a really bad cold with a kind if ineffective vaccine instead of the dangerous killer that it actually is. If coronavirus was as bad as the flu, we'd be in a really bad situation. But it doesnt seem to mutate as easily or have the ability to rearrange its genome on the fly, so a vaccine might be able to defeat it properly and for good.
PhantomGremlin 6y ago
Article focuses on just fatalities. But there is evidence that infection can cause other problems for survivors. For example, rupturing of pulmonary alveoli. How quickly can those regenerate? Wikipedia says we have 300 million alveoli. If you lose 10% to the virus, is that OK? By itself that's probably not a big deal, but let's say you're a smoker who has lived his whole life in a polluted city. So you come out of ICU with only 50% of optimal lung capacity. Is that still OK? What if you only have 20% left, are you barely hanging on? Do you have to walk around breathing supplemental oxygen?
- teruakohatu 6y ago
  It is a very good question, as well as how it compares to damage done by flu or a bad cold how and quickly damage is repaired.
- Fjolsvith 6y ago
  > So you come out of ICU with only 50% of optimal lung capacity. And of course, this gets blamed on coronavirus and not the lifestyle choices.

37 points by mrfusion 6y ago

This is all great, but what is also important here is, how many of those 99.7% of people need hospital care to survive? As far as I understood all the "flatten-the-curve" motions, the plan was to keep hospitals in a state, where they're able to process the patients with complications, so they don't become the 0.3%++. My country (Slovenia) has pretty-much zero new cases per day currently, with 2-4% of people (there was a study) with antibodies.... in a potential second wave with exponential growth, our hospitals are basically fucked.

rpiguy 6y ago
Flattening the curve to preserve hospital beds, ,at least in the US, only really applied to NYC and northern New Jersey. Nowhere else in the country ever even came close to overwhelming the hospitals. If NJ and NYC had shut down public transportation the curve wouldn’t have been so bad - hotspots map 1:1 with subway and train stops. 40% of US counties haven’t had a single COVID death. Most hospitals are so under utilized they are losing money.
- ajsnigrutin 6y ago
  Slovenia has pretty-much opened up everything (masks still required indoors, nightclubs still closed, public events limited to 50 people, but everything else, from stores to gyms open). We did lose like 2 months of business, so if we knew what to expect, it might have gone better (but italy is our neighbor, so we went 'better safe than sorry' way). Closings should basically be linked to number of cases in that area.
- downrightmike 6y ago
  There was a point in Arizona that of the 1019 ICU beds, only 9 weren't being used.
- dopylitty 6y ago
  Birmingham Alabama's hospitals are currently overwhelmed[0]. This thing is going to get even uglier quickly as it tears through rural areas that have little to no nearby healthcare options. And as the people who are out mingling this weekend start to get ill in the next week and start to feel bad enough to need hospitalization in the next two to three weeks we're going to see some scenes just like in NYC with overwhelmed hospitals and morgues being replaced by air conditioned semi trucks (or worse, bodies piling up in the summer heat). It's already happening in some rural areas[1]. The only counties that have no deaths are those with very few people and/or no testing [2]. [0] https://www.businessinsider.com/dire-coronavirus-alabama-icu-no-beds-2020-5 https://www.businessinsider.com/dire-coronavirus-alabama-icu... [1] https://www.sfgate.com/news/article/A-deadly-checkerboard-Covid-19-s-new-surge-15292024.php https://www.sfgate.com/news/article/A-deadly-checkerboard-Co... [2] https://www.texasmonthly.com/news/coronavirus-spread-rural-counties-university-texas-researchers/ https://www.texasmonthly.com/news/coronavirus-spread-rural-c...
  - odkamkfn 6y ago
    I like the slight, vague implication in your post that no testing means no deaths.
  - rpiguy 6y ago
    According to the Alabama department of public health as of 5 hours ago there are only 571 people in ICU. https://www.alabamapublichealth.gov/covid19/assets/cov-al-cases-052420.pdf https://www.alabamapublichealth.gov/covid19/assets/cov-al-ca... According to CDC data, 81% of deaths from COVID-19 in the United States are people over 65 years old, most with preexisting conditions. If you add in 55-64-year-olds that number jumps to 93%. For those below age 55, preexisting conditions play a significant role, but the death rate is currently around 0.0022%, or one death per 45,000 people in this age range. Below 25 years old the fatality rate of COVID-19 is 0.00008%, or roughly one in 1.25 million. I am sure some hospitals will come close to capacity in some areas due to some case clusters, but Florida, Texas, and Georgia have been largely fine after opening up even in rural areas. I expect over the next two to three weeks we will continue to see drops in deaths.
  - masonic 6y ago
- 0d9eooo 6y ago
  The hospitals losing money part is complicated because most of them are losing money because of lost elective procedures. In many hospitals, especially private ones, those elective procedures are the main source of profit and underwrite the acute/ICU cases. So you could have a hospital that's at occupancy with intensive/acute care cases, and still losing a ton of money if electives are being canceled either to quarantine orders, or due to patients canceling them out of fear. There's also essentially a U-shaped curve between intensive care occupancy and profitability. As ICU/acute care cases decrease below some point, profits decrease, and as they increase they increase, but at some point the cases become overwhelming and the hospital loses money again. Beds are not all equal when it comes to profit/financial solvency at hospitals.
- duckMuppet 6y ago
  Yes but it wasn't about a virus, this was about control. https://mobile.twitter.com/EthicalSkeptic/status/1263662455219240962 https://mobile.twitter.com/EthicalSkeptic/status/12636624552...
  - epakai 6y ago
    I tried to understand where these numbers were coming from. The 32,768,000 cases led me to seroprevalence studies (total infection estimates by sampling the population for antibodies). I think this wired article [1] really covers the signifigance of these results. Some of the issues are the results are not nationwide (Santa Clara County only), the studies weren't peer reviewed before the twitter analysis happened, there may be issues with testing method accuracy, etc. In any case these are fairly new results (April 11). Our decisions about lockdown were based on data available at the time which had a worse outlook in part due to lack of early action. I think it's also important to point out the CDC's planning scenarios state "sCFR: ... reflects existing standard of care". I presume this means our current ability, not our ability a month ago when fatalities peaked. [1] https://www.wired.com/story/new-covid-19-antibody-study-results-are-in-are-they-right/ https://www.wired.com/story/new-covid-19-antibody-study-resu...
    - elcritch 6y ago
      > I think this wired article [1] really covers the signifigance of these results. Some of the issues are the results are not nationwide (Santa Clara County only), the studies weren't peer reviewed before the twitter analysis happened, there may be issues with testing method accuracy, etc. For future consideration by others, while the Stanford paper was pre-peer review the deluge of Twitter and media articles denouncing the results based on those limitations is as bad or worse than the drawbacks of the original research. Given that multiple papers from groups around the world have produced similar or more drastic results from serological testing, it seems most likely the Stanford researchers were more correct than wrong and the Twitter crowd (included many scientists) were more incorrect. The wisdom of the crowds is not always that wise. It was largely people disagreeing with the Stanford results for various non-scientific reasons using the limitations as a way to invalidate the entirety of the results rather than understand that the issues would most probably only affect the range of the results but even accounting for that still indicated order of magnitude lower infection fatality rate than assumed from the known case rate at the time. The same reasoning should apply to the early research from Imperial College that had (IMHO) even more serious limitations, but even their results weren't completely invalidated and the current numbers are still within their estimated ranges. Similar logical failures happened when the WHO stated that there wasn't evidence that getting infected would result in immunity. While technically true and worth investigating, taking it as a justification to extend lockdown or denounce states planning to reopen was unfounded given the paucity of data showing people don't develop some degree of resistance to sars-cov-2 virus after infection. Given there are few viruses that can elude the immune system and prevent immunity, the prior knowledge by itself strongly indicates some level of immunity will be formed. None of the traits of a lingering viral infections were there, aside from a few anomalous results from Wuhan or South Korea. Luckily, later research showed those possible cases of relapse were due to false positives, but it was also a highly likely outcome based on the known of prior information about viruses and the immune system.

jimrandomh 6y ago

I'm pretty sure the CDC is just wrong about this, and the true IFR is about 1.0%. The CDC page in question doesn't have much of any information about their methodology, and the claimed 0.3% is not consistent with what other research has found. Given the CDC's extremely poor track record, I don't think it's right to assign this estimate much weight.

pnako 6y ago
I'm pretty sure you are just wrong about this, and the true IFR is about 17.282%. "Believe Science(R). Wait, no, not this one!"
boomboomsubban 6y ago
The CDC is basically testing different methods of creating models based on the available research. They note that these results >Are not predictions of the expected effects of COVID-19 Nothing wrong with their results, just some grunt work that other researchers may use. It isn't worth this media reaction though.
nradov 6y ago
What is your basis for being pretty sure about that?
ericb 6y ago
Could this channel be politically influenced, or do we think this is direct from scientists? I have no idea how to evaluate this possibility.
- boomboomsubban 6y ago
  This is not trying to evaluate the actual fatality rate, they're generating scenarios for use with mathematic models elsewhere in the federal government.
svaha1728 6y ago
Agree. New York fatalities are already at .3% of the population (estimating population at 8.39 million and deaths at 23k). Antibody estimates are that only ~20% were infected.

boomboomsubban 6y ago

By "best estimate" they mean >Parameter values for disease severity, viral transmissibility, and pre-symptomatic and asymptomatic disease transmission that represent the best estimate, based on the latest surveillance data and scientific knowledge. Also, the "best estimate" predicts a .4% fatality rate, below .3% comes from the best performing model using the lower bound for disease severity...

nradov 6y ago
The 0.4% fatality rate is for symptomatic patients. At least a third of infections are asymptomatic, so the actual infection fatality rate is lower.

woodandsteel 6y ago

Interesting, but the question is what it implies for public policy. In particular, a lot of people on the right argue this ifr is not much worse than the flu, so there should be no government dictated restrictions. Well to address that we need to calculate the numbers. And to start, the flu death numbers are as low as they are because we have a vaccine, and every year there is an aggressive campaign aimed at the people most likely to die for the disease. We don't yet have a covid vaccine, and it is highly transmissible, so if we did nothing to prevent its spread it would likely reach a majority of people. In the US that means at least 200 million. Now multiply that by 0.2% and you get at least 400,000 deaths, 8 times the number for a bad flu season. I say that is high enough to merit aggressive governmental action to halt transmission until we get a vaccine. For those who say no, I would ask what, if anything, would be a high enough number?

SpicyLemonZest 6y ago
It's not a binary decision. Every reasonable person agrees the government should take action; the questions are how much is needed and how severe it can be.
- woodandsteel 6y ago
  I agree, it's not a binary solution. But there is a significant portion of the American public, including Trump for a good while, and off and on today, that thinks the government should be doing little or nothing.
vikramkr 6y ago
The flu is also really bad. Making sure you are vaccinated before working in a hospital is a requirement is taken seriously, and the 1918 flu outbreak was one of the deadliest pandemics in human history. Don't underestimate the flu. Sometimes, I wonder if we would respect it more if we had truly or nearly vanquished it like Smallpox and polio. The way we have it now, just barely under control with the constant specter of a new pandemic, such as h5n1, seems to lead to less respect for how dangerous it really is, since we just see it as a really bad cold with a kind if ineffective vaccine instead of the dangerous killer that it actually is. If coronavirus was as bad as the flu, we'd be in a really bad situation. But it doesnt seem to mutate as easily or have the ability to rearrange its genome on the fly, so a vaccine might be able to defeat it properly and for good.

PhantomGremlin 6y ago

Article focuses on just fatalities. But there is evidence that infection can cause other problems for survivors. For example, rupturing of pulmonary alveoli. How quickly can those regenerate? Wikipedia says we have 300 million alveoli. If you lose 10% to the virus, is that OK? By itself that's probably not a big deal, but let's say you're a smoker who has lived his whole life in a polluted city. So you come out of ICU with only 50% of optimal lung capacity. Is that still OK? What if you only have 20% left, are you barely hanging on? Do you have to walk around breathing supplemental oxygen?

teruakohatu 6y ago
It is a very good question, as well as how it compares to damage done by flu or a bad cold how and quickly damage is repaired.
Fjolsvith 6y ago
> So you come out of ICU with only 50% of optimal lung capacity. And of course, this gets blamed on coronavirus and not the lifestyle choices.