4 ms·

Deaths / Deaths + Recoveries gives a 5.7% death rate. Obviously there are mild cases that are never diagnosed, but it seems if you get sick enough for a confir

Deaths / Deaths + Recoveries gives a 5.7% death rate. Obviously there are mild cases that are never diagnosed, but it seems if you get sick enough for a confirmed diagnosis, your risk is higher than generally reported.

vonmoltke 7y ago
This is only true if you assume the unresolved cases will resolve in the same ratio as the currently-resolved cases. This ratio has been dropping as more cases are diagnosed.
Alupis 7y ago
The problem here is, from what doctors seem to be saying, there's a massive amount of the population that has or has had COVID-19 and will never know - since the symptoms in an otherwise healthy adult are similar to the common Flu. So that death rate is likely far, far lower than what it seems. The flu kills between 15-60k people in the US every year, and it's around 1% fatality rate (and nobody seems overly concerned about that).
- DennisP 7y ago
  That's what I mean by the "mild cases that are never diagnosed."
  - Alupis 7y ago
    Well sure, but we could say that exact thing about the common flu too. If you're sick enough to go to the doctor and be diagnosed with the flu, there's likely an increased chance of it being fatal (in that your symptoms are extreme or you're already high-risk, ie. elderly or in poor health). Otherwise, like most folks, you just get better on your own in a few days. With the flu killing 10's of thousands every year, and nobody bats an eye... I'm not particularly concerned about COVID-19. It's mostly a media PR stunt at this point - designed to whip up fear and clicks. Swine Flu anyone?
    - wahern 7y ago
      A case fatality rate (CFR) is calculated based on confirmed cases. The CFR for the flu[1] or any other disease is roughly based on the number of people who present to the hospital and are diagnosed, with or without serological testing. Therefore, in places with drive-by testing the CFR will be artificially lower. Likewise, using total infected as the denominator makes it a totally useless figure for comparison. Knowing that, it seems pretty clear that COVID-19 is more-or-less 10x deadlier[2] than seasonal influenza, as the CFR for seasonal flu is ~0.1-0.2%. [1] It's a little more complicated with the flu because of underreporting. I guess elderly pneumonia is common enough that doctors don't always care why it developed. So apparently the CDC uses a more complex model (e.g. total number of pneumonia cases minus base rate, non-flu season cases). But what they're trying to estimate is the CFR, which is by definition deaths/confirmed. [2] At least from an epidemiological perspective, regardless of the percentage of symptomatic cases / infections. At the end of the day what matters is how many will die, not how scared someone should be if they're infected.
      - Alupis 7y ago
        > A case fatality rate (CFR) is calculated based on confirmed cases. This is not true[1]. CDC projects the numbers and estimates total number of cases, because almost all of the actual cases are undiagnosed. So no, COVID-19 is not "more-or-less 10x deadlier" than the common Flu. Let's be clear. The media frenzy around COVID-19 is not grounded in science, and it purely designed to get panic people, resulting in clicks and eye-balls. Nothing more. [1] https://www.cdc.gov/flu/about/burden/how-cdc-estimates.htm https://www.cdc.gov/flu/about/burden/how-cdc-estimates.htm
- codeulike 7y ago
  South Korea have been testing more people than anyone else - 10,000 per day, drive-in test centres, results in two hours. If you look at the South Korea numbers here (scroll down a bit) https://www.worldometers.info/coronavirus/ https://www.worldometers.info/coronavirus/ you see 43 deaths and 52 'serious/critical' cases from 6,593 confirmed cases, so about 0.75% for both categories. I'm hoping these South Korea numbers are giving us a more accurate picture of whats going on.
  - Alupis 7y ago
    I'm fairly confident if we started testing everyone with Flu-like symptoms, we'd show numbers even lower than the 0.75% in South Korea. It's just not as deadly as was initially reported.
- pieix 7y ago
  Per CDC numbers [0], in the US the death rate of the flu for the 2018-19 season was roughly 0.1%, an order of magnitude lower than the number quoted above. [0] https://www.cdc.gov/flu/about/burden/past-seasons.html https://www.cdc.gov/flu/about/burden/past-seasons.html
  - Alupis 7y ago
    That's interesting! Thanks for the link. It only proves my point further. From your link, 34,200 people died from the Flu in the US last year - and not a single person is panicked about those numbers.
    - dougmwne 7y ago
      I think this is the difference between velocity and acceleration. Flu deaths are known and well studied. Our hospital system has the capacity to treat serious cases. There are existing public health initiatives around vaccination and education. The resources have been allotted and accounted for and flu deaths aren't expect to suddenly 10x or 20x. This virus is an unknown. It is spreading worldwide very quickly and the growth rate is exponential. Within months, hundreds of millions could be infected without interventions. Today the number of dead is very small. If the virus were to become as common as the flu, it could make infectious disease the leading cause of death, more than heart disease or cancer. We still don't know what the long term health complications are or if you can be re-infected. The science will be in shortly, but in the mean time an abundance of caution would seem to be warranted. I will be washing my hands until then.
    - adventured 7y ago
      > and not a single person is panicked about those numbers. That's for a very obvious reason. If 40 million people in the US get Covid, as with the flu, it will probably kill between 500k and 1m people. That's unlikely to happen based on what we know so far. However if two million people get it, that will kill as many as the flu and produce ~200,000 ICU cases. That's far beyond what the US healthcare system can handle on top of routine circumstances. Along with a drastically higher mortality rate, Covid also produces a far higher rate of ICU outcomes than flu does. It swamps healthcare systems, as we saw so dramatically in China.
- dougmwne 7y ago
  In an exponentially increasing epidemic, there are 2 factors at work. One, like you say, is that there is an unknown number of undetected cases. The other is that there is a lag time between detection and resolution of up to 6 weeks. Comparing the known cases today to the resolved fatalities from several weeks ago doesn't work. The undetected case issue can be addressed by intensively studying smaller populations and extrapolating to the general population. The Diamond Princess or other cases where every person in a defined population was tested and monitored would be a good place to start. The exponential growth issue can be solved with cohort analysis, group people by the date of their onset of symptoms and track the resolution for each group. With a few more weeks of data we should start to see these studies, then we'll know just how serious this is.
SkyPuncher 7y ago
10% of patients require ICU with mechanical ventilators. Yes, it skews to the older population, but that doesn't exclude younger people entirely. The biggest challenge is we only have ~20 mechanical ventilators per 100k people[0]. That means our entire nation's capacity for mechanical ventiation (excluding overload from regional clusters) is consumed with just 200 infections per 100k - or 660k infections of the US's 330 million. That doesn't account for baseline ICU/vent usage, which is estimated at at ~70% occupancy rate with 3 in 10 currently needing mechanical vents [1]. A fuzzy summary is 500k infections in the use will utilize all existing mechanical ventilators. [0] - https://www.ncbi.nlm.nih.gov/pubmed/21149215 https://www.ncbi.nlm.nih.gov/pubmed/21149215 - this was 2010, but the most recent data I could easily find. [1] - https://www.ncbi.nlm.nih.gov/pubmed/23963122 https://www.ncbi.nlm.nih.gov/pubmed/23963122
- DennisP 7y ago
  That's been worrying me. If we swamp our treatment infrastructure, the death rate could go up significantly.
- codeulike 7y ago
  South Korea have been testing more people than anyone else - 10,000 per day, drive-in test centres, results in two hours. If you look at the South Korea numbers here (scroll down a bit) https://www.worldometers.info/coronavirus/ https://www.worldometers.info/coronavirus/ you see 43 deaths and 52 'serious/critical' cases from 6,593 confirmed cases, so about 0.75% for both categories. I'm hoping these South Korea numbers are giving us a more accurate picture of whats going on.
  - dougmwne 7y ago
    I am also waiting to see a study on the final serious/fatal case rate on the Diamond Princess. Since that was a heavily monitored population given a high standard of care it should be a good indicator of what western countries are in for. Since the ages of the passengers will be known, they can age-weight the statistics to better compare to the general population. 706 infected passengers is a reasonable sample size.
tunesmith 7y ago
I think people just have to let go of the sense that these numbers are anything other than a snapshot in time. It's not like something gets written down in a book somewhere that the great Covid-19 of 2020 had a case fatality rate of 3.4 / 5.7 / whatever. It's just a ratio between a numerator and a denominator, and both of those numbers are changing at different rates as time passes, as testing increases, as containment practices change, etc.

vonmoltke 7y ago

This is only true if you assume the unresolved cases will resolve in the same ratio as the currently-resolved cases. This ratio has been dropping as more cases are diagnosed.

Alupis 7y ago

The problem here is, from what doctors seem to be saying, there's a massive amount of the population that has or has had COVID-19 and will never know - since the symptoms in an otherwise healthy adult are similar to the common Flu. So that death rate is likely far, far lower than what it seems. The flu kills between 15-60k people in the US every year, and it's around 1% fatality rate (and nobody seems overly concerned about that).

DennisP 7y ago
That's what I mean by the "mild cases that are never diagnosed."
- Alupis 7y ago
  Well sure, but we could say that exact thing about the common flu too. If you're sick enough to go to the doctor and be diagnosed with the flu, there's likely an increased chance of it being fatal (in that your symptoms are extreme or you're already high-risk, ie. elderly or in poor health). Otherwise, like most folks, you just get better on your own in a few days. With the flu killing 10's of thousands every year, and nobody bats an eye... I'm not particularly concerned about COVID-19. It's mostly a media PR stunt at this point - designed to whip up fear and clicks. Swine Flu anyone?
  - wahern 7y ago
    A case fatality rate (CFR) is calculated based on confirmed cases. The CFR for the flu[1] or any other disease is roughly based on the number of people who present to the hospital and are diagnosed, with or without serological testing. Therefore, in places with drive-by testing the CFR will be artificially lower. Likewise, using total infected as the denominator makes it a totally useless figure for comparison. Knowing that, it seems pretty clear that COVID-19 is more-or-less 10x deadlier[2] than seasonal influenza, as the CFR for seasonal flu is ~0.1-0.2%. [1] It's a little more complicated with the flu because of underreporting. I guess elderly pneumonia is common enough that doctors don't always care why it developed. So apparently the CDC uses a more complex model (e.g. total number of pneumonia cases minus base rate, non-flu season cases). But what they're trying to estimate is the CFR, which is by definition deaths/confirmed. [2] At least from an epidemiological perspective, regardless of the percentage of symptomatic cases / infections. At the end of the day what matters is how many will die, not how scared someone should be if they're infected.
    - Alupis 7y ago
      > A case fatality rate (CFR) is calculated based on confirmed cases. This is not true[1]. CDC projects the numbers and estimates total number of cases, because almost all of the actual cases are undiagnosed. So no, COVID-19 is not "more-or-less 10x deadlier" than the common Flu. Let's be clear. The media frenzy around COVID-19 is not grounded in science, and it purely designed to get panic people, resulting in clicks and eye-balls. Nothing more. [1] https://www.cdc.gov/flu/about/burden/how-cdc-estimates.htm https://www.cdc.gov/flu/about/burden/how-cdc-estimates.htm
codeulike 7y ago
South Korea have been testing more people than anyone else - 10,000 per day, drive-in test centres, results in two hours. If you look at the South Korea numbers here (scroll down a bit) https://www.worldometers.info/coronavirus/ https://www.worldometers.info/coronavirus/ you see 43 deaths and 52 'serious/critical' cases from 6,593 confirmed cases, so about 0.75% for both categories. I'm hoping these South Korea numbers are giving us a more accurate picture of whats going on.
- Alupis 7y ago
  I'm fairly confident if we started testing everyone with Flu-like symptoms, we'd show numbers even lower than the 0.75% in South Korea. It's just not as deadly as was initially reported.
pieix 7y ago
Per CDC numbers [0], in the US the death rate of the flu for the 2018-19 season was roughly 0.1%, an order of magnitude lower than the number quoted above. [0] https://www.cdc.gov/flu/about/burden/past-seasons.html https://www.cdc.gov/flu/about/burden/past-seasons.html
- Alupis 7y ago
  That's interesting! Thanks for the link. It only proves my point further. From your link, 34,200 people died from the Flu in the US last year - and not a single person is panicked about those numbers.
  - dougmwne 7y ago
    I think this is the difference between velocity and acceleration. Flu deaths are known and well studied. Our hospital system has the capacity to treat serious cases. There are existing public health initiatives around vaccination and education. The resources have been allotted and accounted for and flu deaths aren't expect to suddenly 10x or 20x. This virus is an unknown. It is spreading worldwide very quickly and the growth rate is exponential. Within months, hundreds of millions could be infected without interventions. Today the number of dead is very small. If the virus were to become as common as the flu, it could make infectious disease the leading cause of death, more than heart disease or cancer. We still don't know what the long term health complications are or if you can be re-infected. The science will be in shortly, but in the mean time an abundance of caution would seem to be warranted. I will be washing my hands until then.
  - adventured 7y ago
    > and not a single person is panicked about those numbers. That's for a very obvious reason. If 40 million people in the US get Covid, as with the flu, it will probably kill between 500k and 1m people. That's unlikely to happen based on what we know so far. However if two million people get it, that will kill as many as the flu and produce ~200,000 ICU cases. That's far beyond what the US healthcare system can handle on top of routine circumstances. Along with a drastically higher mortality rate, Covid also produces a far higher rate of ICU outcomes than flu does. It swamps healthcare systems, as we saw so dramatically in China.
dougmwne 7y ago
In an exponentially increasing epidemic, there are 2 factors at work. One, like you say, is that there is an unknown number of undetected cases. The other is that there is a lag time between detection and resolution of up to 6 weeks. Comparing the known cases today to the resolved fatalities from several weeks ago doesn't work. The undetected case issue can be addressed by intensively studying smaller populations and extrapolating to the general population. The Diamond Princess or other cases where every person in a defined population was tested and monitored would be a good place to start. The exponential growth issue can be solved with cohort analysis, group people by the date of their onset of symptoms and track the resolution for each group. With a few more weeks of data we should start to see these studies, then we'll know just how serious this is.

SkyPuncher 7y ago

10% of patients require ICU with mechanical ventilators. Yes, it skews to the older population, but that doesn't exclude younger people entirely. The biggest challenge is we only have ~20 mechanical ventilators per 100k people[0]. That means our entire nation's capacity for mechanical ventiation (excluding overload from regional clusters) is consumed with just 200 infections per 100k - or 660k infections of the US's 330 million. That doesn't account for baseline ICU/vent usage, which is estimated at at ~70% occupancy rate with 3 in 10 currently needing mechanical vents [1]. A fuzzy summary is 500k infections in the use will utilize all existing mechanical ventilators. [0] - https://www.ncbi.nlm.nih.gov/pubmed/21149215 https://www.ncbi.nlm.nih.gov/pubmed/21149215 - this was 2010, but the most recent data I could easily find. [1] - https://www.ncbi.nlm.nih.gov/pubmed/23963122 https://www.ncbi.nlm.nih.gov/pubmed/23963122

DennisP 7y ago
That's been worrying me. If we swamp our treatment infrastructure, the death rate could go up significantly.
codeulike 7y ago
South Korea have been testing more people than anyone else - 10,000 per day, drive-in test centres, results in two hours. If you look at the South Korea numbers here (scroll down a bit) https://www.worldometers.info/coronavirus/ https://www.worldometers.info/coronavirus/ you see 43 deaths and 52 'serious/critical' cases from 6,593 confirmed cases, so about 0.75% for both categories. I'm hoping these South Korea numbers are giving us a more accurate picture of whats going on.
- dougmwne 7y ago
  I am also waiting to see a study on the final serious/fatal case rate on the Diamond Princess. Since that was a heavily monitored population given a high standard of care it should be a good indicator of what western countries are in for. Since the ages of the passengers will be known, they can age-weight the statistics to better compare to the general population. 706 infected passengers is a reasonable sample size.

tunesmith 7y ago

I think people just have to let go of the sense that these numbers are anything other than a snapshot in time. It's not like something gets written down in a book somewhere that the great Covid-19 of 2020 had a case fatality rate of 3.4 / 5.7 / whatever. It's just a ratio between a numerator and a denominator, and both of those numbers are changing at different rates as time passes, as testing increases, as containment practices change, etc.