5 ms·

> But isn't that what you just did here? No, because I'm not drowning out the discussion, simply pushing back on the usage of hyperbole for pushing an opinion.

> But isn't that what you just did here? No, because I'm not drowning out the discussion, simply pushing back on the usage of hyperbole for pushing an opinion. > The last stats I had were that 5% of people who catch it will die without a ventilator, Your stats are way off and I suggest doing better research. They also don't factor in the idea that the virus is far more widespread than initially thought, and that these death rates only account for people that are sick enough to get tested in the first place.

gwd 7y ago
> Your stats are way off and I suggest doing better research. I've been reading [1] daily more or less since early February, as well as lots of things posted here. I'm certainly no professional, but I think I'm reasonably well-informed for a layperson. If you think there's something I've missed, I think you'd do much better to actually point me to a resource you think is better. I realize that "number reported" is an issue, and varies a lot from country to country. But consider Germany, who have [EDIT: ramped up to 400k+ tests per week] now. The most recent numbers from the situation report says that Germany has 125k reported cases, with 3k reported deaths -- a death rate of 2.3%. EDIT: Also consider Korea, which is alleged to have instituted a massive testing infrastructure as well. Their numbers are currently 10,564 cases, 222 deaths -- again, 2.1% death rate. It's possible that both the German and Korean numbers of cases are massively underreported, but it can't be by much. That puts the death rate with non-overwhelmed hospitals at significantly higher than 0.5%. If you want to convince me that the death rate for overwhelmed hospitals is less than 5%, you're going to have to do better than waving your your hands and saying "but what about people not sick enough get tested". EDIT2: Been challenged on the "500k tests per day", acknowledging that while I look for support for that claim. EDIT3: I slightly mangled my numbers from [2]. The numbers from that are that Germany has been doing 160K tests per week since 20 March; and that the government's goal is 200k per day. [2] According to [3], on the week ending April 5, Germany did 393k tests. [1] https://www.who.int/emergencies/diseases/novel-coronavirus-2019/situation-reports https://www.who.int/emergencies/diseases/novel-coronavirus-2... [2] https://www.spectator.co.uk/article/how-germany-has-managed-to-perform-so-many-covid-19-tests https://www.spectator.co.uk/article/how-germany-has-managed-... [3] https://ourworldindata.org/covid-testing#germany https://ourworldindata.org/covid-testing#germany
- ceilingcorner 7y ago
  Do you have a link for this 500,000 tests per day? Every article I am reading says they only started ramping up testing in the past few weeks. > And the total number of swab tests done by 4 April was well over 1.3 million. https://www.bbc.com/news/health-52234061 https://www.bbc.com/news/health-52234061 The population of Germany is 83 million. I find it hard to believe that they've tested more than 5 million or so people in 10 days.
  - gwd 7y ago
    So I did mangle my numbers somewhat, thanks for checking. As of April 5th they'd done 1.3 million tests, found 95k cases, and had 1.4k deaths [1]. Presumably there are at least some people tested multiple times; so those 1.3 million tests don't represent 1.3 million people. The only way to get the "die with the best medical care" number below 0.1% is if you have a million untested people who are sick. If I was a proper statistician, with those numbers I could probably come up with models or estimates of how many possibly untested people there might be in Germany. My instinct is that number can't be millions, because at least some of those millions would be sick enough to go to the hospital, and then all of their connections would be tested and be found out. But even then, what we really need is the "would have died without medical care" number, which is what the number would look like if we did a "fast burn" and let everyone get sick at once. For that number to be less than 1% (which would still be pretty dire), we need closer to 2 million unknown cases. I just don't see how 2 million positive cases can hide from the kind of testing Germany is doing. The same goes for Korea. [1] https://www.who.int/docs/default-source/coronaviruse/situation-reports/20200406-sitrep-77-covid-19.pdf?sfvrsn=21d1e632_2 https://www.who.int/docs/default-source/coronaviruse/situati...
- klmadfejno 7y ago
  Your mental model is mostly meaningless. There's no such thing as a generic death rate for a disease, because the death rate is dependent on numerous factors. You cannot compare the death rate between countries with different demographic profiles and ask "which is correct". How old are they? How fat are they? The idea that 5% of a person's social network is going to die is not accurate because social networks aren't evenly distributed. Most people don't have many contacts over 80 unless they're over 80 themselves. Most people will see very few deaths to people they're connected to. Old people will see a lot. And it's sad, but most of the people dying are already above life expectancy and close to death from alternative causes. Hospitals being overwhelmed is an exaggerated problem. It's not happening in many places. Most hospitals only need to provide a small amount of care and can actually scale this up rather quickly. You hear about people being triaged in Italy, but you need to remember that the people being triaged are likely to die regardless of whether or not they're admitted to the hospital. Averages are not your friend here. If 50% of ICU patients survive, that does not mean that everyone going into the ICU has a 50% chance of surviving. If we get to a point that triaging is needed (I think dubious in major US cities), that does not mean fatality rates will skyrocket. Social distancing has been pretty darn effective, and it's probably a good idea to cautiously continue. But we also need to be pragmatic.
gwd 7y ago
All models are wrong, but 1) some models are useful, and 2) you can't not have a model. If you don't have an explicit model, you have an implicit model. I presented you a model; yes it's oversimplified, but it results in at least some sort of concrete answer; and you can actually examine where it came from and reason about how valid it is. You can make arguments for some other death rate or herd immunity rate. (I took 80% from the most recent article I heard on the subject, but previously I heard 60%.) You also actually have a model in your head for how many people are going to die (or not die). But I don't know where it came from. I can't examine it or reason about it. If you were an old-timer who'd seen hundreds of pandemics, or a historical epidemiologist who had studied hundreds of epidemics, maybe your "gut feel" would carry some weight. But since (AFAICT) you're a layperson like me, I'm not inclined to think your implicit model is going to be very accurate.
- gwd 7y ago
  Hmm, this one seems unpopular for some reason. I thought it was useful and polite. Anyone want to give me a hint?

gwd 7y ago

> Your stats are way off and I suggest doing better research. I've been reading [1] daily more or less since early February, as well as lots of things posted here. I'm certainly no professional, but I think I'm reasonably well-informed for a layperson. If you think there's something I've missed, I think you'd do much better to actually point me to a resource you think is better. I realize that "number reported" is an issue, and varies a lot from country to country. But consider Germany, who have [EDIT: ramped up to 400k+ tests per week] now. The most recent numbers from the situation report says that Germany has 125k reported cases, with 3k reported deaths -- a death rate of 2.3%. EDIT: Also consider Korea, which is alleged to have instituted a massive testing infrastructure as well. Their numbers are currently 10,564 cases, 222 deaths -- again, 2.1% death rate. It's possible that both the German and Korean numbers of cases are massively underreported, but it can't be by much. That puts the death rate with non-overwhelmed hospitals at significantly higher than 0.5%. If you want to convince me that the death rate for overwhelmed hospitals is less than 5%, you're going to have to do better than waving your your hands and saying "but what about people not sick enough get tested". EDIT2: Been challenged on the "500k tests per day", acknowledging that while I look for support for that claim. EDIT3: I slightly mangled my numbers from [2]. The numbers from that are that Germany has been doing 160K tests per week since 20 March; and that the government's goal is 200k per day. [2] According to [3], on the week ending April 5, Germany did 393k tests. [1] https://www.who.int/emergencies/diseases/novel-coronavirus-2019/situation-reports https://www.who.int/emergencies/diseases/novel-coronavirus-2... [2] https://www.spectator.co.uk/article/how-germany-has-managed-to-perform-so-many-covid-19-tests https://www.spectator.co.uk/article/how-germany-has-managed-... [3] https://ourworldindata.org/covid-testing#germany https://ourworldindata.org/covid-testing#germany

ceilingcorner 7y ago
Do you have a link for this 500,000 tests per day? Every article I am reading says they only started ramping up testing in the past few weeks. > And the total number of swab tests done by 4 April was well over 1.3 million. https://www.bbc.com/news/health-52234061 https://www.bbc.com/news/health-52234061 The population of Germany is 83 million. I find it hard to believe that they've tested more than 5 million or so people in 10 days.
- gwd 7y ago
  So I did mangle my numbers somewhat, thanks for checking. As of April 5th they'd done 1.3 million tests, found 95k cases, and had 1.4k deaths [1]. Presumably there are at least some people tested multiple times; so those 1.3 million tests don't represent 1.3 million people. The only way to get the "die with the best medical care" number below 0.1% is if you have a million untested people who are sick. If I was a proper statistician, with those numbers I could probably come up with models or estimates of how many possibly untested people there might be in Germany. My instinct is that number can't be millions, because at least some of those millions would be sick enough to go to the hospital, and then all of their connections would be tested and be found out. But even then, what we really need is the "would have died without medical care" number, which is what the number would look like if we did a "fast burn" and let everyone get sick at once. For that number to be less than 1% (which would still be pretty dire), we need closer to 2 million unknown cases. I just don't see how 2 million positive cases can hide from the kind of testing Germany is doing. The same goes for Korea. [1] https://www.who.int/docs/default-source/coronaviruse/situation-reports/20200406-sitrep-77-covid-19.pdf?sfvrsn=21d1e632_2 https://www.who.int/docs/default-source/coronaviruse/situati...
klmadfejno 7y ago
Your mental model is mostly meaningless. There's no such thing as a generic death rate for a disease, because the death rate is dependent on numerous factors. You cannot compare the death rate between countries with different demographic profiles and ask "which is correct". How old are they? How fat are they? The idea that 5% of a person's social network is going to die is not accurate because social networks aren't evenly distributed. Most people don't have many contacts over 80 unless they're over 80 themselves. Most people will see very few deaths to people they're connected to. Old people will see a lot. And it's sad, but most of the people dying are already above life expectancy and close to death from alternative causes. Hospitals being overwhelmed is an exaggerated problem. It's not happening in many places. Most hospitals only need to provide a small amount of care and can actually scale this up rather quickly. You hear about people being triaged in Italy, but you need to remember that the people being triaged are likely to die regardless of whether or not they're admitted to the hospital. Averages are not your friend here. If 50% of ICU patients survive, that does not mean that everyone going into the ICU has a 50% chance of surviving. If we get to a point that triaging is needed (I think dubious in major US cities), that does not mean fatality rates will skyrocket. Social distancing has been pretty darn effective, and it's probably a good idea to cautiously continue. But we also need to be pragmatic.

gwd 7y ago

All models are wrong, but 1) some models are useful, and 2) you can't not have a model. If you don't have an explicit model, you have an implicit model. I presented you a model; yes it's oversimplified, but it results in at least some sort of concrete answer; and you can actually examine where it came from and reason about how valid it is. You can make arguments for some other death rate or herd immunity rate. (I took 80% from the most recent article I heard on the subject, but previously I heard 60%.) You also actually have a model in your head for how many people are going to die (or not die). But I don't know where it came from. I can't examine it or reason about it. If you were an old-timer who'd seen hundreds of pandemics, or a historical epidemiologist who had studied hundreds of epidemics, maybe your "gut feel" would carry some weight. But since (AFAICT) you're a layperson like me, I'm not inclined to think your implicit model is going to be very accurate.

gwd 7y ago
Hmm, this one seems unpopular for some reason. I thought it was useful and polite. Anyone want to give me a hint?