4 ms·

The US is at 121K deaths atm. I made an estimate a while back assuming 60% of the population becomes infected and came up with 350K deaths. Bad, but not catas

The US is at 121K deaths atm. I made an estimate a while back assuming 60% of the population becomes infected and came up with 350K deaths. Bad, but not catastrophic. And I don't think we will get that high.

tunesmith 6y ago
60%/350k assumes an R0 of less than 2.5 (2.5 is to get to 60% herd immunity threshold, and then more infections continue to happen as it dies out), and also an IFR of less than 0.2%. I think those are outside of the range of any recent estimates I've seen. Natural R0 is generally believed to be higher, and IFR is believed to be in the 0.5% - 1% range. I also hope we won't hit 350k, but if we stay below that number, it will be because of a combination of continued lockdowns and finding an effective treatment/vaccine before we reach herd immunity.
- tripletao 6y ago
  The 1-1/R0 assumes a homogeneous and well-mixed population, and thus overestimates the fraction infected required for herd immunity. For real, people with lots of contacts will get infected and become immune first, with disproportionate benefit. That effect is hard to quantify, but it's near-certainly big, with many papers speculating that herd immunity may occur around 30% recovered: https://www.medrxiv.org/content/10.1101/2020.02.10.20021725v2.full.pdf https://www.medrxiv.org/content/10.1101/2020.02.10.20021725v... This is before considering overshoot, which can occur just as in the homogeneous case (and is a reason to slow the spread as much as reasonable, even if you expect/fear that the end result will be herd immunity from recovered patients). Reported IFRs are now calculated from serostudies testing for IgG. The best studies are on the high end of your 0.5-1%, but there's reason to suspect they're significant overestimates. Sensitivity for the blood tests was established from cases severe enough to seek medical help, and may be much lower for mild cases. There's also evidence of mucosal immunity in patients who test negative for IgG, and recently of T cell response in such patients too: https://www.medrxiv.org/content/10.1101/2020.06.21.20132449v1.full.pdf https://www.medrxiv.org/content/10.1101/2020.06.21.20132449v... And the number of coronavirus deaths is greater than the actual number of excess deaths, since they're disproportionately elderly. All these second-order effects are uncertain, and difficult to incorporate into a final number; but they're real, and to ignore them completely risks a gross overestimate. Maybe people know that, and think of the overestimate as "conservative"; but considering the tremendous costs of both the coronavirus itself and society's response, I strongly disagree. Finally, since you said "lockdowns"--if you credit lockdowns (which are always vaguely defined; but government-imposed restrictions on movement and assembly, broadly) with all of benefit to the USA, how do you explain Japan? They're sitting at 8 deaths per million, they never significantly locked down, and their contact tracers are missing ~80% of cases. Maybe it's the masks? Even in California, use is waning. No one seems to have the political appetite to enforce the order coercively; NYC just stopped enforcing theirs entirely after an incident. Maybe it's the contact tracing anyways, since missing 80% of cases might be fine if you catch the 20% who are super-spreaders (which you will disproportionately, since each person they infect is a chance to find them tracing back)? I could imagine big public campaigns to explain the importance of cooperating with contact tracers, or websites to help people anonymously advise their contacts (which have been tried for STDs, which have at least as much stigma). But I've seen almost nothing. To be clear, I agree that restrictions on movement and assembly were and are a good idea in some cases, and New York and Lombardy surely wish they'd locked down earlier. I just don't like seeing them discussed as if they're responsible for all the difference between our current situation and a first-order model that we know is significantly wrong.
  - tunesmith 6y ago
    Therefore what? Are you arguing in favor of ending all lockdown-ish behavior, letting the chips fall where they may, and relying on herd immunity to save us? Because that's what I'm arguing against. I feel like I just read a template.
    - tripletao 6y ago
      In my last paragraph, I said explicitly that I supported lockdown-type interventions in some cases. Further, even if the end result is herd immunity from recovered cases, slowing the spread will limit overshoot--so support for lockdowns and expectation the pandemic will end in herd immunity from recovered cases are not mutually exclusive. So my suspicion is that you've skimmed my (long) comment, assigned it to some "template" that you've seen before, and made quick and inaccurate assumptions about my beliefs instead of reading what I wrote. I don't see how you could otherwise believe in good faith that I was "arguing in favor of ending all lockdown-ish behavior, letting the chips fall where they may, and relying on herd immunity to save us". To start, I'm arguing that the first-order death calculations using 1-1/R0 and IFR from IgG are probably significant overestimates. Do you disagree? If yes, why do you think heterogeneity and evidence other than IgG of prior infection are negligible? And again, how do you explain Japan? They found a way to continue mostly operating society, while American schoolchildren without attentive parents fall behind their classmates a day at a time. Wouldn't we be better off trying to understand and emulate their success instead of arguing about "lockdowns"?
      - tunesmith 6y ago
        I don't believe you are "arguing in favor of ending all lockdown-ish behavior, letting the chips fall where they may, and relying on herd immunity to save us". My point is, if you aren't, then why are you acting like you are disagreeing with me? I've seen you ask about Japan in other threads, using much of the same words, without the various counterpoints seeming to register to you, and now you're making the same points here. It just seems weird. Why are you bringing Japan up as if it's some sort of counterpoint, when it's a counterpoint to something that you apparently believe in? Why are you bringing it up if you aren't in favor of ending mitigation and aren't in favor of acting more like... Japan? I don't care if (R0-1)/R0 is off by some negligible percentage - it would still mean there's a huge difference in death depending on whether we do or don't stop mitigation. I'm arguing against the people that don't even seem to have a rough mathematical sense about this, and also the people that seem content to be like Thanos and shrug their shoulders at a million, 350k, or 200k people in the US dying as long as it's not their friends. I mean, back to your first comment, if you think herd immunity maxes out at 30%, and that IFR of 0.5 - 1% is a "significant overestimate", then you're pretty close to projecting our current death toll anyway. 0.25% IFR at 30% would be like 250k in the US, which we'll probably hit with our half-hearted mitigation. But since you're in favor of mitigation efforts, you clearly don't believe the estimates to be that low. So... again, what is the relevance of what you're saying? Are you just being sophist? What's the syllogistic connective tissue in your points? What's your therefore? You believe the R0 and IFR estimates are overestimates; therefore what? Why do I need to explain Japan? I mean obviously something must be happening in Japan that is consistent with what is happening elsewhere, it's not like they are an alternate universe. It's not like Japan is a counterpoint to NYC or Lombardy; they're in the same reality somehow. How do you explain Japan? It's probably some mixture of culture and bad stats in a way that doesn't really have any bearing on what we should do - we already have sufficient data to know that mitigation saves lives, and we're stupid enough that we're about to learn it all over again for the first time over the next six weeks.
jhpriestley 6y ago
Amazing, NYC has already reached twice that death rate so they must have had 120% of their population infected
greedo 6y ago
60% of 330M is 198M. 10% will be hospitalized = 19.8M 10% of those hospitalized will die = 1.98M Even if the IFR isn't 1%, say it's half that, almost 1M deaths. And if you think 350k surplus deaths isn't catastrophic, you must have been nodding your head while watching Chernobyl (3.6 Roentgen? Not great, not terrible).

tunesmith 6y ago

60%/350k assumes an R0 of less than 2.5 (2.5 is to get to 60% herd immunity threshold, and then more infections continue to happen as it dies out), and also an IFR of less than 0.2%. I think those are outside of the range of any recent estimates I've seen. Natural R0 is generally believed to be higher, and IFR is believed to be in the 0.5% - 1% range. I also hope we won't hit 350k, but if we stay below that number, it will be because of a combination of continued lockdowns and finding an effective treatment/vaccine before we reach herd immunity.

tripletao 6y ago
The 1-1/R0 assumes a homogeneous and well-mixed population, and thus overestimates the fraction infected required for herd immunity. For real, people with lots of contacts will get infected and become immune first, with disproportionate benefit. That effect is hard to quantify, but it's near-certainly big, with many papers speculating that herd immunity may occur around 30% recovered: https://www.medrxiv.org/content/10.1101/2020.02.10.20021725v2.full.pdf https://www.medrxiv.org/content/10.1101/2020.02.10.20021725v... This is before considering overshoot, which can occur just as in the homogeneous case (and is a reason to slow the spread as much as reasonable, even if you expect/fear that the end result will be herd immunity from recovered patients). Reported IFRs are now calculated from serostudies testing for IgG. The best studies are on the high end of your 0.5-1%, but there's reason to suspect they're significant overestimates. Sensitivity for the blood tests was established from cases severe enough to seek medical help, and may be much lower for mild cases. There's also evidence of mucosal immunity in patients who test negative for IgG, and recently of T cell response in such patients too: https://www.medrxiv.org/content/10.1101/2020.06.21.20132449v1.full.pdf https://www.medrxiv.org/content/10.1101/2020.06.21.20132449v... And the number of coronavirus deaths is greater than the actual number of excess deaths, since they're disproportionately elderly. All these second-order effects are uncertain, and difficult to incorporate into a final number; but they're real, and to ignore them completely risks a gross overestimate. Maybe people know that, and think of the overestimate as "conservative"; but considering the tremendous costs of both the coronavirus itself and society's response, I strongly disagree. Finally, since you said "lockdowns"--if you credit lockdowns (which are always vaguely defined; but government-imposed restrictions on movement and assembly, broadly) with all of benefit to the USA, how do you explain Japan? They're sitting at 8 deaths per million, they never significantly locked down, and their contact tracers are missing ~80% of cases. Maybe it's the masks? Even in California, use is waning. No one seems to have the political appetite to enforce the order coercively; NYC just stopped enforcing theirs entirely after an incident. Maybe it's the contact tracing anyways, since missing 80% of cases might be fine if you catch the 20% who are super-spreaders (which you will disproportionately, since each person they infect is a chance to find them tracing back)? I could imagine big public campaigns to explain the importance of cooperating with contact tracers, or websites to help people anonymously advise their contacts (which have been tried for STDs, which have at least as much stigma). But I've seen almost nothing. To be clear, I agree that restrictions on movement and assembly were and are a good idea in some cases, and New York and Lombardy surely wish they'd locked down earlier. I just don't like seeing them discussed as if they're responsible for all the difference between our current situation and a first-order model that we know is significantly wrong.
- tunesmith 6y ago
  Therefore what? Are you arguing in favor of ending all lockdown-ish behavior, letting the chips fall where they may, and relying on herd immunity to save us? Because that's what I'm arguing against. I feel like I just read a template.
  - tripletao 6y ago
    In my last paragraph, I said explicitly that I supported lockdown-type interventions in some cases. Further, even if the end result is herd immunity from recovered cases, slowing the spread will limit overshoot--so support for lockdowns and expectation the pandemic will end in herd immunity from recovered cases are not mutually exclusive. So my suspicion is that you've skimmed my (long) comment, assigned it to some "template" that you've seen before, and made quick and inaccurate assumptions about my beliefs instead of reading what I wrote. I don't see how you could otherwise believe in good faith that I was "arguing in favor of ending all lockdown-ish behavior, letting the chips fall where they may, and relying on herd immunity to save us". To start, I'm arguing that the first-order death calculations using 1-1/R0 and IFR from IgG are probably significant overestimates. Do you disagree? If yes, why do you think heterogeneity and evidence other than IgG of prior infection are negligible? And again, how do you explain Japan? They found a way to continue mostly operating society, while American schoolchildren without attentive parents fall behind their classmates a day at a time. Wouldn't we be better off trying to understand and emulate their success instead of arguing about "lockdowns"?
    - tunesmith 6y ago
      I don't believe you are "arguing in favor of ending all lockdown-ish behavior, letting the chips fall where they may, and relying on herd immunity to save us". My point is, if you aren't, then why are you acting like you are disagreeing with me? I've seen you ask about Japan in other threads, using much of the same words, without the various counterpoints seeming to register to you, and now you're making the same points here. It just seems weird. Why are you bringing Japan up as if it's some sort of counterpoint, when it's a counterpoint to something that you apparently believe in? Why are you bringing it up if you aren't in favor of ending mitigation and aren't in favor of acting more like... Japan? I don't care if (R0-1)/R0 is off by some negligible percentage - it would still mean there's a huge difference in death depending on whether we do or don't stop mitigation. I'm arguing against the people that don't even seem to have a rough mathematical sense about this, and also the people that seem content to be like Thanos and shrug their shoulders at a million, 350k, or 200k people in the US dying as long as it's not their friends. I mean, back to your first comment, if you think herd immunity maxes out at 30%, and that IFR of 0.5 - 1% is a "significant overestimate", then you're pretty close to projecting our current death toll anyway. 0.25% IFR at 30% would be like 250k in the US, which we'll probably hit with our half-hearted mitigation. But since you're in favor of mitigation efforts, you clearly don't believe the estimates to be that low. So... again, what is the relevance of what you're saying? Are you just being sophist? What's the syllogistic connective tissue in your points? What's your therefore? You believe the R0 and IFR estimates are overestimates; therefore what? Why do I need to explain Japan? I mean obviously something must be happening in Japan that is consistent with what is happening elsewhere, it's not like they are an alternate universe. It's not like Japan is a counterpoint to NYC or Lombardy; they're in the same reality somehow. How do you explain Japan? It's probably some mixture of culture and bad stats in a way that doesn't really have any bearing on what we should do - we already have sufficient data to know that mitigation saves lives, and we're stupid enough that we're about to learn it all over again for the first time over the next six weeks.

jhpriestley 6y ago

Amazing, NYC has already reached twice that death rate so they must have had 120% of their population infected

greedo 6y ago

60% of 330M is 198M. 10% will be hospitalized = 19.8M 10% of those hospitalized will die = 1.98M Even if the IFR isn't 1%, say it's half that, almost 1M deaths. And if you think 350k surplus deaths isn't catastrophic, you must have been nodding your head while watching Chernobyl (3.6 Roentgen? Not great, not terrible).