15 ms·

Yup. And for far longer.

gnulinux 7y ago
It seems to indicate this disease infected a lot more than what we see, and maybe mortality rate is lower than what we currently assume. Maybe it's around 0.1% to 0.5%. So is it possible this disease appeared in China way before December, maybe a few months back. Then in December it reached a saturation point (i.e. infected hundreds of thousands of people, so deaths started to get noticed by respiratory specialists). Then around January/February this disease spread to the world, but it went unnoticed. Then in March it reached the same saturation point that the deaths became too many not to notice. Is this a realistic scenario?
- raphlinus 7y ago
  No. We know from genomic analysis [1] that "The common ancestor of circulating viruses appears to have emerged in Wuhan, China, in late Nov or early Dec 2019." It's fun to speculate about other possibilities, but I personally like to stick to what the science tells us. [1]: https://twitter.com/nextstrain/status/1248707010750640128 https://twitter.com/nextstrain/status/1248707010750640128
  - gnulinux 7y ago
    Very interesting, thanks for the link. In that case, how should that change our mental model? It seems like more people in Western countries are infected than visible, doesn't that mean the same would apply to China? How would it spread so fast in December?
    - Retric 7y ago
      US is generally not testing people with mild symptoms to significant symptoms. We are mostly telling people to stay home until you have difficulty breathing. That’s why are deaths to cases ratio is so high before the new infection rate has dropped to near zero for weeks. Adjusting for infection vs death curve in China US deaths from people currently infected are likely to hit in the ballpark of 50k to 100k. Assuming we get new infections under control in the next few days and keep it low for the next month.
  - autokad 7y ago
    well, google coronavirus searches in China, specifically Hubei province, spiked in September of 2019. Science is a term being twisted out of its original context. if people keep doing that, its going to be a pejorative.
    - gnulinux 7y ago
      How is Google searches relevant here? Was the news of possible coronovirus outbreak leaked to the public? Also, do you have a source?
    - IAmEveryone 7y ago
      Even if I were inclined to believe such statements without a link, this doesn’t make any sense. For Google searches to spike, information about the disease would need to be public knowledge. And at that point, you wouldn’t need Google search traffic to prove anything. Also: Google is blocked in China, isn’t it?
    - bb2018 7y ago
      Do you have a link to that? I'd love to know more about why this would be the case.
      - raphlinus 7y ago
        Here's my attempt to validate the claim: https://trends.google.com/trends/explore?date=2019-01-01%202019-12-31&geo=CN&q=%2Fm%2F01cpyy https://trends.google.com/trends/explore?date=2019-01-01%202... I'll leave it for readers to determine whether it holds up.
        gnulinux 7y ago
        So there was 0 searches, then ~90 searches suddenly? Yeah no sounds like just noise. There will always be some people searching about corona viruses, they always existed and we always knew of them.
    - FrojoS 7y ago
      Source?
  - gowld 7y ago
    That report appears to based on genomic analysis, which obviously only includes cases that were tested, not undocumented transmissions before the discovery. The article you cite says: "The common ancestor of circulating viruses appears to have emerged in Wuhan, China, in late Nov or early Dec 2019. Accordingly, the majority of sampled and inferred ancestral cases were located in Asia during this early period. "
    - raphlinus 7y ago
      It samples cases that are detected after the fact, from which they are able to reconstruct the phylogeny with impressive clarity. If you are not familiar with the work of nextstrain.org, I highly recommend it, they are brilliant.
- Retric 7y ago
  No for several reasons. The Dimond Princess had over 100 crew of young healthy adults test positive and only 25% where asymptotic long term. Additionally South Korea has a massive effort into contact traction which demonstrated just how rapidly the disease spread. Their mortality numbers looked to be about 0.6% early on, but eventually hit 2% a month after the peak with people spending weeks in critical condition. On top of that it’s spread rapidly across every country outside of China and there uncontrolled growth rate showed a similar doubling every few days.
  - gnulinux 7y ago
    When we say long term here, do we mean within the 14 day incubation period that's been going around? Or does this imply incubation period is likely a lot more than 14 days, maybe in the order of months?
    - Retric 7y ago
      Within 14 days, but most people where evacuated to different countries almost immediately after infection. So, it took longer to gather the data on that population.
  - w0utert 7y ago
    >> Their mortality numbers looked to be about 0.6% early on, but eventually hit 2% a month after the peak with people spending weeks in critical condition. I see this repeated often but this is a misrepresentation of the statistics. Case fatality rate (CFR) is not the same as mortality rate. The CFR is the ratio between deaths and confirmed cases, mortality rate is the ratio between deaths and total number of people infected. The former number is 2% in South Korea now, the latter can not be measured directly without testing 100% of the population continually during the entire timespan of the outbreak, but it is (practically speaking) by definition significantly lower than 2%, unless there is a huge number of unaccounted Covid-19 deaths, which is very unlikely.
    - dh37 7y ago
      Agreed, although a nitpick: I think what you call "mortality rate" is called "infection mortality rate". (Totally agree with your important point that IFR < CFR.)
- sjg007 7y ago
  No.
- redis_mlc 7y ago
  Hospital Patient Zero has been traced back to Dec. 1. Given that most ill people don't get lung complications, that likely means Nov. was the real patient zero. This is close to 100% certain. It should be verified by looking at Chinese sites.
pm_me_ur_fullz 7y ago
I really don't understand why people immediately want to discredit the idea that US patient zero was not the first in US. Or that this virus was around earlier and that there are limitations in contact tracing capabilities. This seems to have the toughest aspect of gaining consensus. Need antibody tests now, and ones that work reliably.
- sjg007 7y ago
  Because it's not true.
  - vmception 7y ago
    okay, now it is true, ready for a discussion? https://www.sfchronicle.com/bayarea/article/Exclusive-Autopsy-report-of-first-known-15226422.php https://www.sfchronicle.com/bayarea/article/Exclusive-Autops... https://www.cnn.com/2020/04/23/us/california-woman-first-coronavirus-death/index.html https://www.cnn.com/2020/04/23/us/california-woman-first-cor... tl;dr first covid death (feb 6) is nearly a month earlier than previously official first US covid death (feb 26), and also from community spread. suggesting it was in the bay area at least half a month earlier than that.
  - pm_me_ur_fullz 7y ago
    Why is that the most preposterous hypothesis that doesn't warrant any review? We have no reliable antibody tests, and then won't give the unreliable ones people because they are so rare. So where does the confidence come from that we can rule out this possibility? What if this is already the "second wave". Something with lower symptoms wouldn't have warranted checking for a new strain, complications and deaths would fit into the normal distribution of last fall's flu season with no outlier spikes. We would then be deep into the second wave and can't even test most for people currently exposed to it, much less having already been exposed. This is enough not to dismiss the hypothesis.
    - MaulingMonkey 7y ago
      > Why is that the most preposterous hypothesis that doesn't warrant any review? It's not. What is preposterous is advancing the hypothesis without any meaningful evidence, to the detrement of more useful topics, or even useful discussion of the same topic. It's one thing to be an epidemiologist saying "hey, let's see how we can double check the exact origins of this", which is a potentially useful line of inquery. It's another to take some hypochondriac's third hand retelling of someone else's flu symptoms back in December and jump to the same untested unproven unscientific fear-driven conclusions they did on little more than their wild speculation. Even if they do somehow end up being right, they'll be right in the "broken clock is right twice a day" sense rather than a bringing anything useful to the table sense. So far, alternative suggestions as to the virus's origins have looked more similar to the latter - with perhaps some undertones of (completely understandable and expected) attempted political diversion (I sure wouldn't want fault for this mess hanging on my neck!) If you have some epidemiologist's proposed study that you're trying to drive funding towards - something that even remotely looks like the former - it'd be a welcome breath of fresh air on the topic, and I'd suggest sharing that as a far more useful and constructive way of advancing the hypothesis.

It seems to indicate this disease infected a lot more than what we see, and maybe mortality rate is lower than what we currently assume. Maybe it's around 0.1% to 0.5%. So is it possible this disease appeared in China way before December, maybe a few months back. Then in December it reached a saturation point (i.e. infected hundreds of thousands of people, so deaths started to get noticed by respiratory specialists). Then around January/February this disease spread to the world, but it went unnoticed. Then in March it reached the same saturation point that the deaths became too many not to notice. Is this a realistic scenario?

raphlinus 7y ago
No. We know from genomic analysis [1] that "The common ancestor of circulating viruses appears to have emerged in Wuhan, China, in late Nov or early Dec 2019." It's fun to speculate about other possibilities, but I personally like to stick to what the science tells us. [1]: https://twitter.com/nextstrain/status/1248707010750640128 https://twitter.com/nextstrain/status/1248707010750640128
- gnulinux 7y ago
  Very interesting, thanks for the link. In that case, how should that change our mental model? It seems like more people in Western countries are infected than visible, doesn't that mean the same would apply to China? How would it spread so fast in December?
  - Retric 7y ago
    US is generally not testing people with mild symptoms to significant symptoms. We are mostly telling people to stay home until you have difficulty breathing. That’s why are deaths to cases ratio is so high before the new infection rate has dropped to near zero for weeks. Adjusting for infection vs death curve in China US deaths from people currently infected are likely to hit in the ballpark of 50k to 100k. Assuming we get new infections under control in the next few days and keep it low for the next month.
- autokad 7y ago
  well, google coronavirus searches in China, specifically Hubei province, spiked in September of 2019. Science is a term being twisted out of its original context. if people keep doing that, its going to be a pejorative.
  - gnulinux 7y ago
    How is Google searches relevant here? Was the news of possible coronovirus outbreak leaked to the public? Also, do you have a source?
  - IAmEveryone 7y ago
    Even if I were inclined to believe such statements without a link, this doesn’t make any sense. For Google searches to spike, information about the disease would need to be public knowledge. And at that point, you wouldn’t need Google search traffic to prove anything. Also: Google is blocked in China, isn’t it?
  - bb2018 7y ago
    Do you have a link to that? I'd love to know more about why this would be the case.
    - raphlinus 7y ago
      Here's my attempt to validate the claim: https://trends.google.com/trends/explore?date=2019-01-01%202019-12-31&geo=CN&q=%2Fm%2F01cpyy https://trends.google.com/trends/explore?date=2019-01-01%202... I'll leave it for readers to determine whether it holds up.
      - gnulinux 7y ago
        So there was 0 searches, then ~90 searches suddenly? Yeah no sounds like just noise. There will always be some people searching about corona viruses, they always existed and we always knew of them.
  - FrojoS 7y ago
    Source?
- gowld 7y ago
  That report appears to based on genomic analysis, which obviously only includes cases that were tested, not undocumented transmissions before the discovery. The article you cite says: "The common ancestor of circulating viruses appears to have emerged in Wuhan, China, in late Nov or early Dec 2019. Accordingly, the majority of sampled and inferred ancestral cases were located in Asia during this early period. "
  - raphlinus 7y ago
    It samples cases that are detected after the fact, from which they are able to reconstruct the phylogeny with impressive clarity. If you are not familiar with the work of nextstrain.org, I highly recommend it, they are brilliant.
Retric 7y ago
No for several reasons. The Dimond Princess had over 100 crew of young healthy adults test positive and only 25% where asymptotic long term. Additionally South Korea has a massive effort into contact traction which demonstrated just how rapidly the disease spread. Their mortality numbers looked to be about 0.6% early on, but eventually hit 2% a month after the peak with people spending weeks in critical condition. On top of that it’s spread rapidly across every country outside of China and there uncontrolled growth rate showed a similar doubling every few days.
- gnulinux 7y ago
  When we say long term here, do we mean within the 14 day incubation period that's been going around? Or does this imply incubation period is likely a lot more than 14 days, maybe in the order of months?
  - Retric 7y ago
    Within 14 days, but most people where evacuated to different countries almost immediately after infection. So, it took longer to gather the data on that population.
- w0utert 7y ago
  >> Their mortality numbers looked to be about 0.6% early on, but eventually hit 2% a month after the peak with people spending weeks in critical condition. I see this repeated often but this is a misrepresentation of the statistics. Case fatality rate (CFR) is not the same as mortality rate. The CFR is the ratio between deaths and confirmed cases, mortality rate is the ratio between deaths and total number of people infected. The former number is 2% in South Korea now, the latter can not be measured directly without testing 100% of the population continually during the entire timespan of the outbreak, but it is (practically speaking) by definition significantly lower than 2%, unless there is a huge number of unaccounted Covid-19 deaths, which is very unlikely.
  - dh37 7y ago
    Agreed, although a nitpick: I think what you call "mortality rate" is called "infection mortality rate". (Totally agree with your important point that IFR < CFR.)
sjg007 7y ago
No.
redis_mlc 7y ago
Hospital Patient Zero has been traced back to Dec. 1. Given that most ill people don't get lung complications, that likely means Nov. was the real patient zero. This is close to 100% certain. It should be verified by looking at Chinese sites.

pm_me_ur_fullz 7y ago

I really don't understand why people immediately want to discredit the idea that US patient zero was not the first in US. Or that this virus was around earlier and that there are limitations in contact tracing capabilities. This seems to have the toughest aspect of gaining consensus. Need antibody tests now, and ones that work reliably.

sjg007 7y ago
Because it's not true.
- vmception 7y ago
  okay, now it is true, ready for a discussion? https://www.sfchronicle.com/bayarea/article/Exclusive-Autopsy-report-of-first-known-15226422.php https://www.sfchronicle.com/bayarea/article/Exclusive-Autops... https://www.cnn.com/2020/04/23/us/california-woman-first-coronavirus-death/index.html https://www.cnn.com/2020/04/23/us/california-woman-first-cor... tl;dr first covid death (feb 6) is nearly a month earlier than previously official first US covid death (feb 26), and also from community spread. suggesting it was in the bay area at least half a month earlier than that.
- pm_me_ur_fullz 7y ago
  Why is that the most preposterous hypothesis that doesn't warrant any review? We have no reliable antibody tests, and then won't give the unreliable ones people because they are so rare. So where does the confidence come from that we can rule out this possibility? What if this is already the "second wave". Something with lower symptoms wouldn't have warranted checking for a new strain, complications and deaths would fit into the normal distribution of last fall's flu season with no outlier spikes. We would then be deep into the second wave and can't even test most for people currently exposed to it, much less having already been exposed. This is enough not to dismiss the hypothesis.
  - MaulingMonkey 7y ago
    > Why is that the most preposterous hypothesis that doesn't warrant any review? It's not. What is preposterous is advancing the hypothesis without any meaningful evidence, to the detrement of more useful topics, or even useful discussion of the same topic. It's one thing to be an epidemiologist saying "hey, let's see how we can double check the exact origins of this", which is a potentially useful line of inquery. It's another to take some hypochondriac's third hand retelling of someone else's flu symptoms back in December and jump to the same untested unproven unscientific fear-driven conclusions they did on little more than their wild speculation. Even if they do somehow end up being right, they'll be right in the "broken clock is right twice a day" sense rather than a bringing anything useful to the table sense. So far, alternative suggestions as to the virus's origins have looked more similar to the latter - with perhaps some undertones of (completely understandable and expected) attempted political diversion (I sure wouldn't want fault for this mess hanging on my neck!) If you have some epidemiologist's proposed study that you're trying to drive funding towards - something that even remotely looks like the former - it'd be a welcome breath of fresh air on the topic, and I'd suggest sharing that as a far more useful and constructive way of advancing the hypothesis.