COVID-19 thread

[Kuchigakatai]

I wonder what the thing on the left in the "three germs talking" frames is supposed to be. The thing in the center is a virus, and the thing on the right is a bacterium, but what's the thing on the left?

It's a couple of superimposed streptococcus strands. Bacteria.

[Raphael]

I wonder what the thing on the left in the "three germs talking" frames is supposed to be. The thing in the center is a virus, and the thing on the right is a bacterium, but what's the thing on the left?

It's a couple of superimposed streptococcus strands. Bacteria.

Ah, thank you!

[WeepingElf]

Meanwhile, drive-in movie is back! (Sorry about the German. Marl is a town in Germany, on the northern edge of the Ruhrgebiet, where a new drive-in movie theater has been set up by cinema proprietors who had to close their theaters because of COVID-19.)

[Kuchigakatai]

Hymn for the current times:

https://www.youtube.com/watch?v=vWrjy7vztMw

I apologize for the bad quality.

[Richard W]

I wonder what the thing on the left in the "three germs talking" frames is supposed to be. The thing in the center is a virus, and the thing on the right is a bacterium, but what's the thing on the left?

It's a couple of superimposed streptococcus strands. Bacteria.

So from left to right, it's a bacterium, a virus and a protozoan comparing notes.

[Moose-tache]

Discouraging news from New York. Over 600 new deaths on April 3. This isn't just bad because it's a higher number than before. Quarantine measures were put in place on March 20, so we knew that deaths would accelerate for the rest of March. But the first week of April is the soonest we could see positive results from the governor's social distancing policies. So far that has not happened. We may still see a sharp drop in the growth of new cases in NY in the next few days, but it's already been two days since the 12-day mark, with no change so far.

[MacAnDàil]

For the confinement measures, I have been hearing on France 24 that we have been following the model of the PRC when we could be following the model of Taiwan, with more testing and not a complete confinement.

For the numbers, I may be just bullshitting, but should we really be defining thus the fatality rate of a pandemic while it is happening? The fatality rate (currently 5.4% and rising; 5.1% two days ago) is comparing the deaths to the number of cases. BUT the majority of cases are still current. So we should be comparing the deaths only to the other end of the disease: healing. And there, the death rate compared to the healings is much less positive: 20% of the time, the end of the disease is death. So what do you think? Clutching at straws? Or a real indication that the final death rate shall be much higher?

[chris_notts]

On the other hand there are claims that some people are almost asymptomatic and therefore the death rate is lower than reported. Without testing absolutely everyone, it's hard to know what the total infected population is, and therefore the true rate. The reported death rate right now is the rate for those with severe enough symptoms to bother testing.

[zompist]

For the numbers, I may be just bullshitting, but should we really be defining thus the fatality rate of a pandemic while it is happening? The fatality rate (currently 5.4% and rising; 5.1% two days ago) is comparing the deaths to the number of cases. BUT the majority of cases are still current. So we should be comparing the deaths only to the other end of the disease: healing. And there, the death rate compared to the healings is much less positive: 20% of the time, the end of the disease is death.

No, that doesn't work. The basic calculation is easy (deaths / cases) but there are several types of missing data:

1. current cases which will recover (you're leaving these out, and thus getting a rate that's much too high)
2. current cases which will die but haven't yet (these mean calculated rates are too low)
3. current cases which aren't reported, because not enough testing is being done-- this means calculated rates are too high

The numbers for (3) are likely to be very high, and will swamp the other two.

The best data is from South Korea, where they test like crazy and trace all cases. It currently has a 1.75% fatality rate.

[chris_notts]

Since the virus is still evolving to better attack its new human hosts and has split into several strains already, and since human populations do differ in genetics, general health, nutrition etc., it's also not obvious to me that the number would be expected to be the same everywhere.

On the number itself, a lot of people are still saying that 1% or 2% sounds small. But to lose 1/50 people to a single cause in a year or two sounds quite significant to me, on top of background mortality. And that's ignoring the effects of dedicating the health service to corona on the death rate. There are reports today in the Guardian of people not receiving essential cancer care because the NHS is diverting resources into preparing for / fighting corona, which will further raise the total mortality of the crisis.

[akam chinjir]

About the Taiwan comparison, one thing that's distinctive here is that the government has apparently been very successful at stopping the virus at the border. (I haven't been keeping notes, but I remember just one case in the last week or so of someone who caught it here and isn't assumed to have caught it directly from a close family member---a spouse or parent---who caught it overseas. This is out of something like 10-25 newly identified cases each day.) I hope Taiwan is ready in case it really starts spreading domestically, but in the meantime I'm not sure what kind of a model it can provide for other countries.

[akam chinjir]

Ugh, one new case of local transmission today (out of seven new cases).

[MacAnDàil]

@zompist: I take into account your relevant points. Only one point I would like to add: I get the impression that South Korea has such a low fatality rate partly because it is one of the few places that tests and tracks so much. The capabilities of the health system seems to be a major factor in the fatality rate. I suspect another factor, not yet known, is to which extent the fatality rate shall increase with the overwhelming of the health systems. There are indeed so many unknowns and testing massively is a key to actually knowing more.

@akam chinjir: You're lucky. Here in Réunion, with 4% of the population of Taiwan, there were 17 new cases yesterday.

[Raphael]

And that's ignoring the effects of dedicating the health service to corona on the death rate. There are reports today in the Guardian of people not receiving essential cancer care because the NHS is diverting resources into preparing for / fighting corona, which will further raise the total mortality of the crisis.

Yeah, I've seen people convincingly argue that the numbers we really should be looking at are the excess deaths numbers, that is, the number of people who die in a given place over a given time from any and all causes compared to the number of people who died an that place over an equal length of time from any and all causes before all this.

[akam chinjir]

@akam chinjir: You're lucky. Here in Réunion, with 4% of the population of Taiwan, there were 17 new cases yesterday.

Definitely, I'm in maybe the only country in the world where the people in charge immediately listened to doctors in Wuhan and not to Chinese officials or the WHO.

[Linguoboy]

Since the virus is still evolving to better attack its new human hosts and has split into several strains already, and since human populations do differ in genetics, general health, nutrition etc., it's also not obvious to me that the number would be expected to be the same everywhere.

The strains differ in minor ways that, according to what virologists have been able to determine, don't really affect its infectiousness or lethality and are mainly useful for tracing its spread (though it's been so undersampled so far even this data needs to be treated very cautiously). Moreover, coronavirus strands tend to mutate less rapidly than the common flu.

Here's a bit more on the perils of overintepretation of the limited data we have: https://science.sciencemag.org/content/367/6483/1176.

Yeah, I've seen people convincingly argue that the numbers we really should be looking at are the excess deaths numbers, that is, the number of people who die in a given place over a given time from any and all causes compared to the number of people who died an that place over an equal length of time from any and all causes before all this.

Doing that might make US figures look worse than Italy's, since Italy had a mild flu season, so at least some of their COVID-19 fatalities represent folks who would have fallen to seasonal flu in a more typical year. But the USA was having a worse-than-average flu season before this all happened (which also had the effect of diminishing health care capacity in advance of the outbreak).

[Raphael]

Doing that might make US figures look worse than Italy's, since Italy had a mild flu season, so at least some of their COVID-19 fatalities represent folks who would have fallen to seasonal flu in a more typical year. But the USA was having a worse-than-average flu season before this all happened (which also had the effect of diminishing health care capacity in advance of the outbreak).

Oh. That's a complication I didn't think of.

[Pabappa]

nearly every country on https://gisanddata.maps.arcgis.com/apps ... 7b48e9ecf6 now has a flattening curve, when looking at the logarithmic graph. i think that although the worst of this is still ahead, the end is in sight and we can plan for a return to normal life soon.

[Raphael]

Another German-language Covid-19 related news story: A small town in the north-east of Germany is temporarily partly being governed from Cape Town, South Africa, because that's where the mayor got stuck as a result of anti-corona measures:

https://www.mopo.de/im-norden/mecklenbu ... t-36515304

(The article itself is SFW, but since it's on the website of a European tabloid newspaper, there may or may not be NSFW stuff elsewhere on the website - I didn't check.)

[KathTheDragon]

Doing that might make US figures look worse than Italy's, since Italy had a mild flu season, so at least some of their COVID-19 fatalities represent folks who would have fallen to seasonal flu in a more typical year. But the USA was having a worse-than-average flu season before this all happened (which also had the effect of diminishing health care capacity in advance of the outbreak).

Oh. That's a complication I didn't think of.

Can't it still be controlled for?