CoronaBro Meltdown/SARS-Covid-19 Spitballing Thread

[catastrophe]

Compared to a few weeks ago I don’t think we’ve exceeded hospital capacity anywhere or come up with better treatments, so deaths should still be the most apples to apples metric we have.

I can guarantee that we are both better equipped and we've learned how to treat the disease much better over the course of the last month. More ventilators, putting patients on their stomachs instead of their backs, etc etc.

So what was the death rate 3 weeks ago and what is it now?

[Kat Kid]

New cases is a really dumb thing to look at imo.  As we test more people (which we are, finally), the rate of new cases is going to continue to rise even if the rate of growth stays the same. 

Much harder to hide a covid death from the data than a covid case.

i will ask again. what do deaths have to do with transmission?

X number of people die from the virus.  If you can get your arms around X, and you know how many people are dying, you could extrapolate a solid estimate or people that have/had it. 

Also, deaths are kind of the point.  We're not shutting crap down to prevent people from getting a cough/fever, we're shutting it down to prevent people from dying.  Stated differently, who gives a eff about the virus if it's hardly killing anyone?

Prior to the last three days, every wednesday had more people die than the last wednesday; every thursday more than the last thursday; every friday more than the last friday.  Three days in a row there have been fewer.  Maybe you're right and there's nothing to be drawn from that, and we should instead stare at the number of new positive tests (despite testing over 50% more people than we did last Friday).

I don't agree with everything dlew thinks but on this he is exactly right. deaths are the only number in the US worth looking at.

There is lots of statistical noise/assumptions/unknowns with all of this stuff-- "are we under counting deaths at home? how do we determine co-morbidities? are we counting only deaths from confirmed cases? are death rates roughly equal when taking in to account population characteristics between states, countries? how much does smoking, obesity, age, etc. have to do with death rates?" etc.

but within the US, deaths are the best apples to apples comparison because they are a lot less noisy and the testing is not representative and based on the samples we have done, we are missing tons of asymptomatic carriers and so while we can't extrapolate exactly how many people have/had it or the exact death rate and testing has just been frankly much too spotty and inaccurate to draw a lot of conclusions on the total population from it.

we can better approximate momentum of spread from deaths than testing with the understanding that this is a lagging indicator.

[sys]

I would think deaths are the better metric.

i agree.  case counts is largely a function of testing.  you can guess at the testing effect if you account for the % positive in tests, and all three give measures give some useful information, but if you can only pick one metric, the death count is the best indicator of the true number of infections in a population.

[sys]

you can have drastically different death rates with the same transmission rate

in similar aged populations with similar medical treatment, i think you probably get a pretty similar mortality rate.  comparing across countries may get a little dicey, but comparisons between states is probably subject to fewer biases than are the case count data.

agree that it doesn't give you any information on the rate of transmission after the time that the cohort currently dying was infected, but you can get some idea of that from other data.

[sys]

deaths are kind of the point.  We're not shutting crap down to prevent people from getting a cough/fever, we're shutting it down to prevent people from dying.  Stated differently, who gives a eff about the virus if it's hardly killing anyone?

if we ever get to the point where we seriously consider just letting the virus run through the population, we should probably also think about long-term organ damage.  but i don't think we're there yet, even in the nyc metro.

[sys]

I don't agree with everything dlew thinks but on this he is exactly right. deaths are the only number in the US worth looking at.

There is lots of statistical noise/assumptions/unknowns with all of this stuff-- "are we under counting deaths at home? how do we determine co-morbidities? are we counting only deaths from confirmed cases? are death rates roughly equal when taking in to account population characteristics between states, countries? how much does smoking, obesity, age, etc. have to do with death rates?" etc.

but within the US, deaths are the best apples to apples comparison because they are a lot less noisy and the testing is not representative and based on the samples we have done, we are missing tons of asymptomatic carriers and so while we can't extrapolate exactly how many people have/had it or the exact death rate and testing has just been frankly much too spotty and inaccurate to draw a lot of conclusions on the total population from it.

we can better approximate momentum of spread from deaths than testing with the understanding that this is a lagging indicator.

yes.

[michigancat]

Also, deaths are kind of the point.  We're not shutting crap down to prevent people from getting a cough/fever, we're shutting it down to prevent people from dying.  Stated differently, who gives a eff about the virus if it's hardly killing anyone?

I've actually been worried about the potential chaos of ICU and ventilator shortage at least as much as the number of deaths, if not more. It's all related though.

[catastrophe]

Also, deaths are kind of the point.  We're not shutting crap down to prevent people from getting a cough/fever, we're shutting it down to prevent people from dying.  Stated differently, who gives a eff about the virus if it's hardly killing anyone?

I've actually been worried about the potential chaos of ICU and ventilator shortage at least as much as the number of deaths, if not more. It's all related though.

Definitely my biggest concern. I mean tons of people die from tons of preventable crap every second (heart disease being a huge part of it). I’m scared of something that can make you go from healthy to dead in a few weeks, but I’m absolutely terrified of being in a situation where you don’t even have a fighting chance in the world’s most advanced and expensive healthcare system.

[DQ12]

Also, deaths are kind of the point.  We're not shutting crap down to prevent people from getting a cough/fever, we're shutting it down to prevent people from dying.  Stated differently, who gives a eff about the virus if it's hardly killing anyone?

I've actually been worried about the potential chaos of ICU and ventilator shortage at least as much as the number of deaths, if not more. It's all related though.

Definitely my biggest concern. I mean tons of people die from tons of preventable crap every second (heart disease being a huge part of it). I’m scared of something that can make you go from healthy to dead in a few weeks, but I’m absolutely terrified of being in a situation where you don’t even have a fighting chance in the world’s most advanced and expensive healthcare system.

Well, and it’s related.  If fewer people are dying from covid, then it suggests that there’s not a mass run on ICU/ventilators - or at least less of a run than it would be suggested if more and more people were dying.

[sys]

kind of interesting.  his big idea seems to be: eff contact tracing, we aren't good at it and people don't want to do it.  so let's just test everyone.  seems kinda dumb, but i guess if people will do it instead of just hanging around not doing anything to fix the situation then maybe it's smart.

come to think of it, that was probably the plan in the harvard report that wanted 20 m tests/day too.

https://twitter.com/paulmromer/status/1253247238446264321

[catastrophe]

Absent google and Apple making contact tracing a requirement to play candy crush, I’m on board with that idea.

I can’t think of anything more American than solving the problem by working a ton and spending a ton of money so you can do whatever the eff you feel like.

[sys]

i enjoyed goolsbee's obvious frustration with our refusal to copy other countries that are successfully dealing with this virus.  good interview.

https://twitter.com/morningmoneyben/status/1253710577140105222

[sys]

I can’t think of anything more American than solving the problem by working a ton and spending a ton of money so you can do whatever the eff you feel like.

yeah, it does seem like something we would do.

[sys]

on the other end of the spectrum - something that we could do right now and which probably works, but that doesn't sound like something we'd do.

https://twitter.com/Comparativist/status/1253898984143204352

https://twitter.com/BrianKelch/status/1236359069004414976

[treysolid]

you can have drastically different death rates with the same transmission rate

in similar aged populations with similar medical treatment, i think you probably get a pretty similar mortality rate.  comparing across countries may get a little dicey, but comparisons between states is probably subject to fewer biases than are the case count data.

agree that it doesn't give you any information on the rate of transmission after the time that the cohort currently dying was infected, but you can get some idea of that from other data.

I'm sorry, but we only get deaths. We don't get the metadata that comes along with it, and that data would be too exhaustive to compile anyway.

[michigancat]

on the other end of the spectrum - something that we could do right now and which probably works, but that doesn't sound like something we'd do.

https://twitter.com/Comparativist/status/1253898984143204352

I don't understand why there's four groups. Anyway, I'm pretty sure there was a guy on MSNBC who went into detail on that way back in mid March. Interesting stuff.

Also, the US not doing what works for other countries goes back to us rejecting the WHO tests to develop our own that sucked. So dumb.

[michigancat]

And with testing count vs deaths, I think testing data is most useful if it's paired with positivity rates. Don't know why that isn't as widely shared

[Bqqkie Pimp]

https://twitter.com/SarahLongwell25/status/1253839064362606595

https://twitter.com/disclosurel/status/1253992595224748032

[sys]

I'm sorry, but we only get deaths. We don't get the metadata that comes along with it, and that data would be too exhaustive to compile anyway.

it's not that the deaths data are so great, it's that the cases data are so bad.

[chum1]

https://twitter.com/thegarance/status/1253499811506118663

It appears that they wish to press on with their coordinated effort to deliver a phony cure in spite of the red flags from studies and official warnings against.

https://twitter.com/Acyn/status/1253869919961137152
https://twitter.com/Acyn/status/1253839719646900225

[sys]

view from a city that didn't surrender to the virus without a fight.

https://twitter.com/Birdyword/status/1253958377656352769

[michigancat]

https://twitter.com/thegarance/status/1253499811506118663

I just read this. Wish they would have jumped through hoops to get KN95 masks approved like they did with Trump's miracle drug

[michigancat]

view from a city that didn't surrender to the virus without a fight.

https://twitter.com/Birdyword/status/1253958377656352769

That seems pretty empty for HK.

[michigancat]

kind of interesting.  his big idea seems to be: eff contact tracing, we aren't good at it and people don't want to do it.  so let's just test everyone.  seems kinda dumb, but i guess if people will do it instead of just hanging around not doing anything to fix the situation then maybe it's smart.

come to think of it, that was probably the plan in the harvard report that wanted 20 m tests/day too.

https://twitter.com/paulmromer/status/1253247238446264321

I also don't think he understands how contact tracing works. Isolating asymptomatic close contacts of confirmed cases is probably more effective than relying on testing alone given the high false negative rates from a lot of the tests (although it would probably lead to a lot of unnecessary isolation). Anyway, you should send him that contact tracing chart.

[Bqqkie Pimp]

Who knew, right?

https://twitter.com/TVNewsHQ/status/1253821036333797376