OT: Corona virus

[Canoerebel]

Already covered.

ORIGINAL: mind_messing

ORIGINAL: Canoerebel

We already covered that very thing in posts in here. You are stating and re-stating things, as though novel, that we've already addressed. We know these things.

ORIGINAL: mind_messing



So one point of statistical awareness for folks is around statistical uncertainty.

That predicted figure of say, 1200 likely came with a LOT of uncertainty. That's because predictive modelling is hard and can't account for everything.

In statistical terms, that would have been reported as 1200 +/- 500 (purely back of envelope here). This means that the model is predicating mortality at 1200, but the actual value may range from as low as 700 to as high as 1700.

Stealing an example from earlier, you can see:

Prediction was 421, but the range was 297 to 593, so a +/- of 172.

In short, don't get excited unless the values is outside that confidence interval.

Then what's the error on the Univ. of Washington predictions, and what is it compared to the true reported value?

[MakeeLearn]

Nearly 800 hospital employees furloughed in last week amid coronavirus pandemic
Tuesday, April 7th 2020

https://fox17.com/news/local/nearly-800 ... s-pandemic



"They are our most valuable asset right now in the fight against the coronavirus: nurses, doctors and medical staff. However, in the last week, nearly 800 staffers in Middle Tennessee have been told they’re not needed - at least not for now.

“The last four weeks it’s been pretty rough,” said Todd Haines, an ER nurse and president of the Tennessee Nurses Association."


"Dr. Feagins says she expects to re-open soon. Meanwhile, nurse Haines says he’ll be waiting for his fellow nurses to return and join him on the front lines."

[Lokasenna]

ORIGINAL: HansBolter

ORIGINAL: Lokasenna

ORIGINAL: HansBolter

So we opened the floodgates of America bashing and we finally see people's true colors.

Good riddance.

More inappropriate.



For the record, since we seem to be self-moderating, that's why I'm tagging these.

And you were apparently voted hall monitor?

What part of self meaning only you is too complex for you?

A shame you can't grasp what you are doing is just as inappropriate.

Sorry, you're incorrect. Nor do you appear to know what self-moderating meant in that context.

Somebody has to keep us children in line, clearly. Somebody needs to play hall monitor to the schoolyard chest-puffing.

ORIGINAL: Cap Mandrake

I'm going out on a limb to predict that COVID will NOT be eradicated like SARS or MERS (apparently). It's much better adapted to human transmission. I suspect it will settle in for the long run like the other 4 human Coronaviruses.

229E (alpha coronavirus)
NL63 (alpha coronavirus)
OC43 (beta coronavirus)
HKU1 (beta coronavirus)

With growing herd immunity and almost certainly a vaccine and improved knowledge on treatment it will not be such a terrible threat. I believe of the other human Coronaviruses 3 are of bat origin and one appears to have come from ungulates. We may have had these things for centuries.

This has been my thought as well.

ORIGINAL: Cap Mandrake

Sometimes the "quants" are wrong. Half of them have Asperger's syndrome and can't understand human behavior of emotions. It is possible to be incandescently smart and utterly clueless at the same time.

It would be like putting up a shark net on a beach and then taking it down a week later because the attacks stopped.

Let's not forget the "quants" told Lehman Bros. that mortgage backed securities were safe. [:D]

Can we not resort to stereotypes please?

FWIW, I think greed and motivated reasoning told Lehman Bros. (and all the other involved parties) that those securities were safe.

ORIGINAL: Encircled

We don't use the word "Ungulants" nearly enough

Reminds me of "unguent", which I prefer...

[MakeeLearn]

Adjusted coronavirus model predicts fewer people in US will need hospitals, but 82,000 will still die by August
Updated: Apr 7, 2020


https://kdvr.com/news/adjusted-coronavi ... by-august/


"(CNN) — An influential model tracking the coronavirus pandemic in the United States now predicts that fewer people will die and fewer hospital beds will be needed compared to its estimates from last week.

As of Monday, the model predicted the virus will kill 81,766 people in the United States over the next four months, with just under 141,000 hospital beds being needed. That’s about 12,000 fewer deaths — and 121,000 fewer hospital beds — than the model estimated on Thursday.

A “massive infusion of new data” led to the adjustments, according to the model’s maker, Dr. Christopher Murray, who serves as director of the Institute for Health Metrics and Evaluation at the University of Washington School of Medicine."

[Lokasenna]

ORIGINAL: Canoerebel

Italy projected to drop very fast...but a couple of cautionary notes. First, the mortality reported today was more than projected by a fair amount (further caveat: this could also be a timing of reporting issue or other irregularity). Also, according to this graph, Italy is only expected to incur about 3,000 more deaths. That would be great but seems so contrary to what we've been seeing.

Bottom line: More time needed to get some level of confidence about the reliability of source for jurisdictions outside the US.

Not a substantive note, but one on presentation and ease of digestion of the information.

Graphs like this and the likelihood of daily "noise" are a great example of how rolling averages can present a clearer picture. For this, I'd use a 3-day rolling average.

Obviously, screenshots from the model don't have that. One would have to make their own spreadsheet.

[Lokasenna]

ORIGINAL: MakeeLearn

Meanwhile in Florida...


"This is a reminder that during COVID 19, please remember to keep at least 1 large alligator between you and everyone else at all times,"

https://edition.cnn.com/2020/04/04/us/s ... syndicated

OK, this was quite funny.

[Canoerebel]

In reply to Mind_Messing's question above, here's one chart posted this morning that included a note about the range in projections.

I've done the same thing multiple times over the past few days, usually in the chart portion of my posts.

ORIGINAL: Canoerebel

I don't think any of us feel that things are set in stone or assured.

Regarding the projected mortality, we're using that number as a useful reference point while not losing sight that it's within a range. Nor overlooking that the range is trending encouragingly downward.

[Lokasenna]

ORIGINAL: Canoerebel

[Not directed at anybody in particular, including recent posts, but rather on my just completed click-train to another site]

For the first time today, I ventured into the main page virus thread. It's locked, I think as of today. It seems to have veered into a permanent state of acrimony.

It was a rough weekend here, most certainly for me. But the community as a whole has rallied and done a remarkably good job of getting back on point and trying not to inflame things. Pretty cool to see. Here's hoping this thread stays open for business.

I admit that thus far in my 2-day catch-up since my last check-in, I've been mildly surprised that we didn't have a few more dust-ups while I was on hiatus.


...how many more hyphens can I squeeze into this post...

[Lokasenna]

ORIGINAL: JohnDillworth

Looks like NYC is under-counting, by a wide margin, the number of dead from Corona-19. It's a big city and typically 20 people die at home every day. In the last few weeks, over 200 people a day are dying at home. Coincidence? I think not. Anyway, if you die at home the body is not tested for Corona-19. So the actual body count may be significantly higher than the reported number. Not sure of other locals policies but I suspect they are similar. At home deaths are not being counted so the 1% mortality rate (so-far) is probably artificially low

Some of those 200 deaths are almost certainly not direct SARS-COV-2 deaths, but things like overdoses, home accidents, other illnesses, and so on. People who would have died out in public anyway (heart attack, stroke/hemorrhage, etc.) but who are staying inside. Indirect effects of remaining at home.

It would not surprise me if a plurality were COVID-19 deaths, however.

[Lokasenna]

ORIGINAL: Cap Mandrake

ORIGINAL: mind_messing


Well, it also looks like you can't have a Nav Sec losing the confidence of every person in the chain of command below them either.

No, that would be bad. If he made inflammatory or vindictive comments that was unprofessional.

Trump can fire the Sec of Navy [:)]

I read the remarks he made and found the "too dumb"/"too stupid" to be highly inappropriate. Not just for the military - no leader should make remarks like that. It's very destructive for whatever team you're on. I'm glad he had the cajones to resign.

Not gonna get into the whole incident, but I think it was poorly handled by basically every principal party.

[Lokasenna]

ORIGINAL: JohnDillworth

I prefer accurate numbers

Agreed, I don't think we will get really good numbers for years. Once the statistical anomalies are all added up you will get better numbers. New Jersey just reported a nursing home that had 22 deaths in the last week, in house. None of those deaths right now are in the stats. Actual numbers don't really mean much right now. As you say, consistent numbers from a controlled source are important because we can use them for meaningful comparisons and trends. In other news I have a co-worker test positive. He works on the same floor and we travel in the sames spaces. As he is on a different shift we have not had personal contact in 2 weeks but the other director I switch days with had recent contact. We are both meticulous about had washing, gloves and masks so I think we should be OK. We have maybe 30 people out of 1,800 in the agency working and even then on split shifts so there should not be much chance for cross contamination. Still, lots of open offices and conference rooms so I'm moving to a different place tomorrow.

This is what I understood Chickenboy's JANAC reference to be about.

[Lokasenna]

ORIGINAL: MakeeLearn

In N.Y.C., the Coronavirus Is Killing Men Twice as Often as Women

More men also are infected than women, and they are hospitalized more frequently, new data show. A similar pattern was seen in China.
April 7, 2020, 6:29 p.m. ET

https://www.nytimes.com/2020/04/07/heal ... k-men.html

"In its inexorable spread across New York City, the coronavirus is exacting a greater toll on men than women. Not only are men infected in greater numbers, new data show, but they are also dying at nearly twice the rate of women."

--------------

NO COMMENT

Wondering how much of this effect is sample bias - men are more likely to take unnecessary risks, so a logical hypothesis would be that they are more likely to become infected.

Also wondering how many more men than women have the higher comorbidity factors.

[obvert]

ORIGINAL: Encircled

Can i just urge everyone to treat UK figures on stuff with a lot of caution.

We are miles behind where we should be in the testing, so our models are based on assumptions, not data (I think Mind messing has stressed this)

We are not alone in that, but we are testing more and more, but are still thousands of people behind the likes of Germany (who appear to have got it right, and have the data to back it up)

Regarding the death tolls, and people going "oh well, its not as bad as originally thought", I suspect that if a close relative gets it then the reality will drive it home more.

Both me and Mrs Encircled parents are perfectly healthy 70+ year olds, and this has us scared to death about them, especially as in the UK our two weeks delay in doing the right thing has could potentially result in many more deaths than if we'd done the right thing at the start.

See attached Reuters report on this

https://www.reuters.com/article/us-heal ... SKBN21P1VF

Regarding the China virus thing, its clearly is from China, but its also clearly being used as something else (not by forumites) by people who should know better. Maybe in the interests of this all staying civil it might be best just to call it Covid-19?

Just a thought.

ORIGINAL: Canoerebel

Encircled, thanks for that note.

Yesterday, we discussed the UK situation and the high mortality projection (66k) from the Univ. of Washington site we're discussing here. There were posts about reasons to hold off on putting much stock in that, at least yet, but folks reading through the thread know the UK is really in the thick of things at the moment and that the forecast is troubling.

We're pulling for you all.

P.S. I posted this a.m., asking for an update about your PM. There hasn't been a reply yet (lost in the avalanche of posts).

I agree whole heartedly with the perspective offered by Encircled. It has actually always hit home here because I know people in Italy, France, Germany, Greece, Spain and even in China. So there was never a point since January when I didn't have an impending sense of dread. Hearing stories of the Chinese in Sian, not even in Hubei, locked down inside and not able to even go out for an exercise or to the store was sobering.

We are still very much in the middle of the worldwide crisis. There are some signs that our worst fears may not become reality, but I'm not popping any bubbly anytime soon. This is going to be a long, difficult road followed by a difficult recovery.

The PM here is apparently stable, didn't have to go on a ventilator, and has been reported to be sitting up in his bed today. Fingers crossed.

[obvert]

The cases chart is showing Italy and Spain doing better.

[Canoerebel]

Italy daily report is up at Worldometers. Mortality the second lowest since March 19 but they're still grinding with stubbornly high numbers.

[Canoerebel]

Per Worldometers, UK mortality 938, up about 150 from yesterday.

But like yesterday, that number comes in under the University of Washington projection and again by about 500.

It is within the range of 550 to 3,276.

[Lokasenna]

ORIGINAL: Sammy5IsAlive

ORIGINAL: Canoerebel

Peak in deaths also dropped by four days.

There is a big movement underway - a major trend. What the nation is doing, and the way the experts project, are combining to bring this into a sharper focus that is much more encouraging. Just one week ago, the nation was riveted on 100k to 200k.

Just one note of potential caution on getting ahead of ourselves here with comparisons to the Washington modelling.

As far as I can tell the model has three main assumptions (I don't use that word in a negative sense - somebody with more of a science background might know of a better word) - 1) a base rate assumption about how quickly the virus spreads and the proportion of people requiring hospital admission and in particular ICU care, 2) an assumption of the extent to which varying levels of lockdown are able to slow the spread of the virus and 3) an assumption of the extent to which reaching the limit of healthcare provision will result in excess mortality.

If the actual numbers are lower than the predicted ones then that suggests that some of these assumptions were not correct. The problem that we still have at this point is that we don't know which or to what extent.

An optimist will look towards assumptions 1) and 3) and suggest that the evidence is showing that either the virus is less communicable and/or (more likely) less lethal than was assumed or that health services are in reality more agile and resilient to stress than was assumed (as an example of the latter point - the UK prediction is using our historic ICU availability of c.800 beds unoccupied - the new hospital built in London has added another 500 of these beds on its own).

A pessimist on the other hand might look more at assumption 2) i.e. that the various lock-down measures have had a stronger than assumed effect on slowing the spread of the virus. They might also worry in more general terms that the model appears to assume that the only way to reverse the rapid increase of cases and consequent deaths is by imposing significant lock down measures on society and the economy - and that it appears to have been accurate in that assumption. In other words from that point of view the predicted deaths potentially have not all been avoided - more kicked down the road to a later date.

I would probably class myself as somewhere in between. I think there is a lot of encouragement to be taken from the reduction in the predictions. But I think the largest political/economic challenge may still lie ahead. I think that national governments are going to need to hold firm with the lockdown measures (and indeed the public will need to hold firm in their adherence), not just until the case numbers are dropping but until they are confident that they have the testing regime in place to identify and squash outbreaks at source and the administrative/logistical/economic structure in place to put down hard lockdowns/quarantines on very specific localities (as the Italians were able to do in the small villages that showed very early cases in the Veneto province, which subsequently had a much better ride of things compared to Lombardy etc). I think that if national governments and the public are premature in going 'back to normal' then we may well find ourselves back to square one and at the start of another 'curve'. So I think there is a long road ahead and the more pessimistic mortality predictions remain very much in play in the longer term.

You can put me on this train as well. It is by far the simplest and therefore most likely.

I'm mildly optimistic that sometime in 2020 we'll be able to do some level of serological testing (as well as rapid deployment of testing for active cases and the ability to do contact tracing + isolation) and know more about what level of immunity is granted from antibodies, as well as how people can get those antibodies. Because of what's needed to get a vaccine approved (and for good reason), we won't have one of those until 2021 at earliest.

[Cap Mandrake]

Statistical models usually provide a "confidence interval"(usually +/-2 SD). These seem to be omitted for presentations to the public.

Interestingly, the National Weather Service does include "probability cones" when they publish hurricane maps. These models, though imperfect, are quite useful for planning purposes..ICU beds, vents, staffing...etc.

[Lokasenna]

ORIGINAL: Cap Mandrake

Sorry to burst your bubble, but it's not really. I'll paraphrase from a much wiser and more respected forumite (you know who you are!)

[8|] <facepalm>

Santa Maria! What on Earth is wrong with you? The US healthcare system shite. The US Navy is shite. America is racist. "I'm smarter than everyone". Give it a rest, man.

Did your mama have to pay kids to come to your birthday parties?

He didn't actually say any of those things.

That last sentence especially doesn't belong here.

[Lokasenna]

ORIGINAL: mind_messing

So that prompted a pair of genuine question from me, and before pops a blood vessel, it is genuine:

- Are private healthcare institutions incentivized or required by legislation to have extra capacity to contribute towards a pandemic response?
- If not, why would they, given the business motivations for their organisations?

Before people start getting frothy, I'm well aware of the scale and that equipment and PPE are not an exclusively American issue.

Required? No.

Why would they? Because being prepared is better for their business than being unprepared. Being unprepared can mean going out of business.

Does that mean I think they were prepared? No, but I don't think it was a failure limited just to health care industries. Models of efficiency (somebody a few pages back mentioned the "just in time" economy, I think John Dillworth, which is exactly what I'm talking about) push in the opposite direction.