75 comments

[ 2.9 ms ] story [ 140 ms ] thread
Honest question -- I am trying to see how she and others arrive at these numbers. (Believe me, I'm not a burying my head in the sand ... quite the opposite. I'm very concerned.)

Italy has 8,500 who are still sick. And yet, reports are that their medical system (which is good) is completely overwhelmed, to the point where treatable elderly are not even diagnosed [1].

Yet the numbers indicate they have about 3 hospital beds per 1000, or 186,000 beds, which are normally 80% full. So that means, 37,000 beds available.

So, 8,500 cases (not all of whom are in the hospital) should not overwhelm 37,000 beds, right?

Perhaps the number of _doctors and nurses_ are overwhelmed? Anybody know where to get reliable stats on that?

Anybody see what I'm missing? Or do you think I've gotten incorrect info?

I'm trying to arrive at a rational prediction of how bad things could be in the US, and how soon.

[1] https://twitter.com/jasonvanschoor/status/123714289107769753...

I would guess that Italy's severe Covid-19 cases and available hospital beds have very different geographic distributions.
The numbers are based on confirmed cases. The task of confirming cases is not a trivial matter -- you need test kits, and someone to administer them, and the virus doesn't wait around for you to test for it.
Yeah, it's not the total number of hospital beds, it's the number of intensive care hosptal beds with associated equipment like ventilators and intubation equipment, chest x-ray capacity, and the number of doctors and nurses familiar with that type of equipment.

In my country we have ~12 000 hospital beds total, but only ~400 intensive care beds. Reportedly our capacity for immediately stretching that number is that we can muster ~1 200 intensive care beds. If the need goes beyond that (and everyone is seriously hoping it will not) then we will be overwhelmed like Italy.

This is the right answer. And it’s also why mortality could potentially shoot up to the +/- 7% of patients that need ICU care, if that capacity is exceeded.
Not only that, but these intensive care beds need to be isolated. Ever been to an ICU? Isolation is not a thing. We can't just go in and tape a bunch of plastic drapes to the walls/ceilings/floors and make these ICU beds isolated. 400 ICU beds is to a first approximation zero isolated ICU beds.
Well theatres can be used instead which will allow for isolation. In reality the main HDU or ITU will become covid zone and theatres used for non Covid admissions.
There is a possibility of further increasing ventilator numbers by hooking multiple people up to the same machine providing they have similar lung volumes and compliance. It’s not ideal but better than leaving people to die. Of note, shouldn’t be risk of infection as piping has filters built in to prevent ventilators becoming infected.
From what I‘ve read, issues arise because the personnel cannot be treated with vaccines. So even if you have 37k beds, that‘s pointless if you don‘t have the people to treat them.

Also, the flu is currently active in europe, which also filled up some of the beds already.

You're missing distributions/concentrations.

37k beds are available across the entire country. However, the outbreaks are currently relatively concentrated. Some areas might still have capacity while others are overflowing.

But if someone needs intubation they should be transported to a different location where they will be taken care of, rather than left alone, no?
Transporting an intubated patient hundreds of miles to an area not experiencing an outbreak and having a hospital equipped to deal with same is not a trivial task. Moreover, these numbers of ICU beds bandied about are meaningless. Zero, or close to, of those ICU beds are isolated (infection control).
Transporting a patient that sick requires a specially trained doctor and nurse, ambulance transfers to helipad/airport, an 8 hour+ return trip, and a high likelihood of giving the disease to a bunch of people at the new centre.

If you had all the resources spare to achieve that, you could just keep them probably.

There are very few resources dedicated to long distance patient transfer at the best of times. If you pulled away regularly ambulance services and somehow equipped them to handle infection control you'd barely be breaking out of the single digits of patient transfers on a good day.
1. Those beds aren't just lying around empty. They are used at close to capacity during normal hospital operation.

2. Most of Italy's cases are in a part of it. They are not evenly distributed among the country. Hospital beds, on the other hand, are.

3. Most of those beds are not ICU beds. If you have COVID-19, and you're sick enough to need to go to the hospital, you need to go to an ICU bed.

A typical ICU bed in a U.S. hospital is not isolated in any way. I've seen four to a room, all in close proximity. This can't work for COVID-19. Number of generic ICU beds is meaningless as almost all of those beds are totally inappropriate for a highly infections disease with high mortality rate. The problem is much worse than you posit.
No need for isolation if all the people in a hospital have COVID-19.

Just move all other patients elsewhere, and use staff who have all already had the infection.

I think the 60–70% is a misquote.

The original is in German, but I can't say what it means exactly. They are currently at 0.0000175%.

I think it's more a warning that an "estimate".

This virus is so easily transmitted that one needs a significant percentage of the population to be resistant before new epidemics won’t happen every time it is reintroduced. So 60-70%.

With a little luck there’s a still larger group of asymptotic infections, because otherwise we are going to have to put up with large numbers of ICU patients for a long time.

The estimate is from Prof. Drosten of the Charité in Berlin, a highly regarded virologe. His lab developed the first diagnostic corona test. Before that they also developed the tests for SARS, Zika and MERS.
Meaning that they are quoting from steady states in a model that uses differential equations. I am saying that the news article is misquoting the science. Merkel didn't make a prediction, she made a comment about epidemiology.
It's not - 70% percent as an upper bound is a function of herd immunity kicking in.
Yes, but as the upper bound for the whole of Germany or for a part of Germany? That is what the issue is about. In terms of likeliness for this is accurate. The chances of the whole of Germany being the improper subset here is the issue.

My point is that they make it sound like a prediction.

Edit: My point is that this is science about epidemiology, it's not a prediction.

I saw that number as the usual spread of the seasonal flu pandemics. My guess is she generalized it to COVID-19.
> I'm trying to arrive at a rational prediction of how bad things could be in the US, and how soon.

As a whole, USA is around 11.5 days behind Italy. (some others say 14 days)

https://mobile.twitter.com/MarkJHandley/status/1237144386569...

But this is not so simple to translate, because of the very different sizes between USA and Italy (population, geography, and pop density), and the resources available (if the federal gov finally reacts).

So, while that chart counts the cases per country, maybe the best thing here would be to consider the evolution in a single state, instead of the whole country.

City by city may be more relevant considering how much less dense USA is outside of high concentration urban areas (and even really inside such areas).
It’s honestly impressive how I seen daily predictions of infections in single countries exceeding the entire confirmed infected global population by 100x.

It’s either the idea that this is a generally mild virus with some random severe cases and basically we’re all already infected, or this is going to be an endless, multi year spread and they’re accounting for that.

> It’s either the idea that this is a generally mild virus with some random severe cases and basically we’re all already infected, or this is going to be an endless, multi year spread and they’re accounting for that.

It's partly the latter, but it's mostly just the result of exponential growth. The number of confirmed cases doubles every ~5 days, if that continues there are over a billion infections in 10 weeks. Whether it plays out that way or not depends on our containment efforts.

The bed count is for all of Italy where as all the seriously sick are in 1 or 2 places. Italy has already reached a point where doctors have to decide who to save and where to spend the resources as they don't have enough. 3 days ago all the beds were occupied but they were still getting 200+ new critical patients a day
How many spare ventilators do they have?
"Merkel went on to tell lawmakers that more events and large gatherings could be canceled, depending on the severity of the outbreak.

Postponing parliamentary proceedings is also a possibility, Merkel said."

wtf. This is what they are planning to do about it?

What's to downvote? I mean they are essentially doing nothing. Allegedly this is also due to the federal system where constituent states get to decide such policies. The federal health minister can only "recommend" stuff but this is not the way to contain the spread of a virus.

Neighbour states such as Austria are closing universities while in Germany they are still playing football matches in arenas with 50.000 people (yesterday).

> yesterday

I hate the "it's a dynamic situation" wording but yes, days matter here. Also yesterday, NRW, one of the most affected states put out the official recommendation to cancel major events with more than 1000 persons. Austria closed some universities at this point but many are scheduled to close next Monday.

Seems to be a prediction at the dramatic end in particular if the quick lockdowns and quarantines are kept up. As China has shown it's very possible to bring the numbers down.

Either way though even if it turns out to be a too high a prediction at least she speaks plainly and I guess electing a physicist who understands how these dynamics can work was a good idea.

>Seems to be a prediction at the dramatic end in particular if the quick lockdowns and quarantines are kept up.

I've read several epidemiologists that said stuff among the lines that, while the objective is not to generate panic, leaders should not minimize the scale of this outbreak, and the consequences of downplaying it are severe. When politicians say stuff like "this is just another flu", then lots of people continue their lives as usual, and thus, it gets almost impossible to slow down the propagation.

Maybe it's OK to throw those numbers, as a wake up call for society. If the government ordered a full lock-down, many people that still think "it's just a flu" would get mad... but if she first announces this, then maybe the rest of society would be more open to accept the consequences of a lock-down. In the end, it's politics.

How long do you propose we maintain this "lockdown"?
I would look into China's experience for guidance. I know the costs would be monumental, so it's not an easy decision, and would not like to be in her shoes right now. But if the alternative is what Merkel suggests...
But this virus won't go away until a significant fraction of the population gets infected and heals. It'll come back in the fall when it's colder and hit again. I'm afraid the only principled solution here is to develop vaccines and working antivirals in the next few months, which by "normal" standards is completely unprecedented.
quarantines will happen anyway way too many people will die even with them
The best way forward is to flatten the curve as much as possible while other means are worked on.
> As China has shown it's very possible to bring the numbers down.

China was good at it because they really enforce drastic measures. People must stay at home, everybody outside wears a mask (they're made in China...), streets, transport & supermarkets are disinfected daily, people who leave hospital have even clothes disinfected etc.

Nothing like this happens in EU or USA.

This is completely unwarranted fatalism. Taiwan and Singapore bent the curve, and Korea is doing it now as well. Learn from them and implement! Really don't understand this logic at all - containment is possible.
(comment deleted)
One way to contain is to take the potential risk serious. If that fails, it's still better to slow down the spread.

Preparing a country for this potential outcome is a good thing. If it didn't end up at 60-70%, it will be in part because of a very strong response.

If you said this last week about Singapore I would agree with you. Have you looked at their graph in the last few days?
It's still mostly contained (most cases are connected to existing clusters) but the government is foreshadowing that containment is becoming a major problem without being able to rely on geographic travel screening. Many imported cases (UK, Italy, Thailand) in the last few days.

The curve is still very much bent when we look at how early this started in the country and it will continue to be flatter than in any place not putting distancing and testing in place on time

Until you bend the curve down all the way to 0 new infections per day, you'll eventually reach 60-70%.
Right, but the longer it takes to get there, the more likely you are to either find a cure. Failing that, you at least reduce pressure on medical infrastructure, which means more people get meaningful care.
That of course assumes all countries can implement this kind of measure as effectively as e.g. Taiwan did. Which gets harder and harder, especially if you can't travel back and time and start it at the right time. The line between fatalism and realism about how much you'll be able to do can be thin. (Although I think that at least parts of Germany would have a fair chance at such measures still)
The estimate is from Prof. Drosten* of the Charité in Berlin, a highly regarded virologe. His lab developed the first diagnostic corona test. Before that they also developed the tests for SARS, Zika and MERS.

He also constantly warned to flatten the curve, spreading infection time as long as possible - preferably over multiple years...

For those speaking german - there is a daily podcast update from Prof. Drosten https://www.youtube.com/playlist?list=PLkKON9te6p3OpxqDskVsx...

*Head of the Institute for Virology at the Charité

> preferably over multiple years...

Reminds me of a tweet I read today:

https://twitter.com/kurteichenwald/status/123742407086352384...

Even if this outbreak is contained, either because of social policies or seasonal change, governments should prepare for a new wave next year (and the southern hemisphere should be preparing for a hit in a few weeks/months).

Exactly what happened with SARS which was contained by the third year and now no longer occurs
Well at least the leader of the remaining Free world is willing to tell the truth to her people...
Would you please stop posting political flamebait and otherwise breaking the site guidelines? You've been doing it repeatedly, and we ban accounts that do that.

https://news.ycombinator.com/newsguidelines.html

Christian Drosten, who's likely the origin of these numbers, has explained them a few days ago in a podcast.

The math is incredibly simple. The idea is that right now you have an estimated transmission rate of 3 people infected by one person. (Of course this number has high uncertainties and can be changed by better isolating patients, but let's assume it's roughly correct.) If you want to stop the transmission this needs to go below one. The major thing that can stop transmission is immunity from patients that already had the disease (assuming, which is probably true, that most patients won't get infected twice, even though it may happen in rare cases).

So getting from 3 to 1 means 2/3 of the population need to be immune.

The Spanish flu “only” infected about 1/4 or 1/3 of people. Why the higher number for Covid?
Probably because people travel much more now.
The spanish flu happened during the first world war when many soldiers were traveling significantly.
(comment deleted)
Would it make sense to do controlled infections of younger medical staff to build immunity against the disease so they can do their jobs when masks and gloves become scarce? Ethics aside just from a pure optimal outcome perspective. Italy is already reporting a shortage of masks.
>> when masks and gloves become scarce

Already there... At least in Germany.

There is nothing called a controlled infection outside of a vaccine. As we saw in Wuhan, several youngish doctors and medical staff died as well. When some of the "controlled infections" turn severe, now you have to take care of them as well(probably with a higher priority since they volunteered for the good of the community), which adds to the burden.
Well, there is. You infect them and keep them isolated until they recover. If you do that before the problem becomes rampant, you will have more healthy doctors available. However, if the virus mutates significantly, they may just get sick again.

It's possible, but probably not ethical.

If the estimate is 60% to 70% of the county will get the disease you could assume every frontline health care person will get infected especially due to the proximity to the disease. You're just controlling when they get infected.