COVID-19

[sTeamTraen]

Germany's cases peaked about 18 days ago, but their deaths are still going up very fast. The last two days have seen the highest numbers of daily deaths since the pandemic started. This suggests either that people take a lot longer to die from COVID-19 in Germany than in many other places, which would in turn imply super-early testing (but their tests/million are half the UK's, so it would have to be *very* targeted), or something else is happening.

Untitled.png (20.05 KiB) Viewed 4086 times

[shpalman]

But their CFR is quite low so they probably found a higher proportion of their cases.

I've been comparing the Ticino and Como numbers and there's an odd feature that the declared recovery numbers are very low, maybe they only note people discharged from hospital and don't pay attention to people who get better at home.

The provinces of Como and Varese have the highest rates in Italy right now and also have borders with Switzerland.

I note that the French are checking the borders to stop people going skiing in the mountainous neutral-during-WW2 country #secondmentions because the Swiss have left their skilifts running.

[sTeamTraen]

I posted the charts of the German numbers (post N-2) and a Twitter contact who is a doctor in Spain sent me the following by direct message (slightly edited for context, as I asked him a couple of questions). May be of interest:

One of the causes [[of the German numbers]], is that they wait for longer until withdrawing support (waiting is somewhat futile in most cases but that's the trend if healthcare is not very strained. Just an hypothesis that might be not the cause but actually occurs

There are biological times after which there's no tissue function improvement, but we must wait until inflammation wears completely out

Think of it as waiting for Notre Dame to stop burning until we can evaluate "what's left"

If what's left is not enough to support life out of the machine, support is withdrawn after group decision and sometimes a few days more

This can take an additional week. Many deaths in ICU COVID are "semi-programmed" don't occur unknowingly to us in the icu, and are as a result of treatment withdrawal due to ethical reasons

Possibly the German cohort is aggressively treated (they don't lack resources) so more likely that the elderly last longer. I might be mistaken about this.

Germany did well from the beginning. I have the feeling it has been the consistently most robust response in Europe. They allocated resources according to the size of the problem (proned patients by themselves clearly signal ICU stays >3 weeks and potential ICU shortages. No other country reacted like this in the beginning...PH plans were first written for bad flu instead...

The sight of a proned patient has no meaning for virtually all physicians (including virologists, PH, Internal Medicine..) except the ICU ones...

The virus was well studied from the Epi point of view...But to design containment measures they needed the clinical aspects of the disease (which condition healthcare stress)

(From me) People from Sweden have told me that (since before COVID) almost nobody in Sweden gets admitted to ICU, as it's considered unethical. I suppose it's a coincidence that it's also cheaper for the system.

Totally...Maybe they offered the most extreme care to the elderly...And the healthier ones surely stand the 3 weeks without dying even if final outcome is death...

It's a sensible measure to be restrictive with ICU care...we should go towards that but reinforcing intermediate care, cheaper and equally effective.... We border unethical managing many times...

To be fair, both elderly and young that die in the ICU, behave in a similar way... and treatment and withdrawal is actually very similar. We have very few patients >75...

Maybe Germany has middle aged patients attached to machines for too long...we wait even more in the young although it is futile in some cases

There's also an unsettling thing in COVID ICU patients... They don't improve until the 3rd week in many cases...(they oscillate randomly up and down only)

So we sometimes don't know how the patient will do until 3rd week from severe symptom onset.

Part of the healthcare worker stress with COVID, apart from hard workloads, unstability, is the mental toll of not seeing what's ahead in the mist, with our patients

It makes you feel like being a kidnapped, tied passenger and not the driver at the bedside

Once we knew this pattern, it was easier mentally

we were lucky due to Galicia being a low strain zone in March...

Due to having no surgeries we could have some days of rest scattered through the calendar. In those days we could use the time as we wanted... We devoted part of this time to planning, logistics... The sense of control is an amazing tool for mind well-being in stress... We did things like setting up decontamination routes, craft PPE pieces...

And we doubled turns in bad times...this is more hours but less mental toll, and it was only for a few weeks...

We tried to regain a sense of control, and we remained united like a small army, this helped much

People solidarity sending PPE also kept morale afloat. I can't imagine how hard it must have been in Madrid...

(When I asked if I could share this, in anonymised form)

Ok! I only think of two caveats

First is that the same pattern of long admission times occurs even if we don't admit people > 75 y.o

Even people w/ 40 yo pose the same challenge for ICU capacity. The elderly cannot be blamed (or only partially blamed even if there's abundant ICU capacity

(the young in the ICU are just as bad as the elderly in terms of complexity and stay)

The other caveat is that the long stay explanation ys very appealing but there might be other causes we don't know about...

[jimbob]

I posted the charts of the German numbers (post N-2) and a Twitter contact who is a doctor in Spain sent me the following by direct message (slightly edited for context, as I asked him a couple of questions). May be of interest:

One of the causes [[of the German numbers]], is that they wait for longer until withdrawing support (waiting is somewhat futile in most cases but that's the trend if healthcare is not very strained. Just an hypothesis that might be not the cause but actually occurs

There are biological times after which there's no tissue function improvement, but we must wait until inflammation wears completely out

Think of it as waiting for Notre Dame to stop burning until we can evaluate "what's left"

If what's left is not enough to support life out of the machine, support is withdrawn after group decision and sometimes a few days more

This can take an additional week. Many deaths in ICU COVID are "semi-programmed" don't occur unknowingly to us in the icu, and are as a result of treatment withdrawal due to ethical reasons

Possibly the German cohort is aggressively treated (they don't lack resources) so more likely that the elderly last longer. I might be mistaken about this.

Germany did well from the beginning. I have the feeling it has been the consistently most robust response in Europe. They allocated resources according to the size of the problem (proned patients by themselves clearly signal ICU stays >3 weeks and potential ICU shortages. No other country reacted like this in the beginning...PH plans were first written for bad flu instead...

The sight of a proned patient has no meaning for virtually all physicians (including virologists, PH, Internal Medicine..) except the ICU ones...

The virus was well studied from the Epi point of view...But to design containment measures they needed the clinical aspects of the disease (which condition healthcare stress)

(From me) People from Sweden have told me that (since before COVID) almost nobody in Sweden gets admitted to ICU, as it's considered unethical. I suppose it's a coincidence that it's also cheaper for the system.

Totally...Maybe they offered the most extreme care to the elderly...And the healthier ones surely stand the 3 weeks without dying even if final outcome is death...

It's a sensible measure to be restrictive with ICU care...we should go towards that but reinforcing intermediate care, cheaper and equally effective.... We border unethical managing many times...

To be fair, both elderly and young that die in the ICU, behave in a similar way... and treatment and withdrawal is actually very similar. We have very few patients >75...

Maybe Germany has middle aged patients attached to machines for too long...we wait even more in the young although it is futile in some cases

There's also an unsettling thing in COVID ICU patients... They don't improve until the 3rd week in many cases...(they oscillate randomly up and down only)

So we sometimes don't know how the patient will do until 3rd week from severe symptom onset.

Part of the healthcare worker stress with COVID, apart from hard workloads, unstability, is the mental toll of not seeing what's ahead in the mist, with our patients

It makes you feel like being a kidnapped, tied passenger and not the driver at the bedside

Once we knew this pattern, it was easier mentally

we were lucky due to Galicia being a low strain zone in March...

Due to having no surgeries we could have some days of rest scattered through the calendar. In those days we could use the time as we wanted... We devoted part of this time to planning, logistics... The sense of control is an amazing tool for mind well-being in stress... We did things like setting up decontamination routes, craft PPE pieces...

And we doubled turns in bad times...this is more hours but less mental toll, and it was only for a few weeks...

We tried to regain a sense of control, and we remained united like a small army, this helped much

People solidarity sending PPE also kept morale afloat. I can't imagine how hard it must have been in Madrid...

(When I asked if I could share this, in anonymised form)

Ok! I only think of two caveats

First is that the same pattern of long admission times occurs even if we don't admit people > 75 y.o

Even people w/ 40 yo pose the same challenge for ICU capacity. The elderly cannot be blamed (or only partially blamed even if there's abundant ICU capacity

(the young in the ICU are just as bad as the elderly in terms of complexity and stay)

The other caveat is that the long stay explanation ys very appealing but there might be other causes we don't know about...

Could you suggest they actually tweet reply that because it would be good to get to a wider public?

[sTeamTraen]

Could you suggest they actually tweet reply that because it would be good to get to a wider public?

I presume they would do so if they thought it was in their professional interest. I might try and gently suggest it, but I don't know the person at all well (even on Twitter terms). I think I am providing a little bit of psychological support by listening.

[sTeamTraen]

993 deaths reported in Italy today. The previous highest was 971, on 28 March. It seems to have all gone to sh.t in the last six weeks, far steeper than most comparable countries.

Have the "We want our freedumb" demonstrations of a few weeks ago died down yet?

[shpalman]

993 deaths reported in Italy today. The previous highest was 971, on 28 March. It seems to have all gone to sh.t in the last six weeks, far steeper than most comparable countries.

Have the "We want our freedumb" demonstrations of a few weeks ago died down yet?

... and Italy was doing so well... back when schools had just gone back, but the cases hadn't started showing up #buildinganarrativefromaweightedrandomnumbergenerator

(although that article discusses the terrible impact of the first wave in the Bergamo area, and that region has just about the lowest incidence of second wave in Lombardy if not the whole of the north... but then it's also estimated that 25% of them have antibodies)

And then we let everything grow exponentially for the whole month of October.

However I do think today's spike is a "date reported" artefact (the seven-day average was 700 per day and falling) since I don't know if we have data by date-of-death. Another thing which seems to have happened in the numbers today is that the screening numbers seem to have been shifted into the main testing figure. (ETA no it's not that, but there seems to have been a typo in the number of cases tested which meant it was suddenly 10 million higher than yesterday; and something has certainly happened which means the screening numbers have gone to zero.)

However the government also wants to impose quite a heavy lockdown for Christmas and New Year.

I'm scheduled to go back to Italy tomorrow morning.

[jimbob]

993 deaths reported in Italy today. The previous highest was 971, on 28 March. It seems to have all gone to sh.t in the last six weeks, far steeper than most comparable countries.

Have the "We want our freedumb" demonstrations of a few weeks ago died down yet?

... and Italy was doing so well... back when schools had just gone back, but the cases hadn't started showing up #buildinganarrativefromaweightedrandomnumbergenerator

(although that article discusses the terrible impact of the first wave in the Bergamo area, and that region has just about the lowest incidence of second wave in Lombardy if not the whole of the north... but then it's also estimated that 25% of them have antibodies)

And then we let everything grow exponentially for the whole month of October.

However I do think today's spike is a "date reported" artefact (the seven-day average was 700 per day and falling) since I don't know if we have data by date-of-death. Another thing which seems to have happened in the numbers today is that the screening numbers seem to have been shifted into the main testing figure. (ETA no it's not that, but there seems to have been a typo in the number of cases tested which meant it was suddenly 10 million higher than yesterday; and something has certainly happened which means the screening numbers have gone to zero.)

However the government also wants to impose quite a heavy lockdown for Christmas and New Year.

I'm scheduled to go back to Italy tomorrow morning.

I think it is *probably* reporting noise:

[jimbob]

Sweden is another place where the trend is always downward:

[monkey]

Sweden is another place where the trend is always downward:

...

Not if you read the Guardian (graph 1/2 way down page) clicky

I saw that this morning and thought "every other Sweden death graph I've seen is corrected for day of death, why not this one?", and then promptly forgot about it and went on to read about Hitler getting elected.

[shpalman]

993 deaths reported in Italy today. The previous highest was 971, on 28 March. It seems to have all gone to sh.t in the last six weeks, far steeper than most comparable countries.

Have the "We want our freedumb" demonstrations of a few weeks ago died down yet?

The number for today is a bit lower, but still high: 814. Lower than last Friday though.

[shpalman]

Italy has passed 60,000 official covid deaths; however I think deaths per day have peaked and will come down next week.

Cases have come down a lot and restrictions are being eased - only Abruzzo is still a red zone. We'll see if cases per day keeps coming down or stabilizes at some low level before the Christmas lockdown comes into force (and whether there will be any obvious effect in January of people travelling or meeting on the last weekend before the lockdown). To be honest I don't think there'll be a strong effect, as compared to people going back to work in January which might lead to a third wave in late January / early February.

(I base this on how you could see infections resulting from summer holidays in the numbers in September, but it was work and school starting in mid-September that led to an exponentially-growing second wave in October.)

[shpalman]

Meanwhile I think the UK's cases per day have started going up again.

[lpm]

Or at least stopped going down.

UK official cases hit 15,000 (7 day average) a week ago - and we've basically stayed flat at that level. Given the lags it should have been a solid reduction as this is all from the lockdown period.

[sTeamTraen]

Could you suggest they actually tweet reply that because it would be good to get to a wider public?

The conversation continued, and the author of the DMs gave me permission to blog a cleaned-up version.

[KAJ]

Meanwhile I think the UK's cases per day have started going up again.

Agreed. Note the zero weighted cases are flagged as incomplete by gov.uk and can only be expected to rise, not go down.

speccases.png (22.55 KiB) Viewed 3696 times

[sTeamTraen]

Or at least stopped going down.

UK official cases hit 15,000 (7 day average) a week ago - and we've basically stayed flat at that level. Given the lags it should have been a solid reduction as this is all from the lockdown period.

Last 21 days, difference between mean new cases per day in the past 7 days and the previous 7 days. (That is, negative means fewer cases.) Most recent first (0 is today).

Summary: First few days of lockdown, not much happened. Week 2 of lockdown, numbers went up a bit. Weeks 3 and 4 of lockdown, solid reductions every day (although still not as substantial a percentage as many other countries). Last few days, the descending line has pulled up and is now basically flat.

Code: Select all

 0    91 
 1 -1570 
 2 -2095 
 3 -3105 
 4 -3488 
 5 -3396 
 6 -4950 
 7 -5349 
 8 -5320 
 9 -5746 
10 -5965 
11 -6707 
12 -6985 
13 -5786 
14 -4940 
15 -3413 
16 -2143 
17  -562 
18  2279 
19  2438 
20  2546 
21  2887 
22  1876 
23  2039 
24  1306 
25   126 
26   512 
27    46 
28  -574 
29   304 
30  -287 
31   426 
32   534 
33   182 
34   814 
35  1389 

[sTeamTraen]

Agreed. Note the zero weighted cases are flagged as incomplete by gov.uk and can only be expected to rise, not go down.

Can I just say how glad I am that I told you about being able to get R on your Chromebook? Your ggplot skills are terrific.

[KAJ]

Agreed. Note the zero weighted cases are flagged as incomplete by gov.uk and can only be expected to rise, not go down.

Can I just say how glad I am that I told you about being able to get R on your Chromebook? Your ggplot skills are terrific.

When I first came across ggplot I found it difficult. When I finally started thinking about charts in the same way - mapping data to aesthetics - it became much easier. I still need to consult the help frequently and indulge in trial and error - a bit like having the basics of a language but needing to look up vocabulary and idiom.

[shpalman]

But their CFR is quite low so they probably found a higher proportion of their cases.

I've been comparing the Ticino and Como numbers and there's an odd feature that the declared recovery numbers are very low, maybe they only note people discharged from hospital and don't pay attention to people who get better at home.

The provinces of Como and Varese have the highest rates in Italy right now and also have borders with Switzerland.

I note that the French are checking the borders to stop people going skiing in the mountainous neutral-during-WW2 country #secondmentions because the Swiss have left their skilifts running.

Your country locked down? Come and ski in Switzerland, we don't give a sh.t



cases look like they're going up again

swiss 2020-12-08.png (18.25 KiB) Viewed 3584 times

[lpm]

That photo reminds me of queuing for ski lifts in Italy. It was always a big crush with kids in particular forcing their way through the scrum.

Later we learned why: the Italian ski instructors were using lesson time to teach the kids pushing-in techniques. They were taught to put a ski on top of your skis so you can't shuffle forward into a gap, then they promptly wriggle past and step on the next person's skis. I ended up admiring how good they were at it.

[shpalman]

That photo reminds me of queuing for ski lifts in Italy. It was always a big crush with kids in particular forcing their way through the scrum.

Later we learned why: the Italian ski instructors were using lesson time to teach the kids pushing-in techniques. They were taught to put a ski on top of your skis so you can't shuffle forward into a gap, then they promptly wriggle past and step on the next person's skis. I ended up admiring how good they were at it.

Yeah and their skis are shorter which means they can get through gaps which adults can't.

Little bastards.

Still not as bad as snowboarders having a sit down in the middle of the piste though.

Right now in the Dolomites there's a maximum avalanche warning as well as the closure of all the lifts.

[sTeamTraen]

That photo reminds me of queuing for ski lifts in Italy. It was always a big crush with kids in particular forcing their way through the scrum.

Later we learned why: the Italian ski instructors were using lesson time to teach the kids pushing-in techniques. They were taught to put a ski on top of your skis so you can't shuffle forward into a gap, then they promptly wriggle past and step on the next person's skis. I ended up admiring how good they were at it.

A mathematically-inclined Italian colleague once told me that his working definition of a sphere was "a group of Italians trying to get on a ski lift". A [hollow] sphere is, of course, a body where every point is the same distance from the centre.

[Bird on a Fire]

If little children get past you while you've got a long metal pole with a spike on the end in each hand, you've only got yourself to blame.

They've got to learn.

[sTeamTraen]

Meanwhile I think the UK's cases per day have started going up again.

20964 cases today, highest since 20 November.

Cases in the last 7 days and the nine 7-day periods before that, most recent first:

Code: Select all

113649
 99572
121306
163061
166998
157857
154873
136845
111807
101637