Page 217 of 258

Re: COVID-19

Posted: Tue Mar 23, 2021 4:37 pm
by Grumble
jimbob wrote:
Tue Mar 23, 2021 12:50 pm
Year running from wk31-wk30.
Why not wk1-wk52?

Re: COVID-19

Posted: Tue Mar 23, 2021 6:21 pm
by jimbob
Grumble wrote:
Tue Mar 23, 2021 4:37 pm
jimbob wrote:
Tue Mar 23, 2021 12:50 pm
Year running from wk31-wk30.
Why not wk1-wk52?
Because the 1999-201x data is from the ONS excess winter mortality report (2015-16 for example) and that all cause death data (fig 4 which you can download but will need to rename to open in excel due to filename length limits) starts in wk31, 1999.

Because of this, and because it also separates the first and second waves, I've started then.

I'm on my phone so can't post links easily.

Re: COVID-19

Posted: Tue Mar 23, 2021 11:37 pm
by badger
jimbob wrote:
Tue Mar 23, 2021 12:50 pm
This is a nightingale plot for England and Wales all-deaths data from wk31, 1999 to wk10, 2021. Year running from wk31-wk30.

Image
Thanks for that. It's probably the most moving and apposite thing I've seen all day, on this anniversary of (mis)rememberence.

Re: COVID-19

Posted: Wed Mar 24, 2021 7:14 am
by badger
In among all the understandable and deserved ire directed yesterday at Boris and his cronies commemorating a disaster partly of their own making, James Annan was also getting annoyed at the subtle but telling rewriting of history with respect to Sage and I think it's worth reading in full, because it does remind us it's not all Boris's fault (first lockdown).

http://julesandjames.blogspot.com/2021/ ... aking.html

When the enquiry does finally come it won't be pretty.
I certainly remember the bafflement on here and everywhere else about the delay in locking down, and I wonder if we'll ever get to the bottom of it, or find a way of minimizing those kind of errors in the brains trust advising the country (where they are apparently so out of step with reality).

Re: COVID-19

Posted: Wed Mar 24, 2021 11:31 am
by jimbob
badger wrote:
Wed Mar 24, 2021 7:14 am
In among all the understandable and deserved ire directed yesterday at Boris and his cronies commemorating a disaster partly of their own making, James Annan was also getting annoyed at the subtle but telling rewriting of history with respect to Sage and I think it's worth reading in full, because it does remind us it's not all Boris's fault (first lockdown).

http://julesandjames.blogspot.com/2021/ ... aking.html

When the enquiry does finally come it won't be pretty.
I certainly remember the bafflement on here and everywhere else about the delay in locking down, and I wonder if we'll ever get to the bottom of it, or find a way of minimizing those kind of errors in the brains trust advising the country (where they are apparently so out of step with reality).
Yup - the whole "we're 4-weeks behind Italy" which was obviously wrong at the time. If your model says that, but the deaths data (which is inherently more robust than testing data, which relies on sampling bias) shows us tracking Italy 11 days behind, it's not reality that's wrong.


Also I was basically working from my memories of the government losing control of the 2000-2001 Foot and Mouth outbreak - where (as my dad said as soon* as reports of it being suspected were in the news) one needed to move quickly and impose restrictions before being certain.


*based on his first year working for MAFF coinciding with the 1967 outbreak.

Re: COVID-19

Posted: Wed Mar 24, 2021 12:10 pm
by jimbob
Grumble wrote:
Tue Mar 23, 2021 4:37 pm
jimbob wrote:
Tue Mar 23, 2021 12:50 pm
Year running from wk31-wk30.
Why not wk1-wk52?
It means I can plot the data like this. Top are weekly numbers. Bottom are cumulative numbers. Left are raw numbers. Right are excess (from baseline of median 1999-2021 for that week number)

Image

This also shows winters as continuous, the yellow line is the winter of 2018, which some Covid deniers said this winter was like.

Re: COVID-19

Posted: Wed Mar 24, 2021 3:18 pm
by jimbob
badger wrote:
Mon Mar 22, 2021 2:21 pm
jimbob wrote:
Mon Mar 22, 2021 12:09 pm
badger wrote:
Mon Mar 22, 2021 11:37 am
Apparently earlier does not necessarily mean better, when it comes to UK lockdown (because earlier first spring lockdown equals bigger second waves around winter, which, erm, could be prevented by, erm, earlier lockdowns?) Odd piece, seems to be no paper behind it:

Argues Cambridge Epidemiologist

Sweden, Denmark, Norway say that's bollocks.

And look at London in the first and second waves too
In the article he gives specific reasons for not using Den/Nor and non-Euro as comparators (mainly because of amount of international travel and/or climate), and am happy to go with that for the sake of argument and see what turns up...

London - yep, that seems to run counter to his argument. Sweden, am not so sure - as I see it they had a slightly higher peak of excess death in second wave, so to me that also undoes his position, but he's really talking about countries which did well first, then didn't.
Well Switzerland or Czechia would be the poster children for this. Switzerland with the opening of ski resorts last autumn.

And the thing about Scandinavia and Finland is that you'd have to posit some magic combination of demographics and climate for Sweden to be predicted to be the worst-affected (by 5x more than Denmark, for example) so it's a good argument for the importance of policy. Sweden's large number of single-adult households would suggest it would be amongst the best placed to deal with something like Covid without any policy changes.

See also the different Canadian Provinces

Re: COVID-19

Posted: Wed Mar 24, 2021 5:56 pm
by jimbob
My plot of February and March cases in England according to demographics, from the government's dashboard

Image

Grouped by "mostly retired" "mostly working age" "mostly children and university age"

Cases are rising in school children - especially secondary schools

Re: COVID-19

Posted: Wed Mar 24, 2021 7:11 pm
by Grumble
jimbob wrote:
Wed Mar 24, 2021 5:56 pm
My plot of February and March cases in England according to demographics, from the government's dashboard

Image

Grouped by "mostly retired" "mostly working age" "mostly children and university age"

Cases are rising in school children - especially secondary schools
Yes, but they started rising before schools went back - when mass testing of pupils started in fact.

Re: COVID-19

Posted: Wed Mar 24, 2021 7:36 pm
by KAJ
Grumble wrote:
Wed Mar 24, 2021 7:11 pm
jimbob wrote:
Wed Mar 24, 2021 5:56 pm
My plot of February and March cases in England according to demographics, from the government's dashboard

<snip plot>

Grouped by "mostly retired" "mostly working age" "mostly children and university age"

Cases are rising in school children - especially secondary schools
Yes, but they started rising before schools went back - when mass testing of pupils started in fact.
Hospital admissions are still declining in an almost pure exponential (with a small day-of-week effect) - my earlier suspicion of a flattening wasn't supported by subsequent data.
admits.png
admits.png (13.54 KiB) Viewed 2722 times

Code: Select all

At 20/03 fit = 346.4 with halving time = 17.4 days. That time doubling in 701.7 days 

Analysis of Variance Table

Response: log(Admits)
              Df  Sum Sq Mean Sq   F value    Pr(>F)    
poly(date, 2)  2 10.4723  5.2361 1705.7153 < 2.2e-16 ***
day            6  0.0832  0.0139    4.5146  0.001918 ** 
Residuals     33  0.1013  0.0031                        
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Coefficients:
               Estimate Std. Error t value Pr(>|t|)    
(Intercept)     6.69408    0.02272 294.594  < 2e-16 ***
poly(date, 2)1 -3.22071    0.05618 -57.333  < 2e-16 ***
poly(date, 2)2  0.02417    0.05542   0.436  0.66563    
dayMon          0.08806    0.03200   2.752  0.00955 ** 
dayTue          0.09462    0.03202   2.955  0.00574 ** 
dayWed          0.07673    0.03207   2.393  0.02257 *  
dayThu          0.09033    0.03212   2.812  0.00822 ** 
dayFri          0.01038    0.03219   0.322  0.74918    
daySat         -0.01370    0.03227  -0.424  0.67401    
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 0.05541 on 33 degrees of freedom
Multiple R-squared:  0.9905

Re: COVID-19

Posted: Thu Mar 25, 2021 12:45 am
by Herainestold
The rapid rise of variants and rising infections in the face of vaccine rollout might mean we are now in a permanent pandemic where things will never go back to normal.
Most epidemics disappear once populations achieve herd immunity and the pathogen has too few vulnerable bodies available as hosts for its self-propagation. This herd protection comes about through the combination of natural immunity in people who’ve recovered from infection and vaccination of the remaining population.

In the case of SARS-CoV-2, however, recent developments suggest that we may never achieve herd immunity. Even the U.S., which leads most other countries in vaccinations and already had large outbreaks, won’t get there. That’s the upshot of an analysis by Christopher Murray at the University of Washington and Peter Piot at the London School of Hygiene and Tropical Medicine.
If this is the evolutionary trajectory of SARS-CoV-2, we’re in for seemingly endless cycles of outbreaks and remissions, social restrictions and relaxations, lockdowns and reopenings. At least in rich countries, we will probably get vaccinated a couple of times a year, against the latest variant in circulation, but never fast or comprehensively enough to achieve herd immunity.
Get used to masks and physical distancing and rotating lockdowns, closed borders, fewer pubs and restos, an absence of social events. The future is now.

https://www.bloomberg.com/opinion/artic ... t-pandemic

Re: COVID-19

Posted: Thu Mar 25, 2021 2:26 am
by jdc
I'm far from certain that we'll be able to achieve herd immunity but I'm not sure that in the event of such a failure we'll necessarily have lockdowns, closed borders, and an absence of social events.

If we can protect ourselves from serious illness and death through vaccination, then that might be good enough for the majority of people.

Aside from the tweaked vaccines targeting variants, there are various next gen vaccine candidates being investigated that target parts of the virus that are less likely to mutate than those tricksy spike proteins - also some that have a number of targets so if one mutates there's another 10 targets that have been conserved. (I think I also read something about them being less reliant on relatively short-lived antibodies and making sure they got decent cellular immunity which is longer-lasting.)

Also, aren't the variants tending to have mutations in common? If there are only so many pathways that allow the virus to both be more transmissible / more immune escapey / whatever and remain viable then maybe we won't have so many variants to chase after.

Also, coronaviruses mutate more slowly than your flu viruses and I think that's plenty of straws for me to clutch at now.

Re: COVID-19

Posted: Thu Mar 25, 2021 7:08 am
by shpalman
Grumble wrote:
Wed Mar 24, 2021 7:11 pm
jimbob wrote:
Wed Mar 24, 2021 5:56 pm
My plot of February and March cases in England according to demographics, from the government's dashboard

Image

Grouped by "mostly retired" "mostly working age" "mostly children and university age"

Cases are rising in school children - especially secondary schools
Yes, but they started rising before schools went back - when mass testing of pupils started in fact.
https://twitter.com/dgurdasani1/status/ ... 46724?s=19

Re: COVID-19

Posted: Thu Mar 25, 2021 7:42 am
by badger
shpalman wrote:
Thu Mar 25, 2021 7:08 am
Grumble wrote:
Wed Mar 24, 2021 7:11 pm
jimbob wrote:
Wed Mar 24, 2021 5:56 pm
My plot of February and March cases in England according to demographics, from the government's dashboard

Image

Grouped by "mostly retired" "mostly working age" "mostly children and university age"

Cases are rising in school children - especially secondary schools
Yes, but they started rising before schools went back - when mass testing of pupils started in fact.
https://twitter.com/dgurdasani1/status/ ... 46724?s=19
She undoes a lot of good work there by being wrong about primary kids.

She says they are only being PCR tested, so rise in numbers in not due to increased testing. But primary kids do voluntary LFT. Kits are available to parents to test household twice a week, and some parents I know are using local LFT centres. In both cases you report test results, so they will be included.

Am not saying that the numbers won't go up because of transmission in schools, they will and we aren't doing enough to stop it, just that she's undoing her own analysis in this instance.

Re: COVID-19

Posted: Thu Mar 25, 2021 7:45 am
by bob sterman
A couple of things that explanation misses...

- The big rise in lateral flow testing in schools was really the week beginning March 8th - not March 1st (see below).

- While primary school age children might not all be routinely tested, many are. Moreover, they often share a house with secondary age siblings. This means families have boxes of lateral flow tests at home for use on the whole family. Moreover, a positive lateral flow test in an older sibling often triggers PCR testing of the whole family. Yes - I know officially the rest of the household should only test if they have symptoms - but people often ignore that and book PCR tests for everyone if one person is positive.

pillars.jpg
pillars.jpg (79.14 KiB) Viewed 2659 times

Re: COVID-19

Posted: Thu Mar 25, 2021 10:05 am
by jimbob
Grumble wrote:
Wed Mar 24, 2021 7:11 pm
jimbob wrote:
Wed Mar 24, 2021 5:56 pm
My plot of February and March cases in England according to demographics, from the government's dashboard

Image

Grouped by "mostly retired" "mostly working age" "mostly children and university age"

Cases are rising in school children - especially secondary schools
Yes, but they started rising before schools went back - when mass testing of pupils started in fact.
Wouldn't that possibly lead to more of a one-off change rather than what looks like a steady increase?

Re: COVID-19

Posted: Thu Mar 25, 2021 10:12 am
by lpm
The step-up would have been blurred by lags in reporting etc, making the step look more like a slope?

Re: COVID-19

Posted: Thu Mar 25, 2021 10:41 am
by Sciolus
Is our zillion-pound testing system really so sh.t that they don't publish like-for-like data, separating out different strands of the testing programme?

Re: COVID-19

Posted: Thu Mar 25, 2021 3:11 pm
by jimbob
Sciolus wrote:
Thu Mar 25, 2021 10:41 am
Is our zillion-pound testing system really so sh.t that they don't publish like-for-like data, separating out different strands of the testing programme?
It's not too easy to see as they generally lump both into pillar 2 as far as I can tell, except for the first bit of data
Screenshot 2021-03-25 150500.png
Screenshot 2021-03-25 150500.png (166.22 KiB) Viewed 2596 times
Screenshot 2021-03-25 150615.png
Screenshot 2021-03-25 150615.png (250.71 KiB) Viewed 2596 times
Screenshot 2021-03-25 150644.png
Screenshot 2021-03-25 150644.png (192.46 KiB) Viewed 2596 times

Re: COVID-19

Posted: Fri Mar 26, 2021 4:06 am
by Millennie Al
jdc wrote:
Thu Mar 25, 2021 2:26 am
I'm far from certain that we'll be able to achieve herd immunity but I'm not sure that in the event of such a failure we'll necessarily have lockdowns, closed borders, and an absence of social events.
I don't know if anyone in authority is planing it this way, but it seems to me what will happen is that reopening schools and relaxing lockdown will allow the disease to spread rapidly in the unvaccinated younger people in the population, presumably giving them immunity. Combined with immunity in older people from the vaccine, it's quite possible we will achieve herd immunity. Of course this depends on people spreading the disease quickly enough so that it doesn't die out leaving large pockets of people who have neither caught it nor been vaccinated. Hopefully such people would eventually be vaccinated, though it would be morally dubious to use vaccines on them which could be sent to parts of the world where lots of people are still dying of the disease.

Re: COVID-19

Posted: Fri Mar 26, 2021 9:18 am
by Chris Preston
Herainestold wrote:
Thu Mar 25, 2021 12:45 am
The rapid rise of variants and rising infections in the face of vaccine rollout might mean we are now in a permanent pandemic where things will never go back to normal.
Most epidemics disappear once populations achieve herd immunity and the pathogen has too few vulnerable bodies available as hosts for its self-propagation. This herd protection comes about through the combination of natural immunity in people who’ve recovered from infection and vaccination of the remaining population.

In the case of SARS-CoV-2, however, recent developments suggest that we may never achieve herd immunity. Even the U.S., which leads most other countries in vaccinations and already had large outbreaks, won’t get there. That’s the upshot of an analysis by Christopher Murray at the University of Washington and Peter Piot at the London School of Hygiene and Tropical Medicine.
If this is the evolutionary trajectory of SARS-CoV-2, we’re in for seemingly endless cycles of outbreaks and remissions, social restrictions and relaxations, lockdowns and reopenings. At least in rich countries, we will probably get vaccinated a couple of times a year, against the latest variant in circulation, but never fast or comprehensively enough to achieve herd immunity.
Get used to masks and physical distancing and rotating lockdowns, closed borders, fewer pubs and restos, an absence of social events. The future is now.

https://www.bloomberg.com/opinion/artic ... t-pandemic
I am not at all convinced this or will be the case. Most of the variants of concern arose prior to the vaccines being rolled. In almost all countries, there is still, insufficient people vaccination to put significant selection pressure on the virus.

Variants are most likely to arise when the virus is being rapidly spread through the community, due to high virus populations. If vaccines are able to greatly reduce the population of virus being spread, then new variants will be much less likely.

Probably the biggest risk of selecting vaccine avoiding variants will be through vaccinating part of the population, but allowing the virus to spread freely among the rest. That will provide a scenario of high virus population regularly meeting vaccinated people. It is a scenario we should try to avoid by not stretching out vaccination for too long and not opening up too soon.

Re: COVID-19

Posted: Fri Mar 26, 2021 9:44 am
by shpalman
Chris Preston wrote:
Fri Mar 26, 2021 9:18 am
Herainestold wrote:
Thu Mar 25, 2021 12:45 am
The rapid rise of variants and rising infections in the face of vaccine rollout might mean we are now in a permanent pandemic where things will never go back to normal.
Most epidemics disappear once populations achieve herd immunity and the pathogen has too few vulnerable bodies available as hosts for its self-propagation. This herd protection comes about through the combination of natural immunity in people who’ve recovered from infection and vaccination of the remaining population.

In the case of SARS-CoV-2, however, recent developments suggest that we may never achieve herd immunity. Even the U.S., which leads most other countries in vaccinations and already had large outbreaks, won’t get there. That’s the upshot of an analysis by Christopher Murray at the University of Washington and Peter Piot at the London School of Hygiene and Tropical Medicine.
If this is the evolutionary trajectory of SARS-CoV-2, we’re in for seemingly endless cycles of outbreaks and remissions, social restrictions and relaxations, lockdowns and reopenings. At least in rich countries, we will probably get vaccinated a couple of times a year, against the latest variant in circulation, but never fast or comprehensively enough to achieve herd immunity.
Get used to masks and physical distancing and rotating lockdowns, closed borders, fewer pubs and restos, an absence of social events. The future is now.

https://www.bloomberg.com/opinion/artic ... t-pandemic
I am not at all convinced this or will be the case. Most of the variants of concern arose prior to the vaccines being rolled. In almost all countries, there is still, insufficient people vaccination to put significant selection pressure on the virus.

Variants are most likely to arise when the virus is being rapidly spread through the community, due to high virus populations. If vaccines are able to greatly reduce the population of virus being spread, then new variants will be much less likely.

Probably the biggest risk of selecting vaccine avoiding variants will be through vaccinating part of the population, but allowing the virus to spread freely among the rest. That will provide a scenario of high virus population regularly meeting vaccinated people...
Opening schools and universities to face-to-face teaching, while teacher-age (or lecturer-age, or parent-age) people have been vaccinated but school-age and university-age children/youths haven't, for example.
Chris Preston wrote:
Fri Mar 26, 2021 9:18 am
It is a scenario we should try to avoid by not stretching out vaccination for too long and not opening up too soon.
I have the impression from over here that, apart from the usual gammon-faced arsed on the tory benches, there doesn't seem to be massive impatience to end the UK lockdown. What's the national mood actually like?

Re: COVID-19

Posted: Fri Mar 26, 2021 9:48 am
by shpalman
40,600 people likely caught Covid while hospital inpatients in England
One in seven patients treated for Covid between 1 August 2020 and 21 March got it while in hospital.
Apart from London (11%), 15-19% of covid cases treated in hospital possibly caught it in hospital.

Re: COVID-19

Posted: Fri Mar 26, 2021 11:19 am
by jimbob
badger wrote:
Mon Mar 22, 2021 2:21 pm
jimbob wrote:
Mon Mar 22, 2021 12:09 pm
badger wrote:
Mon Mar 22, 2021 11:37 am
Apparently earlier does not necessarily mean better, when it comes to UK lockdown (because earlier first spring lockdown equals bigger second waves around winter, which, erm, could be prevented by, erm, earlier lockdowns?) Odd piece, seems to be no paper behind it:

Argues Cambridge Epidemiologist

Sweden, Denmark, Norway say that's bollocks.

And look at London in the first and second waves too
In the article he gives specific reasons for not using Den/Nor and non-Euro as comparators (mainly because of amount of international travel and/or climate), and am happy to go with that for the sake of argument and see what turns up...

London - yep, that seems to run counter to his argument. Sweden, am not so sure - as I see it they had a slightly higher peak of excess death in second wave, so to me that also undoes his position, but he's really talking about countries which did well first, then didn't.
I've just plotted London's all cause deaths for 2015-2021, in my usual 4-plot format.


The excess deaths are already as bad for 2021 as they were for 2020

Edited to remove graph at jimbob's request.

Re: COVID-19

Posted: Fri Mar 26, 2021 11:55 am
by jimbob
Arrgh -- my graph was wrong.

it is actually this:
Image