Loading Notifications...

Friday Night Covid Plotting #7

This topic has been archived, and won't accept reply postings.
 wintertree 08 Jan 2021

Just a brief post for now - the additional reporting lag from New Years Day won't work its way through the system for another day or two; I'll be back with more plots when that's happened.

There really isn't a need for me to say much when it comes to interpreting this.  

  • Notes: 
    • The cases curve stops a couple days further back than usual, due to the sampling lag from New Year's Day.
    • I've included a day of what is normally the provisional window on deaths, as that day is high and can only go up more.
    • See plotting #6 for a description of how cases are shunted around to even out festive sampling lag issues.
  • Comments:
    • The apparently long doubling time from deaths is likely due to noise in the data; I don't expect it to last for long; I think the real value is closer to 20-days
    • As always, the most recent week of data on Plot 9e is provisional.
    • Remember that hospitalisations lag behind cases by some significant period of time. 

Other thoughts:

  • We are likely to exceed combine Pillar 1 and Pillar 2 testing capacity soon enough.  Positivity is already rising well past WHO guidelines.  I think it is wise to consider the possibility that the gap between actual infections and cases is widening, and that this gap could be getting worse.

Previous thread: https://www.ukhillwalking.com/forums/off_belay/friday_night_covid_plotting_6-729493

Post edited at 16:50

3
 wintertree 08 Jan 2021
In reply to wintertree:

Latest update to the plot of people in hospital (x-axis) and ITU (y-axis).  Only about half of the fraction of patients are in ITU as were in March/April.  However, exponential growth of cases and hospitalisations is about to exhaust the slack that created.  Some additional ITU capability has been created since the start of the pandemic.  Exponential growth if not checked very soon will exhaust that as well.


1
 wintertree 08 Jan 2021
In reply to wintertree:

Plots on testing levels.  The sudden up-tick at the end of Plot T1 is also clearly visible on the government dashboard and is notable.  

The doubling time for cases has been consistently shorter than for tests conducted for some time - this is another way of saying that positivity is rising.

Testing capacity as given on the dashboard today is ~ 778,000 tests/day.

Post edited at 17:05

1
 Wicamoi 08 Jan 2021
In reply to wintertree:

I've given you a like to offset.

I'm sorry to see the 4 nations plot is not here (but I understand that England is your primary interest and audience).

 wintertree 08 Jan 2021
In reply to Wicamoi:

> I'm sorry to see the 4 nations plot is not here (but I understand that England is your primary interest and audience).

I'll crank out some more plots tomorrow when the New Years Day lag has passed through.  Don't worry, the 4-nations plot will be back.

Post edited at 17:50
1
In reply to wintertree:

Thanks again for these. 

1
 mik82 08 Jan 2021
In reply to wintertree:

Thanks, although the graphs are still terrifying.

I think this is unfortunately going to be the darkest time of the pandemic in the UK. Even just based on current increases it looks like we'll get to 50% of all acute hospital beds in England being occupied by covid patients and treble the number of current deaths per day. Grim doesn't do it justice.  

 elsewhere 08 Jan 2021
In reply to wintertree:

> Latest update to the plot of people in hospital (x-axis) and ITU (y-axis).  Only about half of the fraction of patients are in ITU as were in March/April.  However, exponential growth of cases and hospitalisations is about to exhaust the slack that created.  Some additional ITU capability has been created since the start of the pandemic.  Exponential growth if not checked very soon will exhaust that as well.

Thanks. That's a very good way of presenting data. Better than I have seen from government or in the media.

 wintertree 08 Jan 2021
In reply to mik82:

> I think this is unfortunately going to be the darkest time of the pandemic in the UK. Even just based on current increases it looks like we'll get to 50% of all acute hospital beds in England being occupied by covid patients

This seems likely, yes.  Except it is not clear to me that 50% of all acute beds can be made available to Covid in time.  

> and treble the number of current deaths per day. Grim doesn't do it justice.  

I make it ~1,400 per day or 2.3x those most recently seen in the plots above.  

As you say, that's all based on the situation to date.

  • I think tomorrow's update will show that most or all regions are some way away yet from hitting R <=1.  
  • This is of course also assuming that there are sufficient medical facilities for all people who will require hospitalisation.
 wintertree 08 Jan 2021
In reply to elsewhere:

> Thanks. That's a very good way of presenting data. Better than I have seen from government or in the media.

Thanks.  Florence Nightingale, John Snow and others showed the power of visualisation in understanding and then combating public health and epidemiological issues over a century ago.  If they looked at a modern weekly PHE report on Covid, made with computers to do the plotting rather than pen, paper and a compass, I think they would be appalled at how so much data is so poorly presented.

Of course, going to the effort to present data as clearly as possible is pointless if the decision makers look at data and science as just another political opinion to be dismissed if they don't like it. 

Post edited at 21:21
 Blunderbuss 08 Jan 2021
In reply to mik82:

> Thanks, although the graphs are still terrifying.

> I think this is unfortunately going to be the darkest time of the pandemic in the UK. Even just based on current increases it looks like we'll get to 50% of all acute hospital beds in England being occupied by covid patients and treble the number of current deaths per day. Grim doesn't do it justice.  

Isn't the massive problem that London and the South East is going to run out of beds I.e it can't be stopped now ....here in Yorkshire we have more spare capacity but how easy it it for hundreds or even thousands of patients to be shipped around the country? 

 mik82 08 Jan 2021
In reply to Blunderbuss:

Yes - I really hope not but London/SE looks like a Lombardy type scenario now. There's no way of transporting 1000s of patients cross-country as all the ambulance capacity is already tied up. Even if there is capacity in Yorkshire, admissions are still increasing nationwide, so how long is that going to last?

 Andy Johnson 08 Jan 2021
In reply to wintertree:

Given how utterly terrifying the situation is, it seems somehow inappropriate to thank you for this work. But thank you anyway. This is quality analysis.

> We are likely to exceed combine Pillar 1 and Pillar 2 testing capacity soon enough.

It also looks like.hospital capacity in London will saturate soon, if it hasn't already. And those curves don't look set to flatten anytime soon.

So. Staring into the abyss for a moment, how do people see this playing out? Field hospitals? People left to fend for themselves? Is there any sign of a plan other than vaccination (which is currently at less than 2% of the population)?

Anything?

Post edited at 23:02
 WaterMonkey 08 Jan 2021
In reply to wintertree:

Given how grim and terrifying the situation looks, particularly for London and the South East, would you not expect the government to be communicating that so that people may be more inclined to follow the rules? Or do you think they will wait a few weeks and then lock us down ‘slightly’ harder?

 mik82 08 Jan 2021
In reply to WaterMonkey:

I don't think there's much more they can do. There's apparently some ads going out encouraging people to stay at home. For London,  SE and E England they've just got to hope that the the effects of Tier 4 and schools breaking up for xmas in December start to feed through into admissions reducing but there's no signs of that as of today.

Post edited at 00:00
In reply to Andy Johnson:

> It also looks like.hospital capacity in London will saturate soon, if it hasn't already. And those curves don't look set to flatten anytime soon.

Leak suggesting London hospitals will run out of capacity within 2 weeks.

https://www.hsj.co.uk/acute-care/exclusive-london-will-be-overwhelmed-by-covid-in-a-fortnight-says-leaked-nhs-england-briefing/7029264.article

In reply to mik82:

There's stuff they could do to lock down harder...

Force more workplaces to close even if people can't work from home - everything apart from essential services. Construction sites for example.

French style system where you have to make a declaration before leaving home and the police check it. Would need the army as well as not enough police!

Total lockdown - remove exercise as a reasonable excuse and actually enforce it plus children not allowed outside at all except young children with parents. Spain did that.

It sounds unlikely but give it another couple of weeks...

In reply to mik82:

> I don't think there's much more they can do. There's apparently some ads going out encouraging people to stay at home. For London,

There is a lot more they can do. They will have graphs the same as Wintertree so why not bung them on TV. Those curves are shocking and seeing a visual representation like that can only help. 

The current yellow background with a voice over saying 1 in 3 people are not showing symptoms is ineffectual crap. Much like the government. 

 davidalcock 09 Jan 2021
In reply to wintertree:

Despite expecting this, I suddenly, genuinely need a poo. Looks worse than I feared. Jesus.

 Punter_Pro 09 Jan 2021
In reply to wintertree:

Great work as always, Wintertree.

Things really have got quite dire as of this week with worse to come I fear, my partner is one of the outpatients managers at Addenbrookes, Cambridge and works with a lot of top consultants in the country.

They have had to cancel most appointments this week because they have had to send the consultants down to London to work on the front line, something she hasn't seen in her 9 years of service....

John Burn-Murdoch has also been doing some good work with visualizing the data against previous years, these video graphs are really good for showing the skeptics how bad things are getting.

https://twitter.com/jburnmurdoch/status/1347200811303055364

Post edited at 06:49
 ablackett 09 Jan 2021
In reply to wintertree:

Plot 7.2e on Hospital occupancy is absolutely terrifying. Given yesterdays record case numbers we can expect that to keep going up for at least the next couple of weeks.  That graph should be on every tv in the country when people turn it on this morning.

Stories of 50%+ of pupils in some primary schools show just how ridiculous the current 'lockdown' is. 

 wintertree 09 Jan 2021
In reply to ablackett:

> Stories of 50%+ of pupils in some primary schools show just how ridiculous the current 'lockdown' is. 

Our local university, with whom I am not yet fully separated, has granted *all* staff critical worker status with messaging from HR that does not recognise the - government determined - risk of transmission into the household by children nor the approaching crisis.  I have used all formal route open to me on this and on Monday I may be sending a wide and career ending email with my full views.  My dept are telling staff not to use it and that others will cover for them if childcare is a problem.  Others are not.

In reply to Punter_pro:

Interesting observation re: staffing, thanks.

> John Burn-Murdoch

He does great work. Thanks for that link - it makes you wonder how anyone can still be doubting this.

 wintertree 09 Jan 2021
In reply to Dax H:

> There is a lot more they can do. They will have graphs the same as Wintertree so why not bung them on TV. Those curves are shocking and seeing a visual representation like that can only help. 

They have a *lot* more access to data than me.  On Dec 14th shortly after the new variant was announced I put together a set of noddy maps on UKC that suggested the failure of lockdown was propagating out over time from Kent.  On Dec 15th I wrote all this up and sent it to everyone I could think off; it has made it so some central labs and other interesting places.  It was clear to me then that the new variant continued to grow exponentially during Tier 4 control measures and by a few days later it was clear that the new strain was everywhere including Tier 1, 2 and 3 areas.

The government had access to more data than me.  They had “pre-discovery” data from the lighthouse labs that was reprocessed out of the archives.  The motor of all pennies must have dropped about Medway and Kent remaining exponential during Tier 4 when they saw this data.

The LSHTM pre print came out on Dec 23 and was astoundingly clear in its analysis and predictions.

The NERVTAG minutes giving a value of about 70% increased transmissibility are dated Dec 23.

When did everywhere move to a tighter tier level than the known-insufficient tier 4? Jan 4th.  After sending children in to schools for one day.

Dec 14th to Jan 4th is 21 days.

Post edited at 08:06
 wintertree 09 Jan 2021
In reply to Andy Johnson:

> So.

Quite.  I don’t like to think about it, and there is nothing I can contribute to that side of things.  As I understand it, there are criteria for triaging access to treatment and that criteria can change.

In reply to WaterMonkey:

> Or do you think they will wait a few weeks and then lock us down ‘slightly’ harder?

I’ve a bad track record at predicting this government’s actions.  I think the next 8 weeks are wide open to possibilities.  Depending on how the media run with this, the cabinet may not outlast the crisis.

Post edited at 08:17
In reply to wintertree:

Yep once more proving how useless our powers that be are. The kids back to school for 1 day was a master stroke, let people mix at home on Christmas day (plus I'm sure lots of rule breaking all week) then let all the kids mix for 1 day just over a week later. This is the sort of thing I would do of my intention was to spread it not control it.

Screw the government, I will carry on following your data and making up my own mind on things. 

In reply to wintertree:

> > Stories of 50%+ of pupils in some primary schools show just how ridiculous the current 'lockdown' is. 

> Our local university, with whom I am not yet fully separated, has granted *all* staff critical worker status with messaging from HR that does not recognise the - government determined - risk of transmission into the household by children nor the approaching crisis.  I have used all formal route open to me on this and on Monday I may be sending a wide and career ending email with my full views.  My dept are telling staff not to use it and that others will cover for them if childcare is a problem.  Others are not.

Is it not possible that you have developed a cough and temperature? And can't taste pot noodles anymore, even the really spicy one?

The game goes both ways.

 wintertree 09 Jan 2021
In reply to bouldery bits:

> Is it not possible that you have developed a cough and temperature? And can't taste pot noodles anymore, even the really spicy one?

For sure, but the situation and the risk were not spelt out to several thousand other staff who were told they could put their children in school - and I include personal, school staff and societal risk in that which was omitted.

I shouldn’t be surprised after their decisions to date.

 minimike 09 Jan 2021
In reply to wintertree:

I don’t know which university you work for but it’s been a common theme across the sector for the last decade that no one in management cares about staff, students or academic credibility. (Unless it happens to affect the bottom line)

I left 8 yrs ago for a combination of these reasons.

 Si dH 09 Jan 2021
In reply to wintertree:

Thanks as usual. Instinctively I agree you would probably be right about testing missing an increasing number of cases. There is additional data available for Liverpool (which I think should be representative for other areas experiencing a rapid rise) because of the LFT testing and the data the council puts up here:

https://liverpool.gov.uk/covidcases

I did a bit of analysis using data from the link above and from the data about total number of tests conducted at UTLA level on the usual dashboard, but unfortunately my tablet ran out of battery. Here is the summary: measured asymptomatic positivity in Liverpool has increased from 1.0% to 3.0% from the week ending 20/12 to the week ending 05/01. The upward trend using all available data in between those dates is not noisy. The total population is ~ 500,000 so if you extrapolated that out - fairly wild but ROM? - it equates to an increase from 5000 to 15000 asymptomatic people in the city (UTLA) who would have been positive if tested in those weeks. Given LFT has a high false negative rate, the true numbers may be higher but the% change about right.  In the same period according to the link above measured symptomatic positivity has increased from 5.9% to 18.9% and the total number of positive PCR tests increased from 618 to 3453 (with LFT, the total positive cases found in Liverpool went from 844 to 4450.)

So we can see that the total number of cases detected through PCR changed from ROM 618/5000 = 12% to ROM 3453/15000 = 23%. 

This would actually suggest that the PCR programme is missing a lower % of cases, but a higher absolute number, as infection rates increase. I'm trying to write this while making family breakfast so might have made an error or missed an alternative explanation (I'm confident in the raw case and positivity data though.)

Post edited at 09:29
 wintertree 09 Jan 2021
In reply to Si dH:

A very interesting set of calculations; thanks.  

It does make me wonder - how well targeted is PCR testing?

Given the large number of symptoms not on the list for testing, and the number of people that do not display the symptoms on that list, it could be viewed as a mix of random and symptomatic inputs streams for testing; with a rising prevalence of infections, the positivity of the random component of the input stream would rise, but not of the symptomatic component of the input stream.

Hopefully the ONS will soon produce a new estimate of daily infections to give another comparator.

 wintertree 09 Jan 2021
In reply to minimike:

> I left 8 yrs ago for a combination of these reasons.

I hope you find your choice has been rewarding.

I’m on a year’s leave to focus on something different; I look at it more as a suspended sentence...  The response of sensor management across the sector to this crisis will I think be a powerful help to businesses and other organisations looking to make technical/scientific appointments.  I hope to be hiring some of them soon enough.

As well as senior management issues sector wide, there is a creeping privatisation going on under many different guises and that creep is starting to look more like a brisk walk to the starting blocks.  There is a lot more change to come to the sector before I am on the parental/student end of the admissions process but with the direction it is going, I am thinking of alternatives including the Nordic systems or part time work and OU distance learning study.  I’m certainly not going to be raising Jr with the expectation that going to a UK university is the only or necessarily best option.

Post edited at 09:46
 minimike 09 Jan 2021
In reply to wintertree:

> I hope you find your choice has been rewarding.

absolutely. The best career decision I ever made. I’m basically an academic in the nhs now, having far more real world impact and without the teaching or financial pressures of HE (and on a permanent contract).

> I’m on a year’s leave to focus on something different;

I doubt you’ll go back tbh. I suspect you shouldn’t! Best of luck with the business.

 Offwidth 09 Jan 2021
In reply to minimike:

Anyone else here think most people have no idea what is coming?

I know case rates are not as accurate as they could be over xmas and real rates are different (probably worse)  but the indications are daily deaths are pretty much locked in to at least double and that's assuming the seriously ill can get the same level of care they have now. Hospitalisations look locked in to go up a third assuming an immediate step change at lockdown  (lockdown measures normally take a week or more to have a proper affect). If hospitals systems start to crumble death rates of the seriously will increase fast (around three times from NY and Lombardy data) and large numbers of people with other serious illness will die needlessly.  I've just been chatting to my parents who are being really good (and my dad got his second jab) but they say lots of people are worried because an obvious minority in their sizable village are still visiting each other in houses and some of them are publicily call the situation a hoax.

 mik82 09 Jan 2021
In reply to Offwidth:

>Anyone else here think most people have no idea what is coming?

Yes. It's not helped by articles in papers like the Daily Mail. Headline today "Boris tries to scare us into lockdown". The same papers that have published things saying hospitals are no busier than usual for winter. A lot of people just don't believe the news on what's actually happening.

We reported the 2nd highest absolute number of cases and deaths in the world yesterday, only behind the USA. Only 4 Eastern European countries have a higher 7 day average death rates/population currently - and a quick google reveals a story about the main crematorium in the Czech Republic running out of capacity despite running 24hrs per day.

 Blunderbuss 09 Jan 2021
In reply to Offwidth:

> Anyone else here think most people have no idea what is coming?

Yep, doesn't help with the Public Health Director of London doesn't answer a question from the BBC about what will happen if hospitals run out of beds and instead goes on about increasing capacity....anyone watching him could have had the impression it's not so bad, we can just increase capacity.

 wintertree 09 Jan 2021
In reply to Blunderbuss, mik82 and Offwidth:

> anyone watching him could have had the impression it's not so bad, we can just increase capacity.

He's right. 

As long as he can increase capacity exponentially for at least another 2-3 weeks.

 wintertree 09 Jan 2021
In reply to minimike:

> absolutely. The best career decision I ever made. I’m basically an academic in the nhs now, having far more real world impact and without the teaching or financial pressures of HE

The teaching is the part I'm missing the most, even though it invariably consumed far more time than contracted.  It is always good to see someone take their skills to somewhere that can cause direct and recognisable effect on the wider world.

> (and on a permanent contract).

[I deleted a rant the social media spotting brigade on the HR payroll would get all stroppy about.].  

Good.

> I doubt you’ll go back tbh. I suspect you shouldn’t! Best of luck with the business.

Thanks!  It is not an easy time to be trying to get a business off the ground and grow it, but it is better than sitting around complaining, and opportunity knocks.

Post edited at 12:02
In reply to Blunderbuss:

> anyone watching him could have had the impression it's not so bad, we can just increase capacity.

Didn't you get the memo? There are tens of thousands of doctors and nurses plus intensive care beds sat in a warehouse ready to be deployed at a moments notice

 wintertree 09 Jan 2021
In reply to Offwidth:

> but they say lots of people are worried because an obvious minority in their sizable village are still visiting each other in houses and some of them are publicily call the situation a hoax.

Yes.  We seem to leave it too late and then over legislate for under compliance.  I mostly avoid people lately, but a conversation with an increasingly incoherent and angry neighbour suggests that the lunatic views seen on here from the pop up misinformation troll are somewhat widespread these days.

 wintertree 09 Jan 2021
In reply to thread:

I am not confident that today’s data release has fully resolved Jan 4th and 5th, which are needed to even out the sampling low from New Year’s Day on Jan 1st and the weekend immediately after.  Tomorrow’s update should be enough and then I’ll run the rest of the plots out.

The “tests done” data on the dashboard is rising even more alarmingly than when I noted it yesterday.  I’d not really looked at this before; it’s interesting because it’s reporting is only lagged by about 1.5 days and, being demand driven, it somewhat represents the state of cases - at least short term trends in testing presumably correlate to short term trends in symptomatic infections.  Perhaps the increased media coverage is driving more people to go for tests.....  Some of it may also be displaced demand from the most recent bank holiday weekend.  

 neilh 09 Jan 2021
In reply to wintertree:

Any current stats on the age range for deaths and people also age range of people in hospital ?

 jonny taylor 09 Jan 2021
In reply to wintertree:

> The “tests done” data on the dashboard is rising even more alarmingly than when I noted it yesterday.  I’d not really looked at this before; it’s interesting because it’s reporting is only lagged by about 1.5 days and, being demand driven, it somewhat represents the state of cases - at least short term trends in testing presumably correlate to short term trends in symptomatic infections

I've never been sure if/how that can be interpreted on its own - I feel I don't know enough about who all the [majority of] people are who are getting *negative* tests. How many are routine screening, how many are individuals who genuinely think they have it, how many are individuals who think they are getting a "precautionary test", etc. My half-paying-attention impression is that the % positive seems to fluctuate fairly independently of the absolute number of positive tests, at least in Scotland...

 wintertree 09 Jan 2021
In reply to neilh:

> Any current stats on the age range for deaths and people also age range of people in hospital ?

The last demographic hospitalisations data I pulled down from the API was on Jan 1st and is in plotting no 6 [1].  The subject came up on a thread in The Pub this week where someone had grossly misunderstood by 7x the admissions numbers for working aged people in their quest to convince us to “let it rip”.  The calculation I did for that from the dashboard source [2] is that about 48% of admission on that day were under 65 years of age.  My questionable demographic model fitting suggests about 42% were under 60 years of age but the data is not released with that level of detail.  A plot I haven’t shared suggests this ratio had been relatively stable for weeks up to that point.

I think there is a breakdown of deaths by age out there somewhere but I haven’t used it - and would happily take a link to it.  I think it’s in some NHS spreadsheets that can be downloaded?  

[1] https://www.ukhillwalking.com/forums/off_belay/friday_night_covid_plotting_6-729493?v=1#x9366625

[2] https://coronavirus.data.gov.uk/details/healthcare?areaType=nation&areaName=England

 wintertree 09 Jan 2021
In reply to jonny taylor:

> I've never been sure if/how that can be interpreted on its own - I feel I don't know enough about who all the [majority of] people are who are getting *negative* tests. How many are routine screening, how many are individuals who genuinely think they have it, how many are individuals who think they are getting a "precautionary test", etc. My half-paying-attention impression is that the % positive seems to fluctuate fairly independently of the absolute number of positive tests, at least in Scotland...

Indeed.  I *think* most of the routine testing is pillar 1 and the data is broken down by pillar.  This sudden rise looks to be in pillar 2 which is demand driven - although that’s a funny demand being a mix of symptomatic testing and LFT selected asymptomatic testing (this I think has been small in number though).   Pillar 2 may also include non state regular screening?

I think precautionary submission for pillar 2 testing could be rising now as more and more people get spooked.  Which it shouldn’t if false negative rates, particularly of symptom free people, were well communicated along with the implications for being symptomatic but testing negative.

Really though it’s all flights of fanciful interpretation on my behalf.  I certainly don’t know enough to pull the information in with other sources to make more current quantitative readouts of the situation.

I look at it a bit like a Native American listening to a railway rail; they don’t really know what the sound means but they know there’s a train coming.  This one doesn’t sound like it’ll pass quickly.

Post edited at 17:01
 Tallie 09 Jan 2021
In reply to wintertree:

A couple of questions for the more knowledgeable (which includes but is not limited to WT):

It's my understanding that we detected the new variant and are able to trace it because the PCR test the UK uses looks at 3 sections of the viral DNA/RNA (not exactly sure which) and the new variant changed 2 of the 3 such that the PCR test no longer detected them.  This allows us to track the new variant across the country.  What happens if the virus mutates again such that the final detectable section changes?  Has the PCR test already been changed to look for sections of DNA/RNA which remain unchanged in both variants of the virus?  Is such a change a manufacturing issue or just a change to the processing equipment?

I  ask as I've had a number of colleagues display 'classic' CV symptoms but test negative in the last week; clearly a cold / flu is the most likely answer but...

Secondly, given we now appear to be sleep walking into a 'let it rip' scenario through a combination of government incompetence, public apathy / selfishness and (arguably) western individualism, at what point does some sort of herd immunity effect begin to limit the R rate?  65-75K cases per day rising exponentially + vaccinations + those infected in the first wave (I seem to recall an estimate of 8M?) must begin to have an impact at some point?

 mik82 09 Jan 2021
In reply to neilh:

You can pull these stats off the dashboard. Last 7 days deaths in age ranges:

  • 0-9:      20
  • 10-19:  1
  • 20-29:  6
  • 30-39:  29
  • 40-49: 60
  • 50-59: 165
  • 60-69: 425
  • 70-79: 938
  • 80-89: 1597
  • 90+:     863

Note that 18 of the  0-9 age group deaths are in the 0-4s. Not sure what is going on here - could be deaths with an incidental positive test within 28 days rather than direct deaths from covid - this group would include perinatal mortality.

Post edited at 17:15
 mik82 09 Jan 2021
In reply to neilh:

Admissions last 2 weeks:

  • Under 65: 15924
  • 65-84:       16312
  • 85+:          8076
 wintertree 09 Jan 2021
In reply to Tallie:

Hopefully one of the biologists will come along to answer the PCR stuff.  As interesting as it is, I don’t think it’s challenging to adapt the “primers” used to lock the PCR in to this or that variant.  Although PCR is one of the dark arts of molecular biology.

I think the new variant was detected by sequencing from the estuary problem areas, and then “pre-discovery” data was processed out of the lighthouse labs quality control records and/or raw data.

Herd immunity starts the moment one person is immune.  It is sufficient to stop the spread when - statistically speaking - (the number of people you would have transmitted the virus to without immunity) x (fraction of people immune) is less than 1.  R of the old variant was about 3 with no control measures, so we’d need (66.6% immunity x 3 people) to get R < 1.  

The bad news is that the new variant has increased R by a lot, although I’ve seen no estimates of the R for a “no control measures” scenario.  I would guess we need 80% immune to let all control measures be dropped.

There is a discussion of this on plotting #6; the view I think is that we’re nowhere near herd immunity levels yet , even making the very dodgy assumption that mild/asymptomatic infections confer useful levels of  immunity let alone against high viral loads.  If you took the most infected 5-year age range from the most infected part of the country and isolated them, they’d still probably have R>1 with this new variant and Tier 4 as it was until Jan 4th.

I don’t buy the estimate of R on the government dashboard right now - at all - but my gut is that we need about 40% immune to start significantly moderating the growth currently seen - but that will hopefully reduce with the effects of the lockdown measures introduced on Jan 4th.  At the current rate of infection and immunisation including time to take effect, that could be > 12 weeks away.  These are really crude estimates.

Both the vaccines and past infections increase immunity.  Perhaps there is scope to be smart about minimising the intersection there to stretch the vaccines out more.  The vaccines yet to be given don’t work quickly enough to change much about the next 4-6 weeks perhaps.m, unless there is a step increase in vaccination rates almost immediately.  

> clearly a cold / flu is the most likely answer but...

Not all infections are PCR detected.  If flu is rising that could explain some of the apparent spike in testing rates.  Good news / bad news that with the pressure on hospital beds.

Post edited at 17:43
 minimike 09 Jan 2021
In reply to wintertree:

Typo?

 should be R0*(1-f_immune), surely?

 Si dH 09 Jan 2021
In reply to Tallie:

I thought (?) it was one of the three PCR targeted parts of the covid make-up that was missing in the new variant, rather than two of the three. They call it S gene target failure (SGTF) (although the mutation that causes it is a deletion of two amino acids.) It's discussed in this report and is only picked up by a proportion of our PCR test sites, not all of them:

https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/950823/Variant_of_Concern_VOC_202012_01_Technical_Briefing_3_-_England.pdf

I have a supplementary question if anyone knows. Would we expect this South African variant if present in the UK to be picked up as an SGTF case too? Does it have the same missing gene pair?

Thanks

Post edited at 19:16
 wintertree 09 Jan 2021
In reply to minimike:

> Typo?

>  should be R0*(1-f_immune), surely?

Brainfart rather than typo.  Good catch.  For my post’s noddy example, the 66% immunity is correct but the calculation should obviously use the remaining 34% not immune!  To fix myself:

> so we’d need 66.7% immunity so that (33.3% non-immune x 3 people) gives R < 1.  

 minimike 09 Jan 2021
In reply to Si dH:

You’re right, it’s one of 3 PCR targets which fail with the UK variant due to mutation del69-70 in the spike (S) gene. Not all our labs use the S target, so not all can detect this variant.

The SA variant doesn’t have that deletion so AFAIK there is no S gene target failure signature and it can currently only be found through sequencing. 
 

wiki has details of the mutations. 
 

https://en.m.wikipedia.org/wiki/501.V2_variant

 bruxist 09 Jan 2021
In reply to Tallie:

On the herd immunity question: worth remembering to keep an eye on the one place where the virus was unquestionably left to let rip last year: Manaus, in Brazil. Despite estimates that 76% of the population were infected inbetween March and October, Manaus is now deep into a second wave:

https://www.telegraph.co.uk/global-health/science-and-disease/hopes-herd-immunity-wane-manaus-city-declares-state-emergency/ 

(apologies for the Telegraph link but it's the most readable recent English-language account I can find).

 Blunderbuss 09 Jan 2021
In reply to bruxist:

Surely that means either it was a lot lower than 76% infected or immunity has been lost in a large proportion of those previously infected? 

 wintertree 09 Jan 2021
In reply to Blunderbuss:

> Surely that means either it was a lot lower than 76% infected or immunity has been lost in a large proportion of those previously infected? 

Or that some of that 76% who had only mild, asymptomatic infections didn't get much meaningful immunity from it.

 wintertree 09 Jan 2021
In reply to minimike:

> wiki has details of the mutations. 

I followed the link, but my peril sensitive sunglasses kicked in.

 RobAJones 09 Jan 2021
In reply to bruxist:

That was my understanding, I think it was based on this report.

https://science.sciencemag.org/content/early/2020/12/07/science.abe9728

 bruxist 09 Jan 2021
In reply to Blunderbuss:

I just don't know. I hope the estimates were wrong, but they'd have to be out by quite a large margin; I fear that natural immunity didn't last longer than 6 months. Either way it indicates that we need vaccine-induced immunity to last much longer than that, or we need to hope that Covid becomes a seasonal problem.

 RobAJones 09 Jan 2021
In reply to Blunderbuss:

from the report

In conclusion, our data show that >70% of the population has been infected in Manaus approximately seven months after the virus first arrived in the city. This is above the theoretical herd immunity threshold. However, prior infection may not confer long-lasting immunity (30, 31). Indeed, we observed rapid antibody waning in Manaus, consistent with other reports that have shown signal waning on the Abbott IgG assay (14, 32). However, other commercial assays, with different designs or targeting different antigens, have more stable signal (14), and there is evidence for a robust neutralizing antibody response several months out from infection (33). Rare reports of reinfection have been confirmed (34), but the frequency of its occurrence remains an open question (35)

 wintertree 09 Jan 2021
In reply to bruxist & RobAJones:

Fading antibody levels does not necessarily imply fading immunity; after a while the information persists only in a specialised kind of memory cell, ready to build more antibodies if needed.  Some promising pre prints on this were shared a couple of weeks back.

I wonder how many of the infections now circulating in Brazil are being sequenced?  Tentatively I don’t think it’s looking very good for global travel for the foreseeable; the number of variants is rising over time and it feels like immune escape - proved to be possible - is only one incoming traveller away.  Until we have essentially achieved near elimination we can’t hope to track to track down imported variants until it’s too late.  I’m not convinced one negative PCR tests is anything like good enough, NZ’s managed isolation and quarantine approach seems saner than ever.

 minimike 09 Jan 2021
In reply to wintertree:

Indeed. I think it’s safe to say it’s probably rather widespread given it will rarely be detected (only when sequencing is done).

 bruxist 09 Jan 2021
In reply to wintertree:

Are you thinking that this second wave in Manaus might be down to a variant? I'm very much hoping it isn't that. But I guess Manaus will be the first place to start reporting significant rates of reinfection if so.

 minimike 09 Jan 2021
In reply to bruxist:

I don’t think anyone is suggesting that. It’s possible of course but as far as I’m aware there’s no evidence of it at all.

 wintertree 09 Jan 2021
In reply to bruxist:

> Are you thinking that this second wave in Manaus might be down to a variant?

It was a much more wide open speculation than that.  

  • If this is simply "bog-standard" covid at work, every country that sustains high case rates for a long time, as Brazil and the UK are, presents a risk as a source of more new variants.   That many places are having widespread infection without much sequencing is not a happy thought.  It's another reason I think we need MIQ on incoming travellers, to just a single negative Covid test.  
  • If people there have a bit more social contact than in the UK at present, and the new variant sweeping the UK is also present there, a lower bound on the past infected fraction in Manaus might not be enough for herd immunity against the new variant.  The more relevant piece of information is what the final size of their new wave of infections is, rather than that it is happening.  This could be picked up by their RT-qPCR testing with the right choice of primers, as with the UK.

>  But I guess Manaus will be the first place to start reporting significant rates of reinfection if so.

I imagine that it's a site of great interest to the professionals at the moment.

 bruxist 09 Jan 2021
In reply to minimike:

Brazilian press do report the UK variant confirmed in São Paulo; plenty of other variants in Manaus but not the UK one yet: https://www.acritica.com/channels/coronavirus/news/analise-de-genoma-da-covid-de-dezembro-ira-identificar-se-nova-linhagem-ja-esta-no-am. I'm aware evidence will trail theory. I'm less interested in whether or not there's evidence at this point.

In reply to wintertree:

> A very interesting set of calculations; thanks.  

> It does make me wonder - how well targeted is PCR testing?

> Given the large number of symptoms not on the list for testing, and the number of people that do not display the symptoms on that list, it could be viewed as a mix of random and symptomatic inputs streams for testing; with a rising prevalence of infections, the positivity of the random component of the input stream would rise, but not of the symptomatic component of the input stream.

> Hopefully the ONS will soon produce a new estimate of daily infections to give another comparator.


Which if I understand correctly what you mean - ( a greater % of those with compatible symptoms actually have the disease now, and so test positvity rate will increase) - fits with patterns of positive test results in other diseases that I've seen in the past - I'm not quite sure why a higher % positive test rate in and of itself is seen as a bad thing (the high number of cases being detected clearly is!)

There will certainly be behavioural effects on those going for tests associated with the current situation - I've developed a dry cough this weekend and am now self isolating after getting a test - probably in the summer I'd have felt it extremley unlikely to be Covid and not felt much pressure to get a  test, now it seems more likely and also potentially with the new variant around more consequential (we have our students back on monday and I do a lot of fairly close contact teaching).

Happy to report in the Bristol area at least loads of test capacity, and a very slick opperation - over 1500 tests available to me within half an hour of reporting on the NHS App on a saturday evening.

 wintertree 09 Jan 2021
In reply to Dr.S at work:

> Which if I understand correctly what you mean - ( a greater % of those with compatible symptoms actually have the disease now, and so test positvity rate will increase) - fits with patterns of positive test results in other diseases that I've seen in the past -

I think we are on the same page.   I made a sketch (below) of what I mean.  This is purely illustrative and the proportions do not convey information beyond making sure we mean the same thing. (Said for other readers).  The sketch represents people who go for testing, and doesn't include the false negative rate of testing.

  • The big circle is people who have one or more of the 3 symptoms that are listed as reasons for testing in the UK.  
  • The small circle with the bold outline is people who have Covid.
  • Blue shading is people with one or more of the "3 symptoms" but not Covid
  • Red shading is people with Covid but without any of the "3 symptoms"
  • Black is people with the one or more of the "3 symptoms" and with Covid.
  • Positivity (assuming all tests results are correct) is the black area expressed as a fraction of the area of the large circle.

The top half of the diagram shows a time when Covid is low prevalence.   The bottom half of the diagram shows a time with more Covid prevalence.    As prevalence rises the number of people with any of the 3 symptoms and with Covid rises, meaning that positivity rises.

Excepting perhaps taste, the "3 symptoms" are clearly a lot less specific than Covid; likewise we know a lot of people with Covid aren't symptomatic.  

> I'm not quite sure why a higher % positive test rate in and of itself is seen as a bad thing (the high number of cases being detected clearly is!)

Yes; the guidance of a fixed positivity as being appropriate is hard to understanding in all the different possible permutations of symptomatic and asymptomatic testing.  Where somewhat generic symptoms are used to pre-filter the input to testing as with pillar 2, I don't see much sense in it now I sit and think about it.

Post edited at 21:36

In reply to wintertree:

Yes - thinking about it a bit more, given the symptoms overlap with common disease conditions, which will have fairly well known incidence - let’s say A, then if your test positive rate is low  then it could be taken to suggest you have an effective screening program in place, and that the incidence of the disease you are looking for (let’s say C) is low in the population - hence the “target” low test positive rates?

 wintertree 09 Jan 2021
In reply to Dr.S at work:

That's a rater tail-wagging-dog way of thinking but it also makes sense.  I guess I'm going to have to dig out and digest the WHO guidance...

In reply to wintertree:

Sounds like light reading!

I expect like most things you need to look at a constellation of indicators that the system is working, any one indicator alone can lead you astray

In reply to wintertree:

> Depending on how the media run with this, the cabinet may not outlast the crisis.

The media should be full of stories from hospitals being overwhelmed. There's a bit of that but not enough.

Instead of Whitty droning on about things, put up graphic ads of people in oxygen masks struggling to breathe on prime time TV.

I suspect the Cabinet will do just fine but that's a different story.

Incidentally, apparently the director of my Dad's institute has said people should come in because it's Covid secure. This is to do stuff which is far from essential, some of which can be done at home, some can't be that could surely wait a couple of months. Some people just don't seem to understand the concept of a workplace outbreak. My Dad actually got a fairly bad case of Covid back in November which, realistically, he could only have picked up at work as he was going there a couple of times a week (recovered ok after a few weeks and didn't go into hospital but still wiped out for a couple of weeks).

In reply to wintertree:

A thought occurred to me. We now have 3m official cases - a small number of these may be duplicates but that's immaterial. So the real number of cases could be more like 6m (the ONS estimates of the number of daily infections have consistently been around double the official numbers - sometimes a bit more, sometimes a bit less but 2x seems a rough rule of thumb). So let's say at most 10% of the population. Call it 20% given that testing wasn't really sorted until after the first wave had receded.

Rather puts paid to the herd immunity argument...

In reply to Offwidth:

Some people just don't get - until they get it...

In reply to wintertree:

>clearly a cold / flu is the most likely answer but...

> Not all infections are PCR detected.  If flu is rising that could explain some of the apparent spike in testing rates.

We had a briefing from our government on Friday, there hasn't been a single case of flu identified in the island this autumn, so correspondingly I'd suspect UK flu numbers will be very low.  Another interesting tidbit was that having no covid in the community means it's really easy to vaccinate people - no PPE required or associated changing of it, and no problems with having to rearrange testing for people having the virus or self-isolating. They plan to get 42k doses administered by the end of March, which is the equivalent of 40million for the UK.

 neilh 10 Jan 2021
In reply to wintertree:

Thanks 

 Offwidth 10 Jan 2021
In reply to Toerag:

Hancock was boasting today on Andrew Marr about how well they are doing on vaccinations... lets see shall we.... they have around 10 million to do to meet their pledge in  a little over a month's time. He also boasted about how lateral flow tests had done so well in Liverpool.... Marr pointed out his own scientists disagreed (a fact) and he just shook his head.  He also explained that 'schools are safe' means safe for kids (stuff the teachers and the kids' families and the further infection chains then)... terrible muddled messaging and an idiotic epidemiological response in the current situation.

What needs to happen on popular shows like Marr is someone needs to explain deaths will pretty much double the current average daily rate, to 2,000+ a day, in three weeks, irrespective of what we do, as they are locked in (and if hospitals start 'breaking' it will be worse). Then if we don't lock down as hard as March RIGHT NOW, things will be even worse. Prof Susan Michie (on SAGE) came close, this am on BBC News Breakfast, but backed off a bit when asked if ten thousand extra unnecessary deaths is really likely if we don't increase restrictive measures (she said it's her worst estimate). Even Ferguson is losing it a bit ....he said this weekend we will have deaths over 100,000 if we are not careful.... it's bloody certain already on official deaths and on ONS we will be there in a week!!

Lets say any single week delay on a really hard lockdown will involve an extra 10% or so growth in a week (before Boris and Hancock realise: oops we better lock down harder). If that locks in an increasing level from 200 extra to a peak and a then slower decline than March (as the virus has a higher R rate) we must be looking at an average of 300 odd extra deaths a day... we need just a month from the start of Feb to get to Susan's extra 10,000

Lets also assume 1 in 5 old/vulnerable folk won't want the jab (similar to the flu jab).. that's a lot of the most vulnerable people when added to the percentage of the various demographics under 70s who would likely end up in hospital if infected. Restrictions will be around into summer I think and social distancing probably for a year.

I really hope I've missed something but the S African strain and hospitals 'breaking' might mean things are even worse. Prof Bell said maybe we need to bend NHS rules and massively increase vaccination rates... I have to agree.

Post edited at 17:03
 wintertree 10 Jan 2021
In reply to Offwidth:

The big picture isn’t looking good.  

Hopefully lots of updates tonight from me when I can have the peace and quiet to upload them.  In the mean time, the Isle of Wight stood out...

https://coronavirus.data.gov.uk/details/cases?areaType=ltla&areaName=Isle%20of%20Wight

Rather obvious but tragic.  If Tier 4 wasn’t containing the new variant around the Thames estuary, Tier 3 obviously wasn’t going to.

Similar comments for Herefordshire

https://coronavirus.data.gov.uk/details/cases?areaType=ltla&areaName=Herefordshire,%20County%20of

 Offwidth 10 Jan 2021
In reply to wintertree:

Also away from London and the SE have a look at Carlisle, Cherwell, Bedford, Halton, Knowsley, Wolverhampton.... all over 1000 per 100,000 cases on official tests.

Another simple way of looking at the date to get people to realise how serious this is, is for every official new case now there will be over 3% unavoidable deaths in 3 weeks time. (eg current UK official average cases per day at around  60,000, means around  2000 average deaths a day in 3 weeks).

Post edited at 17:30
 RobAJones 10 Jan 2021
In reply to Offwidth:

>He also boasted about how lateral flow tests had done so well in Liverpool.... Marr pointed out his own scientists disagreed (a fact) and he just shook his head. 

How useful are they? Birmingham Uni did PCR tests on all positive cases and a sample of negative cases before sending students home at Christmas. From over 7000 students they think they correctly identified 2 and missed about 60 cases? I spoke to a primary head friend, today,  he was complaining about them being shut, on the basis that 16000 students in Cumbria had been tested on their first day back and only a "handful" were positive. If this is true, it seems extremely unlikely to be accurate, given the case numbers in Cumbria. I can see how they would be useful in some circumstances.  Using Jon's example,  if someone was going to get an eye test anyway, then having to take a LFT before is better than nothing, but no where near as effect as making the appointments emergency only.   

 Si dH 10 Jan 2021
In reply to RobAJones:

> Using Jon's example,  if someone was going to get an eye test anyway, then having to take a LFT before is better than nothing, but no where near as effect as making the appointments emergency only.   

This is the key point. They are a useful extra line of defense for activities deemed necessary that involve some moderate enhanced risk , where other ways to reduce risk or a PCR test are not reasonably practicable. This is good practice. They shouldn't be used to screen people with the intention of allowing additional risky activity that otherwise wouldn't take place.

I got annoyed with Marr this morning because he didn't understand this and therefore kept just quoting things out of context. With some understanding he could have tested whether or not the new proposed application was sensible (I was none the wiser afterwards.) A blunt LFT are good/bad approach does noone any favours.

Edit to add, they apparently picked up 40% of all positive cases in the Liverpool trial, 50-60% for the people who were thought to be in their most infectious phase (I don't know how they determined this.) Even now about 25% of the case count in Liverpool is coming from LFT so they are obviously having a positive effect, as long as the test centres aren't becoming centres of transmission.

Post edited at 18:34
 wintertree 10 Jan 2021
In reply to thread:

4 Nations case rates.  No interpretation necessary.  De-weekending and shunting used as discussed on previous thread to try and improve quality of analysis against reporting lags/spikes.  

Post edited at 18:40

 wintertree 10 Jan 2021
In reply to thread:

Plot 16 - still getting worse.  From a quick glance I noticed that the Isle of White has risen spectacularly.  Offwidth notes some other areas above.  Less than Tier 4 in the face of the new variant has not worked well.

Plot 17 - you can see the growth rate increasing in "blue" areas and slackening off in the "red" areas.  Lots of discussion of this initaly-fast-then-slower exponential behaviour on previous threads.  

Plot 18 - the right side of this is a bit more provisional than normal given the New Years day extended weekend passing trough.  I think the coming Friday is the next opportunity for a "best possible" update to this plot.  

  • Cases - The line plots are slightly filtered (as per the captions) from the heat maps as otherwise is jagged-line city and hard to interpret.  From the heat maps it seems the "red" regions have just hit slightly negative exponential rates meaning cases are decreasing.  But - this is the provisional window and is sketchier than normal this week.  The growth rate for cases haven't finished increasing in "blue" regions I think, although for most of them the doubling times are increasing as growth slacks off.   Promising, but some way to go before the exponential growth rate hits zero - which means a lot more growth in actual cases ahead.  Being positive, this growth is all for infections that happened before the new lockdown started on Jan 5th, so I hope to see all rates tip over in to decay soon.
  • Hospitalisations - the distribution of lag times from detection to hospitalisation blur out features from the cases.  I don't think these rates are done rising for the "red" regions yet, and are going to remain positive (meaning the number keeps growing) for at least 2 weeks.
  • Deaths - it looks like all regions are now in growth.  As with hospitalisations, there is a lot of growth locked in on these.

Hopefully I'll be back later with demographic breakdowns for London and England.

Post edited at 18:53

In reply to Si dH:

> I got annoyed with Marr this morning because he didn't understand this and therefore kept just quoting things out of context. With some understanding he could have tested whether or not the new proposed application was sensible (I was none the wiser afterwards.) A blunt LFT are good/bad approach does noone any favours.

I also shook my head just before Hancock did. Marr was being a bit of a nob about this.

In reply to Offwidth:

> I really hope I've missed something but the S African strain and hospitals 'breaking' might mean things are even worse. Prof Bell said maybe we need to bend NHS rules and massively increase vaccination rates... I have to agree.

They also need to make sure they don't turn vaccination into a spreader event by chasing targets.  There are pictures on Twitter of a huge queue of old people waiting for their jags.  We know this strain is more infectious and you don't get protection until a couple of weeks after getting the vaccination.

https://twitter.com/P_Demetrious/status/1347550913024708610

 wintertree 10 Jan 2021
In reply to wintertree:

Demographic breakdowns.  The source of demographic data cuts off a day sooner than the data I use for Plots 16-18.  Errorbars on plot D3 are the standard deviation of measurements for the most recent 5 days.

Some observations:

  • Doubling times a lot longer compared to a couple of weeks ago.  
    • Most of this change happened before control measures and seems to be an intrinsic characteristic of this pandemic in the UK.  
    • Doubling times initially surge them back off as cases grow, independently of further control measures.  Susceptible sub-population?  Local response to local hospitalisations and illness mediated via social grapevines?  A visualisation of this here [1] from plotting #6.  
    • So - as control measures kick in, we should see exponential rates drop more and fall well in to decay (halving times).  Good.
  • Ages 10-15: These didn't drop as much in England (Plot D1) or London (Plot D2) as in other age bins during the November lockdown when schools remained open.  Now that schools have been closed over the Christmas vacation, exponential growth rates are clearly lower in this bin - generally being blue, meaning decreasing cases, in London.
  • Ages 20-25: These remain the peak of growth in working aged people.
  • Ages 80+: These have risen significantly for England from last week and dropped a bit for London.

A note of caution - the last week in this data is more provisional than usual with all the irregularities from the New Year's Eve long weekend and falling after the irregularities from the Christmass long weekend.  I would say the data has a lot of promise in it - but we have been here before, and whilst cases may be about to exit growth in some age bins in London, that isn't the case for those with the highest hospitalisations, and there is a lot of difficulty locked in for the next few weeks by infections that have already incurred.

[1] https://www.ukhillwalking.com/forums/off_belay/friday_night_covid_plotting_6-729493?v=1#x9370898

Post edited at 20:03

In reply to wintertree:

> Doubling times are down 

Up??? Rate constant down?? Think (hope) you did that easily done thing again....?

How different does all this look before the SG filtering? It's probably unfounded, but I can't help thinking the far RHS of these curves, where there's currently a lot going on, could be really distorted by the library SG end-fudging algo that we've covered before. Though you'd probably have mentioned it if it was. So I'll stop not-helping again.

Tl;dr: thanks again.

Post edited at 20:02
 wintertree 10 Jan 2021
In reply to Longsufferingropeholder:

> Up??? Rate constant down?? Think (hope) you did that easily done thing again....?

God damned it.  Thanks for pointing out the mistake.  I fixed it.  Rate constant down, doubling times longer.

Here's plots with and without the filter.  On a desktop (not mobile) you can use the arrow-keys on the gallery to flip-book between them.

Perhaps counter-intuitively, the SG more closely tracks the shape of the raw curves in the end windows, because it's fitting the same polynomial order to fewer datapoint, so it will naturally conform more to the details of the curve.  I like it because it behaves gracefully at the extremes of the data.  The reason I couch the plots in uncertainty is because by next week, the data there is then going to influence the more refined filtering at the end of the current data, that will be possible next week.

SG filtering is applied to the line plots but not the heat maps.  It mainly mops up micro structure from residual weekend effects and also the xmas weekend.  It’s very weak filtering given the order and number of data points; the length increases with each successive measure as the statistical noise gets worse.

>  So I'll stop not-helping again.

No, it's good to think about these things.  I often plot with/without filtering and the deweekending/shunting before posting to check that nothing untoward appears to be happening.

Post edited at 20:13

 Wicamoi 10 Jan 2021
In reply to wintertree:

Many thanks for the updates, appreciated as always.

I expect you've noticed, but some of the regions' hospitalisation data are missing (plot 18).

 wintertree 10 Jan 2021
In reply to Wicamoi:

Thanks.  I’m still planning to incorporate your scoring in a variant of plot 18.

The missing regions were covered a few weeks back - the NHS regions by which that data is reported don’t correspond to the English regions.  I need to sort that out more gracefully than just omitting those without a direct correspondence...

In reply to wintertree:

Cheers for that. Satisfies a curiosity.
Call that noisy? You should see some of the s**t I have to work with...

 Wicamoi 10 Jan 2021
In reply to wintertree:

I thought you'd have given up on trying to use that scoring. At least the more recent devolved nation measures are a bit simpler. Sorry, must have missed the previous explanation about the data gaps. I'll stop not-helping too!

 Wicamoi 10 Jan 2021
In reply to Longsufferingropeholder:

>

> Call that noisy? You should see some of the s**t I have to work with...

Ditto.

 Si dH 10 Jan 2021
In reply to wintertree:

> Thanks.  I’m still planning to incorporate your scoring in a variant of plot 18.

> The missing regions were covered a few weeks back - the NHS regions by which that data is reported don’t correspond to the English regions.  I need to sort that out more gracefully than just omitting those without a direct correspondence...

If you want to fix them properly then in the interests of pedantism, it's worth noting that the NHS definition of East of England is also quite different from that used for cases. There are quite a few differences that you wouldn't realise without knowing, it's a bit annoying.

https://www.gettingitrightfirsttime.co.uk/hubs/ shows the NHS regions clearly.

 Si dH 10 Jan 2021
In reply to wintertree:

Thanks for all the updates.

There are a couple of key points for me at the moment.

In London and the South East it looks tentatively like we have seen the vast majority of case rises and hope that things will soon start to fall. So, we have a very big problem in that area, but it is hopefully now at the stage where one could predict the size of it with slightly more confidence than before when everything just continued to go up. I have no idea whether the current baked-in hospital load is already unmanageable (I notice today the first London trust has started cancelling non-covid urgent care), but people will at least be able to plan around something.

In other areas, hot spot behaviour is quite localised within the regions. In some areas rates are matching London and in other areas not far behind, and I assume they mostly won't have as much hospital capacity and ability to flex as London either. But, in most cases, these most-badly-hit locations also have other cities close by with which they might be able to share hospital capacity if necessary who have not yet hit those huge rates but are starting to move in that direction. So, it seems critical to me that we start to see rates in these areas turned over by lockdown some time next week, rather than them continuing to increase to the very highest levels, such that there is some possibility still of us maintaining net positive hospital capacity overall. We can't afford (I doubt) for everywhere to hit the 7-800+ cases mark that most of the South East did and some other places have now followed.

Post edited at 20:56
 wintertree 10 Jan 2021
In reply to thread:

Plots 20a & 20r

  • Hospital admissions in absolute numbers (a) and rates per 100k (r).  Different y-axes are used for under and over 18s due to the big difference in numbers.  
  • Whilst 85+ has the highest per-100k admissions rate, 18-to-64 and 65-to-85 each contribute far more absolute numbers to the hospitals.
  • Admissions are higher in the 0-to-5 range than 6-to-17; this corresponds to an earlier observation on this thread from mik82 about deaths.  It is however very low absolute numbers, and instinctively I suspect testing is applied differently in these ages so I'm not minded to speculate on interpreting it.

Plot 21 - an estimate at refining the coarse age bins from the hospitalisation data to the 5-year bins used by cases data.  This presents the number of people who tested positive on the x-axis date who have gone or are going to go to hospital.   Of all the plots I've done, this is the one most likely to generate significant back-and-forth with reviewers if it went in for peer review, so take it with a pinch of salt.  It's done by fitting an age dependant hospitalisation rate model and an age dependant lag model to cases and data against hospitalisations data.  I include it because it clearly illustrates that this is not just an old person's disease in terms of severe effects.

An update to the new hospital occupancy plot.  It looks like we've hit the same intensive care occupancy levels as back in March/April and are close to double the total hospital occupancy.

Post edited at 20:55

In reply to all:

At a tangent here, I noticed on the main gov dashboard that the number of 28 day deaths for England (reported daily) was greater than the number of death certificate deaths (reported weekly).

i.e. the death certificate deaths are generally accepted to be higher, but there are now enough deaths/day to exceed the lower frequency and greater lag for it to be overtaken. Not a good sign methinks.

 wintertree 10 Jan 2021
In reply to Si dH:

> If you want to fix them properly then in the interests of pedantism, it's worth noting that the NHS definition of East of England is also quite different from that used for cases.

Thanks; technically there's nothing to "fix" other than my understanding as the captions correctly describe this - but in the future I can and will draw attention to the fact the regions don't fully correspond.

> There are quite a few differences that you wouldn't realise without knowing, it's a bit annoying.

Really, I think it's worse than annoying.  The mash-mash of geographic and demographic bins severely curtails the quality of real time analytics that can be pulled out of the data.  Unless the people feeding in to SAGE have access to some alternative set of secret, more coherent bins then it frustrates the flow of data.  Mind you, given what happens after SAGE, that's not the biggest barrier to timely, informed policy I suspect.

 wintertree 10 Jan 2021
In reply to Longsufferingropeholder and Wicamoi:

> Call that noisy?

Now now.  If we're going to get in to it...  One of the bits of code I've got lurking around waiting for a use is an extended object tracker that gives sub-pixel precision with SNRs down to 3 or so.  Not much light when you're imaging at 250,000 frames per second (and that was back in 2010).  

These days I'm mainly working in high signal regimes and focusing on other areas than the data itself.  Like making machines that give better data...

In reply to Wicamoi:

> I thought you'd have given up on trying to use that scoring. 

Not given up, but waiting for a bolt of inspiration.  It's hard because you need colour to separate the nations and and the level of control measures.  I've been thinking about denoting change of control measures (↑ for stronger , ↓ for weaker) with markers on the curves.  Then you can look for the changes that lag an arrow.   

 mik82 10 Jan 2021
In reply to Si dH:

>I notice today the first London trust has started cancelling non-covid urgent care

The actual article I read said that the Royal London has not been able to do any non-emergency surgery since 23rd December, and they were prompted to speak out by an incorrect statement from an NHS London boss that urgent cancer surgery was continuing.

In reply to Michael Hood:

The initial large disparity between the total covid + deaths and death certificate deaths is also coming down - in part a result of much more testing? Compared to the first wave its now very unlikely that you would not have a Covid test if you had Covid.

Over 500,000  PCR tests conducted on the 7th, amazing capacity, and turn around times can be fast - I head a test taken at 19:30 last night and got the result back at 13:30 today.

 WaterMonkey 10 Jan 2021
In reply to wintertree:

Hi WT, hope you’re well.

in your plot 17 am I right in thinking there is no slow down of cases in London and the SE after the November national lockdown? This is very worrying if true because we we’re in tier 4 lockdown over Christmas and would have hoped that would have shown through by now. Doesn’t bode well for any areas not in tier 4 over Christmas does it?!

 wintertree 10 Jan 2021
In reply to WaterMonkey:

Golly, it's complicated to get this precise.

> in your plot 17 am I right in thinking there is no slow down of cases in London and the SE after the November national lockdown? 

Cases in London went in to decline during the November lockdown, then started to grow again about half way through.  I think this was driven by two different behaviours; the old variants of the virus went into decay from initially high numbers and continued decaying, and the new variant stared with low numbers and rose exponentially throughout.   We see the sum of the cases from each - one decaying exponential, the other rising.  The result is a curve that goes down then up.

After national lockdown, when London was released in to Tier 2, the cases increased at ever shorter doubling times (the red curve on Plot 18 cases going up, equal to a higher exponential rate constant).  Unsurprisingly, if lockdown didn't hold against the new variant, Tier 2 definitely didn't.   From around Dec 20th when to Tier 4, cases (Plot 17) continued to rise but the exponential rate starts decreasing - so whilst the absolute number of cases keeps going up, the time it takes for them to double is getting longer and longer.  This is a necessary step between fixed exponential rise and zero growth leading to decay.  

> This is very worrying if true because we we’re in tier 4 lockdown over Christmas and would have hoped that would have shown through by now.

It does it you look at plot 18.

  • The more orange the heat map is, the faster cases are doubling. This doubling time (or the exponential rate constant - the two axes on the right plot link them reciprocally - a bigger constant means a shorter doubling time) is what policy goes on to affect - lockdowns make it longer or turn it in to a halving time (cases decreasing).  This is where to look for cause and effect of policy and behaviour; growth in cases is a result of both this doubling time and the absolute number of cases.
  • Plot 18 is a lot less orange for London and my other "red" regions and is perhaps soon to turn blue - meaning cases are shrinking.  However, because absolute cases are very high, even a long doubling time leads to a large daily growth.

>  Doesn’t bode well for any areas not in tier 4 over Christmas does it?!

It really doesn't.  This can be seen by looking at curves for various UTLAs in the government dashboard with some suggestions up thread from myself and Offwidth.  I think the decision not to put the whole country in to Tier 4+ by about Dec 18th (when it was bloody obvious that the new variant grew exponentially under Tier 4 and had spread nationwide) deserves a key focus in the public enquiry.

I am positive thought; moving London to Tier 4, increased awareness of the new variant and increased news coverage seem to have sent the growth rates heading towards 0 in the "red" regions on my plots.  If that trend continues to develop over the next week we'll know that Tier 4 and serious awareness is enough to hold cases level, giving good reason to hope that the new lockdown is going to see a significant decrease.  But it feels like an awfully long time between that happening, that happening for the "blue" regions and all of that translating in to a drop in hospitalisations.

 wintertree 10 Jan 2021
In reply to Si dH:

> So, it seems critical to me that we start to see rates in these areas turned over by lockdown some time next week, rather than them continuing to increase to the very highest levels, such that there is some possibility still of us maintaining net positive hospital capacity overall.

I'm watching the red vs blue curves on plot 18 with interest.  The time from zero-crossing to peak on the exponential rate looks similar for red and blue regions, but the blue ones seem to be peaking at a lower value.   I'm very much hoping that continues to be the case come this Friday's update.

> We can't afford (I doubt) for everywhere to hit the 7-800+ cases mark that most of the South East did and some other places have now followed.

I think it's going to be exceptionally close to the wire.  

In reply to wintertree:

> I think it's going to be exceptionally close to the wire.  

Hopefully we can re-use "a damned close run thing" in a few months time.

 WaterMonkey 10 Jan 2021
In reply to wintertree:

Thanks for the long explanation.

Here in the south East we’re seeing ambulances with blue lights going almost every time we drive somewhere.

I suspect (or hope at least) that the government make the lockdown stricter in the next few days as a damage limitation exercise. Hospitals will be overwhelmed still but the duration will be less hopefully.

 wintertree 10 Jan 2021
In reply to WaterMonkey:

> suspect (or hope at least) that the government make the lockdown stricter in the next few days as a damage limitation exercise.

As always I suspect improved adherence would go a long way, but we seem to tend towards over legislating for under compliance.

There is a big media push going on around awareness and compliance which will help.

I’m not really sure where I can see our government and their back benchers going in terms of stricter rules.  Look at how most of the Kent MPs voted on the post-lockdown tier legislation...  

 Si dH 10 Jan 2021
In reply to wintertree:

Just came up on the Guardian - new analysis of % population infected to date in each UTLA based on deaths and IFR (it's not clear if IFR was applied demographically in age bins, I think so given the data in the reference). Headline national figure is 20% but peak UTLAs up around 40% and Barking just over 50%. Interesting.

https://www.theguardian.com/world/ng-interactive/2021/jan/10/one-in-five-have-had-coronavirus-in-england-new-modelling-says?CMP=Share_AndroidApp_Other

 WaterMonkey 10 Jan 2021
In reply to wintertree:

> I’m not really sure where I can see our government and their back benchers going in terms of stricter rules.  Look at how most of the Kent MPs voted on the post-lockdown tier legislation...  

I think if they limited exercise to 1 hour it would help a lot and would get rid of a lot of the rule flexers. That and improve mask wearing and social distancing when shopping. 

3
 wintertree 10 Jan 2021
In reply to Si dH:

Interesting; thanks.  

What’s notable there is that the infection:detection ratio for the current “wave” is apparently much higher in the worst hit areas than the national average.  That does change the discussion around cumulative infections quite a bit from the last thread.  Not enough but quite a bit.

I thought it was more than a bit disingenuous of the guardian to implicitly compare test and trace numbers with total infections including the first wave which predates T&T.  A fair comparison is damning enough...

In reply to RobAJones:

Funnily enough I got a letter from Boots the other day saying my eye test is due and I should book because places are filling up and my eyesight is important. Which it is but I think I can wait till the summer...

In reply to wintertree:

You mentioned “From the heat maps it seems the "red" regions have just hit slightly negative exponential rates meaning cases are decreasing”

Did you meant the growth rate in cases is decreasing or cases are actually decreasing? I’d be surprised if cases are decreasing anywhere. 

In reply to Misha:

> Funnily enough I got a letter from Boots the other day saying my eye test is due and I should book because places are filling up and my eyesight is important. Which it is but I think I can wait till the summer...

Barnard castle can get busy

In reply to WaterMonkey:

> I think if they limited exercise to 1 hour it would help a lot and would get rid of a lot of the rule flexers. That and improve mask wearing and social distancing when shopping. 

Those things would help a bit. Stopping every man and his dog from calling themselves a key worker and having schools a lot better than 50% closed would help a lot.

 Offwidth 11 Jan 2021
In reply to Misha:

Be surprised.... most of the hard hit areas in S Wales and some areas in the SE and one in London (Richmond).

https://www.theguardian.com/world/2021/jan/08/coronavirus-uk-covid-cases-and-deaths-today

 wintertree 11 Jan 2021
In reply to Misha:

> Did you meant the growth rate in cases is decreasing or cases are actually decreasing? I’d be surprised if cases are decreasing anywhere. 

In the data, daily cases have just tipped over to decreasing in London and are very close to decreasing in the other “red” regions.

You can see this if you look at London cases on the gov dashboard.

Its the leading edge of the data, so what happens over the next few days could reverse the trend over even a short timescale.  Further, the data is messed up by New Years although I’ve done what I can about that it’s still nosier than normal.    So it’s not clear if infections are genuinely falling to match the data.  I think it will be clear one way or the other by Friday.

The good news is that if we’re wondering if cases are falling or not, they’re definitely not rising fast.  As this drop in exponential growth rates proceeds the latest lockdown, and as London is ahead of the “blue” regions on the new variant, it suggests the looming disaster for healthcare isn’t as utter as it could have been.

 neilh 11 Jan 2021
In reply to wintertree:

Maybe its about time BBC news did a report from afew covid pateients for those under 16 instead of concentrating on those around 60 plus.

Perhaps it might ram the message home to parents.

There again getting parents premission for filming  might be extremely difficult to say the least

In reply to WaterMonkey:

I agree with Chris Witty that what we have is lots of people breaking the rules is small ways and we shouldn’t be going ‘let’s blame footballers’. It is more ‘collective’ than that. The only thing the government could do is more police enforcement but I doubt they will do that in a meaningful way. My optimism supply is at an all time low. Vaccines and immunity through already having it seems to be the plan. 

In reply to WaterMonkey:

> Hi WT, hope you’re well.

> in your plot 17 am I right in thinking there is no slow down of cases in London and the SE after the November national lockdown? This is very worrying if true because we we’re in tier 4 lockdown over Christmas and would have hoped that would have shown through by now. Doesn’t bode well for any areas not in tier 4 over Christmas does it?!

The ‘no slow down in London/SE’ was in part due to New Covid (and in my view less compliance).

I expect Si dH to disagree, but not shoving us all in tier 4 over Xmas was totally wrong. Todays info:
Liverpool: rate = 935, weekly increase 113%
Manchester:  440 and 60%

It was as plain as the nose on your face that New Covid would spread everywhere. 

 Si dH 11 Jan 2021
In reply to mick taylor:

> I expect Si dH to disagree

Not at all, ever since the new variant was discovered and shown to be spreading fast at the turn of November/December it has been fairly clear that everywhere needed locking down to avoid it getting everywhere in significant numbers. Unfortunately it's way too late for that now.

Even within the government's own decision framework, when compared against other areas, Liverpool should have been put up to Tier3/4 in the review round two weeks before it was. Being in Tier 2 for too long is likely responsible for the rates rising as fast as they have in recent weeks.

My argument on the other thread was more about what happened in the autumn and how we got to where we were under the old variant, rather than about action since the new variant was discovered. It's only important if we try to learn lessons from what went before.

Edit to add, I am currently in a bit of a fluster that the rise in North West cases could be partially driven by this new South African variant. Apparently one of the initial two cases found was somewhere in the region, so hopefully more sequencing is happening to determine this. The rise has been so fast spreading in recent weeks from N Wales into Cheshire and Merseyside (now worst here) that I can't believe it is all the old variant. Yet the most recent PHE report on the UK variant suggested that it was still lower prevalence around here than in many other areas of the country according to SGTF numbers.

Post edited at 11:22
In reply to Si dH:

Thanks for clarifying, think I slightly misunderstood your point on the other thread. Interesting point about the NW. 

When I recall NW case rates and corresponding death rates in Oct/nov in NW I have actually been surprised we haven’t seen more deaths in the SE. It may be that a higher % in SE get tested, it might be higher amounts of poor health outside of the SE, or it could be me making this up, but the possibility of the South Africa variant PLUS high rates PLUS my perception of higher death rates in the NW is very worrying. 
 

 RobAJones 11 Jan 2021
In reply to Si dH:

> so hopefully more sequencing is happening to determine this. 

I assumed this means they are, at least for the UK variant in Cumbria?

"more than 85% of cases in the Carlisle area are the new variant of Covid-19 that is spreading extremely quickly in the country. Revealing the data, Colin Cox, Director of Public Health at Cumbria County Council, said the variant is "leading to widespread community transmission" of Covid-19, revealing it arrived in Cumbria "earlier" than across many parts of the North of England."

it would seem strange if they weren't doing it for the SA variant as well? At least this seems to one area where we seem to be genuinely " world leading".

 AJM 11 Jan 2021
In reply to RobAJones:

I thought we got lucky with our home-grown new strain, in that the mutation is in one of the PCR regions so to some extent the variant information falls out as a byproduct of the test.

I'm not sure how quickly we would have been able to monitor it's spread otherwise nor whether the SA variant is as fortuitously easy to track.

 RobAJones 11 Jan 2021
In reply to AJM:

Are we now sure it was home-grown? This from before Christmas

A new variant of the virus causing Covid-19, first spotted in Kent, could already be circulating - or have originated from - outside the UK.

But it was spotted here because of the strength of the UK's surveillance system, scientists have said.

The Covid-19 Genomics Consortium (Cog-UK) has tracked the genetic history of more than 150,000 samples of Sars-Cov-2 virus.

That equates to about half the world's genetic sequencing of coronavirus.

Prof Sharon Peacock, head of Cog-UK, said the UK's high level of genomic surveillance meant "if you're going to find something anywhere, you're going to find it probably here first".

https://www.bbc.co.uk/news/health-55413666

In reply to RobAJones:

> Are we now sure it was home-grown?

No, not at all. We'll never know where it first emerged.

https://www.reuters.com/article/us-health-coronavirus-germany-variant/new-coronavirus-variant-in-germany-since-november-die-welt-idUSKBN2921TE

Post edited at 13:39
 Si dH 11 Jan 2021
In reply to RobAJones:

I think where it originated might still be uncertain but what AJM says about the detection of it is correct. Basically, the UK variant (for want of better term without remembering the proper name) has been found to be reliably trackable by PCR testing because one of the three targets of many of our PCR tests has been found to be missing, because of one of the mutations. This is how it is now being tracked around the country (see the PHE report I linked above) and mass genome sequencing is not required. However, that same mutation is not present in the South African variant, so it would show up in a PCR the same as the "original" variant, meaning a full genome sequencing activity is required to track it.

I understand you are right about our national capability being comparatively very strong, but there are still only a relatively small proportion of cases (5-10% I think but varying around the country, and with some delay) that get sequenced.

Post edited at 13:43
 AJM 11 Jan 2021
In reply to RobAJones:

No, it was just a lazy identification method on my part...

In reply to wintertree:

> > Did you meant the growth rate in cases is decreasing or cases are actually decreasing? I’d be surprised if cases are decreasing anywhere. 

> In the data, daily cases have just tipped over to decreasing in London and are very close to decreasing in the other “red” regions.

>    So it’s not clear if infections are genuinely falling to match the data.  I think it will be clear one way or the other by Friday.

> The good news is that if we’re wondering if cases are falling or not, they’re definitely not rising fast.  As this drop in exponential growth rates proceeds the latest lockdown, and as London is ahead of the “blue” regions on the new variant, it suggests the looming disaster for healthcare isn’t as utter as it could have been.

My 'even noddier than yours' stats show the national increase in live cases as a percentage of the live case count has been decreasing steadily since 27th December - 3.4% today from a peak of 4.5%. This shows the rate of infection slackening, but as we know local fast rises can be masked by declines elsewhere.  The absolute increase in live cases seems to have stabilised for about a week at ~22-24k more each day, but isn't noticeably declining yet.

To me this shows the craziness of the current UK restrictions - they're crippling the economy and society, but don't offer any way out of the mess like a proper lockdown would. Restrictions will need to remain for quite some time because the tiniest slackening will crash the NHS, yet relief given by vaccinations will take time to provide sufficient herd immunity. As intimated in this and other recent threads, test capacity is looking like being exceeded which is only going to make things worse. God knows how well T&T is doing these days but I suspect the answer will be 'badly'.

I dunno if you can get comparable data to pump into your stats mangle, but Eire looks to be suffering even more than the rest of the UK judging by their worldometers graphs.

In reply to Si dH:

> I think where it originated might still be uncertain but what AJM says about the detection of it is correct. Basically, the UK variant (for want of better term without remembering the proper name) has been found to be reliably trackable by PCR testing because one of the three targets of many of our PCR tests has been found to be missing, because of one of the mutations.

I'm wondering how they'll tweak the PCR to deal with this - if there's a deletion mutation of the two remaining indicators the tests will fail to identify anything. I suspect they're going to have to either look for an additional indicator or change the one identifying the Kent variant.  There would seem to be little use in being able to identify the Kent variant now that it's established everywhere in the nation.

 RobAJones 11 Jan 2021
In reply to Si dH: and AJM

Thanks, I understand the difference now. As has been said up thread there is a big problem with  how quickly this type of information is acted on. It was two weeks after the new variant had been identified in Cumbria before we were put in tier 4. Give it a week or two and I think we will probably be needing to use hospitals in Lancaster or Newcastle, if they have any spare capacity by then.

 wintertree 11 Jan 2021
In reply to RobAJones:

>As has been said up thread there is a big problem with  how quickly this type of information is acted on

The problem was evident before that information was even gathered I think, and the appropriate action to take was obvious.  Daily cases tipped over to growth (R>1) in Cumbria between the 27th and 30th of November.  I put a plot together showing this on Dec 14th; the data was there by Dec 10th.

If control measures responded locally in response to changes such as cases flipping from decay to growth, rather than to high rates of growth of cases, perhaps things could have gone very differently. 

https://www.ukhillwalking.com/forums/off_belay/friday_night_covid_plotting__3-728848?v=1#x9355746

 minimike 11 Jan 2021
In reply to wintertree:

Whilst I agree in this case, a slight counterpoint and discussion for future use of exponential rates for policy formation..

how many times would there have been false positive signals in that indicator at UTLA level over the past 6 months? Genuine question.

Post edited at 14:36
 wintertree 11 Jan 2021
In reply to minimike:

It's a good question.  

If the noise on cases was effectively Gaussian (large number Poissonian) and temporally uncorrelated., I could directly translate that through to a uncertainty in the rate constant. (𝛘² for me, a P-value for you and probably the target audience...).

 Even with a bodge for weekend sampling effects in-place, the noise is may times what you'd expect for a Poissonian process and is temporally correlated, so... The answer depends on the statistical distribution and temporal power spectrum of the noise at UTLA level.  

The worse the noise, the longer a time-period of data that needs to be used to estimate the change in direction of cases.  This will depend on the population and infection levels with the UTLA.

I would think that with the sheer quantity of data available, a basic noise analysis could be used to give a reasonable rule of thumb on either how many days of data to use, or to derive a significance criteria for the observation that R is > 1 on any particular UTLA.  This should both be derivable from the current corpus of data and from an analysis of residuals in "live" data fit locally against a polynomial.

The noise at UTLA level is bad.  The data can be made to show whatever one wants through choosing the right days and using too-small a time window, and indeed this has been the basis for false qualitative claims of falling cases on UKC more than once. 

A definitive answer seems to be very difficult without a "gold" reference for UTLA level data.  I can see ways of qualifying this however.

My gut feeling is that you're not going to get an "R>1" alarm with good confidence at the individual UTLA level until 10-12 days after the actual turning point in Infections.   However, the trend is very much that the rate constant starts -ve and gradually increases towards the zero-crossing and then +ve growth, and that this trend can be picked up sooner.  There is definite inertia/momentum like behaviour to the exponential rate constant (and hence R), even accounting for the de-local fitting.  This is partly due to the dispersion of lag from infection to case detection, but also partly behavioural I think - and likely also influenced by household infections spreading, where increased control measures make little to no difference once the virus is in.  If this sounds a bit like phenomenology it is, but I think there’s some serious scope for data driven phenomenology here.

Post edited at 15:04
 minimike 11 Jan 2021
In reply to wintertree:

Correction: I do not consider a p-value as a measure of uncertainty! 

;-p

but thanks for the reply, makes a lot of sense. I guess I’m wary of effectively windowing over data which is derived by a non local fitting process. It feels like doing derivative analysis on LOESS fits which has known issues.

Post edited at 15:08
 wintertree 11 Jan 2021
In reply to minimike:

Fair enough!  I still haven’t shifted my head fully in to the mindset of biologists....  I shouldn’t call it an uncertainty as it isn’t; it’s a mathematically related value derived from the uncertainty.  I think there’s a way to go before qualifications of certainty have much link to policy.

What is the P value for 9.2 standard errors from the control anyhow?...  

> I guess I’m wary of effectively windowing over data which is derived by a non local fitting process. It feels like doing derivative analysis on LOESS fits which has known issues.

Yes, no analysis using non locality to deal with noise is without its pitfalls.  But as a barometer that detailed human scrutiny should be applied to the region, rather than an automatic trigger for policy, I struggle to think of better that can be applied at such a local level.   It would be a *lot* better if testing didn’t have a 7-day cadence and I retain my nagging suspicion a lot of this is about date of swab vs date of entry to the lab.

Really someone should suck up the maths and figure out the convolutional kernel that measures the exponential rate constant from the data over a given locality.  That’s the basis for a robust analysis of the method (as well as a computational implementation that isn’t embarrassingly inefficient) although much like the issues another poster has raised with SG this will have to fall back to fitting a function in the critical leading edge of the data.

Post edited at 15:16
 RobAJones 11 Jan 2021
In reply to wintertree:

> If control measures responded locally in response to changes such as cases flipping from decay to growth, rather than to high rates of growth of cases, perhaps things could have gone very differently. 

I completely agree, that would have resulted in changes to restrictions in September, but it would have been difficult to "sell" that to most people (probably impossible given previous decisions/Cummings etc.) Even with the new variation there was initially a lot of "it is only in SE why worry yet" There are still head teachers in the county that can't understand why their schools are not fully open. 

 minimike 11 Jan 2021
In reply to wintertree:

> What is the P value for 9.2 standard errors from the control anyhow?...  

-llareggub, as the welsh might say.

Post edited at 15:26
 wintertree 11 Jan 2021
In reply to RobAJones:

> I completely agree, that would have resulted in changes to restrictions in September, but it would have been difficult to "sell" that to most people 

Yes.  The moment our PM talked about shaking hands in March all hope was lost of bringing the people onboard through a consistent, intelligent messaging strategy.

Perhaps that hope could be recovered now there are contemporary examples of other countries accepting restrictions sooner and suffering less severe consequences, but I'm not going to bet on it.

 Si dH 11 Jan 2021
In reply to Offwidth:

> Be surprised.... most of the hard hit areas in S Wales and some areas in the SE and one in London (Richmond).

Far more of the South East is showing a reduction in 7-day average case rate after data updates from today, including about 10 of the UTLAs in London (I forget exactly which of them are counted as London and which just outside.) Most of the remaining UTLAs in London are up single digit %s and none are up more than 16% (which would equate to a doubling time of 4-5 weeks.)  That looks positive but things were showing a reduction in Kent for a few days just before Christmas, then they all went back up, so it's worth being cautious.

However...the weeks now being compared on the latest weekly-average data (hence, the %changes presented by the dashboard map and usually picked up by media) are 31/12-06/01 and 23/12-30/12. The vast majority of cases missed on 25/12 got picked up on 29/12 or 30/12 so I think it's fair to say that currently the Xmas and New year dips are both dealt with "in week" and this is now comparing apples with apples. That will still be the case tomorrow. In recent days up to yesterday the comparison was biased (conservatively) because the previous week contained the Christmas low but not all of the subsequent days in which the test data caught up. From Wednesday's data drop, the New Year's Day low will move in to the "old" week in the comparison, Christmas day will move out of it and interpretations will be difficult for at least another 3 or 4 days due to these days biasing the weekly averages.

Having worked this through I'm more confident the reductions seen using the data today (and hopefully tomorrow) are real, which corroborates Wintertree's inference from the leading edge of his exponential constant graph and also allows a picture of the geographical spread of change to be understood. After the Wednesday drop, the weekly average data will be poor until at least the weekend (once the 1st-4th Jan are all in the same week in the comparison again.) Hopefully by then we might see some more geographical broadening of the decline, that would tell me we are on the right track.

Post edited at 21:17
 wintertree 11 Jan 2021
In reply to Si dH:

Yup.  

I've just been putting today's data release through the pipeline.  It has helped me tune the shunting of cases over both Christmas and New Year which perhaps improves the estimates of growth rates a little.  It looks to me a bit like each of the South East, London, East of England and the North East are about to see clearly falling cases, with the slimmest of hints of that in all other regions excluding the North West.

We've been here before though.  The most recent weekend now dominates lagged reporting over the next few days.  It'll be really interesting to see where things stand this coming Friday when the data comes "good" again.

It also looks to me like the rate of hospitalisations in London may be about to level out.   This fits as the exponential rate constant for London has been heading towards zero for about a month now, with the biggest decrease towards very long doubling times having happened by mid-December which by now should feed through to hospitalisations.  It's worth remembering however that as people spend quite some time in hospital, there is another lag between the admissions rate levelling off and the occupancy levelling off.  

The rate constant curves for cases all have a correlated downwards buckle in them around 2021-01-04 which I think is the effect of residual issues from New Year's Day; perhaps once more data has cleared through this buckle will flatten out a little and cases will turn out to have not actually been falling - but it's close, and the effects of the new lockdown will start to show in this data very soon.  This plot is for data up to 2021-01-08.

So far, IMO this is looking far more promising than we had any right to expect. Good.  But it can all change on a knife-edge.

Edit:  I really need to get my maps sorted so I can post a time-sequence of the exponential rate constant / doubling time, preferably with a "zoom" of the London region as well.  

Post edited at 21:32

 minimike 11 Jan 2021
In reply to wintertree:

Great, thanks. Question then becomes ‘what have we missed?’

why are the rates falling again across the country when there should be no reason for the non-London/SE regions to be. The new variant hasn’t risen as far as it did in London and we weren’t even in t4 at that point..

following your connectivity theory, is it possible that the hyper connected sub populations are smaller outside the high population density Home Counties and London? Is that enough to explain the drops?

 RobAJones 11 Jan 2021
In reply to minimike:

Schools being shut, people not working over Christmas?

 Si dH 11 Jan 2021
In reply to wintertree:

Do you mean 8th? The data won't be complete for that yet. They've generally been ok for 4 days out recently but that only gives you the 7th. If you are using the 8th it will give you an artificial droop.

 wintertree 11 Jan 2021
In reply to minimike:

> Question then becomes ‘what have we missed?’

I was asking myself that when the November lockdown got off to such an astounding start.  I didn't say anything as I didn't want to jinx it...

> following your connectivity theory, is it possible that the hyper connected sub populations are smaller outside the high population density Home Counties and London? Is that enough to explain the drops?

That's where I've got no actual science to back up intuition...  The peak age for spreading in the "red" regions was 20-25 and 25-30 was close.  It could well be house shares combined with high risk jobs are behind a lot of this; and that situation is far worse close to London than elsewhere.

It would be interesting to segment the data in my "exponential rate constant vs cases/100k" graph into "red" and "blue" regions for before the new variant emerged, and to see if the moderation of exponential rates happens for lower case rates in blue regions.   If you look at the curves for cases on plot 18 between about 10-06 and 11-17 the "red" regions are generally higher than the "blue" ones, despite this being before the emergence of the new variant - but there wasn't a lag. 

I was wondering about doing a PCA of the exponential rate constant curves for each UTLA and assembling groupings that way rather than through the pre-defined regions.  I think that might bring out the differences a bit better. 

As RobAJones says, schools and workplaces closing are a big change, and perhaps the amount of media attention over London and the new variant helped with behaviour nation wide; that's certainly the impression I get from various posters on UKC.

 wintertree 11 Jan 2021
In reply to Si dH:

> Do you mean 8th? The data won't be complete for that yet. They've generally been ok for 4 days out recently but that only gives you the 7th. If you are using the 8th it will give you an artificial droop.

Good spot; I should have truncated at the 7th.  Doing a plot truncating a day earlier has a bit more of a definite rise on the last point on the curves.   I should add the last day used to the title of the plot.  

We'll have to wait and see.

In reply to wintertree:

Thanks, encouraging but let’s see if the trend lasts. Schools being shut and people voluntarily semi self isolating before seeing family at Xmas may help to explain things. 

 Blunderbuss 12 Jan 2021
In reply to Misha:

Just checked the latest NHS COVID-19 stats and there are 1073 in London ICUs, an increase of 32% over the last week.

Allowing for 5% headroom they have 158 spare beds after utilising surge capacity......looks like more and more difficult triage decisions on G&A wards will need to be made very soon.

 Offwidth 12 Jan 2021
In reply to wintertree:

I think you are looking slightly too much for that hope. The November lockdown shows the timeline of what obvious resultant changes in cases should look like. London and the SE have been under roughly the same rules now since xmas and cases would be decreasing if the current restrictions are tough enough.  The deceleration of increase should be starting to show from Jan 4th elsewhere. I think most people in the lower tier areas would have been being very careful so the fast growth away from London and the SE is scary. R needs to be less than 1 and all the indications for the new variant is that it won't be (especially compared to R in the lockdown in March). This is quite likely to mean we face the destructivly high levels of hospitalsation due in a week until vaccination effects kick in.

I think the March lockdown case changes are a real distraction... most cases back then were only detected at hospitalsation.

Post edited at 09:47
 jonny taylor 12 Jan 2021
In reply to Offwidth:

> I think most people in the lower tier areas would have been being very careful

As a general statement, sadly I don't recognise that. We live right on what was a tier 2-tier 4 boundary (WTF!?) and there was a sharp contrast in [illegal] behaviours seen (milling around in tight groups outside pubs, anecdotally home visits, etc). In Scotland it has very much seemed to be the [formerly] lower tier areas that have shot up first in this dec/jan peak.

In reply to wintertree:

> That's where I've got no actual science to back up intuition...  The peak age for spreading in the "red" regions was 20-25 and 25-30 was close.  It could well be house shares combined with high risk jobs are behind a lot of this; and that situation is far worse close to London than elsewhere.

My daughter leaves in Bexley, SE London: one of the most infected areas, and I can quote her:  'Dad, its been a free for all.', little compliance to the rules.  When I took her back after NY, she was actually upset at the lack of mask wearing on Public Transport when she went to get her Covid test (x2).  So i reckon a key driver is behaviours of young adults driven the feeling of invincibility and a lack of connectedness to their wider community, totally compounded by a government that has failed in its leadership (zero credibility with this age group and in this region in particular)..  Good to see infections starting to go negative in these areas.

I think house shares is an issue (others on here said this).  Not convinced about high risk jobs as this was being used as a reason for high rates in Grter Manc etc. which has very different employment.

On another point, the BBC news last night focused on the number of young adults in hospital with Covid.  The high % of young adults has to be a key reason behind what I perceive as a lower death rate in London/SE than in other hard hit areas (and health inequalities).

Post edited at 10:24
 Offwidth 12 Jan 2021
In reply to jonny taylor:

I think its easy to overblow the level of idiocy in the population from such observations. The experts say most people are behaving, a much smaller number are not.

The areas shooting up in Scotland were from a very low base. Overall Scotland looks way better than England on the map.

https://www.theguardian.com/world/2021/jan/11/coronavirus-uk-covid-cases-and-deaths-today

Post edited at 10:40
 wintertree 12 Jan 2021
In reply to Offwidth:

> I think you are looking slightly too much for that hope. The November lockdown shows the timeline of what obvious resultant changes in cases should look like

I'll hold on to it until Friday's data release, then see where we stand.   I could be completely wrong, and I'm hoping rather than predicting. 

>  London and the SE have been under roughly the same rules now since xmas and cases would be decreasing if the current restrictions are tough enough

It looks to me like the restrictions - both formal (change to T4) and informal (school and workplace closures over the holidays) have just about stopped growth in cases.

>  R needs to be less than 1 and all the indications for the new variant is that it won't be (especially compared to R in the lockdown in March). 

I think it can still be brought < 1.  Plot 18 has been heading in that direction for the last week.  The very fast doubling times (high Rs) seen at the start of this new variant I think are like the very fast doubling times seen in other areas when case rates per 100k are low - the phenomenology says they self-moderate to a degree as cases rise and that seems to be happening here.  Not good news for areas where case rates are still low, but gradually we're running out of such areas to keep driving top level growth figures.

I think that increased enforcement of illegal gatherings - large and small - will go a long way towards this.  I think the new variant has tipped relatively safe but illegal events over the edge.  Cressida Dick was making encouraging noises on the radio this morning.   

The good news as I said earlier is that if we're discussing whether or not cases are falling, they're definitely not rising fast any longer (at a national level and likely regional too - although not uniformly at sub-regional levels as Si dH observes up thread from their examining of UTLA level data).

> This is quite likely to mean we face the destructivly high levels of hospitalsation due in a week until vaccination effects kick in.

I think that's more or less unavoidable now.

> I think the March lockdown case changes are a real distraction... most cases back then were only detected at hospitalsation.

Indeed.  I truncate the start of all my plots much later in the year in our current testing regime, and look at hospitalisations back then as a proxy for cases.

> I think its easy to overblow the level of idiocy in the population from such observations. The experts say most people are behaving, a much smaller number are not.

That's the success of this virus though - it weaponises a small number of idiots.

Post edited at 10:40
 wintertree 12 Jan 2021
In reply to mick taylor:

Thanks for all the comments.

> I reckon a key driver is behaviours of young adults driven the feeling of invincibility and a lack of connectedness to their wider community, totally compounded by a government that has failed in its leadership (zero credibility with this age group and in this region in particular)..  Good to see infections starting to go negative in these areas.

That seems a good fit to the plot D2 over the most recent 3 threads.  

> On another point, the BBC news last night focused on the number of young adults in hospital with Covid.  

I think this is a very good thing for the news to do.  I wish the government would release actual data in fine age bins, as I have tried to back out of the data in my plot 21.  Their age bin of "18-64" leaves far too much room for the young to assume it's all people in their 50s and 60s going to hospital in that age bin.  Not so, I strongly suspect.

 Offwidth 12 Jan 2021
In reply to wintertree:

I agree with all of that but I'm thinking of the implications of R at almost 1 (as it likely is): this means no fast decrease in hospital numbers unlike what we saw in March. I shudder to think how hopitals will cope with the carnage of being even worse than now for a couple of months.

I also think blaming people lets the government and their even more lunatic backbenchers and press cheerleaders off the hook. As soon as the R rate for the new variant was known we should have seen a lockdown at least as strict as March across England.

Given attitudes towards things like peppermintteagate I doubt Cressida can do much without being all over the headlines as overreacting (they have been pretty busy cracking down on known lethal idiocy indoors for some time but the police are very overstretched due to budget cuts, illness and self isolating).

Post edited at 10:53
In reply to mick taylor:

I was reading an article in the Guardian yesterday, based on modelling by Edge Health, which was suggesting that (data up to 3rd January) about 22% of people in England had at some point had Coronavirus, but that in some areas, particularly London and the North West this was much higher.

It suggested that 54.2% of people in Barking and Dagenham had the virus at some point, nearer to home it gave figures of 38.8% and 38.6% in Liverpool and Manchester (I have a sibling in each, they have both had it). Rochdale doesn't fair much better (38%), but something jumped out at me and I did some sums.

The number of deaths per million population in Barking and Dagenham is 1,283 (272/212,000), whereas in Rochdale it is 2,200 (488/220,000). Which presumably says a lot about age demographics. It should be noted that the data on deaths is up to Christmas Day and sadly I think we might see a surge in the Barking and Dagenham numbers over the last few weeks.

 Offwidth 12 Jan 2021
In reply to wintertree:

I'd also like more finer detail on age groups.

Another point that seems a bit overlooked is access of press to hospitals has been very controlled (way beyond sensible H &S limits)... only a few like the BBC have been allowed in specific hospitals where things are likely more 'palalable' for public viewing. There seems to be increasing concerns that hospital staff are being pressured not to report desperate situations to the press.

 Si dH 12 Jan 2021
In reply to The New NickB:

Interesting point you make about the demographics, I saw that report yesterday but hadn't appreciated there was that level of disparity in deaths per million whereby some areas with lower modelled infections have had many more deaths. The modelling is obviously extremely sensitive to predictions of IFR as a function of age. I'm a little sceptical of some of the results from another perspective - they predict significantly different case/infection ratios (ie testing pick-up) for adjacent areas, which doesn't really sit right with me.  However, it's an interesting study.

 Offwidth 12 Jan 2021
In reply to Si dH:

The issue I have with that potential demographic difference is it is potentially bad news. At 50% you might be getting herd immunity levels in that age group and any case data drops in those hard hit areas of London might just be hiding the true R rate of the new variant, and how it will impact in a more typical UK area infection demographic.

Post edited at 11:28
 wintertree 12 Jan 2021
In reply to Offwidth:

> I agree with all of that but I'm thinking of the implications of R at almost 1 (as it likely is): this means no fast decrease in hospital numbers unlike what we saw in March

Indeed.  I don't think we're going to see total occupancy fall much for 4 weeks at least.  

> I also think blaming people lets the government and their even more lunatic backbenchers and press cheerleaders off the hook.

For sure.

Edit:  I also think the drop in occupancy you ascribe to the March lockdown was partly down to the business led shutdowns in the two weeks before national lockdown.  I think something similar is happening here, with the Christmas closures and the shift to Tier 4 as well as increased awareness.

Post edited at 11:39
In reply to The New NickB:

> The number of deaths per million population in Barking and Dagenham is 1,283 (272/212,000), whereas in Rochdale it is 2,200 (488/220,000). Which presumably says a lot about age demographics. It should be noted that the data on deaths is up to Christmas Day and sadly I think we might see a surge in the Barking and Dagenham numbers over the last few weeks.

The 'death certificate deaths' info is up to 25th Dec. I also look at (every day for key LA's in SE and NW) the 'deaths within 28 days of a positive test' and this info is up to 4th Jan and it does show a big increase, but the increases are significantly less for SE than NW equivalent (by equivalent I mean looking ate a period of high cases in say Wigan last October and looking at deaths a few weeks later).  I think the big disparity in the figures for Dagenham compared to Rochdale is mainly due to the autumn wave 2 not really hitting the SE, as well as demographics and inequalities.  

 minimike 12 Jan 2021
In reply to wintertree:

A week or so ago, you made a nice plot of rate constant vs. CurrentCases/100k which showed the trend that case rate acts to reduce its own growth rate in a self limiting way. 
We discussed whether it might actually be total cumulative cases rather than case rate which was mediating the growth.. did you ever make that plot? (No pressure, just interested!)

Post edited at 14:42
 wintertree 12 Jan 2021
In reply to minimike:

No, that completely slipped my mind.  It's a plot I want to do. I think the problem was that cumulative cases is unreliable at UTLA level as testing was so low in March/April.  This could be done at a regional level, using hospitalisations, but the regions for cases and NHS regions don't all correspond.  

I should just do one for cumulative cases from the start of mass testing.  That'll be simple enough.  

 minimike 12 Jan 2021
In reply to wintertree:

Most estimates seem to think the first wave cases peaked at about 100k/day nationally, so you could just scale the pre mass testing data arbitrarily to fit that. You wouldn’t be far off and it would only act as an offset anyway..

 Offwidth 12 Jan 2021
In reply to minimike:

Isn't it hard to scale as testing capacity was increasing fast during the first peak? I'd work backwards from hospitalsation data or deaths.

In reply to wintertree:

> Indeed.  I don't think we're going to see total occupancy fall much for 4 weeks at least.  

Doesn't it necessarily have to keep increasing for at least 2.5 - 3 in many areas?
How long is the median stay? (genuinely. Anyone know? I can't find good numbers on that)
It's only [time from detection to hospitalisation + that long] after cases peak that we see departures~=arrivals.......

The demographic stuff has got me thinking about what happens next. But that's probably not for here. There's some grounds for optimism, well, non-pessimism, well, backtracking from abject hopelessness at least, in the charts above.

 WaterMonkey 12 Jan 2021
In reply to minimike:

Presumably hospitalisation rates haven't changed from the first wave to now so surely you could estimate the cases per day based on hospitalisations per day (taking into account the lag time)?

 wintertree 12 Jan 2021
In reply to WaterMonkey:

> Presumably hospitalisation rates haven't changed from the first wave to now so surely you could estimate the cases per day based on hospitalisations per day (taking into account the lag time)?

Indeed; However, the behaviour we’re interested in is desynchronised at national and even regional level, so needs analysing more fine grained to understand it.  The publicly available summary data uses incompatible geographic reporting units for cases and hospitalisations at the regional level (English regions vs NHS regions) and local level (UTLAs vs individual NHS trusts).

If anyone can furnish me with groupings of compatible UTLAs & NHS trusts I can use that but it didn’t look like a trivial job to me...

 minimike 12 Jan 2021
In reply to wintertree:

Perhaps a middle ground.. could use the time-dependent ratio of hospitalisations to cases in wave 1 at a national level to correct cases for detection rate at the UTLA level?

In reply to wintertree:

1200+ deaths reported on a Tuesday, I think the numbers reported this Wed, Thu, Fri & Sat (the traditional high numbers in the week) are going to be pretty gruesome - and if the lag between cases and deaths is approx 3 weeks, that means the numbers are going to climb for at least another 3 weeks.

Although your (& Si dH & others) analysis is showing that there might be some light at the end of the tunnel, there's going to be a large amount of death and suffering before that becomes reality.

So unnecessary, government failure to act at the earliest possible opportunity, waiting instead until it's impossible not to act.

Post edited at 17:27
 wintertree 12 Jan 2021
In reply to Longsufferingropeholder:

>Doesn't it necessarily have to keep increasing for at least 2.5 - 3 in many areas?  How long is the median stay? (genuinely. Anyone know? I can't find good numbers on that)

To the best of my knowledge, longitudinally determined demographic breakdowns of time dependant distributions for detection to hospitalisation, and from hospitalisation to death or discharge have not been made public by the NHS.  Given that some of the recent LSHTM pre-print lead authors are on SAGE, and how that works by model fitting to non-longitudinal daily rates, I'm wondering if this information is even being gather (if not, why not?!?).

I assume that all these distributions are lagged, dispersed and demographically dependant. I can get values by fitting models to the measurables, but I don't trust them.

I don't think hospitalisations have to keep increasing for 3 weeks; the lag from detection to hospitalisation will be less I think for older people where it's more likely to be a Pillar 1 test on admission to hospital.  Further, the most aggressive growth in cases was in the 20-25 age bin it seems, where hospitalisations are lower, so the net cases curves at UTLA or regional level could be quite misleading.  I want to make a version that's weighted by my noddy demographic hospitalisations model to better understand this.  

> The demographic stuff has got me thinking about what happens next

Indeed.  You're not the only person to be thinking about that I reckon.

> There's some grounds for optimism, well, non-pessimism, well, backtracking from abject hopelessness at least, in the charts above.

I think so.  That's also what I thought 2 weeks in to the November lockdown and we got blindsided there by the new variant.  Unpredictable things can happen, and where we are now, the unpredictable seems quite likely to happen. 

> It's only [time from detection to hospitalisation + that long] after cases peak that we see departures~=arrivals.......

I agree - but both of those times are broadly dispersed distributions I think, and both have demographic dependancies.

In reply to Offwidth:

> Isn't it hard to scale as testing capacity was increasing fast during the first peak? I'd work backwards from hospitalsation data or deaths.

The problem is that this analysis needs to run at sub-regional level and there's no good correspondence between the geographic reporting units used for cases and hospitalisations.  I think after this crisis the reporting units (geographic and demographic) need sorting out to reduce the barriers to taking a data driven approach to informing good old fashioned infection control practices in a timely manner. 

In reply to minimike:

> Most estimates seem to think the first wave cases peaked at about 100k/day nationally, so you could just scale the pre mass testing data arbitrarily to fit that. You wouldn’t be far off and it would only act as an offset anyway..

Could do...

> Perhaps a middle ground.. could use the time-dependent ratio of hospitalisations to cases in wave 1 at a national level to correct cases for detection rate at the UTLA level?

That feels a bit more defensible.  I can see a way to a proper statistical model with the seroprevalence estimates and so on, but a quick job is probably all I'll manage.  Home schooling has rather over-constrained everything!

In reply to Michael Hood:

>  there might be some light at the end of the tunnel, there's going to be a large amount of death and suffering before that becomes reality.

Yes, and the danger here is that as soon as cases start to fall, without careful messaging people will think the problem is behind us.  It's important for cases to fall as fast as possible to inject some slack into the hospital pipeline.  A few posters have alluded to the relatively strong apparent media blackout on the hospitals.  I've noticed that more senior doctors are giving really quite angry statements now about people not following the rules.  I think they need to be given constructive help and air time including from inside hospitals and staff interviews to clearly communicate the reality of the situation.

Post edited at 18:28
 Punter_Pro 12 Jan 2021
In reply to wintertree:

Looking at the Zoe data today, it looks like for the first time in this current wave that symptomatic cases/deaths have started dropping. Hopefully a sign that we are now on a downwards trend and that the restrictions are starting to take affect.

Unfortunately we probably won't see this positive affect on the NHS for another 4-6 weeks but it is the first positive data I have seen in a while, I will await your take on it all before I get too excited however.

 SDM 12 Jan 2021
In reply to Punter_Pro:

I haven't looked at other areas but I was surprised to see the sudden switch in the Zoe estimate here.

It was rising very rapidly up until two days ago. Then it was level for a day, then it started dropping just as rapidly.

I think small sample sizes must be playing a role, I don't believe things can have improved that dramatically in such a short space of time.

Should become clearer in the next few days.

 minimike 12 Jan 2021
In reply to wintertree:

Methods:

‘The model was constrained by blah, blah, blah and home schooling.’ !!!

love it.

 wintertree 12 Jan 2021
In reply to Punter_Pro & SDM:

Interesting.  I’m not a fan of ZOE data but I’ll add it to the glimmer of hope list.

 wintertree 12 Jan 2021
In reply to wintertree:

A few stills from a new movie I'm working on that shows the exponential rate constant and characteristic time for English UTLAs, rather than case rates.

The 2021 plot is provisional as it's on the leading edge of the data.

Squint at the first plot and you’ll see Medway in orange for growth.

Post edited at 22:23

 Offwidth 13 Jan 2021
 wintertree 13 Jan 2021
In reply to Offwidth:

That is promising; the most recent updates to plot 18 show cases in those areas in very slow decline, and hospitalisations close to level (and presumably heading for decline).  

From the article:

>  The rolling seven-day total of admissions for the capital on 10 January was 5919, this was a fall of 131 on the previous day. 

The standard deviation of the expected statistical noise on a random large-numbers Poissonian variable would be about sqrt(5919) = 77, and the measured standard deviations (from a trnedline) on admissions seems to be about twice that, or ~155; so a fall of 131 days in that number has no statistical significance in itself, but it does give a high confidence that admissions aren't rising appreciably.

Glimmers of hope indeed.  I think we'll see them continue to drop very slowly over the next 10 days at least, give or take stochastic variation.  Hospital occupancy will continue to rise though given the lag from admission to departure.

More tentatively good news, more need than ever for clarity of messaging about this and the need for measures to be maintained until case rates are much lower.

In reply to wintertree:

> More tentatively good news, more need than ever for clarity of messaging about this and the need for measures to be maintained until case rates are much lower.

I think we can pretty much guarantee that's not going to happen. The government will see the NHS not get completely crashed and think they can get away with running hot and will release restrictions just too much just too soon.  People will be vaccinated and think everything is hunky-dory and stop behaving sensibly, and Hey Presto! Lockdown 4 before the summer.

In reply to Toerag:

> People will be vaccinated and think everything is hunky-dory and stop behaving sensibly, and Hey Presto! Lockdown 4 before the summer.

This is what I'm getting increasingly worried about. Not just the government part in it, but the almost inevitable shift in attitudes that I can see being a massive issue. Warning: gross generalisations on the way here. Apologies for those but I hope you can see past them to my points here...

The "me nexts" have already started from the teachers, supermarket workers and today the customs people. But... what happens when the over 60s have been immunised and (some, not all, but I bet there'll be enough to get noticed) decide they're invincible and start doing what the hell they want regardless of the rules? Then what happens when the millennials see the people who we've been doing all this to protect with the houses they don't have and the pensions they don't have enjoying the freedoms they don't have?
I've been trying to concentrate on the current one, but also looking forward that shitstorm. Can't see lockdown 4 being a viable option. This one really needs to get us out of the woods.

Post edited at 14:29
1
In reply to Offwidth:

Good news. The huge population in the SE and the downward trend should  push national statistics down. The NW is very worrying. Knowsley (Liverpool) nudging towards a rate of 1400 and still rising (but at a slightly lower rate of 70%). Plenty areas bad (but not as bad). And we do seem to have much higher death rate than the SE. I was hoping to see bigger reductions - a week of lockdown one did have a bigger impact than this one. Been out and about with work stuff - traffic is perhaps half as busy at it’s best. 

In reply to Longsufferingropeholder:

I think once the 15 million vulnerables are vaccinated then it’s chocks away. Vast % of less/not vulnerables have always felt lockdowns and rules were there to protect the vulnerables anyway. The government will be immense pressure to reopen everything and they will , in stages  

 neilh 13 Jan 2021
In reply to Longsufferingropeholder:

Sometimes its better just to ignore  the freedom issue when you look at what is going on in other countries. Its like leaping out of the frying pan and into the fire elsewhere.You cannot reaaly migrate away from a global pandemic.

In reply to mick taylor:

> I think once the 15 million vulnerables are vaccinated then it’s chocks away. Vast % of less/not vulnerables have always felt lockdowns and rules were there to protect the vulnerables anyway. The government will be immense pressure to reopen everything and they will , in stages  

It won't be. It can't be. That would still easily overrun the hospitals. We've done the maths loads of times before, but to recap: a really small percentage of everyone else is an enormous number.
It needs to be staged and careful, and gov are saying it will be, but I just don't see people playing along. Not without intense division and resentment.

 wintertree 13 Jan 2021
In reply to mick taylor:

> I think once the 15 million vulnerables are vaccinated then it’s chocks away. Vast % of less/not vulnerables have always felt lockdowns and rules were there to protect the vulnerables anyway. The government will be immense pressure to reopen everything and they will , in stages  

That’s what I worry about because it has the potential to be an utter disaster if you look at inferred demographic hospitalisation number and actual demographic death numbers.  Exponential growth ate up the 50% headroom on ITUs from dexamethasone etc in two weeks.  It’ll do similar with headroom from vaccinating all > 65s if allowed to.

In reply to neilh:

> Sometimes its better just to ignore  the freedom issue when you look at what is going on in other countries. Its like leaping out of the frying pan and into the fire elsewhere.You cannot reaaly migrate away from a global pandemic.

Not really sure what you mean there...
I'm happy to but I can see a growing number having different feelings as immunity, either true or perceived, drives a wedge into society.

Post edited at 15:58
 wintertree 13 Jan 2021
In reply to Longsufferingropeholder:

> I'm happy to but I can see a growing number having different feelings as immunity, either true or perceived, drives a wedge into society.

They are the wedge.  

They'll be driven by the same malevolent forces that have been driving a lot of the push back on control measures that have century old evidence showing their efficacy for public health and for mitigating economic harm.

Still, we have to get to that point first.  Another poster whose I think has shown consistently very sound judgement over the years has reached a similar conclusion to you, as have a couple of my offline contacts.

In reply to wintertree:

> Another poster whose I think has shown consistently very sound judgement over the years has reached a similar conclusion to you, as have a couple of my offline contacts.

It's one of the few bits that I'm having trouble seeing past. Typically (normally isn't the word) in this saga, with a bit of thought it's kind of obvious what happens next, and normally I'm not the only person I know who's already there, but this issue just looks like an unavoidable and massive societal kraken that nobody's talking about yet.
I've alluded to it in a couple of other posts but chose not to say it out loud because I figured no good could come of pointing it out to anyone who can't see it coming, but..... I could really use some views on how it gets resolved because my own thoughts aren't getting me anywhere.

Post edited at 16:04
 wintertree 13 Jan 2021
In reply to Longsufferingropeholder:

> I could really use some views on how it gets resolved because my own thoughts aren't getting me anywhere.

Likewise.  

I think a true and honest projection of different scenarios will be critical to informed decision making on the unlocking process.  The modelling groups have much more time critical things to deal with right now, but I hope this is something that can be modelling and presented in ~ 4 weeks assuming the exponential rates all do what I think they're doing and about to do.  

This is all assuming we get to that point without one of the growing number of variants with multiple changes to the RBD changing the situation.  Again.

In reply to wintertree:

> > I could really use some views on how it gets resolved because my own thoughts aren't getting me anywhere.

> Likewise.  

> I think a true and honest projection of different scenarios will be critical to informed decision making on the unlocking process.  The modelling groups have much more time critical things to deal with right now, but I hope this is something that can be modelling and presented in ~ 4 weeks assuming the exponential rates all do what I think they're doing and about to do.  

Someone must be thinking about it. Not a sniff on the RAMP forums yet, which is why I'm getting worried that maybe nobody is. But somebody must have thought of it. It wouldn't make sense to make it a thing, for obvious reasons, but... yeah, we're on the same page here.

> This is all assuming we get to that point without one of the growing number of variants with multiple changes to the RBD changing the situation.  Again.

Or worse..... There was an interesting Science Vs on this. Read about "polybasic cleavage sites" then sleep well.

 neilh 13 Jan 2021
In reply to Longsufferingropeholder:

The pont is that the position here is not particularly unique. You can go to any part of the globe ranging from " Communist " China to "freedom loving" USA to lockdown heavy Australia  and you will find lockdowns etc a pretty common tool to control the pandemic. All have little variations but broadly its the same. Arguing  about freedoms being curtailed and inter generational issues is just noise and gets you nowhere imho.

Basically you cannot migrate yourself away from the virus to a better world. Its not good  everywhere.

Post edited at 16:29
1
 wintertree 13 Jan 2021
In reply to neilh:

> Arguing  about freedoms being curtailed and inter generational issues is just noise and gets you nowhere.

I think you're misunderstanding the concern that they and I have.  It's not about "freedom", it's about a coming storm driven by a push to unlock faster and sooner than is actually safe, either for the NHS or for many younger individuals, once what is perceived to be a critical number of the more vulnerable are immunised.  

>  All have little variations but broadly its the same

Pretty major variations in my view.  

In reply to neilh:

> Basically you cannot migrate yourself away from the virus to a better world. Its not good  everywhere.

Yeah what I'm saying will be the same everywhere. It's not particular to the UK but it's looking like we'll get there first.
I'm not convinced that we disagree here.

Post edited at 16:34
In reply to wintertree:

> I think you're misunderstanding the concern that they and I have.  It's not about "freedom", it's about a coming storm driven by a push to unlock faster and sooner than is actually safe, either for the NHS or for many younger individuals, once what is perceived to be a critical number of the more vulnerable are immunised.  

My concern was sort of the same but at the same time not quite the same; regardless of how quickly or slowly the laws are relaxed, I was thinking about what happens when a strong-minded minority of the vaccinated go against guidance and take it upon themselves to make the decision because they don't see why they shouldn't go out and play, and everyone else sees them going out to play. Best case is that that minority quickly grows, because "if they can why can't I?". Then the section of society that already resents that demographic will not take it well.

Edit: maybe it is the same thing.

Post edited at 16:41
 wintertree 13 Jan 2021
In reply to thread:

Today's data update feels quite promising to me in terms of case numbers nationally and for the "red" regions around the Thames on my regional plots.  More of a mixed message for the "blue" regions elsewhere.  It's hard to interpret the release fully as it's reporting mainly on weekend days.  

Latest hospital plot.   18 days ago on this plot, hospitals in England had as many covid patients as wave 1, but with only half the number in ITU.  Now, hospitals are twice as full and ITUs are as full as they were in April.

The lesson here is that "halves admission to ITU" sounds phenomenal - and it is for the individuals directly helped by it, but it bought us only 2.5 weeks grace under exponential growth, - and that was growth subdued by widespread Tier 4 than national lockdown.

Alan AtKisson explained it better than I can a couple of decades ago.

https://www.youtube.com/watch?v=bghbxemp4kQ&

Post edited at 16:48

 wintertree 13 Jan 2021
In reply to Longsufferingropeholder:

> Edit: maybe it is the same thing.

I think we are thinking of the same precipitating events , but we've got different views of how it could go wrong.

I think expectation management and messaging are going to remain key tools of competent leadership through until June at least.  So that's alright for us then.

In reply to wintertree:

> The lesson here is that "halves admission to ITU" sounds phenomenal - and it is for the individuals directly helped by it, but it bought us only 2.5 weeks grace under exponential growth, - and that was growth subdued by widespread Tier 4 than national lockdown.

I know you know this but it pays to point out for anyone who doesn't that care has changed since the first go-around.
One of the lessons was that going to a ventilator often doesn't help, and NIV (CPAP) is now the standard of care for an awful lot of people who would have gone to the ITU in April. So number in ITU is not an indicator of how many people are really really sick.

Post edited at 16:55
In reply to wintertree:

> > Edit: maybe it is the same thing.

> I think we are thinking of the same precipitating events , but we've got different views of how it could go wrong.

Yeah, sounds like your "push to unlock faster" and my resentful, rioting millennials are manifestations of the same thought process through two different imaginations.

 wintertree 13 Jan 2021
In reply to Longsufferingropeholder:

> One of the lessons was that going straight to a ventilator often doesn't help, and NIV (CPAP) is now the standard of care for an awful lot of people who would have gone to the ITU in April. So number in ITU is not an indicator of how many people are really really sick.

Indeed.  

11 times out of 10 I’d take CPAP over sedation and ventilation though.  Might ask for some Valium to help pass the time though, and some decent in-ear headphones.  Ventilation is brutal; it’s notable almost nobody over 85 is being admitted to ITU (from an NHS spreadsheet I may be misremembering)

 Si dH 13 Jan 2021
In reply to Longsufferingropeholder and wintertree:

There's obviously potential that, whether through designed government policy or or through civil disobedience and government capitulation, we end up removing restrictions too quickly once all of the most vulnerable people are vaccinated. However, there is also a potential we do it too late. A really sensible and balanced argument has to be constructed here because the best answer is not obvious; it is not good enough on anyone's part to present the argument as one-sided as this just leads to division.

While large numbers of people are dying and many more are afraid to go about their lives anyway for fear of infecting their relatives, the best response is clearly always to reduce infections to low levels, as we have discussed many times - the economy/health balance that gets presented by some people is a false one. However, once few people are dying that is no longer the case. Putting aside overwhelming of the NHS for a moment (see below) it becomes a trade off between how many people become seriously long term sick due to releasing restrictions and how many people lose their livelihoods or/and become mentally ill as an indirect result. That is an important trade off that in my opinion will have to be considered. People like some of us in our late 30s and 40s (or even maybe 50s) cannot expect a default position that millions of people stall their livelihoods to prevent us getting sick, in the way that if large numbers of people were still dying, we would.

The really strong argument to maintain restrictions will be if healthcare overload is on the way. We have discussed this potentially happening due to an overload of people in their 40s and 50s, but I have never seen any numbers that explored the possibility properly. Let's assume people start to relax significantly once all the over-60s have been vaccinated and hospital occupancy has dropped substantially in, say, April. If we all went back to Tier 1 immediately, how long do we think it would actually take to infect enough people in their 50s, 40s, 30s to overwhelm the NHS again? Of course if everyone in the country was infected immediately, it would happen quickly, but that's not plausible. With Tier 1 last summer, with no vaccinations but with the original variant, infections grew slowly for the most part (fast in some areas causing local problems of course), the key point being that overall NHS capacity was never close to being threatened in that period. Can we make an estimate of what R might be given that situation in late spring with the new variant, and then calculate how long it would take to infect enough people in the age range 30-50 to reach levels whereby we seriously risk hospital capacity, given exponential growth at the predicted rate and an assumption about the length of time people spend in hospitals? I'm far from convinced this modelling would give us the answer we fear, furthermore it would be very conservative because it would ignore the self-moderating behaviour we have discussed recently in how infection rates seem to progress and it would ignore behavioural effects of people in their late 40s and 50s if they really did see hospitals starting to max out with people their age.

Bit of a long post and I hope you don't take this as a 'let it rip' post; it isn't. What I'd like to see is a reasoned argument supported by evidence on the healthcare overload topic, if we think we could put one together.

P.S. I'm not convinced yet that infection rates are starting to drop across the board. They have definitely started slowly dropping on average in London, SE and East of England now but elsewhere it could still be anything from a slowing down of the increase, to a plateauing, to the start of a slow reduction. It's difficult to tell yet. The strong weekend effect appears to be coming back with a fairly high number of infections looking like it could build on Monday 11/01.

Edit: I wrote this just before your 16:45 post which makes similar points about the latest data.

Post edited at 17:29
In reply to Si dH:

I fully agree with all of this.
Would add one point:

> People like some of us in our late 30s and 40s (or even maybe 50s) cannot expect a default position that millions of people stall their livelihoods to prevent us getting sick, in the way that if large numbers of people were still dying, we would.

Employers cannot expect people like some of us (healthy late 30s here but valid for the range you mention) to leave the house while that shit's going off.
So.... just reverberating, would not want to be making these decisions.

 wintertree 13 Jan 2021
In reply to Si dH:

I take the point of your post, but my view fundamentally hasn't changed since March when I said this: "This virus could be a relatively benign precursor for something worse.  The fewer people it infects, the lower the probability of that mutation occurring." [1]

We can trade off opening the economy vs hospital occupancy, but I think this is a myopic trade-off, as it invariably results in running with large numbers of cases.  The better we get at shielding and protecting the vulnerable, the hotter It runs in younger, less vulnerable people.  

The more cases there are, the higher the probability of a worse variant emerging, and the harder it is to identify, contain and eliminate that variant.  The last month makes that abundantly clear.

The "Kent" variant does not appear to evade immunity acquired from infection by previous variants or given by the current vaccines.

For some of the other emerging variants, that immunity is an open question with a scramble too answer it.

We are playing with fire, and doing so has become normalised in many individual's and institution's minds.  But the risks just over the horizon are as real as ever, and the horizon is suddenly very close.

Playing with fire is not a good strategy for the long term health of either the people or the economy.  This virus is perhaps only 10% as bad as states it probably has evolutionary access to.  Immune evasion through mutation has already been demonstrated in a (hopefully secure) lab setting.

[1] https://www.ukhillwalking.com/forums/off_belay/is_it_worth_it-717284?v=1#x9154383

Edit: To give a rather OTT and fictional example of where this is going, the Rick and Morty episode Rick Potion #9 springs to mind.

Post edited at 17:33
 wintertree 13 Jan 2021
In reply to Si dH:

> P.S. I'm not convinced yet that infection rates are starting to drop across the board. They have definitely started slowly dropping on average in London, SE and East of England now but elsewhere it could still be anything from a slowing down of the increase, to a plateauing, to the start of a slow reduction. It's difficult to tell yet. The strong weekend effect appears to be coming back with a fairly high number of infections looking like it could build on Monday 11/01.

Yes, beyond my "red" areas the picture is very mixed.  Perhaps we can tell when that Monday clears through the reporting lag.

>  We have discussed this potentially happening due to an overload of people in their 40s and 50s, but I have never seen any numbers that explored the possibility properly.

I reckon the red line healthcare for “no covid control measures” falls somewhere around 56 assuming mass immunisation above that age.  It’s done backing out a demographic model from the 18-64 year age bin in the hospitalisation data and it’s not a robust or peer reviewed model so take it with a pinch of salt.  It’s a model that needs to be run by a professional modelling group pretty soon I think.

Post edited at 17:40
In reply to wintertree:

> I reckon the red line healthcare for “no covid control measures” falls somewhere around 56 assuming mass immunisation above that age.  It’s done backing out a demographic model from the 18-64 year age bin in the hospitalisation data and it’s not a robust or peer reviewed model so take it with a pinch of salt.  It’s a model that needs to be run by a professional modelling group pretty soon I think.

Presumably neglecting the cost of long covid and/or other as yet undiscovered long term effects?

It's still a no thanks from me. I'll see you all in 2023, once we know whether your dick flies off a year after infection.

 wintertree 13 Jan 2021
In reply to Longsufferingropeholder:

> Presumably neglecting the cost of long covid and/or other as yet undiscovered long term effects?

You noticed that too, huh?  I’m particularly interested in the cumulative effects of getting hit by repeat infections as the virus evolves.  

What I want is hard borders, military run MIQ, and a push for near elimination this summer followed by low rates going forwards.  I’m 100% certain I won’t get what I want.  It’s not all doom and gloom in my head though, more and more continues to be learnt about the damage mechanisms of this type of virus, and how they suppress immunity until it’s too late.  I think we’re going in to a multi-years arms race with this virus but hey, at least it’s happening in 2021 not 2001.  That makes a big difference.  If this had hit in the 1990s, Christ what a mess.  If this doesn’t serve as a warning to humans about our utter unpreparedness for the next chance meeting with the great filter, I don’t know what does.

In reply to wintertree:

My next charitable donation will be to Giant Fire-breathing Lizard Research UK.

Not many Hollywood disaster movie standards left to come true.

Post edited at 18:52
In reply to wintertree:

Apart from pandemics there are 2 biggies (that I can think of straight away) that could seriously hurt (i.e. put the survival of the human race in doubt) that humans are totally unprepared for:

1. Some big rock happily floating about the sun finds a small wet blue planet in the way - various disaster movies have rather made most people think that we could deal with this - no chance. First of all we've got to see it coming and even with all the advances in the last few years there's a fair chance we would have bugger all warning (which might not be a bad thing).

2. Yellowstone goes pop - again, disaster movies have shown "escape" from this - ha-ha, to where? it would be global.

And then there are the ones that would just do rather a lot of economic damage (as well as killing a few million) - like the side of the mountain on one of the Canaries (or is it Madeira?) just sliding off into the sea which it is almost certain to do one of these days - would make a bit of a splash that would happily propagate across the Atlantic into a rather nasty tsunami along the whole of the eastern seaboard of the USA.

Oh I'm full of hope tonight 😁 - you can see why Elon Musk wants to get a viable self-sustaining colony on Mars - we really do (as a species) need to get some of our eggs out of this basket.

In reply to Michael Hood:

> Oh I'm full of hope tonight 😁 - you can see why Elon Musk wants to get a viable self-sustaining colony on Mars - we really do (as a species) need to get some of our eggs out of this basket.

And right on cue, static fire... (Sorry, wrong thread, but so appropriate)

 wintertree 13 Jan 2021
In reply to Michael Hood:

That’s the Cumbre Vieja in La Palma in the canaries.  I tried jumping up and down on it a decade ago but it didn’t budge.  

Not so long now before we have giant orbital laser arrays to deflect space rocks; there’s serious money being spent on that for Breakthrough Starshot.

Post edited at 19:42
 bruxist 13 Jan 2021
In reply to wintertree and Longsufferingropeholder:

I can envisage a number a factors that might come into play over this year to influence popular opinion about restrictions. One is that as long as our infection rate is high, Brits will remain on other nations' travel ban lists, and that's not something that the UK Gov can relax. Not being able to jump on Easyjet to Ibiza might persuade working-age Brits that getting the infection rate down is a good idea - indeed it might be more persuasive than the death rate.

On the other hand, as long as holidays abroad are off the table, we'll continue to see a lot more travel around the UK - and idiots in the hills - and that's surely going to keep moving infection around the country, helping to keep the rates high, increasing the risk of further mutations, and further stressing an already exhausted cadre of NHS staff thereby impacting availability of healthcare.

I've seen a lot of chatter from younger Brits about what happens once the elderly vulnerable are vaccinated, and how they're not going to stay under restrictions once they can see all the old folk partying it up. But I can't conjure a realistic picture of this happening. Are 20-year-olds really going to become suddenly insanely jealous because Mecca bingo halls are full again?

1
 wintertree 13 Jan 2021
In reply to Si dH:

> If we all went back to Tier 1 immediately, how long do we think it would actually take to infect enough people in their 50s, 40s, 30s to overwhelm the NHS again? Of course if everyone in the country was infected immediately, it would happen quickly, but that's not plausible. With Tier 1 last summer, with no vaccinations but with the original variant, infections grew slowly for the most part.

I was contemplating this over dinner.

Cases grew slowly in absolute terms in England over the summer, but in exponential rate terms, it was the most extreme growth at national level we've seen since mass testing began, and that includes the current growth of the new variant.  The high point for cases on Plot 9e is around 08-29 where they were doubling faster than at any point with the new variant and T2/3/4 control measures (when looked at at national level).  If everything went back to Tier 1 and stayed there with say > 60s vaccinated, I think we'd see quite rapid doubling of cases in line with this summer.  Moderated a bit by vaccinated health workers and infection acquired immunity.   As we are seeing now, it only takes a few weeks of exponential growth to un-do all the beneficial factors.

If the government go down the route of unlocking as fast as possible, I hope that they do this with a tier system that responds locally to the exponential rate and not the case rate, so that growth Is moderated before it becomes a major problem.  I think everyone left working in healthcare by the summer could really do with a break, rather than dealing with more fall out of running things hot.  I somehow don't think we'll see this.

Post edited at 21:18
In reply to wintertree:

Well the first problem is getting the government to understand that it's not the absolute numbers that count, it's an exponential increase that is the indicator that action needs to be taken straight away. I have my doubts about the average IQ of the cabinet being high enough to understand that.

Secondly, that message needs to be consistently given to the public until it's understood by enough people, and I think properly explained, enough of the public would understand. I rate their average intelligence above the government's.

 RobAJones 13 Jan 2021
In reply to Michael Hood:

Unfortunately I think you are correct. I was regarded as a bit of a "lock down nutter" (except here) when I expressed concern in Sept/early Oct that the current restrictions weren't sustainable. I'm not sure whether lessons have been learned since then. I also think "it's just like flu" argument will resurface, all be it covid at 40 = flu at 80 

 Si dH 14 Jan 2021
In reply to thread:

Some data emerging from the SIREN in to likelihood of covid reinfection. I first spotted it on BBC news:

BBC News - Past Covid-19 infection may provide 'months of immunity'
https://www.bbc.co.uk/news/health-55651518

...but when trying to find the actual study, discovered the results have so far only been briefed verbally, which inevitably has led to different people reporting slightly different things. I found this page which shows various scientists takeaways:

https://www.sciencemediacentre.org/expert-reaction-to-a-preprint-from-the-siren-study-looking-at-sars-cov-2-infection-rates-in-antibody-positive-healthcare-workers/

TLDR; previous infection appears to convey immunity with efficacy of about 80% against infection, and 90% against symptomatic infection, for at least 5 months, in healthcare workers (around 20000 were followed for the period, including over 6000 with prior infection). Similar rough order impact to the vaccines. No discussion over whether there was any difference between people who had a more or less severe prior infection, or between people in/more less vulnerable groups.

Post edited at 06:52
 wintertree 14 Jan 2021
In reply to Si dH:

Interesting, thanks.  It’s good to have some actual numbers; the speculation on here on re infection has been a bit woolly.  80% is only just enough perhaps to stop the new variant from circulating widely without control measures.  

It will be interesting to see what commonality can be found by re-infected people and people infected some time after immunisation.

So, with effective immunisation as well it can probably be pushed to endemic, seasonal status.  A ticking time bomb perhaps.

I wonder what will happen to the lighthouse lab capacity after this crisis is over?  Work through major diseases one by one pushing for elimination?  Made available to industry for other uses?  Mothballed until the time bomb goes off?

In reply to wintertree:

> I wonder what will happen to the lighthouse lab capacity after this crisis is over?  Work through major diseases one by one pushing for elimination?  Made available to industry for other uses?  Mothballed until the time bomb goes off?

Sold off cheaply to their mates.

In reply to wintertree:

Giant fire-breathing lizard research

In reply to bruxist:

> I've seen a lot of chatter from younger Brits about what happens once the elderly vulnerable are vaccinated, and how they're not going to stay under restrictions once they can see all the old folk partying it up. But I can't conjure a realistic picture of this happening. Are 20-year-olds really going to become suddenly insanely jealous because Mecca bingo halls are full again?

Not universally and not restricted to a single demographic, but I can see it happening in big enough numbers to make compliance with the remaining distancing measures a massive problem.

 neilh 14 Jan 2021
In reply to bruxist:

I cannot imagine the current tranche of 80 year olds " partying it up". 50 years plus being let lose is a good thing as it has a by product of opening up hospitality which employs alot of younger people.There seems to be a bit of a disconnect on the wishes of various generations stoked by ignorance.

As you say though its not what the Uk does that counts.I suspect that France, Spain and Itay for example are not going to accept huge swathes of UK citizens irrespective of testing.You are going to have to have a good reason to go other than a holiday ( the Dutch have already snet back Uk citizens unless they have a good reason to be there). USA is off the cards,so is Aus and New Zealand. Forget Thailand and Vietnam( unless you want to lock yourslef in a hotel for 2 weeks).

Greece or Turkey.Even then if they are being hit by the new variant its debateable they will let loads in.

Uk is going to be the main fall back again and let us hope its a good summer.

Post edited at 09:16
 jonny taylor 14 Jan 2021
In reply to wintertree:

> If the government go down the route of unlocking as fast as possible, I hope that they do this with a tier system that responds locally to the exponential rate and not the case rate, so that growth is moderated before it becomes a major problem.

Of course, if they actually go for near-elimination, like Scotland could almost start to claim over summer 2020, then they wouldn't be looking at either R or the case numbers - they would be responding to individual outbreaks with a functioning test-and-trace system. For a short while it seemed as if Scotland almost had that working - seemed to manage and shut down a couple of local outbreaks. Then Aberdeen happened. I don't know if that was too big to contain, or if it just marks the time when things got out of control again across the country.

Edit: and Aberdeen would also seem to show that if you want things contained you need to be very firm in enforcing the reduced set of rules that still remain in place.

Post edited at 09:26
 jonny taylor 14 Jan 2021
In reply to Longsufferingropeholder and wintertree:

Harking back to yesterday's posts about what happens in the summer when the older generation are vaccinated, and whether anybody is planning for that...

Apparently Sturgeon dropped a hint earlier this week about one week lockdowns over the coming summer, which seems to me rather odd. Doesn't seem like long enough to do anything. But (putting 2 and 2 together to make about 10) I did start to wonder whether this could be a hint of a modified strategy to try and keep a bit of a lid on things once compliance with months-long lockdowns has been completely lost.

In reply to jonny taylor:

Similar techniques have been used in Australia. They recently applied a local 3-day (I think?) super-strict lockdown in response to a handful of cases.
Not sure if the supporting logic is the same in the posited situation but it's definitely a thing in other countries.

In reply to Longsufferingropeholder:

3 days may not contain an outbreak but if you've already traced and quarantined the initial contracts it gives competent authorities enough time to find out (with reasonable certainty) whether it has been contained or not.

Competent authorities is the tricky bit in the UK.

In reply to Michael Hood:

> 3 days may not contain an outbreak but if you've already traced and quarantined the initial contracts it gives competent authorities enough time to find out (with reasonable certainty) whether it has been contained or not.

That seems to be the mechanism, yes. It's potentially long enough to find out if you caught them all.

> Competent authorities is the tricky bit in the UK.

We're not the only country that's lost faith in their government. One thing I've taken from reading a few other national sources is that they're either all just as bad, or everyone around the world enjoys shitting on their government as much as we do. Or is it both....?
We seem to look to Germany fairly often as an example of how to deal with this, but if you read DW you can quickly realise it's all subjective.
It'll get better eventually, like testing did, and treatments did, and vaccination is about to.

 

Looks like it's starting already:

https://www.bbc.co.uk/news/business-55654127

In reply to Longsufferingropeholder:

> Looks like it's starting already:

Hmm, I wonder how "the public" are going to react when we hear about hospitalisations (and maybe deaths) of people who have been vaccinated.

And will that reaction depend on whether those cases have had just 1 or both jabs?

I think there's a lot of (understandable) misunderstanding about vaccines; i.e. that if enough are vaccinated then society as a whole will end up pretty close to 100% protected but individual protection is a lot less than that.

 Offwidth 14 Jan 2021
In reply to Si dH:

Presumably thus study was all on the old variant?

In reply to Michael Hood:

> Hmm, I wonder how "the public" are going to react when we hear about hospitalisations (and maybe deaths) of people who have been vaccinated.

> And will that reaction depend on whether those cases have had just 1 or both jabs?

> I think there's a lot of (understandable) misunderstanding about vaccines; i.e. that if enough are vaccinated then society as a whole will end up pretty close to 100% protected but individual protection is a lot less than that.

It's looking like protection from severe illness/hospitalisation is actually going to turn out pretty close to 100%. But you're absolutely right; even with 99.a lot of nines there's going to come a day when someone dies despite being immunised and it will be all over the papers.
I suspect the story will fizzle out amongst a background chorus from 'let it rippers' who change their tune and want compensation when they realise their dick has flown off
(That's a metaphor in case anyone hasn't realised yet. It'll realistically be chronic lung damage or something)

 Blunderbuss 14 Jan 2021
In reply to Michael Hood:

> Hmm, I wonder how "the public" are going to react when we hear about hospitalisations (and maybe deaths) of people who have been vaccinated.

> And will that reaction depend on whether those cases have had just 1 or both jabs?

> I think there's a lot of (understandable) misunderstanding about vaccines; i.e. that if enough are vaccinated then society as a whole will end up pretty close to 100% protected but individual protection is a lot less than that.

I don't think there'll be much reaction.....as a result of COVID-19 is has become far more wildly known that an average of 8000 a year die from flu and no one really gives a shit about that in the general scheme of things. People will accept a few thousand deaths for the sake of opening up society.

1
 wintertree 14 Jan 2021
In reply to Longsufferingropeholder:

> That's a metaphor in case anyone hasn't realised yet. It'll realistically be chronic lung damage or something)

I don’t know.  I’m not yet disuaded from my earlier view that this is really a vascular and endothelial disease.  Impaired flow may not bake your dick fall off, but rumour has it it’s not unimportant to its normal function.

(Yes, that’s a geometric allusion.)

In reply to Longsufferingropeholder:

The way I’m feeling at the moment I don’t want anyone travelling in or out of the UK.  Seriously. I’d also like to keep our cash within the UK (which normally doesn’t bother me but I think we are so screwed).

ed: I dread some carefree vaccinated holidaymaker importing a vaccine tolerant mutant mega fast virus. 

Post edited at 10:59
 wintertree 14 Jan 2021
In reply to jonny Taylor and Longsufferingropeholder:

> Of course, if they actually go for near-elimination, like Scotland could almost start to claim over summer 2020,

I don't think it would have been an impossible push in England either, excepting obvious and critical problems around messaging and resultant public support (or lack thereof).

The lesson from NZ is that there is never perfect elimination, and that the moment a case is detected, a hard and wide lockdown net needs to be cast, with immediate full sequencing of the cases(s) to track their lineage as part of high priority, informed contact tracing.   This would need to be done by actual local public health teams on the ground and not the national level system.  This needs pre-prepared action plans and a next-day response (same day on measures like masks and open windows in workplaces and public transport), with the PM directly in the loop and the nation clearly prepared to accept these measures as an alternative to another 30,000 dead, almost everything but Covid cancelled from hospitals and children out of school for months on end and endless disruption to business.  Given the extensive examples of awful consequences, there's a guano mine of examples to illustrate why it's worth accepting short, hard restrictions when times are good in pursuit of elimination.

> Edit: and Aberdeen would also seem to show that if you want things contained you need to be very firm in enforcing the reduced set of rules that still remain in place.

Yes; you have to limit the reach of super-spreaders.  It seems like a lot of the loss of control in the North West and in County Durham were a few super-spreader events.  The one in Durham as an indoor party at a WMC after a charity football match.  I think that requiring vaccination or medical exemption for entry to venues associated with a high risk of super-spreading would allow for more relaxed rules that are more akin to a pre-covid normal.

Update: Mick Jagger just phoned this in -  https://www.youtube.com/watch?v=oqMl5CRoFdk&

1
 wintertree 14 Jan 2021
In reply to mick taylor:

> ed: I dread some carefree vaccinated holidaymaker importing a vaccine tolerant mutant mega fast virus

But they've got to go skiing for their mental health!  Which is going to be far more severely compromised by the effects on society caused by a few million people flying out and back this summer.  Remember - most of the world isn't sequencing anything like we are, and we didn't identify the threat posted by the new variant until it was too late. PCR testing on arrival or departure probably misses half off cases.  The world is riddled with Covid right now, and variants are cropping up like bad smells in a curry house bogs. 

I'm worried that my disconnect between how I see all this and the view of government and the masses is widening and widening - but then I also gather from someone there that the emergence of more and more variants is a main driver of concern in NZ where by and large the whole country is aligned with my take.  It's notable there that when we had the early "let it rip" posters on UKC and appearing in the media, investigative journalists dug in to their funding and political background in NZ and shot them down.

 neilh 14 Jan 2021
In reply to Longsufferingropeholder:

Itr s super strict in Aus and that was after impostion of super strict lockdowns and rigorious imposition of barely anything moving.

 neilh 14 Jan 2021
In reply to Longsufferingropeholder:

Yep. I recommetd that people broaden their reading to include European and other news.

Reading Le Monde or DW gives a good perspective. It is not good and people need reminding it  is global.

In reply to wintertree:

All:  just been looking at my ‘sample’ of LAs in the NW and SE of the country. Daily deaths going up everywhere and by some big numbers.  My fag packet maths makes me think we will see huge increases (areas outside of SE appear to have higher death rates and since these area still increasing in cases I fear this will offset reductions in SE (which will only kick in in a week or so).  Gloomy and depressing. 
 

Only the other day I was saying the govt wouldn’t bring in stricter lockdown. Not so sure now. Any sense of massive increase in deaths may force their hand into action MORE than the threat of increasing cases, because it looks worse from a PR perspective. 

In reply to mick taylor:

> ed: I dread some carefree vaccinated holidaymaker importing a vaccine tolerant mutant mega fast virus. 

It's not a popular point to make but it's arguable that as things stand we're more likely to export one than import one at the moment. That will of course change as we get on top of things.

In reply to wintertree:

Elimination isn't a viable option while there will always be a reservoir somewhere in the world. It would be a prohibitively expensive goal to pursue even if possible.
The whole world will end up going for a strategy of vaccinating it down to a tolerable level, mostly of mild cases. Even countries that have 'eradicated' it won't be able to sustain the economic harm of what it takes to keep it that way, especially when the rest of the world is open for business.

 minimike 14 Jan 2021
In reply to wintertree:

I’m literally just back from a ski.. in the park! I poled from the front door so I’m claiming it as legit solo local exercise. Fresh powder turns in Leeds!

I can confirm it was excellent for my mental health and I may bend the rules by repeating later if it keeps snowing.

 jkarran 14 Jan 2021
In reply to bruxist:

> I can envisage a number a factors that might come into play over this year to influence popular opinion about restrictions. One is that as long as our infection rate is high, Brits will remain on other nations' travel ban lists, and that's not something that the UK Gov can relax. Not being able to jump on Easyjet to Ibiza might persuade working-age Brits that getting the infection rate down is a good idea - indeed it might be more persuasive than the death rate.

But it's not really a choice individuals make, it's about whether government continues to support people and businesses whose activity/income is restricted in order to keep infection levels down. If the government policy remains 'run hot' once the most vulnerable are immunised then the choices of individuals to not participate in that barely matter, plenty more will make different choices, prevelence will be high and the necessities of life (once state support is withdrawn): attending or seeking work, access to food and services mean individuals attempting to make different choices will be exposed anyway.

> I've seen a lot of chatter from younger Brits about what happens once the elderly vulnerable are vaccinated, and how they're not going to stay under restrictions once they can see all the old folk partying it up. But I can't conjure a realistic picture of this happening. Are 20-year-olds really going to become suddenly insanely jealous because Mecca bingo halls are full again?

There's some discussion upthread about how hot we can run once the vaccine is deployed above age X, what the maximum prevalence we can stand without overloading hospitals is for different values of X. It's a reasonable question to ponder but it does slightly overlook the fact the hospitalisation to death ratio is unlikely to change barring the discovery of new theraputics. If you run hospitals at full capacity you run deaths at the associated rate too which probably maxes out somewhere between 1000-1500/day nationally if you want to keep non covid hospital services barely functioning. As has been pointed out it may be somewhat self limiting but our ability to moderate our behaviour personally to control the virus at population level given the complexity of the situation and the significant lags would be strongly influenced/limited by the quality of the data and interpretation available to us through the media (assuming government largely steps aside which is the scenario we're considering). Those would be overwhelmingly poorer working people and parents of children dying in that 4th wave so the consequences and narrative would be different.

It's a dangerous moment in the run of the epidemic, one that (alarmingly) requires clear, principled leadership for us to navigate safely.

jk

Post edited at 11:38
 wintertree 14 Jan 2021
In reply to minimike:

Shovelling 15 cubic meters of snow has also proved excellent for every aspect of my health.  I’m thinking of heading out on my snow shoes later...

 TomD89 14 Jan 2021
In reply to wintertree:

> > Of course, if they actually go for near-elimination, like Scotland could almost start to claim over summer 2020,

> I don't think it would have been an impossible push in England either, excepting obvious and critical problems around messaging and resultant public support (or lack thereof).

Not impossible but improbable IMO. You'd be looking at getting to the end of this strenuous 3 month winter lockdown and then convincing the public to extend lockdown into the spring/summer for another 2-3 months possibly with even stricter rules despite daily infection/death figures being at a theoretically all time low.

If the vaccination program is on schedule and all/most of the vulnerable have had the jab, I don't think you'd be able to generate the impetus to push for elimination even if it was the best option. The other issues caused by lockdowns will be more stark by then as a counter-incentive.

> I think that requiring vaccination or medical exemption for entry to venues associated with a high risk of super-spreading would allow for more relaxed rules that are more akin to a pre-covid normal.

Aren't we being told vaccination and/or existing immunity from previously being infected does not prevent spreading? Can't square that current understanding with your suggestion.

https://www.gov.uk/government/news/past-covid-19-infection-provides-some-immunity-but-people-may-still-carry-and-transmit-virus

In reply to TomD89:

> Aren't we being told vaccination and/or existing immunity from previously being infected does not prevent spreading? Can't square that current understanding with your suggestion.

This is solidly in the "we don't know yet" pile. And quite near the top of the "we really need to know" pile given its policy importance. RAMP forums are awash with preprints on this. No consensus.

 wintertree 14 Jan 2021
In reply to TomD89:

> Not impossible but improbable IMO. 

I think I'd see an elephant fly before the UK pushes for elimination.  One of those rare times I'm going to have to find a way to move on despite disagreeing utterly with the government on something I feel really matters, and it's notable that employers big and small are taking their queues from government over working practices this time round - much more (ab)use of key worker status, much less willingness to tolerate constraints on staff time from childminding if you don't want to be a link in the chains of infection taking the NHS to the brink.

> The other issues caused by lockdowns will be more stark by then as a counter-incentive.

The argument is that elimination may be the best way to avoid future lockdowns due to new immune-busting variants.  But that's the long game, and nobody in charge has been thinking that way since the start, and the media have barely started thinking about it a year on.

> Aren't we being told vaccination and/or existing immunity from previously being infected does not prevent spreading? Can't square that current understanding with your suggestion.

It's in the "absence of evidence" category at the moment for obvious reasons - your link is about naturally acquired immunity which is going to have a broader spectrum of responses I think than vaccine induced immunity.  From what's know about most diseases, vaccination and immunity is going to help and therefore requirement of immunity for entrance to high risk venues is going to help.  Better to put it in from the start.  If the evidence ends up emerging that it doesn't help, well, we're going to have bigger problems I suspect.

But I don't think we'll even see this level of protection.  Running with this virus rife has become normalised for an awful lot of people, and whilst I have my views on this (playing with fire, and it's potential to become a lot worse) it looks like I'm just going to have to hope this doesn't come true.

 wintertree 14 Jan 2021
In reply to Longsufferingropeholder:

> Even countries that have 'eradicated' it won't be able to sustain the economic harm of what it takes to keep it that way, especially when the rest of the world is open for business.

The dilemma NZ is going to find itself in is very interesting.  I envy them being in the position of facing that dilemma however; it will be very interesting to see how they solve it.  

In reply to Longsufferingropeholder:

> Elimination isn't a viable option while there will always be a reservoir somewhere in the world. It would be a prohibitively expensive goal to pursue even if possible.

> The whole world will end up going for a strategy of vaccinating it down to a tolerable level, mostly of mild cases. Even countries that have 'eradicated' it won't be able to sustain the economic harm of what it takes to keep it that way, especially when the rest of the world is open for business.

This is where the 'expanding bubble' of elimination comes in.

Step 1: eliminate within a jurisdiction with strong border control.

Step2: Repeat somewhere else within travel distance

Step3: Join those two safe jurisdictions with a travel bridge

Step4: eliminate in additional jurisdictions and add them to the safe area until global elimination is achieved.

It should be entirely feasible to Eliminate within the UK, then Europe, then further afield.  The reality is that most people travel within their continent, thus travel restrictions on destinations further afield aren't onerous to most people.

2
In reply to wintertree:

> The dilemma NZ is going to find itself in is very interesting.  I envy them being in the position of facing that dilemma however; it will be very interesting to see how they solve it.  

I can't see them doing anything but what you'd expect; they'll vaccinate everyone then blow the dam.
Amongst other consequences, it will produce a wealth of invaluable data that won't be attainable any other way.

In reply to Toerag:

You just can't do it though. There's always one or two cases that will slip the net and ruin it. Especially when you factor in the ridiculous exemptions for flight crew and the like which make the whole endeavour a farce.

Post edited at 12:09
In reply to Longsufferingropeholder:

> It's not a popular point to make but it's arguable that as things stand we're more likely to export one than import one at the moment. That will of course change as we get on top of things.

Totally, unless someone holidays in the Brazilian jungle and brings that variant back here...

 TomD89 14 Jan 2021
In reply to wintertree:

> I think I'd see an elephant fly before the UK pushes for elimination.  One of those rare times I'm going to have to find a way to move on despite disagreeing utterly with the government on something I feel really matters, and it's notable that employers big and small are taking their queues from government over working practices this time round - much more (ab)use of key worker status, much less willingness to tolerate constraints on staff time from childminding if you don't want to be a link in the chains of infection taking the NHS to the brink.

With the best will in the world people do need to earn a living to support themselves, their family and continue to keep the country functioning. People trusted early on that shutting up shop and taking a temporary hit was for the greater good, but that simply isn't the case on a longer timeline when you have no money coming in. We're lucky we even have the option to support people during lockdown for as long as we have, places like Indonesia simply couldn't do it.

> The argument is that elimination may be the best way to avoid future lockdowns due to new immune-busting variants.  But that's the long game, and nobody in charge has been thinking that way since the start, and the media have barely started thinking about it a year on.

As you say, maybe. Quite how it could be quantified and demonstrated in a convincing way to the public at large, taking all other important economic and social factors into consideration, I don't know. Trying to implement a UK, leading to world, wide viral elimination strategy requiring every man jack to comply seems less attainable the longer I think about it. Even with police state level autocracy you'd have a job, ignoring all the arguably worse consequences of such an approach.

> It's in the "absence of evidence" category at the moment for obvious reasons - your link is about naturally acquired immunity which is going to have a broader spectrum of responses I think than vaccine induced immunity.  

The .gov article mentioned..

"experts cautioned those with immunity may still be able carry the virus in their nose and throat and therefore have a risk of transmitting to others."

Not an expert, but I couldn't logically see why a person with vaccine acquired immunity couldn't carry the virus in their nose/throat in the same way someone with naturally acquired immunity could, admittedly an assumption so stand to be corrected. 

 Si dH 14 Jan 2021
In reply to TomD89:

> Not an expert, but I couldn't logically see why a person with vaccine acquired immunity couldn't carry the virus in their nose/throat in the same way someone with naturally acquired immunity could, admittedly an assumption so stand to be corrected. 

Everyone up to and including ministers have been quite open that we don't yet know how much the vaccines will reduce infection and transmission, only how much they should reduce disease.  The study I linked above suggested that in the case of natural infection with the virus itself, you were 80% less likely to be reinfected (and therefore transmitting it to others) than someone who had not had the infection; 90% less likely to get symptoms. Very rough numbers as the reporting was inconsistent.

In reply to Longsufferingropeholder:

> Elimination isn't a viable option while there will always be a reservoir somewhere in the world. It would be a prohibitively expensive goal to pursue even if possible

> The whole world will end up going for a strategy of vaccinating it down to a tolerable level, mostly of mild cases. Even countries that have 'eradicated' it won't be able to sustain the economic harm of what it takes to keep it that way, especially when the rest of the world is open for business..

I agree with this. I think the only feasible way forward is mass vaccination, followed by regular updates of the vaccine as and when mutations occur that require it, similar to management of annual flu but required on a bigger scale.

Once that sort of system is established I don't see any major problem with it. It seems a lot more desirable to me than a series of perpetual restrictions to keep the virus close to elimination, and I'd consider myself to be more informed about the need for, and more in favour of restrictions than, most of the population.

The difficult question is how you manage cases from the point at which deaths drop here until the point at which cases drop a lot anyway due to mass vaccination of all adults. And then how low do you try to keep through next winter if there is a small resurgence then despite the vaccination program. The logic here being, that you need cases to be very low and your testing/sequencing capacity sufficiently high that if a mutation occurs, you can discover it, modify a vaccine and distribute it before the mutated variant rips through too many people. If cases are still moderate like we had this last summer, we won't be able to do that quickly enough.

Edit to add, it also strikes me that 'Global Britain' should be sharing our sequencing knowledge/skills with other countries and helping fund development of sequencing capabilities in poorer or middle income countries around the world. Since, it's something the whole world needs and which we are quite good at - and ultimately we can't succeed in isolation on this.

Post edited at 13:49
In reply to Longsufferingropeholder:

> This is solidly in the "we don't know yet" pile. And quite near the top of the "we really need to know" pile given its policy importance. 

Common sense tells me that if I have natural or vaccinated immunity, then I won't make a very good host for the virus to replicate, so the amount of virus that I have in my nose and throat is likely to be less than someone without immunity (*) - so I should be less of a transmission "threat" than someone without immunity.

(*) - it's possible I may have as much virus as a non-immunised person at a particular time, but over time (i.e. the integral), the total amount I have should be less.

But conversely, if I have natural or vaccinated immunity am I more likely to get Covid asymptomatically and spread a smaller amount of virus around but for a longer time (being totally unaware) than someone without immunity who is more likely to develop symptoms and then remove themselves (by isolation) from being a spreader of the virus.

 wintertree 14 Jan 2021
In reply to Si dH:

> I agree with this. I think the only feasible way forward is mass vaccination, followed by regular updates of the vaccine as and when mutations occur that require it, similar to management of annual flu but required on a bigger scale.

The difference is we have a pretty good idea of the sorts of annual changes that flu viruses undergo, and we really don't know how this virus is going to change yet.  

The fatality rates for SARS-CoV-2 are an order of magnitude higher than for common flus - at all ages - and yet they are an order of magnitude below what the virus' closest relations achieve.  Further, there is a seasonal pattern to flu changes and we can look to nations ahead of us in the cycle to give time to pre-prepare a vaccine, where as the emergence of a new Covid strain is a largely random event, meaning that all vaccine creation, testing and deployment happens after it's too late if the variant turns out to have accessed SARS-CoV-1 or MERS-CoV levels of lethality.  Most of the population of the world still hasn't caught this yet the number of worrying variants is on the up.  I don't see us getting to a point where most of the world population is ever getting vaccinated against this.  I'll be surprised if we get to 80% of the population vaccinated in the UK, which leaves 100% - (80% x 90% vaccine efficacy) = 28% of the population susceptible.  Across Europa and North America alone, that ~ 330 million people for it to circulate in.  To date, the virus has swept through perhaps 250 million people, and several variants have arisen that raise red flags on immune evasion; so far nothing has been confirmed as emerging with more lethality (I believe) but it seems likely that relying on vaccination to control hospital overload is going to leave a very large breeding ground for more variants.

I am not at all happy about this as a model for the future but realistically it's the best possible option that I think is going to be committed to.  Luck may protected fools, small children and ships named Enterprise but it's not doing much for pandemic management.  The vision you paint of the future is going to need exceptionally well run screening, sequencing, rapid response hard lockdowns around new variants and ground pounding, sequencing assisted contact tracing.  The sort of thing that NZ have, and we don't.  We did have a lot more capability in some ways for this before the merger of many local public health observatories in to PHE; this was reckoned by some from the former observatories to be a very dogma led decision and is one of the first places I would look to other parts of the world and our past to identify best practice for managing the future you paint.  To manage this without ending up back in damaging nationwide lockdowns and healthcare overload, we're going to need either a run of luck for the next 5 years or so, or much much better, more responsive systemic and localisable response to outbreaks and variants. 

What worries me is just not seeing the political or organisational capability to put this kind of stuff together in the UK.  I think the formation of PHE led to a bit of a brain drain as well.

> The logic here being, that you need cases to be very low and your testing/sequencing capacity sufficiently high that if a mutation occurs, you can discover it, modify a vaccine and distribute it before the mutated variant rips through too many people. If cases are still moderate like we had this last summer, we won't be able to do that quickly enough.

Indeed.

I think we have to pre-emptively contain future new mutations given that we know it's not very transmissible and so has the potential to become more so, and given that we know it's nowhere near the lethality potential of its family of virus. 

This means cases need to be low enough for sequencing to be completed, and contact tracing to be started within ~18 hours of a sample being taken - which I think needs a radical improvement on where we were last summer, both in terms of case rates and the organisational responsively of the testing systems.

I said "5 years or so" above - I think in time sufficient will be learnt about this type of virus that more powerful, less intensive pharmacological and therapeutic care is going to take the sting out of its tale, vaccine or not.  Near where I work there used to be a Cat 3 lab built specifically to study SARS; it was shut down as of little interest about 7 years ago.  Perhaps it'll be re-opened along with many more soon.

> Edit to add, it also strikes me that 'Global Britain' should be sharing our sequencing knowledge/skills with other countries and helping fund development of sequencing capabilities in poorer or middle income countries around the world. Since, it's something the whole world needs and which we are quite good at - and ultimately we can't succeed in isolation on this.

This thread isn't alone in worrying about the lack of sequencing capability in many countries; I know this troubles the public health people in one distant land.  I suspect business will be booming for Oxford Nanopore for the foreseeable.  We can certainly fund and train sequencing centres elsewhere. 

In practice I think the way this virus is varying, and the practicalities of vaccinating everyone, mean that this virus is going to drive a wedge between less and more developed countries, just as it is driving a wedge into British society along fiscal lines.

In reply to Longsufferingropeholder:

> You just can't do it though. There's always one or two cases that will slip the net and ruin it. Especially when you factor in the ridiculous exemptions for flight crew and the like which make the whole endeavour a farce.

That's why you have strong borders - test before travel, test on arrival, compulsory isolation and test to exit it.  Don't forget, once other jurisdictions eliminate, those sources of cases disappear.

 TomD89 14 Jan 2021
In reply to Si dH:

> Everyone up to and including ministers have been quite open that we don't yet know how much the vaccines will reduce infection and transmission.

Exactly, hence why suggesting actions like allowing entry to venues for vaccinated and exempt only is premature. We simply don't know enough yet.

> I agree with this. I think the only feasible way forward is mass vaccination, followed by regular updates of the vaccine as and when mutations occur that require it, similar to management of annual flu but required on a bigger scale.

Repeated vaccination of the entire planet (x2 doses per person) or even just the UK is much bigger undertaking than the annual flu jab. I suspect we'll need a more pragmatic approach going forward once we can ascertain hospitalisation/death rates in the not vulnerable.

 Luke90 14 Jan 2021
In reply to Longsufferingropeholder:

> We're not the only country that's lost faith in their government. One thing I've taken from reading a few other national sources is that they're either all just as bad, or everyone around the world enjoys shitting on their government as much as we do. Or is it both....?

> We seem to look to Germany fairly often as an example of how to deal with this, but if you read DW you can quickly realise it's all subjective.

Well, not entirely subjective. Our Covid deaths per capita are well over double Germany's, and presumably racing even further ahead over the coming weeks.

Edit: Fixed quote formatting

Post edited at 15:28
 wintertree 14 Jan 2021
In reply to TomD89:

> Exactly, hence why suggesting actions like allowing entry to venues for vaccinated and exempt only is premature. We simply don't know enough yet.

"We don't have certainty yet" has been used as an excuse to delay measures that would have been effective time after time.  We don't know, but there are strong evidence led hunches, and we can take precautionary measures first and relax them - or not - as the evidence develops.

Plunging in head first and waiting to see what happens has turned out to be a disastrous strategy time after time in this pandemic.

If it turns out that vaccination is sufficient to allow high-risk-of-transmission venues to open sooner than otherwise, I think we'll be seeing a business led call for vaccine passports.   It's not just about lowering transmission risk but lowering the risk to the patrons from the virus, on which we have very hard and certain evidence.

I could well be wrong, and it doesn't affect me either way as I'm going to be giving such venues a wide berth for a long time yet.

Post edited at 15:07
 wintertree 14 Jan 2021
In reply to Toerag:

> That's why you have strong borders - test before travel, test on arrival, compulsory isolation and test to exit it.  Don't forget, once other jurisdictions eliminate, those sources of cases disappear.

Looks like you and NZ may have a travel corridor - I assume Guernsey doesn't have a long enough runway for 747s though!

What's the felling on the ground there about opening up in the long run?  Not a dilemma faced by NZ alone.

 neilh 14 Jan 2021
In reply to Luke90:

I sometimes wish in current envionment I could read a South Korean or Taiwanese national paper.........Its all well and good comparing us with Germany,,but they are still a long way behind on vaccination at the moment..and when Merkel goes it is going to be interesting

In reply to Luke90:

Not disagreeing with that, but I think you've only read half the post there....
They're as pissed off with their handling of it as we are with ours. That's the subjective part.

 jkarran 14 Jan 2021
In reply to wintertree:

> The dilemma NZ is going to find itself in is very interesting.  I envy them being in the position of facing that dilemma however; it will be very interesting to see how they solve it.  

I'm guessing once they've vaccinated down to a low level of predicted vulnerability they'll relax border restrictions from mandatory quarantine to passporting and screening while we all figure out if that's enough to get on with life again. Tough call for the government as it inevitably leads to newsworthy deaths but they can't stay closed forever.

Worrying about a nasty new mutant is reasonable but there are millions of pathogens out there mutating all the time, another one will come for us sooner or later. If it's covid derived at least we'll be starting with one hell of a toolkit and some residual vigilance.

jk

 AJM 14 Jan 2021
In reply to wintertree:

> The difference is we have a pretty good idea of the sorts of annual changes that flu viruses undergo, and we really don't know how this virus is going to change yet.  

> The fatality rates for SARS-CoV-2 are an order of magnitude higher than for common flus - at all ages - and yet they are an order of magnitude below what the virus' closest relations achieve. 

Musing aloud to some extents here, but I wonder the extent to which common flu and Spanish Flu (the 1919 variety) are related, versus the relationship between Covid-19 and MERS say.

Agree that we know a lot more about run of the mill flu, but the obvious thought occurs that we have the same head in the sand approach with regards to "running hot" on flu (we give it a pretty large potential population to infect every year and only vaccinate those most at risk from the current strain) - with presumably the same conceptual risk (no idea about likelihood) that by giving it a lot of opportunities we increase the risk of a real nasty cropping up.

I tend to agree with those who say they can't see us trying to drain the pool of potential mutation opportunities via a push for elimination - the risk is too conceptual for most people to buy into (even if we have the recent examples to look to which should make it real - but I can easily imagine the "what, so we've got to stay in lockdown to avoid the risk of having to go into lockdown" argument, as obviously flawed as it is, gaining traction especially amongst those who consider the risk to themselves low and the inconveniences high) and we have the historic precedent of how we deal with what might eventually be a similar variable outcome virus (i.e. most of the time it could end up being squashed to an endemic nuisance by vaccination etc but the risk of a monster mutant being ever lurking). Which isn't to say that's the sensible idea, but it feels far and away the most likely.

In reply to AJM:

I guess we could all agree not to order our bats medium-rare from now on

 mik82 14 Jan 2021
In reply to wintertree:

Anyone seen any more information on the Brazilian variant that's led to the travel bans from Portugal and South America? Only thing I can find is media reports about it being out of control in Manaus, where 76% were antibody positive after the first wave, so presumably some antibody resistance and possibly a problem for current vaccines. I suppose given a lot of trials were done in Brazil it should be possible to get data about vaccine efficacy easily.

Post edited at 17:08
 wintertree 14 Jan 2021
In reply to mik82:

I've not fully digested the pre-print yet but, it's the third significant variant (in terms of spread or suspected spread) with mutations to the receptor binding domain which is enough reason IMO to shut borders.  The problem is that shutting borders with Brazil only delays the inevitable without negative screening then travel then MIQ on our borders for travellers from everywhere.

https://virological.org/t/phylogenetic-relationship-of-sars-cov-2-sequences-from-amazonas-with-emerging-brazilian-variants-harboring-mutations-e484k-and-n501y-in-the-spike-protein/585

Perhaps 2020 was just the bootloader for 2021.

Paging Dave Garnet, Kathrynrc and cb264 - do you think there has been some previous mutation which has now allowed the spike to better explore substitutions and deletions?  Is it possible now that the virus is so widespread that a higher replication error rate could be tolerated on individual virions, exploiting the widespread prevalence to mutate faster than simple probability would suggest, allowing for example degradation of the ribosome's error checking to propagate?  Or is this a simple case of the probability of a beneficial mutation occurring, and spreading has crossed a threshold from "unlikely" to "basically certain" with the scale of current infections?

Post edited at 17:36
In reply to wintertree:

Addendum to that: have all the variants of concern identified so far carried the N501Y change? If so, does that indicate that it's a clear upgrade that will evolve convergently all over the place?

 mik82 14 Jan 2021
In reply to wintertree:

Thanks. It really wouldn't surprise me if this variant has been seeded across Spain and Portugal, so perfect set-up for a repeat of the tourism related imports of last year.

 wintertree 14 Jan 2021
In reply to Longsufferingropeholder:

I think N501Y has cropped up in all 3.  I don’t know if it’s convergent evolution or a common ancestor that’s not been found.  Well outside my competence zone at this point.

In reply to wintertree:

Everything I've seen or read states very confidently there's not a common ancestor

 SDM 14 Jan 2021
In reply to AJM:

> Agree that we know a lot more about run of the mill flu, but the obvious thought occurs that we have the same head in the sand approach with regards to "running hot" on flu (we give it a pretty large potential population to infect every year and only vaccinate those most at risk from the current strain) - with presumably the same conceptual risk (no idea about likelihood) that by giving it a lot of opportunities we increase the risk of a real nasty cropping up.

There are currently trials ongoing for universal flu vaccines. They might have the potential to change this?

I don't know any more about them other than that the trials exist so I don't know anything about how likely they are to succeed or how far they are from potentially hitting the market.

 bruxist 14 Jan 2021
In reply to neilh:

Quite. Neither can I; yet somehow one (probably small sub-section of one) demographic can. Perhaps they imagine their grandparents off the leash, roaming the bars and clubs of Shoreditch or wherever, whilst they're forced to sit at home as if they're grounded teens again. It strikes me as an imaginary tension that will never arise.

 bruxist 14 Jan 2021
In reply to jkarran:

Agree with all you say here, but wonder if it's a response to what you've quoted - it doesn't seem related. I was trying to make sense of factors that might influence popular opinion on restrictions. On the first point, you seem to be expressing the same view, i.e. that those who do not have a choice, whether that choice be to go to work or to take a holiday, will be more exposed; I agree entirely, but am interested in what might shape or change the opinions of those who are intrinsically more exposed, because opinions are just as infectious and connected as viral hosts.

On the second point I'm afraid I can't see the connexion at all - sorry! I'm not sure that younger working-age people consume 'the media' at all any more, and I'm certain that fear of death, whether one's own or another's, hardly registers until at the earliest the forties.

 jkarran 14 Jan 2021
In reply to bruxist:

I started out responding to the bit I quoted but then yeah, I digressed. 

I agree covid growth regulation by media feedback is unlikely to be very good. I don't really think the government will get to step away from this as much as they might like to once the third wave deaths slow.

Jk

In reply to mick taylor:

> I think once the 15 million vulnerables are vaccinated then it’s chocks away. Vast % of less/not vulnerables have always felt lockdowns and rules were there to protect the vulnerables anyway. The government will be immense pressure to reopen everything and they will , in stages  

This will be a big issue. I suspect they will reopen too early and too fast, before the baseline of cases drops low enough for test & trace to be able to keep up. That would keep hospitalisations and deaths higher than they could otherwise be.

 Offwidth 14 Jan 2021
In reply to jkarran:

On the subject of the media ... the Telegraph and Toby Young have been rebuked for covid misinformation.

https://www.theguardian.com/media/2021/jan/14/daily-telegraph-rebuked-over-toby-youngs-herd-immunity-covid-column

In reply to Longsufferingropeholder:

I'm not convinced the vulnerable groups will start going crazy with their (perceived) freedom. Fistly, many are pretty old and/or have health conditions - they aren't the people going to illegal raves! Secondly, I suspect many people will remain cautious as long as they know that cases are high. The vaccines aren't 100% effective. 

 wintertree 14 Jan 2021
In reply to Offwidth:

> On the subject of the media ... the Telegraph and Toby Young have been rebuked for covid misinformation.

Well that’s a shitter for “cp123” and his taunting threads (deleted for being misinformation).  

1
In reply to wintertree:

> Looks like you and NZ may have a travel corridor - I assume Guernsey doesn't have a long enough runway for 747s though!

We had one with the IoM in the summer which was quite useful for the tourism sectors in both islands, then we had an outbreak which caused it to be closed. Both islands wanted to reinstate it and I think that time was close then the IoM had its current outbreak (which they've locked down for to nip in the bud). We'd have liked one with Jersey but they went down the road of opening up travel whilst relying on ToA which predictably ended in tears once the prevalence in the UK rose in September - they were in an equivalent of tier 3/4 over Christmas.

> What's the felling on the ground there about opening up in the long run?  Not a dilemma faced by NZ alone.


Rewind to the summer and there was a fair bit of 'we can't stay locked away forever', especially once Jersey opened up in July. Some of the finance industry shifted clients to their Jersey operations as Jersey was sending out 'open for business' signals which caused a little bit of consternation, although everyone I spoke to in the industry said they'd not suffered due to restrictions at all - it was only the touting for new business that was suffering, so a business with low client churn wasn't worried. One wealthy businessman started to sue our government for being 'imprisoned' then withdrew when they tweaked something. Now it's the opposite - people have realised it's much better to be unrestricted within the community now they've seen how everywhere else is suffering second/third waves.  Jersey has equal travel restrictions to all intents and purposes, and the UK is being travel-sanctioned as Sweden was.  We just want the rest of the world to wake up and smell the coffee that trying to 'live with the virus' hammers the economy and society way more than an elimination strategy. People are now starting to want to travel to see family overseas now and are following the state of the UK and effectiveness of vaccination on it's situation (all our flights go to the UK).

By the way, we were following a 'flatten the curve' strategy in the spring but had to lock down when we couldn't get test results back fast enough to t&t properly - we couldn't risk 'running blind' as we couldn't just ship patients down the road like London could.  Our lockdown was harder than the UK's and consequently we ended up in an elimination situation after about 3 weeks which we've pretty much managed to maintain - we've imported cases and contained them with 14 day self-isolation apart from one which led to a ~10 case outbreak which t&t dealt with quickly. We have no restrictions within the Bailiwick whatsoever - it's like the virus has never happened apart from the hand sanitisation stations and temporary queuing areas outside the supermarkets which aren't in use. Tourism has suffered although staycations have been popular and really helped, especially in the outer islands.  Now our vaccination programme is kicking in, the tourist industry is asking government for predictions on opening borders, but of course that depends on the situation elsewhere (UK mainland).  We got to the stage of 10 day s.i. +exit test in the autumn for low prevalence countries, but as that came in the UK & France's prevalence rocketed and nullified it.  As of last Friday we've gone to 'essential travel only' with test on arrival and exit test at 13 days (or 21 days s.i. without test).  People doing s.i. with a traveller also have to do the day 13 test.  We're only PCR testing and are picking up about 1 case  every other day - about half on arrival, and half when they show symptoms whilst in SI.  We did pick up some people with day 13 tests, so that was made 'compulsory'.  We should have vaccinated all the >65s and vulnerable by the end of March, so I think there will potentially be some relaxation of the travel restrictions then, but it all depends on some unknowns - UK prevalence levels, risks of new variants, and dangers of 'long covid'.  I suspect the Kent, SA and Brazilian variants will be focussing every governments' minds on the significant risks of the current 'live with the virus' strategies - no-one wants their people to be travel-sanctioned or have a worse variant develop in or enter their country.  It wouldn't surprise me if the EU makes moves towards an elimination or strong suppression strategy soon.  Vaccination progress and effect will be key, but the risk of mutation causing an undetectable variant will keep borders relatively tight for some time, especially if a jurisdiction starts having success at supressing prevalence levels and relaxing - they'll be hell-bent on keeping it out once they've got rid of it.

 Offwidth 15 Jan 2021
In reply to Toerag:

Thanks for the comprehensive report.

Back on data, the Guardian on the rise map has gone mostly blue (decreasing) across much the UK for the Jan 10th update on weekly changes (the Liverpool area still looks bad). On the surface this looks like good news from nearly a week ago but it looks odd to me as Jan 10th cases were near identical to Jan 3rd.  I wonder if this is about late case assignment Jan 3rd maybe a bit high as they were still catching up on xmas/new year data. Hospitalisations should be dropping overall very soon if this is real (yet hospitalisation data doesn't indicate any slowing).

https://www.theguardian.com/world/2021/jan/13/coronavirus-uk-covid-cases-and-deaths-today

Post edited at 08:31
 wintertree 15 Jan 2021
In reply to Offwidth:

> I think you are looking slightly too much for that hope

Looking in the right places though it seems.   I'm looking forwards to putting this evenings update together for the first time in a month.  Looking quickly at last night's much delayed update, cases are definitely falling at a national level and many regions - as could be inferred from thinking of the traces on plot 18 as having "momentum" and "velocity" headed in the right direction even before lockdown started to bite.

But we have been here before back in November.  

It'll be interesting to see if my weekend case reassignment works with falling cases.  It should...

I'm really heartened to see how quickly cases are falling; with the new variant having largely evaded Tier 4 controls in its early UTLAs I was quite concerned.  As with the November lockdown, this seems really quite potent at the starts.  I think the tragic numbers locked in to healthcare pathways over the coming weeks will keep the pressure on through government and the media.  

 wintertree 15 Jan 2021
In reply to Toerag:

Interesting take; thanks.  A lot to think about.

It's really good that we have several places that have achieved and held elimination despite occasional outbreaks - at least the option is on the table which anchors discussions a bit.

Post edited at 09:09
In reply to wintertree:

One man's occasional is another man's repeated, persistent, inevitable and never-ending. It's not a strategy until it's the world's strategy.
There's a growing movement, but it needs to be universal to stand a chance and 100% support for anything is beyond optimistic, especially when you look at what it takes to work.

https://www.dw.com/en/zerocovid-campaigners-urge-europe-wide-shutdown/a-56230732

 neilh 15 Jan 2021
In reply to Toerag:

I know when very well placed business man who seems to fly in and out of Jersey regulalry. How on earth does he get away with it?

He goes for the day and has meetings at the airport. Suggesting that there are great big holes in the strategy for those with the right connections.

Post edited at 09:34
 jkarran 15 Jan 2021
In reply to Offwidth:

> On the subject of the media ... the Telegraph and Toby Young have been rebuked for covid misinformation.

I'm shocked. Unlike the toadmeister to be wrong or the Telegraph to provide entitled little pricks an uncritical megaphone.

jk


This topic has been archived, and won't accept reply postings.