Friday Night Covid Plotting #51

New Topic
This topic has been archived, and won't accept reply postings.
 wintertree 06 Nov 2021

Four Nations Plots

England, Wales and Northern Ireland are all showing decay incases; Scotland - not so much.

Squinting at the various plot 6s and plot 9x, it seems that England, Wales and Northern Ireland have generally followed a more similar curve to each other with Scotland doing its own thing.  

Link to previous thread - https://www.ukhillwalking.com/forums/off_belay/friday_night_covid_plotting_50-...


2
OP wintertree 06 Nov 2021

England - I

The decay of cases stalled for a bit last week, but then resumed.  It looks like typical weather associated variation in rate constants to me.  Big news if the “weather wobbles” are no longer taking us to growth at their peaks, I dig in to that a bit more with demographics in the next post. 

It looks like the decay is mostly real and not down to half term sampling artefacts.

  • Total test numbers are up for the most recent days after the half term week, yet cases are still falling.
  • Hospital admissions and deaths both look to have turned to decay over the last half-week, and plot 9e agrees with their rate constants (in the provisional zone) heading for decay.

OP wintertree 06 Nov 2021

England - II

The last few wobbles in the week-on-week method rate constant plot seem to be descending towards an average of decay.   The very last data point looks a bit suspect to me, could be subject to reporting lag; we’ll see what the next few days of data shows.  Be nice if it turns out to be real.

Plot D1.c has been changed to use a week-on-week method on raw data, instead of the usual de-weekending, exponential fitting and polynomial filtering method this is; a much more “pure” method but more noisy.  I like it - it shows fine structure very clearly.

These plots show how much of the recent decay in cases was in school ages, although there is also decay emerging in the oldest adults; likely the effects of the booster program?  Often the oldest ages are some of the last to turn to decay, rather than the first; presumably because people in this range often experience infections downstream of the people upon whom they are more dependant,.

I’ve added an annotated version of plot D1.c that shows a correspondence between recent periods of stand-out growth and more recently stand-out decay in school aged children and their parents’ age range.  It seems almost certain that this is largely household transmission from school aged children to family members.  If immunity in school aged children has reached the point transmission can’t be sustained within these aged bands, this correspondence is a clear sign that one cause driving infection in adults aged 30-55 is going away. If we also have boosters bringing decay descending the age range, as they continue, decay will hopefully continue sweeping down to the parent aged block.  That would then only leave adults aged 18-40 or so to really drive spread, and there’s been no sign of university outbreaks this term. 

The contrast between the situation in England and that in much of Europe is large, and is growing by the day.   Could this be it?  Are we really nearly there now?  Answers on a delay mail post card to two weeks from now.  


OP wintertree 06 Nov 2021

Scotland - I

All measures are doing a pretty good impression of a steady state in Scotland.  

Post edited at 20:18

OP wintertree 06 Nov 2021

Scotland - II

Plot D1.c shows cases turning to decay in the oldest ages, as with England.  Perhaps this is the booster roll out at work.

The standout difference to England is that children and adults under 24 are showing a clear turn to growth; best guess is that this is seasonality manifesting first in the demographics with the least immunity; whilst Scotland had early rapid growth in school aged children at the start of their autumn term, it did not develop to the off-the-charts scale of the English outbreak; as we’re seeing with Europe, delta is riding in everywhere ahead of winter and won’t easily be kept at bay; the risk of having a lot of ability to spread delta in children is that it will drive cases and worse in adults.  There’s a clear chance I’m over-interpreting the differences here, but it all falls in to a familiar pattern….

I don’t normally look at the population normalised rates as I’m more interested in trying to interpret the changes going on than where we have been, but as the international plots show in the next post, there are big differences in the rate the virus is spreading in different nations, and it seems likely that infection granted immunity is as much a part of this as differences in vaccine uptake - including demographic differences.

Plot Rates.a shows case rates and death rates (/million /day) for the four home nations.  During most of the lat few months, Scotland has had a worse case rate and much worse death rate than England despite what I perceive as a generally more cautious approach to dropping control measures.

Plot Rates.b steps back in time and shows the same data over the last year, so including the Kent/alpha wave from last winter.   It’s clear from this that Scotland had much lower spread of the virus back during that wave.

This hints at - but does not prove - that infection granted immunity is helping to moderate the current spread of the virus.  This is a trend that comes up time and again within demographic data and international comparisons, and is a large part of why it's not paradoxical that some places with more robust control measures are now having more Covid.

The big change between waves  is that the death rates are much lower in all 4 nations now than a year ago, despite comparable or higher case rates.  Nothing mysterious there - delta looks a bit worse than alpha, but vaccines make everything an awful lot better.

In terms of area under the deaths curve in each of the waves - that is the total number of deaths in tat wave - if we take as a given the interpretation that infection last winter can be held off to infection this winter (or alternatively infection that happens last winter induces immunity reducing infection this winter), it’s clear that whilst having had much less infection last winter makes it harder to control the spread of the virus now, the net cost in terms of deaths is vastly lower.  When modelling the shifting of infections forwards or backwards along the time axis, it's important to recognise that the fatality rate is not time invariant, both with the vaccines and potentially now with a raft of therapeutics coming online.


1
OP wintertree 06 Nov 2021

International Plots

These plots were introduced last week, building on some plots and discussions from the previous few weeks.

The phase space plot compares where different nations are in terms of cases (x-axis, per million people per day) and their normalised rate of change (y-axis, exponential rate constant for cases).

  • The rate of change is measured over a 2-week period to keep noise down, and so lags the most immediate change in cases a bit.
  • The plot is a bit cluttered this week, but we see that a clutch of EU nations have higher case rates (cases/million/day) than the UK, and that they’re on rising trajectories vs the UK decaying.  From skimming the news, more control measures are kicking in in various parts of the EU, but case rates have been rising rapidly as delta meets winter.  
  • The way the UK continues to see cases stagnate or fall (depending on age) vs rapid exponential growth across many other nations hints strongly at the immunological differences.

The plot is a bit cluttered this week - a couple of nations zig-zag more due to problems with the data I think (see below).  

Cautionary notes:

  • Belgium has really glitchy data over the last week in the OWiD data set
  • Denmark looks to have a glitch on the last day in the OWiD data set falsely making it look like cases are turning to dacay.
  •  I’ve put in the plot I use to check the input data - this is not for communicating results but for internal self-checks, so all axis tick marks are removed.  Raw data is red ‘x’ marks, and whilst for presenting results a line joining the dots would not be appropriate (I used to give student’s join-the-dots drawing pages if they gave me a plot like this to mark), it’s a great way of visualising the day-of-week structure in the data.  Black curves are a 7-day rolling average (implemented with a decreasing window size to include the leading edge.). This plot makes the glitchy nature of recent data for Belgium and Denmark clear.  I don’t normally bore people with the details but there’s normally a couple of plots like these behind anything I post, that I just glance at to keep an eye on the raw data going in to the output plots, as part of checks to make sure I’m happy that the analysis being used for the output plots is processing sane data.  It’s far too easy to make a bunch of analysis for something and get invested in what the results mean when it turns out to be suffering from a GIGO problem.

The plot titled "this plot is not a prediction".

Left-side plot

  • There’s a lot going on in this plot.  If I haven't explained it well, please ask.
  • This plot is not a prediction.  This is an attempt to condense a lot of information in to one plot to understand the present moment and the general direction things are currently heading in.
  • The left plot shows what I consider recent, reasonable bounds for the case fatality rate, measured using lags of between 7- and 28- days between cases and deaths statistics.  This represents the deaths rate as a fraction of detected cases, not true infections.

Right-side plot:

  • The current death rate averaged over the last week of data is drawn as a vertical bar, with a whisker showing the change from the previous week - these look like a sideways letter “T”, and are purple for growing death rates and grey for falling death rates.
  • The shaded grey bars show an estimate of the death rate that would be locked in right now if the current daily case rate is subject to the most recent estimate of the CFR  As the CFR is bounded to a range, so is this estimate, hence it being a bar of some width.
  • This is an extrapolation based on nothing changing - it is not a prediction. 
    • We know that the demographics of infections are always changing (sometimes for the better, sometimes for the worse),
    • We know hospitals can become overloaded or exit overload,
    • We know new therapeutics with immense potential are coming online imminently in the UK and are pending approval in the EU. 
    • All these and more are  why this grey bar is not a prediction.  Note however that it’s based on the current case rates, so is not affected by future transmission control measures.  Likewise, the unknown relationship between cases and true infections "drops out" of this analysis, assuming that said relationship is remaining time invariant.
  • The whiskers on the grey boxes show an extrapolation of where the middle of the grey bar would be in another week's time if the cases keep changing at their recently measured exponential rates.  In reality, this will change based on predictable factors (e.g. more Covid > more control measures,) and unpredictable events.  So, definitely not a prediction.  But it’s a quick way of understanding how things are changing.

Notable points on the plot

  • The Netherlands, Germany, Ireland, Belgium and Austria now have extrapolations for higher death rates than the UK "locked in" - and cases are currently still rising for all of them.
  • Hungary looks to be in real trouble.
  • Greece is looking very worrying. 

Norway appears to have shot up in CFR since last week; looking at the raw data, they had had almost no deaths when cases were very low; something similar was seen in New Zealand where the CFR was exceptionally low when cases were very low; I suspect this arrises from the distribution of cases not fully exploring the demographic range when cases are low but I haven’t looked for national data to verify this.  Both Norway and New Zealand are now showing CFRs in line with other nations.

[1] https://www.ukhillwalking.com/forums/off_belay/friday_night_covid_plotting_50-...

Post edited at 20:40

1
 Si dH 07 Nov 2021
In reply to wintertree:

> England - II

> The last few wobbles in the week-on-week method rate constant plot seem to be descending towards an average of decay.   The very last data point looks a bit suspect to me, could be subject to reporting lag; we’ll see what the next few days of data shows.  Be nice if it turns out to be real.

I'm feeling very positive about the provisional data. The best way to judge it is usually this chart from the dashboard because significant changes in testing lag are rare. As you can see, the new data for each day is lower than that for the previous day a day earlier, showing increasing decay. The decay in the provisional data up to at least this Thursday just gone therefore looks like it is going to be faster than the decay in the previous week that your graphs mostly measure. Let's hope it's sustained; I guess this will depend partly on whether more school age transmission happens after being back at school for a week. My son has a cold, again....4th time I think in the last 2 months.

I noticed that Zoe are also now showing a turn to decay in cases as of this last week - the first time they have done so for quite a while. Spector seems quite optimistic which is unusual these days!

Edit to add, unfortunately it looks like the provisional data in Scotland is heading the opposite way.

Post edited at 08:45

 Si dH 07 Nov 2021
In reply to wintertree:

> Right-side plot:

I get the big picture you are showing here, but for future reference / clarification, two things are slightly confusing:

1) not separating out your two x-axis labels - it would help a lot if one was at the top, as neither looks like a CFR scale?

2) I think you have mis typed grey instead of green in your description

Post edited at 09:13
OP wintertree 07 Nov 2021
In reply to Si dH:

> I'm feeling very positive about the provisional data.

Yes, if there's no more reporting lag than normal, looking at the provisional data you've noted, we're going to see even faster halving times emerge, perhaps ~14 days by tomorrow evening. I'm very curious to see what the demographic behind this looks like when it's revealed in a couple of days time.

> I noticed that Zoe are also now showing a turn to decay in cases as of this last week - the first time they have done so for quite a while. Spector seems quite optimistic which is unusual these days!

More digital chicken bones are aligning.  Excellent...

> Edit to add, unfortunately it looks like the provisional data in Scotland is heading the opposite way.

Oof.  I can't un-see that plot now I've looked.  The week-on-week rate constant does look like it's heading for growth again, but it's within the bounds of noise over the last couple of months.  Looking at a CFR measurement (below), whilst cases have been relatively level in Scotland of late, their fatality rate has been getting consistently worse and is stand-out bad for the four nations.  Presumably this is down to demographics but the use of coarse bins for the Scottish data frustrates a proper comparison.

> My son has a cold, again....4th time I think in the last 2 months.

It's crazy just how much snot they can produce.  It's mad what's going around the school circuit it seems vs other years; our area is currently hit by a nasty tummy bug as well as another cold.  I've started to see people hacking and coughing in public again recently, something noticeably absent for a long time...


OP wintertree 07 Nov 2021
In reply to Si dH:

> 1) not separating out your two x-axis labels - it would help a lot if one was at the top, as neither looks like a CFR scale?

Not quite sure what this comment means; the right plot has a single x-axis but I've put the numbers on in exponential annotation (bottom) and normal (top); the value shown is the product (cases/day x CFR) equivalent to an estimate of "locked in death rate".  I probably need a better axis label to make it less muddling.

> 2) I think you have mis typed grey instead of green in your description

Yes, yes I have.  Thanks.  

 Si dH 07 Nov 2021
In reply to wintertree:

> > 1) not separating out your two x-axis labels - it would help a lot if one was at the top, as neither looks like a CFR scale?

> Not quite sure what this comment means; the right plot has a single x-axis but I've put the numbers on in exponential annotation (bottom) and normal (top); the value shown is the product (cases/day x CFR) equivalent to an estimate of "locked in death rate".  I probably need a better axis label to make it less muddling.

Sorry, looking at it again I understand. I had internally misread 'death rate' in your first bullet point as CFR, rather than deaths per day. That confused me. So all of my own making.

OP wintertree 07 Nov 2021
In reply to Si dH:

> Sorry, looking at it again I understand. I had internally misread 'death rate' in your first bullet point as CFR, rather than deaths per day. That confused me. So all of my own making.

The Sunday morning effect, I'm never my sharpest at this time of the week....  It's Wintertree University here (80s cartoon classics).  I've re-worked the plot a bit during the Gummi Bears slot - a few more nations, horizontal grid lines to help read them off, expanded the x-axis range to include the full bracketing range for the antipodean nations, dropped the CFR side and added an orange marker for the all-time high death rate for the UK as seen in January 2021.  

Back last January we were seeing the consequences of rapid spread without vaccination.  Now, despite vaccines, two countries on this plot have worse actual death rates than us back then, two more have worse extrapolated rates locked in, Greece is about one week from hitting that level "locked in" by extrapolation, and Austria two weeks or so away.   It's pretty shocking how fast this is snowballing in many nations.  The stakes around the new antivirals and MAB therapeutics are looking pretty high.

Edit: Some relevant plots in this FT article - https://www.ft.com/content/c08951d3-4a88-4edd-9219-a2089879bd07

Post edited at 10:06

In reply to Si dH:

> I'm feeling very positive about the provisional data. 

I’m also feeling positive about the UK for the same reasons you are. Take a step back and the implications are terrifying.

China and the Far East, with very low levels of natural immunity have to take the same path as we have. A dodgy? Chinese vaccine, plus close to 0 natural immunity (if we believe the Chinese figures) means that the worlds factory is in for a very bumpy ride over the next 12/18 months.

Time to buy gold bullion and stick it under the bed methinks.

OP wintertree 07 Nov 2021
In reply to Si dH:

> I'm feeling very positive about the provisional data. 

Yes, the final data point from yesterday's thread barely moved with today's update and today's is showing more decay - at the best rate for two months; so looking very real and not a bit of extra reporting lag.

This period of decay corresponds to a period of higher temperatures in the weather passband; that runs out in a few days (very cold Guy Fawkes night); On the passband plot, you can see the temperature diving down ahead of cases for the latest cold spell.  Let's hope the baseline - that this weather modulation gets superimposed on - is low enough to keep in decay.  The dive down might not be so extreme when another couple of days of data land.

Scotland is still hovering around slow growth.

I think my interpretation of all this is that England lost the ability to host rising infection in late August, then schools returned - this is where there's a sampling anomaly creating a "false" rise in cases around the August bank holiday (last "matched" annotation on the cases plot); then things remained in decay at the top level as cases rose (from a low baseline) in school ages; eventually this took over driving both top level cases and infections in adults; now we've exhausted the ability for school ages to drive infections, and so infections in adults are back to the decay they'd started in late August.  Now we appear to have the effect of the 3rd dose kicking in in older adults; perhaps decay through to Christmas isn't such wishful thinking although the really grotty winter weather hasn't kicked in yet.

Cautious optimism is definitely in order.  Well, for England at any rate; lots of food for thought in the emerging differences perhaps.  I don't think there's anything particularly complex behind it, but it seems to be a contentious topic.  

In reply to VSisjustascramble:

> A dodgy? Chinese vaccine

If I understand right, Sinovac-CoronaVac is an inactivated virus vaccine without and adjuvant.  It seems possible that adjuvanting it with the same TLR9 agonist as Valneva could make big improvements to its efficacy.  This seems like a way to improve things for them without having to make material changes to their biological production plants, more an addition at a later stage.  No idea what Dynavax's production capability is like. 

Post edited at 17:24

In reply to wintertree:

> Back last January we were seeing the consequences of rapid spread without vaccination.  Now, despite vaccines, two countries on this plot have worse actual death rates than us back then,

But the countries near the top - particularly Latvia - have small populations relative to the UK.  Latvia is less than 2 million.   A comparison of short term death rate for a 2 million country against a 66 million one isn't fair.  You'd be able to find 2 million person regions in England with a higher death rate than England as a whole at the peak of the first wave. 

The data for the larger country is smoother - you can actually see this in England vs Scotland comparisons as well - so the fair way to compare would be over a long time period.

Post edited at 22:22
15
OP wintertree 07 Nov 2021
In reply to tom_in_edinburgh:

For anyone reading who isn't tom_in_edinburgh, I think their entire post here directly contradicts various recent contributions from them.  I give it no heed.  Feel free to skip over the detailed comments below and do something (anything) more constructive instead of reading the rest.

It has become astoundingly clear to me they have nothing to contribute to these threads but noise. 

Tom, two threads ago you said this:

> The second bias which makes England look better than it is, is the use of rate constant plots rather than absolute numbers scaled for population. 

https://www.ukhillwalking.com/forums/off_belay/friday_night_covid_plotting_49-...

Now, when looking at absolute numbers scaled for population (which you previously claimed to be the way to a non-biassed comparison), you say:

> But the countries near the top - particularly Latvia - have small populations relative to the UK.  Latvia is less than 2 million.   A comparison of short term death rate for a 2 million country against a 66 million one isn't fair. 

Make your mind up.  To me, it looks like you're trolling pure and simple, coming along with a contrarian view point to anything I say that isn't even consistent over the space of two weeks.

You have claimed to be an engineer multiple times.  If so, you should understand a fair amount about noise statistics, and if you did understand the most basic points of noise statistics, you wouldn't be using the difference between a population of 2 million and 66 million as a reason to discount a comparison in the situation where rates are high.

https://en.wikipedia.org/wiki/Shot_noise

> The data for the larger country is smoother - you can actually see this in England vs Scotland comparisons as well - so the fair way to compare would be over a long time period.

  • The case rate phase space plot compares things over a longer period of time for various reasons including those you give.  The plot you comment on in isolation is an aid to contextualise the case rate plot as CFRs differ wildly between nations (even between England and Scotland) and so case rates aren’t a fair comparison alone.
  • The difference between the extrapolated rates of deaths "locked in" is close to 10x.  At no point have separate geographic regions of England been that far apart.  The statistical noise from the population differences is nowhere near enough to create this kind of transient bias.  Google News is a thing - go and do some reading on it.  
  • Once again, you're flat out contradicting yourself.  Here you claim that the fair way to compare is over a long period of time, yet two weeks ago you cherry picked data points from two separate, single days to mis-represent reality by a factor of 5x - https://www.ukhillwalking.com/forums/off_belay/friday_night_covid_plotting_49-... -  Your deceit their was so transparent it was called out by a bunch of posters.  Yet, you spent a series of increasingly indignant posts defending yourself against accusations of cherry picking.  But here you make it clear you did understand the danger of using too small a point in time.  So go judge yourself, ey?

At this point you're contradicting so much of what you've had to say in recent weeks it's hard to see you as anything but a deliberate troll.

I've gone to a lot of effort to eliminate obvious causes of bias from this plot, and to qualify what it does, and does not show.  It is what it is, a way of understanding the now - where different nations are and which way things are moving.  

You give the impression you want to be taken seriously, but you regularly contradict your previous posts, and you show absolutely no understanding of the crudeness of the data, let alone their subtleties.  It raises a question for me - What are you trying to achieve here?  What do you want?

Edit: At least two regular readers will understand the attached image...

Post edited at 23:07

1
In reply to wintertree:

> You have claimed to be an engineer multiple times.  If so, you should understand a fair amount about noise statistics, and if you did understand the most basic points of noise statistics, you wouldn't be using the difference between a population of 2 million and 66 million as a reason to discount a comparison in the situation where rates are high.

Actually, although it may come as a shock to you, electronics and computing is a big field.  Your claim makes about as much sense as saying  I couldn't be a neurosurgeon unless I knew a lot about haemerroids.

How about some evidence the variation in Covid cases/deaths have anything like the same statistical properties as shot noise?     Looks to me like you grabbed a random page from Wikipedia.

 It's pretty obvious the national curves for Northern Ireland and Wales are spikier than those for the larger nations of the UK, because only one major city needs to have an outbreak to get a high national rate.  Latvia has roughly the same population as Northern Ireland.  

18
 AJM 08 Nov 2021
In reply to tom_in_edinburgh:

Hungary is about 20m, Greece, Austria, Romania all looked to be about 9-10m.

You could probably toss small differences (but only up to a certain size) for Latvia into the statistical noise bucket. That's a lot harder for Hungary. Onus is on you to prove that's noise, I think.

 Si dH 08 Nov 2021
In reply to AJM/Tom/Wintertree:

Tom is right that sub populations in the UK of a similar size to Latvia may well see (and in some cases have seen) a significantly different curve of detected cases from the country as a whole. This is nothing to do with statistical noise though. It's because the overall cases in a large country like the UK are far from a single outbreak that act homogeneously - for long periods different areas are partially disconnected from each other and behave in different ways. The overall rate is just an arithmetic mean of the case rates in the contributing regions, which in turn are means of the rates in their cities, towns, rural areas etc. It's only once you look at single major urban areas that, in my view, it becomes a more reasonable approximation to consider them as a single outbreak. So I don't think his argument is completely without merit. However, as we know and as one of the plots higher up partially shows, Latvia has had high case rates for quite a while so it's spurious in this instance.

OP wintertree 08 Nov 2021
In reply to Si dH:

I agree with your comments on outbreaks and cases; Tom was commenting on the actual death rates however…

The convolution of response times from infection to death is quite a broad kernel and blurs out most of the structure in cases caused from outbreaks in infection; once cases are as high as they are in Latvia, the prominence of individual outbreaks is greatly reduced even for populations down at that level, and then the convolution with the response function to the deaths curve further removes information on outbreaks from the signal.

Because cases and deaths have both been high and rising consistently for some weeks there’s no doubt in my mind that this is a fair representation of the situation.

For the home nations - including those with lower populations, the noise on the daily death numbers (by actual date, not reporting) is about 2x shot noise - I showed this about 30 threads ago; this holds even down at low absolute daily cases when outbreaks are dominating the behaviour of infections.  Noise on case numbers was about 20x shot noise, in no small part down to the day of week effect.  

> However, as we know and as one of the plots higher up partially shows, Latvia has had high case rates for quite a while so it's spurious in this instance.

Indeed.  Another angle to dismiss the claim is to assemble a “pseudo-country” with a 10x larger population centred geographically or politically/socially around Latvia.  Doing so is I think going to show something similar.  To try and dismiss the situation in Latvia as some statistical quirk - given the wider news coverage of the whole region - is heading for outright denialism.

Post edited at 08:25
OP wintertree 08 Nov 2021
In reply to tom_in_edinburgh:

Points in my post you did not address

  • The disparity between you claiming normalised rates would be the unbiased way of comparing things, then as soon as I use them and you don’t like what they show, complaining they are biassed…
  • The disparity between you previously cherry picking a single day of data and now claiming longer term rolling averages are biassed due to noise over too short a period…

You seem to have flipped your position about face on two different issues practically overnight. 

In reply to wintertree:

> Points in my post you did not address

> The disparity between you claiming normalised rates would be the unbiased way of comparing things, then as soon as I use them and you don’t like what they show, complaining they are biassed…

> The disparity between you previously cherry picking a single day of data and now claiming longer term rolling averages are biassed due to noise over too short a period…

I'm happy to admit that I'm often wrong.  The point where I start to take offence is when you start using words like 'lying' and 'cherry-picking'.

I'm also happy to admit I have a political agenda.  It is pretty f*cking obvious.  So do you.  It is less obvious but if you look at what you post from a non-English perspective it is pretty strong.  Far too much respect for the UK Government, the UK scientific establishment and specifically Oxbridge.  Not enough respect for the US, Germany and China.  

16
 AJM 08 Nov 2021
In reply to Si dH:

I don't think you're describing anything significantly different from what I mean. A grotesquely simplified way to model what you've described is to say that each country has n urban areas and that each area has either the outbreak rate or the baseline rate (a binomial distribution effectively - grotesquely simplified but for the purpose of explaining my thought process). The average case rate for the country at any point in time is the average rate across the n areas. As n falls, the volatility of outcome goes up.

My point is that of you're going to use this to explain away differences as just being volatility, some idea of how big a difference it could reasonably explain becomes necessary, particularly when all of the other countries at the top of the rankings are not the same size as the minnow that Tom initially highlighted! 

OP wintertree 08 Nov 2021
In reply to tom_in_edinburgh:

>  The point where I start to take offence is when you start using words like 'lying' and 'cherry-picking'.

You were cherry picking.  Pretend otherwise all you like.  I did offer the possibility that it was not deliberate, but was a result of gross incompetence at understanding the data filtered through your usual xenophobic bias.

I'm just surprised that you've gone from being so super-naive 14 days ago that you didn't understand the dangers of using a single day - or two different days for different nations - with day-of-week reporting effects to now dismissing multi-week trends through a moving average because of short term effects.  I'm not sure how you can be so wrong in two opposite directions basically at once. 

>  So do you.  It is less obvious but if you look at what you post from a non-English perspective it is pretty strong.  Far too much respect for the UK Government, the UK scientific establishment and specifically Oxbridge.  Not enough respect for the US, Germany and China.  

You may have missed it but I absolutely tore in to the UK government over our early response and I have consistently noted how unhappy I have been with many aspects since then, including many aspects of the present - messaging about risks for the more vulnerable, legal protections for the more vulnerable, honesty and transparency over the current approach, the testing lab scandal.  I've said almost nothing about the response of other governments except NZ, where I have had nothing but the most positive comments about their approach and whose approach I continue to hold as the gold standard we could at least have tried towards.

I have absolutely torn in to the University of Oxford as hosting and platforming - repercussion free - not one but two of the most transparent denialists throughout this crisis.

In short, Tom, I think your head is so far up your ass I'm surprised you can see.

 Offwidth 08 Nov 2021
In reply to tom_in_edinburgh:

I just choked on my morning pint of tea... oxbridge bias??? ....wintertree has been most vocal of all in criticising  two Oxford profs and their employer.

OP wintertree 08 Nov 2021
In reply to AJM:

> The average case rate for the country at any point in time is the average rate across the n areas. As n falls, the volatility of outcome goes up.

It's interesting; I get the sense that the "volatility vs n" relationship is cranked up a notch over the last six months vs last autumn; I think with vaccine uptake being geographically and demographically structured it preserves the outbreak-y nature to much higher n.  I haven't tested this.  If that is an effect, it's much less so in the countries with poorer vaccine uptake. 

> My point is that of you're going to use this to explain away differences as just being volatility, some idea of how big a difference it could reasonably explain becomes necessary,

Yes, it would be nice if the poster could test their hypothesis rather than going off on a tangent.

> particularly when all of the other countries at the top of the rankings are not the same size as the minnow that Tom initially highlighted! 

Surely you're not suggesting they cherry picked a country with a small population and ignored the presence of others with larger populations showing the same thing?  Checks calendar, confirmed: Day ending in -y.

 AJM 08 Nov 2021
In reply to wintertree:

> Plot Rates.a shows case rates and death rates (/million /day) for the four home nations. During most of the lat few months, Scotland has had a worse case rate and much worse death rate than England despite what I perceive as a generally more cautious approach to dropping control measures.

I have to confess I've been waiting since reading this to see in what way it was either:

A) a meaningless and invalid comparison, or

B) the fault of the Tories

It seems that this week we're running with A), based on the idea that comparisons between countries of different sizes are invalid. 

I can't quite put my finger on who it was that used to point at figures for Ireland (significantly smaller than Hungary, and everyone else at the top of the list except Latvia) and muse about how the quality of outcome was correlated with the distance from London and how there were some other parts of the UK - I forget which now - that could have done just as well had they not been tied to Tory policies from Westminster.

I hope they never read Toms post of this week, they'll be really disappointed to know that those sorts of comparisons don't work...

OP wintertree 08 Nov 2021
In reply to VSisjustascramble:

> China and the Far East, with very low levels of natural immunity have to take the same path as we have. A dodgy? Chinese vaccine, plus close to 0 natural immunity (if we believe the Chinese figures) means that the worlds factory is in for a very bumpy ride over the next 12/18 months.

I've been mulling this over.  I've decided to file away in the draw "don't borrow trouble from the future" for now.

 Offwidth 08 Nov 2021
1
 Si dH 08 Nov 2021
In reply to wintertree:

The China situation is interesting.

We don't know how good their vaccines are yet; the data I've seen suggested not as good as ours.

There are more than one and they might have been further developing them as they go; we probably wouldn't know if they were. It would seem unlikely they've been sitting back.

They seem to be getting pretty good take-up I think.

The population generally is much more inclined than ours to follow rules and the Government is far more able than ours to do big things fast (see early images from Wuhan's lockdown or the implementation of mass testing programmes for millions of people over a few days.)

Given the size of the country, they have done a remarkable job of keeping cases out over the last 18 months while simultaneously growing their economy. I wouldn't put it past them to continue doing this for quite a lot longer. If they have to maintain travel restrictions, quarantine and testing capabilities as they currently are, will that really hurt them in the grand scheme of things? Will other countries suffer more disruption from China's rules than they do? The CCP certainly seem to have a long term strategic direction and I'd be surprised if their actions over the last 18 months don't turn out to support it more than not.

Power shortages might cause them more problems than covid.

Having said all of the above, I have seen reports that some parts of the Chinese population are starting to get fed up with the current situation. No idea how reliable they are.

Post edited at 13:32
OP wintertree 08 Nov 2021
In reply to Si dH:

> The China situation is interesting.

Isn't it just.  The CCP clearly has a lot in its favour in terms of getting the population to adhere to stricter measures - more stick, less carrot etc, and I think they benefit from a much better integration of science and scientists in to their decision making - but it won't take much of a gap in their vaccine efficacy to open the door to basically uncontrollable spread of Delta if it gets in and loose.  Stopping that would I think hurt them quite badly.  High stakes stuff.

>  Will other countries suffer more disruption from China's rules than they do? The CCP certainly seem to have a long term strategic direction and I'd be surprised if their actions over the last 18 months don't turn out to support it more than not.

Yes.  They're never ones to squander an opportunity.  

> Power shortages might cause them more problems than covid.

Certainly changes the balance on some options in sealed drawers I suspect.

I was going to do an update of the week-on-week method rate constant plot today before I hit the road out of the labs, but the dashboard has this to say "Because of network issues, today's update is delayed".  The dashboard has gone from being basically perfect for month on end to having a lot of delays lately.  If it gets much worse the next Friday Night thread is going to slip from Saturday to Sunday...

OP wintertree 08 Nov 2021
In reply to wintertree:

> Yes, if there's no more reporting lag than normal, looking at the provisional data you've noted, we're going to see even faster halving times emerge, perhaps ~14 days by tomorrow evening

Yup, halving time is now sub 14-days in England.  An updated D1.c using the week-on-week method shows the fall is almost across the board - all ages over 10 - and still shows three ages leading the decay - secondary school aged, their parental / career age and the very elderly.  

I didn't get Plot 18 out in time for the start of the thread; the latest version is below and shows all measures in decay or turning to decay in all the English regions.  Something Si dH noted a few threads back - and that cuts to the other discussion on this week's thread - is that the data has been pretty homogenous over all of England recently; other than the SW cases debacle there's unusually tight grouping of the regional traces on the plots for the last couple of months.

Big changes coming to England.  Not so much to Scotland which remains in low exponential rate growth.

Driving home there were several radio adverts on the subject of opening windows at home and at work occasionally to clear the air of virus loaded particles; the voice was rather channeling the spirit of dark and lonely water I thought.  

Edit: To put the current decay in to context - assuming that the latest data marker holds up against unexpected reporting lag - this is the most impressive exponential decay rate by the week-on-week method since mid-April, if we discount misleading decays from a bank holiday and from the end of the football driven transmission.  In April most cases were the previous, less transmissive variant, we still had control measures in place and the weather was getting better.  Now, we have a more transmissive variant, less control measures and the weather is getting worse.  If the decay rate keeps dropping like this I just might wrap the threads up with #52 after all...

Post edited at 18:21

 Misha 09 Nov 2021
In reply to wintertree:

Thanks. Things are clearly in decline (I kept an open mind last week given the impact of half term). The question is how far and for how long. I fear there are a number of factors working against us - winter, Xmas socialising and waning immunity (including in younger age groups). So I’d be surprised if it really drops off much before Easter but if it stabilises at 20-30k cases per day, that’s better than 40-45k…

 Misha 09 Nov 2021
In reply to tom_in_edinburgh:

Things got pretty bad in Latvia. They had to implement a strict lockdown a couple of weeks back and hospitals were already running out of capacity. You are right that comparing short term death rates can suggest misleading conclusions (especially when using one day’s data, now who would ever do that 😉) but in this case the short term data reflect the sad reality. Vaccine uptake seems to be an issue (not surprising given that a third of the  population are ethnic Russians and uptake in Russia is very low). Another factor could be that they got away relatively lightly in the first wave (less so last winter).

 Misha 09 Nov 2021
In reply to tom_in_edinburgh:

> Far too much respect for the UK Government, the UK scientific establishment and specifically Oxbridge.

I’m not sure many of the regular contributors to these threads can be accused of having much (if any) respect for the UK government. WT has repeatedly said he’s appalled by the government’s Covid response. I think most contributors would echo that. I certainly would.

I do think that ‘the scientific establishment’ (whatever that is - I assume you mean SAGE, the modellers and AZ) deserve respect. They don’t and can’t get everything right but they do the best they can in a difficult situation. I do think the JCVI was too slow opining on vaccinating children and gave the ‘wrong’ answer due to a rather narrow remit (probably due to the way it’s been set up for ages). However that’s the only major slip up on the scientific side of things which I would call out.

 Misha 09 Nov 2021
In reply to wintertree:

Of course it could just be due to another dodgy lab scandal… I say in jest, hopefully.

Scotland is an interesting one. Could it be a forerunner for the rest of the UK? Colder climate and earlier half term (at least a week earlier in most areas) meaning that they are seeing now what we will see in a couple of weeks’ time? Or lower historic infection rates? Vaccination rates are similar or higher. 

OP wintertree 09 Nov 2021
In reply to Misha:

> Of course it could just be due to another dodgy lab scandal… I say in jest, hopefully.

I've been wracking my brains about what one-off factor could be behind it leading to another round of false hopes; I'm coming up with nothing.  There's no sudden burst in the gender signal like there was for the football, although the imbalance towards female cases in younger adults is higher than last year.

> Things got pretty bad in Latvia. They had to implement a strict lockdown a couple of weeks back and hospitals were already running out of capacity. [...] in this case the short term data reflect the sad reality. 

Indeed, and sadly it's not a reality confined to one nation in that general region.  You can see their lockdown showing in cases really clearly now, and the extrapolation plot for deaths is now showing a fall - but a fall from a level very close to that the UK hit pre-vaccine.   

The plot is updated below with the sideways "T"s flipped to be more intuitive with the vertical marker at current death rate.  It's very worrying how rapidly this plot is changing.  

  • The main way I look at the extrapolated death rates is as an indicator for the level of healthcare overload; anything approaching the peak level from the UK pre-vaccine is going to be awful for healthcare; different countries have different levels of provision but not wildly so vs exponential growth.
  • To preface the next points - they are not predictions, but extrapolations; the expectation is for increasing use of control measures to pull countries off their current trajectories and away from healthcare overload - although it's sadly clear that that doesn't always work out in time.
  • Hungary is now showing an extrapolated death rate twice the pre-vaccination peak in the UK and they're still rising.
  • Slovakia, Greece and Poland only have two weeks at current rates before hitting this level, and they're still rising.
  • Belgium and Austria have only about two and a half weeks at current rates before hitting this level.  These are the ones really surprising me; I'd expected more control measures by now - the exponential rate constant (growth rate) for Austria is still getting worse, it's hard to tell for Belgium from the OWiD data given the erratic sampling effects.

The phase space plot is becoming a dog's dinner but can be squinted at.  It really needs a Web 2.0 thing to select / deselect individual nations in real time.


OP wintertree 09 Nov 2021
In reply to thread:

Another day of week-on-week decay for England.  The top level halving time has stopped getting better, but in the more lagged demographic data, the halving time for adults is still reducing (getting better).  

An update to Plot 9e confirms that all measures for England are now in decay; so this is definitely a real fall in infections in adult ages. 

The big question now will be if the next weather wobble remains in decay even at its peak...

The right side of the low frequency band on the LMH plot is a happy curve.  Long may it continue.

This is not a prediction.  To give another context to the current decay; if things continued on this route and without considering the emerging anti-viral therapeutics, this would have us down to around  2,000 cases, 100 admissions and 10 deaths a day by Christmas.  If the new antivirals can be well targeted (easier with lower cases I suspect) that could more or less eliminate deaths and reduce hospital admissions to insignificance.

Scotland - cases still rising.   Wales (no plot) is staring to show decay much like England; NI continues to stagnate at a high level.


 Misha 09 Nov 2021
In reply to wintertree:

There’s also behaviour - people see the news and modify their actions to an extent. But even if the death rates are moderated, anything near our worst point is bad news. Not least because of how deaths relate to hospitalisations now vs last winter. I assume fewer deaths as a % of hospitalisations now (even in countries with lower vaccine uptake). So same death rate as the UK’s worst point would suggest a higher hospitalisation rate. As you say, hospital capacity is better in some countries - and worse in others. 

 Offwidth 10 Nov 2021
In reply to wintertree:

I'm in shock after just checking how bad vaccination rates still are in some local inner city areas (mid 30s percent). I'm sure there will be some NIMs denominator student effects depressing the numbers, especially in Radford and Hyson Green,  but even so. What's worrying for the next months is I've seen no indication from exceptional cases, hospital and deaths in those areas that nearly all of the unvaccinated have infection immunity. Hospitals in Nottingham and Leicester currently just don't have spare capacity if we have a local surge in infections.

https://www.nottinghampost.com/news/local-news/communities-nottingham-among...

Post edited at 08:29
2
 Offwidth 10 Nov 2021
In reply to wintertree:

Belgium now at UK hospitalisation and death levels per capita and still growing, so they are heading for trouble and will need their better margin of healthcare capacity (the indications of a potential slow down didn't pan out). Rolling average CFRs on OWiD indicated their testing improved as case numbers took off.... hence maybe why hospitalisations are not quite as bad as their case increases might have indicated. Not much obvious sign of major political action though.

https://www.thebulletin.be/coronavirus-belgium-information-covid-19-questio...

It's bad but less serious in Germany, although cases, hospitalisations and deaths continue to rise (again some indications of a slow down didn't pan out) on a higher and more slowly decreasing OWiD rolling CFR.

Post edited at 12:05
OP wintertree 10 Nov 2021
In reply to Offwidth:

Belgium - as I said before, that fall looked like reporting jitter to me, and the data is getting worse not better in that regards.

CFRs - the OWiD plot uses a 10-day lag which is not very robust and to my eye the changes in Germany mostly look to be in the noise; I've put my plot in below giving a range from 7- to 28-days; this obviously is less current given the longer lags, but it shows a generally increasing trend in most nations, not inconsistent with the idea that testing works better at low case rates than at high case rates.   

Big picture - it looks like anything short of signifiant control measures is not going to prevent cases from rising to the point they break healthcare for much of Europe, where-as the UK continues to see real falls in all measures despite a more lackadaisical approach to controlling the virus.  This seemed to cause a lot of consternation when I first suggested it but it's panning out pretty much as expected.  Today's international news is not happy reading.


 Si dH 10 Nov 2021
In reply to wintertree:

It's starting to look like SAGE's recommendations for the UK to have an exit wave in summer/autumn to avoid one in winter might end up being justified after all... I just hope those countries with increasing rates now don't get caught out by waning immunity in their vulnerable populations at just the wrong time.

OP wintertree 10 Nov 2021
In reply to Si dH:

Yes.

As they say, understanding is a three edged sword.  I hope there’s some more understanding out there now.

I wasn’t expecting the situation in much of Europe to degenerate as rapidly as it is.  A stark warning from someone in Germany today - https://www.bbc.co.uk/news/world-europe-59234443 - act soon or risk their worst wave of deaths yet.

 Misha 10 Nov 2021
In reply to Si dH:

May be. Time will tell. An alternative interpretation is that releasing the brakes slower could have mitigated the healthcare and mortality impact to date and flattened the curve both up to now and going forward. We didn’t actually have a really significant spike post Freedom Day, except among children. This suggests to me that where we are now is less to do with what’s happened since Freedom Day and more to do with our trajectory up to that point, including getting hit by Delta earlier than most other countries apart from India.

It is certainly true that the worst (or even mid range) projections made around July did not come to pass. Again, that suggests to me that immunity levels (from vaccines and/or infections) plus people moderating their behaviour is what really made the difference, rather than opting for an earlier wave. So I’d say the jury is still out on this. Perhaps the reality is that, for the UK by late July, it didn’t really matter in the grand scheme of things what the policy was as we were post Alpha, post Delta, post football and post most 30/40+ being vaccinated. It still had an impact of course and hence we are having 1,200 deaths a week now but it was never going to be 1,200 a day again.

Cases seem to be starting to plateau again (by eye), as are admissions (with some variations), occupancy and ICU numbers. I’d love to be proven wrong but I suspect we won’t see major falls until the spring…

8
 Misha 10 Nov 2021
In reply to wintertree:

The article says Saxony is the worst for cases and vaccination level. Saxony was in East Germany, so this doesn’t surprise me. The shadow of the USSR looms large… It is no coincidence that most of the European countries which are currently in the worst situation used to be in the Soviet Block (of course there have been some other consistent bad performers like Belgium and indeed the UK). Partly due to being less wealthy and partly due to having lower vaccine uptake, but both of these factors stem from the Soviet legacy. The vaccine issue is an ex-Soviet mindset thing (a lingering distrust of authorities, plus selfish individualism displacing the former veneer of collectivism). Ironically, vaccines used to be pretty much mandatory in the USSR.

Why Belgium though? High population density? Something to do with the Flemish / French divide and political mess in recent years? It can’t just be bad luck several times in a row!

2
OP wintertree 10 Nov 2021
In reply to Misha:

> Cases seem to be starting to plateau again (by eye), 

> Perhaps the reality is that, for the UK by late July, it didn’t really matter in the grand scheme of things what the policy was as we were post Alpha, post Delta, post football and post most 30/40+ being vaccinated

Looking at the annotated version of plot D1.c (annotated) from the opening of the thread, I wouldn't be so sure.

> Cases seem to be starting to plateau again (by eye),

The week-on-week decay rate is slackening off - nothing unexpected as it always wobbles around its baseline (mostly weather I think); the big question is "has the baseline dropped enough that the two-week average rate remains in decay?" and the cherry-on-top question is "has the baseline dropped enough that the rate remains in decay at all points in a wobble?".  Hazarding a guess I'd say yes and maybe.

> as are admissions (with some variations), occupancy and ICU numbers.

I hope that's some pessimism creeping in to your view... I don't think admissions can be meaningfully interrogated for change on short term timescales in the way cases can; smaller numbers and more statistical noise in a week-on-week comparison.  Occupancy measures change much more slowly as they're convolutions with broad kernels.  We might expect a small plateau in the decay of admissions from the period end of October where the decay faltered in adult cases briefly during the last wobble, but it's only just going to start landing in the data  now; I doubt they'll return to growth giving the blurring out of structure from cases.

Latests week-on-week plots below, the slowing decay hasn't landed in the demographics yet, but recently the swings have been much larger in the child ages; if that's the case again, cases will remain falling more where it matters most.  We'll see clearly by the next thread I hope.


In reply to Offwidth:

> I'm in shock after just checking how bad vaccination rates still are in some local inner city areas (mid 30s percent).

My MSOA had stalled at just under 50% first dose for months. In the last couple of weeks, it has crawled past 50%, and is at 55% today (55/48% 1st/2nd). I'm not sure if this is real vaccination (my MSOA includes most of a University...), or massaging of population figures (i.e. correcting/weeding Uni population figures).

Looking at some of the London MSOAs, there are some weird anomalies; you might think the poorer areas would have lower uptake, but Fitzrovia/Bloomsbury is amongst the lowest.

On the whole, London seems better than most of the other major city centres; Birmingham, Manchester, Leeds, Bradford, Sheffield have worse figures. Leeds Harehills South looks to be the worst at 35.3/28.3%

It's also interesting that some places where vaccination is low also have low case numbers; Birmingham & Oxford, for example.

Post edited at 23:12
 Misha 10 Nov 2021
In reply to wintertree:

There has certainly been a large peak among children but that’s about it. Overall cases have been at significant but not overwhelming levels and have essentially pogoed around (and I think they are still doing that). I’m adopting a helicopter view and looking at the whole period since the football peak passed. I know it’s important to look at the detail but if you zoom out and look at the overall picture, we have essentially been in a steady state situation (not a good steady state but not disastrous either). All the metrics have pogoed around within a range which is a lot worse than in April and a lot better than in January.  

My point (and I may be completely wrong) is that I’m not convinced that triggering an early autumn wave to avoid a worse wave in winter has really worked (if that was ever the strategy), because there hasn’t really been an early autumn wave among the more vulnerable. There has been a wave among children but it did not cause a wider wave. It did however raise the baseline, which is causing significant but not overwhelming healthcare strain and death.

By extension, I wonder whether releasing the brakes more slowly would have been better (if there was never going to be a wave among the vulnerable, there was nothing to be gained by raising the baseline - in other words, the baseline was raised for no real medium to long term gain). That was one of my original concerns with Freedom Day (why pay tax now in the hope of avoiding paying tax later, when you might never have to pay that tax anyway).

Again, by extension, I’m not convinced that the picture in (some) European countries is worse now because of our trajectory since the end of July. I think it goes back much further. Germany got through the first wave relatively well and could now be paying for it (or at least unvaccinated people will be).  

Of course it could be that releasing the brakes later would have caused a wave among the more vulnerable even though it didn’t in early autumn, due to immunity fade. Then again, releasing the brakes after a significant proportion of children have been vaccinated would have moderated the wave.

Lots of possibilities. May be someone is modelling this properly.

5
 Si dH 11 Nov 2021
In reply to captain paranoia:

> My MSOA had stalled at just under 50% first dose for months. In the last couple of weeks, it has crawled past 50%, and is at 55% today (55/48% 1st/2nd). I'm not sure if this is real vaccination (my MSOA includes most of a University...), or massaging of population figures (i.e. correcting/weeding Uni population figures).

> It's also interesting that some places where vaccination is low also have low case numbers; Birmingham & Oxford, for example.

Most city centre msoas have been showing surprisingly low case rates for several months now. I think the same population data anomalies that cause poor vaccination rate predictions also effect data on local case rates. Basically they are over estimating the total population in those areas significantly due presence of lots of students, which artificially reduces the calculated case or vaccination rate. In posh bits of London it will be second home owners having the same effect.

Personally I don't think it's worth worrying about vaccination rates below regional level now. It was interesting during the big push for 1st/2nd doses to see how well different areas were performing in offering them, but there has never been any obvious correlation between low vaccination and high cases even when we looked really hard for it (eg the Bolton outbreak of Delta).

Post edited at 08:00
OP wintertree 11 Nov 2021
In reply to Misha:

> Overall cases have been at significant but not overwhelming levels and have essentially pogoed around (and I think they are still doing that). I’m adopting a helicopter view and looking at the whole period since the football peak passed. I know it’s important to look at the detail but if you zoom out and look at the overall picture, we have essentially been in a steady state situation

I don't think we've ever been steady state.

The pogos are mostly the weather modulating the rate constant on the timescale of weeks; this is shown to almost 4 sigma - beyond a gold standard for detection in the life sciences.

If we look at the behaviour of the rate constant on a longer term than the pogos, it is decreasing - by eye, and by the LMH plot's low frequency band.  We just don't see it easily by eye because the pogos are larger in scale than the change in the baseline - but each pogo is becoming more negative at minimum and less positive at peak.

It's by understanding the detail that you can sweep it aside and get an accurate look at the big picture.

Other things I think are going on in that period are a step change in rate constant when schools returned which took a couple of weeks to build up enough cases in children to then drive the top level figures, and a more immediate blip from the switch on of more testing when schools returned.  

> There has been a wave among children but it did not cause a wider wave. It did however raise the baseline, which is causing significant but not overwhelming healthcare strain and death.

The closest comparator is Scotland which has had a similarly dismal pre-vaccine death rate to England (compared to much of western Europe), and has then had a much lower peak rate and area under the curve in school aged infections in recent months, and is now remaining in growth as England increasingly turns to decay.

> Again, by extension, I’m not convinced that the picture in (some) European countries is worse now because of our trajectory since the end of July. I think it goes back much further. Germany got through the first wave relatively well and could now be paying for it (or at least unvaccinated people will be).  

For sure - but its an ongoing process, and perhaps the last few months brought the final change that tipped the scales over to decay in England; places with a worse starting point and lower case rates in recent months obviously aren't going to hit that tipping point without more restrictions - quite a lot more I think given the way seasonality is apparently panning out in places with very mild restrictions.

In terms of what's making it worse elsewhere I think historic cases are a small factor compared to vaccination; people (not just the unvaccinated) are paying the price for less effective roll out - when examined spatially and demographially rather than just by top-level percentages.

>  Then again, releasing the brakes after a significant proportion of children have been vaccinated would have moderated the wave.

Although the 1-dose regime under the JCVI guidance does little to moderate the spread at school ages perhaps; something missed by a lot of the arguments made around schools and spread.  The vaccination is intended to minimise the - already very small - direct risks to children of both the vaccine and the virus.  

On that note - Moderna is being recommended against for under 30s in both France and Germany.  Once again I find it surprising that there's none of the FUD and astroturfing going on over this that we saw from European politicians, segments of the media and a couple of UKC posters over AZ.  

https://www.forbes.com/sites/roberthart/2021/11/10/germany-france-restrict-... 

Post edited at 08:18
In reply to wintertree:

“Area under the curve”

I suspect this might be the phrase of 2022 in the western world.

Removed User 11 Nov 2021
In reply to wintertree:

I notice with interest the beeb are slipping out the fact that vaccine appears to have little or no effect in reducing transmission beyond any reduction associated with a decreased likelihood of infection in the first place.

The case for vaccinating children appears completely unjustified at this point.

4
In reply to wintertree:

I see the EU are buying Valneva Covid vaccine and their shares are back up despite Javid trying to f*ck them.   

https://ec.europa.eu/commission/presscorner/detail/en/ip_21_5784

If Scotland doesn't get an uptick in cases because of COP it will be a very good sign.  

It's funny that Johnson wears a mask in a train station in Scotland but doesn't see the need in a hospital in England.  

7
 elsewhere 11 Nov 2021
In reply to Removed User:

Decreased* likelihood of infection alone is a reduction in transmission.

Hence whilst vaccination of children may be unjustified (I tend to disagree) it is not completely unjustified.

Additionally if the child doesn't get Covid they don't kill granny, which would be pretty bad for the kid. As well as for granny.

It looks like Covid is or will be endemic. Getting infected before vaccination is never lower health risk than the combined health risk of vaccination and Covid infection after vaccination. I think that's true for all ages including children. 

*vaccination reduces risk of infection by 50-60% and risk of death by 90-95%. Approximate figures. 

Post edited at 09:35
1
 Si dH 11 Nov 2021
In reply to elsewhere:

I am personally uncomfortable with the idea of vaccinating young kids just to reduce transmission. If they get vaccinated it should be for their own benefit. This is especially true with vaccines for which we still have limited long term safety data, which the regulators have asked to continue seeing updated for at least another couple of years.

> Getting infected before vaccination is never lower health risk than the combined health risk of vaccination and Covid infection after vaccination. I think that's true for all ages including children. 

It's quite clear from the various minutes available that JCVI do not see enough evidence about the effects of vaccination to share your view for younger teenagers. This is why they didn't recommend vaccinating them and instead came out with a conclusion that allowed the CMO to overrule them on the grounds of disruption to education.  For kids under 12 the judgement would undoubtedly be more clear against. Since those minutes were released there has also been another big round of natural infection in kids and studies have been reported showing that young kids produce strong antibody responses even though they have mild symptoms (sorry, I can't remember the reference for this but I think it was one of these recent threads where it came up.) Overall, I just can't see a justification for vaccinating young kids against covid in the short or medium term, and if it was pushed through politically before there is long term data available I would be reluctant to do my own son.

Definitely better to spend our doses on boosters for the vulnerable and on people in more need elsewhere in the world.

(Edit to fix a typo)

Post edited at 09:51
 Si dH 11 Nov 2021
In reply to Misha:

> It is certainly true that the worst (or even mid range) projections made around July did not come to pass. Again, that suggests to me that immunity levels (from vaccines and/or infections) plus people moderating their behaviour is what really made the difference, rather than opting for an earlier wave. So I’d say the jury is still out on this. Perhaps the reality is that, for the UK by late July, it didn’t really matter in the grand scheme of things what the policy was as we were post Alpha, post Delta, post football and post most 30/40+ being vaccinated. It still had an impact of course and hence we are having 1,200 deaths a week now but it was never going to be 1,200 a day again.

The time-integrated number of Delta cases since the end of July is significantly greater than prior to the end of July. See graph attached.


Removed User 11 Nov 2021
In reply to elsewhere:

I'm no Covid denier (and I'm double vaccinated) but I think the messaging on vaccines stinks and has 'stunk' all the way through this. Coercion and propaganda all the way through, full of unjustified statements and assertions. Even for those of a scientific mind, trying to assess the real risks is nigh on impossible without a substantial literature trawl. A horrific inability to be able to debate some of the less clear aspects of policy and strategy i.e. those outside the strictly medical arena e.g. vaccine passports etc.

Maybe it's for the greater good but I feel like we are living in a dystopian nightmare...

6
OP wintertree 11 Nov 2021
In reply to elsewhere:

>  Getting infected before vaccination is never lower health risk than the combined health risk of vaccination and Covid infection after vaccination. I think that's true for all ages including children. 

The gods themselves know how much effort I've spent presenting a case against mis-informed and deceitful opinions that there are any circumstances where an adult is better off getting their first exposure by live infection rather than by vaccination.  In my view the evidence is clear that vaccine first is always the best way forwards for an individual adult, using the most age appropriate vaccine.

With younger children, the risks from both infection and vaccination are so low in absolute terms that gathering evidence to say which is a lower risk is a slow and demanding process.  I file this under the class of problems labelled "good problems to have" - certainly compared to alternative possibilities.

The current JCVI guidance for children aged 12-15  and for those aged 16-17 is for a single dose only for now, pending future decisions on a potential second dose much further down the line.  I recommend reading their statements on this.

> *vaccination reduces risk of infection by 50-60% and risk of death by 90-95%. Approximate figures. 

Vaccination with two doses does that.  As stated, aged 12-17 for now, it's single dose only unless some vulnerability criteria apply.  This confers far less protection from infection but confers protection from serious illness.

Various posters have been criticising the dropping of restrictions in schools before more vaccination has happened on the assumption this would significantly reduce transmission, but my view is that this is not a valid criticism as it completely fails to account for the consequences of the JCVI guidance.  IMO to advance the discussion, either the JCVI guidance needs to  be recognised and its implications clearly accepted, or a cogent challenge to the guidance needs to be presented.

Post edited at 10:01
1
 Offwidth 11 Nov 2021
In reply to Si dH:

I totally disagree with that. Two areas with 36% and 38% in Nottingham is a local vaccination organisational disaster. We have a lot of students but they are mostly strongly concentrated elsewhere (where the student age vaccine numbers are more obviously depressed due to NIMs denominator issues). There is no way student populations are more than 10% in those areas. They are the most ethnically diverse parts and amongst the  poorest parts of the city. It's very much something to worry about.

If you missed a correlation in Bolton than something is going wrong with the analysis, given a clear correlation of poorer outcomes in the linked factors of deprived areas and higher proportions of ethnic groups where vaccination rates are worse.

2
OP wintertree 11 Nov 2021
In reply to Si dH:

100% agree with this post.

 mondite 11 Nov 2021
In reply to Removed User:

>  A horrific inability to be able to debate some of the less clear aspects of policy and strategy i.e. those outside the strictly medical arena e.g. vaccine passports etc.

Really? vaccine passports have been repeatedly discussed as to its usefulness.

1
OP wintertree 11 Nov 2021
In reply to Removed User:

> Maybe it's for the greater good but I feel like we are living in a dystopian nightmare...

The way I see it is that Covid gives the clearest illustration yet of the developing cracks being driven in to our society, cracks that could make things worse for a lot of us in a far wider sense than Covid.

 Si dH 11 Nov 2021
In reply to Offwidth:

> If you missed a correlation in Bolton than something is going wrong with the analysis, given a clear correlation of poorer outcomes in the linked factors of deprived areas and higher proportions of ethnic groups where vaccination rates are worse.

I didn't miss it. I did some investigation and found that it did not exist. The initial outbreak occurred in Rumworth, an area with a lot of Indian travel connections (unsurprisingly) and coincidentally low vaccination rates, but there was no clear correlation between the rate of spread to other areas of Bolton and their local vaccination uptake. At the time I found this surprising. Since then we have seen lots of data showing how easily double-vaxxed people can still get infected and still transmit the disease with Delta, so it's far less surprising. Vaccination rates are of course far more important for determining health outcomes but there is no good data in the dashboard to determine this correlation below a regional level. Better off looking at the national-level data in the UKHSA vaccine surveillance reports.

Edited to moderate my language slightly and to provide a link to where I initially reported this: about a quarter of the way down. Spread out of Rumworth did not seemingly correlate with either local deprivation, ethnicity or vaccine uptake.

https://www.ukhillwalking.com/forums/off_belay/friday_night_covid_plotting_27-...

Post edited at 10:16
 Offwidth 11 Nov 2021
In reply to Removed User:

Maybe you and others who have been jabbed  feel that way because the formal population political messaging has too often been at odds with the medical and epidemiological messaging, a situation which is genuinely confusing and is significantly poorer overall compared to the idea of consistent public health messaging. Alongside that, social media is full of dangerous bs funded by bad actors and often involving rogue medical staff... information that can be anything from unnerving to terrifying for people lacking basic medical understanding.

The vaccines followed standard safety approvals and have very high efficacy. In the areas where any known vaccine risks lie they are very small risks and the exact same risk is usually much higher (and always higher) if you catch covid unvaccinated. The mass data following the vaccination process clearly shows vaccination benefits are there for all ages and are stronger the older you are.

We are indeed living in a dystopian nightmare, that's what such a major pandemic is, especially with an underfunded NHS struggling to cope.... we were unlucky that our politicians chose to make it worse than it needed to be.

Where vaccinations are being forced by employers (care homes) it's ethically dubious and may blow up in our face: tens of thousands of workers will lose their care positions today in a sector where we have mass vacancies leaving too many homes at operationally dangerous levels. The systems should be in place to prevent infection from staff in close to 100% of cases (using testing and PPE). The biggest threat by far of being unvaccinated as a care worker is to themselves... pretty high right now; it only increases risk of transmission to residents by a fraction of a percent. The contact time lost due to unvaccinated care workers infected or kicked out by this policy will add significant stress to an already too often overstressed environment....overall a much larger risk to residents. Its also standard human psychology that forcing undermines the public health message.... not a good tactic with too high proportions of vaccine hesitancy in roles with a crisis in unfilled posts.

2
 Offwidth 11 Nov 2021
In reply to Si dH:

The outcomes I'm talking about and worried about in these inner city areas (that are far from student ghettos) are hospitalisations and deaths. If we are talking cases I might agree but cases in themselves are not the major concern.  I've not noticed strong correlation of particular ethnicity with  catching covid but rates of hospitalisation and deaths can be much higher than average. These areas are exactly where you don't want continuing low vaccination rates going into winter, especially if the local hospitals are already in trouble. It's serious.

Post edited at 10:49
3
 Si dH 11 Nov 2021
In reply to Offwidth:

> The outcomes I'm talking about and worried about in these inner city areas (that are far from student ghettos) are hospitalisations and deaths. If we are talking cases I might agree but cases in themselves are not the major concern.  I've not noticed strong correlation of particular ethnicity with  catching covid but rates of hospitalisation and deaths can be much higher than average.

No argument with any of that. I just don't think the dashboard data for vaccinations is very reliable at MSOA level, and hospitalisation data only exists at trust level so there is no way of determining a correlation anyway.

You might be right that specific areas of Nottingham have low vaccine uptake, you obviously know them far better than me. I just wouldn't personally read too much in to some of the numbers on the dashboard.

 TomD89 11 Nov 2021
In reply to Offwidth:

My personal dystopian nightmare fears are not driven by the fringe opinions about vaccines, masks etc but the feeling of growing centralized control and ulterior ideological motives shoehorned in under the guise of consensus science.

However it is refreshing to see that you consider coercing NHS/care staff into vaccination is wrong and clearly not logical when they are reportedly so close to overwhelm. The cost/benefit does not add up at all. It certainly fuels the hesitancy and feeds mine and others paranoia.

I wonder if this could be one of the first points people from all sides could firmly agree on? Anyone for forcing care/NHS workers out of jobs? If no public support and minimal to negative PH benefits, what exactly is driving this move?

Post edited at 11:06
6
In reply to Removed User:

> I'm no Covid denier (and I'm double vaccinated) but I think the messaging on vaccines stinks and has 'stunk' all the way through this. Coercion and propaganda all the way through, full of unjustified statements and assertions. Even for those of a scientific mind, trying to assess the real risks is nigh on impossible without a substantial literature trawl. A horrific inability to be able to debate some of the less clear aspects of policy and strategy i.e. those outside the strictly medical arena e.g. vaccine passports etc.

> Maybe it's for the greater good but I feel like we are living in a dystopian nightmare...

I know many people criticise government messaging for not being clear, but I’m of the view that it’s been too effective in the case of vaccines and the risks posed by Covid to younger people.

The government inadvertently backed themselves into a corner and now there’s no room for intelligent discussion. The media are of course just as culpable as they love a good scare story.

Should children be vaccinated against Covid? Many people (including Offwidth, Elsewhere ect) will say Covid bad, vaccine good, so of course they should be vaccinated. The reality is of course far more complicated, but viewpoints have been set in stone and won’t be changed.

Given where the UK is compared the the rest of Europe, hopefully by spring 2022 the rest of the world will be able to point at the UK as the way out of the pandemic.

5
 Offwidth 11 Nov 2021
In reply to Si dH:

We don't get local area maps of deaths or hospitalisations, despite these being the key factors in this pandemic 

The following is anecdote but it almost certainly contains some important truth. I am a middle class person surrounded by friends and family who take covid seriously. I know many thousands of people (the demographic tends to be weighted at the younger adult and older ends)  but I am still not aware of anyone who has died of covid (and it's unlikely for me not to have been told in most cases...... I've been invited to 7 funerals during the pandemic - all deaths due to other causes) and only a small percentage of those I know have caught covid (mostly school kids, or adults in their 20s and 3Os.... but I'm much less likely to be told about this). I'm only aware of a few adults unvaccinated (young women frightened of fertility issues thanks to misinformation). I know quite a few who have died who are relatives or friends of people I know. When I talk to friends who live in these inner city deprived areas, and are connected in the community, they know many who have died and the majority seem to have caught covid but they know significant numbers of individuals being very careful (the risk is vry obvious to them). Covid is brutally exposing inequality.

Post edited at 11:44
4
 Offwidth 11 Nov 2021
In reply to VSisjustascramble:

As too often is the case you are expressing dangerous opinions. Why don't you line up these so called complexities we are ignoring for us to debate.

Vaccines are not a good in themselves... they have known side-effects so you wouldn't take them if there was no disease risk.

These covid vaccines in a covid pandemic are surprisingly good...good efficacy in preventing infection, very good efficacy in preventing hospitalisations, excellent efficacy in preventing death. All the major side- effects of the vaccines are at a higher individual risk level if you catch covid unvaccinated. There are a tiny number of people with particular conditions who have been advised not to be vaccinated (or vaccinated again). There is no down side I can see to vaccination for adults. The reason I oppose compulsion is it can have unwanted secondary effects (tens of thousands of care workers out of a job in a sector facing a recruitment crisis) and can reduce uptake, as it feeds hesitancy in the malign social media environment we live in.

For children, vaccination is more of a risk balance but the risk is a still significantly higher to them without vaccination (that's what JVCI said, in my view after being rather over-cautious). There are also major additional benefits in reducing school disruption (much reduced in other countries where they vaccinated earlier and had more proportionate restriction control) and in cutting population spread.

Post edited at 11:40
1
Removed User 11 Nov 2021
In reply to Offwidth:

I just think we have a completely fcuked interpretation of risk through all this. Risk may be higher in a given circumstance but that risk may also the same time be essentially negligible, as appears to be the case for Covid in healthy children.

Post edited at 11:46
OP wintertree 11 Nov 2021
In reply to Offwidth:

> For children, vaccination is more of a risk balance but the risk is a still significantly higher to them without vaccination (that's what JVCI said, in my view after being rather over-cautious)

That is not quite my interpretation of what JCVI said.

They are recommending one dose only for now.

> There are also major additional benefits in reducing school disruption (much reduced in other countries where they vaccinated earlier and had more proportionate restriction control) and in cutting population spread.

These "additional benefits" apply with two doses and their benefit against transmission is reduced by delta.

Do you object to the reasons JCVI are recommending one dose for now?  If not, why do you continue to compare outcomes against countries recommending two doses?

Apples and oranges.  

As we're seeing ever more clearly across Europe, double vaccination in children is not a magic panacea to the spread of the virus, the reality is far more complex.

Post edited at 12:05
 Offwidth 11 Nov 2021
In reply to wintertree:

I think one dose is maybe over-cautious but its not an unreasonable position to take as risks increased noticably with the second dose. Those other countries have the same information and their equivalents to the JCVI have said the benefits were there with two doses, their decision was also reasonable if slightly less cautious.

It's not true that the benefits only accrue with two doses. There are just bigger benefits with two doses. There is some infection reduction with one dose and so some reduction in subsequent infections. There is significant reduction against severe infection for children with one dose.

https://www.sciencemediacentre.org/expert-reaction-to-uks-chief-medical-off...

I don't see anyone claiming child vaccination is any panacea but it does help a bit and small differences are important right now in the UK with R (effective) stuck around 1 and a way too low vaccination rate in that age group.

4
OP wintertree 11 Nov 2021
In reply to Offwidth:

> It's not true that the benefits only accrue with two doses.

Which is why I did not say that "the" benefits only accrue with two doses.  

What I said was: 'these "additional benefits" apply with two doses' directly after quoting your post defining those additional benefits as applying to curtailing spread of the virus.

>  There is some infection reduction with one dose and so some reduction in subsequent infections.

But however not by a lot in the face of delta's capabilities, and certainly not enough IMO to make a difference about if we had a full blown large scale pandemic in school ages or not within the timescale of one term.  One dose changes the rate of this a little bit which, with the current isolation requirements, doesn't really affect net total education disruption.  It would have stretched and lowered the peak in adults a little bit perhaps. 

> There is significant reduction against severe infection for children with one dose.

Yes, that much is obvious.  I haven't claimed otherwise as you can see from a proper reading of my previous post.

>  with R (effective) stuck around 1

Is it really?  I don't actually think it is.

Post edited at 13:03
 Offwidth 11 Nov 2021
In reply to wintertree:

Give me your odds on the next assessment of R for England as being below 0.8 to 1.0. When was the last time it wasn't close to 1.0? I'm fully aware this is on delayed data but the bad decisions to  not do something to slightly reduce R were consistent throughout the last months.

https://www.gov.uk/guidance/the-r-value-and-growth-rate

This is why a reduction in R is more important for other medical emergencies than it is for covid. The NHS desperately needs reduced pressure:

https://www.bbc.co.uk/news/health-59237935

Post edited at 13:35
4
OP wintertree 11 Nov 2021
In reply to Offwidth:

> Give me your odds on the next assessment of R for England as being below 0.8 to 1.0. When was the last time it wasn't close to 1.0?

I think we've been pretty clear for a long time that the "R" measurement both lags reality by a lot and is one of the least useful data outputs, for two different reasons.

Reason 1 - Wide Error Bars.  

The assumptions made in translating from concrete measures of "live" infection (ONS) or "new" detected infection (P1+P2) to an estimate of the R number introduce such wide errorbars as to make the results basically useless for understanding.

Reason 2 - We don't expect R<1 in the end

Stepping back, if we move to endemic Covid, we don't expect the endpoint to be R<1 - we expect it to be R≈1, with a time-variant pattern moving between periods of R>1 and R<1 due to bunching effects of the weather, school terms, major events and seasonality.

If you're expecting a true measure of R to be consistently less than 1 you're looking for a unicorn at this point.

What matters is that rising immunity levels in the population (from all sources) manifest in a decrease in the number of cases causing illness, which if achieved will manifest to a great degree in the reduction of hospitalisation and death rates, and to a smaller degree in the reduction of symptomatic case rates, and not at all in the rates of true infection.

As we settle down to an endemic situation, the absolute case rate is going to look like (total population) / (mean time between re-infection), wildly estimating this for England, it's going to look like ROM 20,000 to 100,000 true infections (not detected cases) / day on average, much less in summer, much more in winter.  But very few of these would hopefully meet the current gating criteria for going for a PCR test and almost none would be hospitalising people.

Currently...

Back in the grounded world...

  • What are PCR detected cases doing now?  Falling.
  • What are hospitalisations doing now? Falling.
  • What are deaths doing now?  Falling.

It remains to be seen if all these falling will endure, but it seems to me that as the situation evolves, so does our thinking about which measures actually matter.

Edit:

> Give me your odds on the next assessment of R for England as being below 0.8 to 1.0.

Surely 0.8 to 1.0 means falling....  So why are you asking for below that?  Not that I'm interested in the official estimates of the R number very much for reasons explained above.

Edit to your edit made after my reply:

> This is why a reduction in R is more important for other medical emergencies than it is for covid. The NHS desperately needs reduced pressure:

Yes, this situation has had much discussion lately and it’s good to see it finally as headline news coverage.

Reducing Covid’s contribution to the far wider pressures on healthcare doesn’t need R<1, it needs hospital admissions falling.  Which mercifully they are.  Forcing R<1 through more control measures is one way of reducing Covid’s contribution in the short term.  The data for Europe makes it clear IMO that that approach is not sustainable with delta and seasonality.

Off topic but the problems hammering the NHS are not going to go away I think even if covid did; certainly not going in to a winter flu season.  Really worrying times ahead.

Post edited at 14:00
 elsewhere 11 Nov 2021
In reply to Si dH:

I think you are probably right, we shouldn't vaccinate children for the benefit of others. Although avoid psychological damage of killing granny is a major benefit.

Anyway, other countries are coming to different conclusions to the JCVI about vaccinating children and that's not dictatorships with dubious ethics.

OP wintertree 11 Nov 2021
In reply to elsewhere:

> Anyway, other countries are coming to different conclusions to the JCVI about vaccinating children and that's not dictatorships with dubious ethics.

... although they are countries in very different situations to England, as the OWiD plots over the last few weeks are hopefully making very clear.

Different situations, different guidance.  

Post edited at 14:29
 Offwidth 11 Nov 2021
In reply to wintertree: 

I'm glad you deleted the bit on double vaccination given they don't own crystal balls when they decided in the summer and that some European countries had a higher vaccination rate than the UK at the time.

OP wintertree 11 Nov 2021
In reply to Offwidth:

>  > I'm glad you deleted the bit on double vaccination given they don't own crystal balls when they decided in the summer

I can't remember what I said now; I culled most of my reply to Elsewhere to avoid yet another torturous round of misunderstanding rather than because I didn't stand by all my points.

> and that some European countries had a higher vaccination rate than the UK at the time.

If you discount demographics.  Demographics matter.  They matter more than almost anything else.  Lots of orders-of-magnitude differences in real world outcomes derive from differences in the demographics that don't show in top level data.

The people making the decisions on single and double vaccination are not fools (about as far from it as possible), and they see the demographic data and they make decisions accordingly.  The demographic data on hospitalisation and fatality rates combined with the demographic data on vaccine uptake is one face of a crystal ball.

 elsewhere 11 Nov 2021
In reply to wintertree:

> ... although they are countries in very different situations to England, as the OWiD plots over the last few weeks are hopefully making very clear.

> Different situations, different guidance.  

Also situations and guidance from a different era in the ancient past. Months ago is many links in the chains of Covid infection ago and far removed from the current situation. 

The decisions I'm aware of were made on the basis of studies earlier in 2021 (spring/summer???, see FDA link dated 29 Oct) or circumstances in May (see Guardian link dated 28 May).

https://www.fda.gov/news-events/press-announcements/fda-authorizes-pfizer-b...

https://www.theguardian.com/world/2021/may/28/covid-german-children-12-16-p...

 elsewhere 11 Nov 2021
In reply to VSisjustascramble:

You may be projecting your set in stone views as you don't include "I think ..." like I did. I thought was enough to imply that I know that I am possibly wrong.

2
 Toerag 11 Nov 2021
In reply to wintertree:

Something to add to your weather-based predictions - the power usage from gridwatch. Quite a jump in the past 2 weeks as it's got cold which will almost certainly be reflected in people spending more time indoors.  We just need a 'rainwatch' and 'windwatch' equivalents and your data inputs will be complete.


OP wintertree 11 Nov 2021
In reply to thread:

Another day of faltering week-on-week decay in English PCR cases; will it cross over in to rise tomorrow, or will the wobble break back to more decay?  Your guess is as good as mine...

Plot D1.c is quite encouraging in terms of where most of the decay is falling and what that hopefully means for hospital admissions.  The last wobble where decay almost failed at the top level (entered around Oct 30) maintained decay in the most elderly;  we'll see how the next one shapes up in a few days.

Scotland is now in to its longest period of sustained growth in cases since August, and is having signifiant actual growth - low exponential rate but a high baseline.  Rate is getting worse, as well.  A quick look at Travelling Tabby suggests this is being driven by an explosion of cases in school aged children which we might reasonably expect to start driving cases in parents very soon.


OP wintertree 11 Nov 2021
In reply to Toerag:

Good thinking.  If I can get the data from the eccentric download page on gridwatch I'll chuck it in to the correlation-a-tron.

Something I keep meaning to note is that despite some skepticism from one poster a while back, the short term weather link is holding up remarkably well in the data.   

The other daft idea I had was to see how well the English temperature data correlates with the rate constant for cases in other nations.  The weather gradually decorrelates over physical distance, so the theory would be that the further away a country is from England, the less the correlation coefficient (R value) between their cases and the English weather, and the smaller the statistical significance of the finding vs the null hypothesis of different year's English temperature data (σ value).

One might almost call this the "distance from Westminster" test.

  • R = 0.59, σ=3.15 (English cases)
  • R = 0.41, σ=2.84 (Welsh cases) 
  • R = 0.39, σ=2.39 (Scottish cases)
  • R = 0.36, σ=3.00 (Northern Ireland cases)

I'm convinced.  It'll take more brainpower than I have left by this time of day to run some European nations through this analysis but I reckon the link will be obliterated by around Germany. 

Edit:  Re increased energy usage; the subject of rising gas demand for heating and electrical generation due to schools and workplaces following HSE & CIBSE guidance on ventilation is something I mentioned a while back.  Not a good time for someone to threaten to cut off major gas feeds in to Europe - https://www.bbc.co.uk/news/world-europe-59246899

Post edited at 18:29
 BusyLizzie 11 Nov 2021
In reply to Si dH:

> I am personally uncomfortable with the idea of vaccinating young kids just to reduce transmission. If they get vaccinated it should be for their own benefit.

But transmission by children contributes to the over-loading of the NHS. Children benefit directly from a functioning health service.

2
OP wintertree 11 Nov 2021
In reply to wintertree:

>  I'm convinced.  It'll take more brainpower than I have left by this time of day to run some European nations through this analysis but I reckon the link will be obliterated by around Germany. 

A good slice of cake rescued my brain power.  Here's a plot of the R-value for the passband correlation - bigger numbers mean the changes in temperature are more linearly correlated with the changes in the rate constant, smaller means less.

Distances are walking distances from Google maps, except for Portugal where I took 100 miles off because it arced around.

Not exactly robust science this for many, many reasons (e.g. not population weighted centroids for distance calculations,  no attempt has been made to control for other variables that might obscure a real correlation - although reporting noise is generally rejected from the passband used) , but I like it; the link carried on our past Germany so I was wrong on that; I suppose we all live under the same jet stream, so the wobbles of hotter and colder weather cycles over the timescale of a couple of weeks aren't that localised.

Post edited at 21:12

In reply to Offwidth:

>  some European countries had a higher vaccination rate than the UK at the time.

I'm not sure you can say this with any confidence. Those figures are really amongst the least reliable and least comparable of all the stats. Brace yourselves, this is another denominator rant.
As you know, the UK until recently reported against NIMS, and that this massively overestimates the number of unvaccinated people. Possibly by a factor 2 in some age groups. So we can be pretty sure the UK headline percentage was a lower bound.
We also know some of the countries with the "higher vaccination rates" are definitely wrong the other way. Just picking examples from the freely available ECDC numbers, Portugal's total includes 103% of 80+, 106% of 70-79 and 103% of 60-69s. Spain is similar. No country seems to know how many people they have to vaccinate.
Pinch of salt when you compare those figures. Or better, don't. 

Post edited at 22:06
 Dr.S at work 11 Nov 2021
In reply to Longsufferingropeholder:

i dont really understand how in countries with ID cards there is so much uncertainty about population numbers - how does this arise?

In reply to Dr.S at work:

Beats me. I never thought twice about the error on population estimates until this year. I always just assumed it was a solved problem, but it turns out it really isn't.

In reply to Offwidth:

> The NHS desperately needs reduced pressure:

Not sure who has opened the floodgates, but both BBC and ITV news are leading with crisis in the NHS. Whistleblowers being mentioned.

 Misha 11 Nov 2021
In reply to wintertree:

Steady state was a poor term for me to use. What I meant is the numbers have been up and down with a certain range. Not low but not maxed out either. As you say, gradually the pogos get smaller and the range reduces but it’s got a fair way to go yet. Cases now starting to bottom out at a similar low to previous troughs, though obviously would need a couple more weeks of data to be sure it’s not just a blip (would be nice if it was). As you say, weather is important and if we were going into the summer, no doubt we’d see more decay. That should help NZ get through their way, incidentally. 

In reply to TomD89:

> I wonder if this could be one of the first points people from all sides could firmly agree on? Anyone for forcing care/NHS workers out of jobs? If no public support and minimal to negative PH benefits, what exactly is driving this move?

I think the vaccine should be compulsory for NHS workers in contact with patients for several reasons.

a. there are already some compulsory vaccinations, Covid is a bigger threat in the UK than e.g. TB.  The basic principle that medical workers need to get some jags is established.

b. about 7 billion Covid jags have been given out worldwide.  If they were dangerous we would know.  I wouldn't have been in favour of compulsion 6 or 9 months ago when they were new but there's been more than enough doses given by now.

c. this is all in people's heads.  Once they are jagged and f all happens they'll realise there was nothing to worry about.  A lot of them will be feeling happy the morning after because they're less worried about being hospitalised or dying of Covid.

d. NHS staff should be setting an example.  If they are crazy enough to chuck their job rather than get a vaccine maybe the NHS is better off without them, the last thing we need is medical staff with weird ideas about vaccines passing them on to patients while they have the credibility of a uniform and badge.

1
 Misha 12 Nov 2021
In reply to Si dH:

> The time-integrated number of Delta cases since the end of July is significantly greater than prior to the end of July. See graph attached.

Fair point but most of that was presumably in children, with seep through into parents and grandparents but moderated by vaccines and prior infection.

2
 Misha 12 Nov 2021
In reply to Si dH:

Re vaccinating children. Other countries are rolling out programmes to vaccinate down to about 5 year olds. They’ve got the same data, presumably. As a dispassionate observer with no children, I think it makes total sense to vaccinate children.

Vaccines don’t tend to have long term effects. Real viruses do.

Edit - whether it’s better to send doses elsewhere is another question and a very good one.

Post edited at 01:22
4
 Misha 12 Nov 2021
In reply to TomD89:

> My personal dystopian nightmare fears are not driven by the fringe opinions about vaccines, masks etc but the feeling of growing centralized control and ulterior ideological motives shoehorned in under the guise of consensus science.

This is the most serious public health emergency for 100 years. It calls for a degree of centralised response, don’t you think? If you think lockdowns etc are ideologically motivated, you’re deluded. The Tories hate control measures but even they had no choice.

> I wonder if this could be one of the first points people from all sides could firmly agree on? Anyone for forcing care/NHS workers out of jobs?

I would do what Italy have done. Green pass (vaccine, paid for test in last 48h or recent infection) for all jobs, as well as most indoor venues. No pass, no job, no indoor socialising and so on. This actually has majority support in Italy from what I’ve read. 

2
In reply to tom_in_edinburgh:

I don’t think compulsory vaccinations for NHS and social care has been thought through at all.

Given that Covid doesn’t seem to be the main driver of the current NHS problem (instead it’s a widespread systemic demand and issues with moving people to social care) there’s a real chance that more people could die from delays than be prevented from dying from Covid.

Why put more pressure on the NHS at the moment? It’s just madness. Uptake is already high. What’s gained from pushing 90% uptake to 95% and losing 60k staff?

1
OP wintertree 12 Nov 2021
In reply to VSisjustascramble:

The NHS is in a crisis where staffing in both the NHS and the care sector are contributing factors.  

I’m all for vaccine mandates in public health jobs where it is in the interest of patients - once and only once the vaccines are out of conditional authorisation, as was the case with mandates in the US.

But, we are where we are and that is a precarious place with unsustainable pressure on the NHS.

> Why put more pressure on the NHS at the moment? It’s just madness. Uptake is already high. What’s gained from pushing 90% uptake to 95% and losing 60k staff?

I’m trying to hold off from extreme cynicism here; I would be interested in seeing a timeline of the legal advice the government have undoubtedly sought and received over this issue before jumping to conclusions.

Post edited at 08:17
 TomD89 12 Nov 2021
In reply to tom_in_edinburgh:

> d. NHS staff should be setting an example. 

..and if they don't they should be forced to set an example or forced out of employment despite their heroic efforts throughout the pandemic. Understood.

> If they are crazy enough to chuck their job rather than get a vaccine maybe the NHS is better off without them, the last thing we need is medical staff with weird ideas about vaccines passing them on to patients while they have the credibility of a uniform and badge.

Great, consider your 'complain about NHS on the brink of collapse' card revoked. Where are we sourcing all these non-crazy medical staff from to replace them by the way? It's not like there aren't a lot of very competitive vacancies already needing to be filled in other sectors, without such requirements, for probably better pay and far less stress.

Post edited at 08:18
4
 TomD89 12 Nov 2021
In reply to tom_in_edinburgh:

From what I'm reading Scotland and Wales have no plans to introduce a vaccine requirement for NHS staff. Huh, I guess your on the side of England and the Tories on this one? It's refreshing to see you in agreement for once.

1
In reply to VSisjustascramble:

> Why put more pressure on the NHS at the moment? It’s just madness. Uptake is already high. What’s gained from pushing 90% uptake to 95% and losing 60k staff?

Do you really think there are 60k patient facing staff in the US who would give up their job rather than have a jag?  That's about 4% of the total workforce.  We've got a real problem if 4% of the NHS workforce are that crazy.

In reality I think there's a very big difference between the number of staff who have not had a Covid jag and the number who would lose their job over it.  Probably a lot of the refusers have reasons for thinking they are OK like already catching it or being young and are just not bothered rather than hardcore anti-vax.

1
In reply to TomD89:

> From what I'm reading Scotland and Wales have no plans to introduce a vaccine requirement for NHS staff. Huh, I guess your on the side of England and the Tories on this one? It's refreshing to see you in agreement for once.

Most of the EU and the US have vaccine requirements, I'm on their side.

There are actually many things I disagree with the SNP on.  Their new requirement for interlinked fire alarms for example.  I've got no interest in forking out 200 quid before February.  I disagreed with them when they banned smoking in pubs and I wasn't sure about minimum alcohol pricing.

One of the reasons I respect Sturgeon so much is that in several cases where I disagreed with her she turned out to be correct including the alcohol pricing and smoking ban.  The more that happens the more I reckon she knows what she's doing and the more I'm willing to trust her judgement.

In contrast whenever I think the Tories are a corrupt shower of sh*t as a new story comes out it always turns out to be worse than I thought.

Post edited at 09:12
2
 Offwidth 12 Nov 2021
In reply to Misha:

I'd be OK with measures like Italy as long as it doesn't get too expensive.  Yet I'd prefer to put the onus on the vaccine hesitant.... if you really care so much about vaccination that you would risk your job, jumping through hoops and a bit of extra expense isn't going to be such a problem.

Instead of looking for a pragmatic solution the government have dumped tens of thousands of care workers we desperately needed, based on Daily Fail style opprobrium, and are busy picking fights with NHS staff groups in the middle of a major crisis. News seems to be front line NHS staff will need a first jab by the end of Feb.....can anyone see how they will fill the already yawning staffing gap since the threats didn't work so well in the care system?

2
In reply to tom_in_edinburgh:

This is the Government’s impact assessment on how many unvaccinated staff will be left:

https://assets.publishing.service.gov.uk/government/uploads/system/uploads/...

(Apologies that I’m linking to something from the Tory scum government)

I hope this is just a stick and come Feb 2022 the plans will be withdrawn.Mandating something rarely helps with hesitancy.

There are of course valid clinical arguments to mandating vaccines, but whether the risk around staffing outweighs them is something I’m really sceptical about.

 Offwidth 12 Nov 2021
In reply to Longsufferingropeholder:

I think we can be confident that vaccination rates in Spain and Portugal were better than the good UK numbers, just not as much better as the number gap appears. There has been plenty of research  on why there is less vaccine hesitancy in those two countries compared to most of western Europe: too much other evidence for it to be a data phantom. 

Such arguments are detail though, compared to the big issue. I thought we needed Plan B from the autumn: more because it was obvious the NHS was in a mess than because of covid on its own.... it's so bad now, we need crisis action. Roy Lilley is saying today the only way out of this catastrophe is smart thinking: yet I just can't see how that sort of flexibility can operate given the increasing government top down pressures on trusts. There will inevitably be 'blow back' on government if things unravel any more but by then it will be too late.

We desperately need stronger covid restrictions because it is a lever we can pull, to help now. We are currently looking daily at hundreds of stroke patients not able to get to hospital quickly despite being caught early (and missing the 3 hour anti coagulation medication window). Ditto for avoidable death and damage from delays in heart emergencies. I'm watching for a rise in excess deaths (another lagged UK OWiD plot)... .covid deaths by proxy .... we have been sitting at an average of 13% above the average baseline since the beginning of August, with no significant flu around. It's dystopian that we are killing and maiming people unnecessarily to save face on covid measures at the same time as threatening trusts about release of bad news.

6
 Offwidth 12 Nov 2021
In reply to VSisjustascramble:

That document is a tory fantasy that might make Nero blush given the current state of the NHS. Unlike the situation in Care Homes, they can't just sack NHS staff due to the strength of the unions and professional bodies. If they go ahead there has to be a fudge or there will be industrial chaos in a crisis. Any fix will just shift staff we desperately need on the front-line to somewhere else.

2
 TomD89 12 Nov 2021
In reply to tom_in_edinburgh:

> Most of the EU and the US have vaccine requirements, I'm on their side.

..and on the side of England and the Tories, just say it how it is.

> There are actually many things I disagree with the SNP on.  Their new requirement for interlinked fire alarms for example.  I've got no interest in forking out 200 quid before February.  I disagreed with them when they banned smoking in pubs and I wasn't sure about minimum alcohol pricing.

Being as you're so for removing personal choice to protect perceived public health interests, I can't see why you'd be strongly against the things above. Doesn't seem consistent. 

> One of the reasons I respect Sturgeon so much is that in several cases where I disagreed with her she turned out to be correct including the alcohol pricing and smoking ban.  The more that happens the more I reckon she knows what she's doing and the more I'm willing to trust her judgement.

So why don't you trust her judgement on not implementing a vaccine requirement for NHS/care staff? Or is she wrong on this one?

4
 Toerag 12 Nov 2021
In reply to Offwidth:

> We desperately need stronger covid restrictions because it is a lever we can pull, to help now. We are currently looking daily at hundreds of stroke patients not able to get to hospital quickly despite being caught early (and missing the 3 hour anti coagulation medication window). Ditto for avoidable death and damage from delays in heart emergencies. I'm watching for a rise in excess deaths (another lagged UK OWiD plot)... .covid deaths by proxy .... we have been sitting at an average of 13% above the average baseline since the beginning of August, with no significant flu around. It's dystopian that we are killing and maiming people unnecessarily to save face on covid measures at the same time as threatening trusts about release of bad news.

The stats are very telling - https://www.bbc.co.uk/news/health-59237935

If people knew they were looking at an average 50 minute response time for things like strokes and heart attacks now I bet they'd wear masks. I also think that anyone claiming they can't wear a mask should wear a dunce's cap.

3
In reply to Toerag:

Would that just push the issue back to next winter though?

It’s very clear that the NHS is close to crisis point, but not clear at all that Covid is directly causing it.

If we make a concerted effort to reduce the spread of seasonal respiratory viruses would we just see the mother of all flu seasons + Covid next winter? Causing another crisis in the NHS? And then if we put measures in again next year, would that just push it back another year?

It doesn’t seem like there’s an easy way out of this. Covid restrictions have health consequences (good and bad). One option seems to be to tough it out this winter, the other seems to be to repeat the current shot show ad infinitum.

The obvious thing to do might be to reduce the burden on hospitals by forcing care homes to take on patients, but that wasn’t a very popular decision the first time around.

5
In reply to Toerag:

> If people knew they were looking at an average 50 minute response time for things like strokes and heart attacks now I bet they'd wear masks

You think? They already know that wearing a mask might help save others from a horrible death (unless they're deniers, of course). Why do you think they would care about others dying from stroke or heart attack? I mean, covid, strokes and heart attacks only happen to other people, surely? Not to the healthy mask refusers...

2
 FactorXXX 12 Nov 2021
In reply to tom_in_edinburgh:

> One of the reasons I respect Sturgeon so much is that in several cases where I disagreed with her she turned out to be correct including the alcohol pricing and smoking ban.  The more that happens the more I reckon she knows what she's doing and the more I'm willing to trust her judgement.

Sturgeon is always right?

1
In reply to VSisjustascramble:

> The obvious thing to do might be to reduce the burden on hospitals by forcing care homes to take on patients, but that wasn’t a very popular decision the first time around.

If by 'unpopular', you mean 'murderous', then yes. But that was because they were discharging patients without testing for covid.

NHS is in crisis for a number of reasons: covid occupancy, covid-delayed backlog, post-brexit staffing (NHS  and care), post-covid staffing, etc.

I saw a stat about 999 calls being up something like sevenfold. That seemed unbelievable until it dawned on me that I suspect many are repeats, as people desperately try to summon a delayed ambulance.

In reply to FactorXXX:

> Sturgeon is always right?

No.  But on the occasions where we have disagreed she's been right at least as often as me.  So when we disagree I don't automatically assume she's the one who is wrong.

3
In reply to VSisjustascramble:

> This is the Government’s impact assessment on how many unvaccinated staff will be left:

Looking at their table the big issue isn't the NHS it is social care.

My guess is the pay is sh*t and turnover is high so there's not much leverage in the 'get jagged or lose your job' threat.

When people are skint and you want them to do something then you give them money and the skinter they are the less money it takes to bribe them.

So have say £200 as a Christmas bonus but only for staff who have been jagged. 

Post edited at 23:13
In reply to TomD89:

> So why don't you trust her judgement on not implementing a vaccine requirement for NHS/care staff? Or is she wrong on this one?

After looking at the table

I'm not sure she is wrong. 

My first post was specifically about patient facing NHS staff.  I think I was basically correct in saying a vaccine requirement made sense.  However, when you look at the numbers the vaccination numbers are already really good in the NHS.

I didn't comment on social care at all in my first post.  When you look at the table, that's where the main non-vaccination issue is.  It's also a low paid job with a lot of turnover which means 'no jag, no job' isn't much of a lever when there are staff shortages in retail and hospitality.  

The better approach would be to provide a financial incentive to get jagged.

 FactorXXX 13 Nov 2021
In reply to tom_in_edinburgh:

> No.  But on the occasions where we have disagreed she's been right at least as often as me.  So when we disagree I don't automatically assume she's the one who is wrong.

Sounds like you put a lot of faith and trust in 'The Leader' of the SNP...

Post edited at 01:51
1
 neilh 13 Nov 2021
In reply to VSisjustascramble:

Clearly you have never been in a care home from that daft comment. 

you also completely fail to grasp that a lot   of patients are discharged back to their homes with social care support at home. . 

Post edited at 07:39
OP wintertree 13 Nov 2021
In reply to Misha:

> Steady state was a poor term for me to use. What I meant is the numbers have been up and down with a certain range. Not low but not maxed out either.

The big question for me; is that range incidental or does it have deeper meaning?  It's not so far from the level we might expect once Covid is truly endemic, but it's very hard to judge with the different gating methods of PCR testing.

> As you say, gradually the pogos get smaller and the range reduces

I'd say the pogos are becoming more negative, but the range of a pogo is relatively consistent. 

> but it’s got a fair way to go yet.

Does it?  For true infections that depends strongly on what endemic Covid is going to look like; the key data for that is a demographic "mean time between reinfections" dataset.  What we do need to see is increased distance between infection and symptoms, particularly severe symptoms.  This will manifest to some degree in testing data and hopefully a larger degree in admissions data - conflated by the anti virals coming on line.

> Cases now starting to bottom out at a similar low to previous troughs

For England, it's been the biggest sustained fall in 3 months which is quite something; I think we're now seeing another pogo - this one a weather wobble as the cold weather around Nov 5th raises the rate constant.  Yesterday evening's update had England in mild week-on-week growth; looking at the provisional data there's another such day in the pipeline, then hopefully it's back to decay - if so, this pogo is one in a series of four descending wobbles, but the big picture is a trend to fall.  I'm getting ahead of myself though and we need the data to see if this is right!

> As you say, weather is important and if we were going into the summer, no doubt we’d see more decay.

To be repaid as growth in Winter, as the OWiD plots are showing

For us, there's no doubt colder and grottier weather is yet to come; if we don't see the baseline drop a fair bit more (about another -0.025/day?), the weather wobbles will keep bringing us to growth.   

> That should help NZ get through their [wave?], incidentally. 

Yes, they have picked about the best time to start dropping control measures and allowing the virus to spread.  I hope they can make the most out of their summer - in every way.

Back on Wednesday I said I hoped your observation on hospital admissions was pessimism; pessimism or not they are levelling off for England more clearly now; no obvious reason for that - especially with the demographic structure of decay and rise recently - so hopefully it's back to decay soon and this is just noise...  Maybe....



New Topic
This topic has been archived, and won't accept reply postings.
Loading Notifications...