UKC

Post 1 - Four Nations

I’ve changed the data range on a bunch of plots to focus on more recent times - starting around the football peak in the summer of 2021 and onwards from there.

The cases data is taking off for England and Scotland. and that’s visible in the provisional doubling times on the far right of plot 9.  The growth doesn’t look as dramatic on these log-y axis plots as on the government dashboard, but dramatic it is.  Not disastrous just yet, as later posts will look at.  Northern Ireland and Wales are further behind it seems.

In a recurring theme this week, there’s more growth in the provisional zone of dashboard data that these plots don’t use.  By the time that falls through to the plots, we’ll set an all-time high for detected cases.

Links to previous thread:
1. https://www.ukclimbing.com/forums/off_belay/friday_night_covid_plotting_56-...
2. https://www.ukclimbing.com/forums/off_belay/covid_plotting_56_continued-742...

Post 2 - London

We’re starting with a look at plots for London today, as that’s ahead of everywhere else in terms of the growth of the new variant.

Raw cases

• Absolute case numbers for 3 rather arbitrary age ranges I’ve been using in recent weeks.  All 3 bands are growing, but it’s clear that 65+ is currently the least affected by the new variant, and that younger adults are leading the way 

Plot D1.c

• Technical note -  I’ve used the week-on-week method this week (not exponential fitting) which is cleaner and more comparable with plot 18, but has a lot more noise in the oldest ages.  To mitigate this, ages over 60 have been through a 7-day moving average first, and the final heat map has a 3-day moving average applied.  The line plots show data without the 3-day moving average for more immediacy.
• Heat maps - the left plot shows cases really concentrating around ages 25-30 (white and gold blob on the right side), with correspondingly dark orange bits for high growth rate constants on the right side plot around ages 20-40. 
• There’s a secondary bold orange growth blob emerging around ages 45-65 on the right side plot, although all ages are now seeing some growth.
• Time slices - these show the rate constants and doubling times for the most recent day of data, and for the day two weeks before.  So, for now, the “good” news is that whilst cases are doubling around every 3.5 days in adults aged 25-30, they’re doubling a lot more slowly in older folks.  But, if we look at how much the rate constant has increased by the gap between the blue and red curves, it’s increasing pretty rapidly in ages from 40 to 70 as well.  At the moment it maxes out just short of a week for doubling time, but if it keeps getting worse, that’s going to be bad news for hospitalisations; one way or another.  Well, even if it doesn't get worse, if it's allowed to play out as increasing cases it becomes bad news soon enough.
• Time-series - this shows the time evolution of the rate constant for the most rapidly growing bins.  We’re at the point the behaviour of omicron should dominate these measurements, and they’re not at the sub 2-day doubling times that’ve been mooted by some documents submitted to SAGE for the variant.  My best guess is that the most aggressively growing band is going to max out at about 2.5 days and then back off as the increased control measures and precautionary behaviour feed through to the data.
  • Only 2.5 days!  It's like boiling a lobster...

Post 3 - England 1

Plot 18 - Regional Rate Constants

• Technical note - I’ve re-worked this to use a week-on-week method for measuring the rate constants to be more consistent with some other plots.  This is much less bias prone to day-of-week effects and less twitchy to noise at the leading edge than exponential fitting to the data.  Data can be pretty noisy down at the regional level, so a moving average is applied to the results (3 days for cases, 5 for hospitalisations, 9 for deaths).  There’s always been more filtering on the plots lower down, which trades immediacy in cases for slightly less responsive clarity on the bottom ones.   I’d like to move the plot 9 curves to this method as well for better equivalence of what’s being measured…
• I’ve also had to expand the y-axis range of the top right plot…. But not the others.  Yet…?  Or ever…?  Well, that’s the big question.
• The doubling time continues to rise for London ahead of all other regions and is pretty bananas.
• The North East and South West continue to be furthest behind London.  
• Provisional data on the dashboard makes it clear most regions are going to see more rapid growth land in the next few days; coming as it does later on vs London - particularly after the latest policy tightenings, increased messaging and surging 3rd dose rates, hopefully we’ll see more moderate growth rates in these areas than in London.
• Hospitalisations in London are increasing, but they are doubling much more slowly than cases; at the national level we’ve seen cases continue to increase for two months with hospitalisations at times decreasing and at times doubling much more slowly than cases - in part rising immunity in the population and in part the demographic shift to younger ages.   As the London post shows, cases are growing most in the youngest so we expect hospitalisations to rise more slowly as long as that holds, although this is going to be the most sustained rise in hospital admissions since delta broke though.
• Deaths in most regions appear to be moving to faster decay; perhaps related to the rate vaccines have been landing at in older adults whilst their cases were predominantly omicron.  We wait to see what happens post-omicron.

Growth in Plot D1.c is mostly being driven by growth in London I think.  That will all change soon enough.

Post 4 - England 2

Top level cases are rising, mostly reflecting the growth in London.  Provisional data on the dashboard (excluded from these plots) makes it clear growth is going to hit other regions soon, so the top level hasn’t found it’s peak growth rate yet.

Hospitalisations have been pretty constant, but occupancy is rising.  I presume this is related to the demographic shifts leading up to omicron emerging.

The rate at which deaths are falling seems to be speeding up.   It’s notable that this apparent decay is happening despite hospital occupancy rising, suggesting to me that going to hospital is a critical step to not dying.  Not exactly an Earth shattering revelation, but it does underline one of the many reasons this all goes south really quickly if we let healthcare get slammed.

Reflecting on all the plots in posts 2, 3 and 4:

• Right now, the demographics of the growth are a lot less panic inducing than the quotes of sub 2-day doubling times for Omicron.  To me mind, the key things to look at now are how the doubling times evolve in ages 40-65 and 65+, and how hospitalisation signals develop over the next week - particularly in London.
• But, slow the rate constants down as much as we like, If they remain positive, all measures grow, and we could  reach a point that the hospital admissions levels are completely unsustainable.  
  • There seems to be a developing schism between scientists in the official pipeline and ministers...  On the bright side, the public have been round this loop a few times and are getting better at reading between some lines perhaps, certainly a lot of news stories about cancellations etc.

Edit: Forgot the plots...

Post 5 - Scotland

Cases - rising for about ten days now.   I do wonder how much that Steps concert kicked all this off ahead of the other English regions and other home nations beyond London.  One way to find some fame I suppose.

Admissions continue to fall - I think the trendline is led astray a bit by the noise in the data.  The fall in occupancy is levelling off, somewhat akin to the situation in England with occupancy vs admissions.  Deaths continue to fall, although again perhaps the trendline is over-enthusiastic at the leading edge..

Plot 9s is quite something - the difference in cases (growing, lots) and the other measures (decaying, lots) is opening up.  Looks like the 3rd dose would have slammed the doors down on delta.  Interesting that hospital admissions aren’t rising yet, probably a lot of conflating factors but the demographic data for Scotland is too coarse really.

Post 6 - International Plots

I’ve only done the phase space plots this week; so much is going to change with omicron that the basis for extrapolating death rates feels unjustifiable right now.

Over the last couple of weeks this plot was looking quite swirly, as various countries went round different bits of their orbits driven by going in and out of lockdown to control delta in the absence of sufficient immunity.

 Now it’s starting to look like everything is making a move for the upper-right quadrant - high cases/million and cases doubling quickly.

On first look, Denmark stands out, but they have a similarly very low CFR to the UK suggesting (like for us) better testing and/or younger demographic than most other places.  Looking at the “sanity checking” plot and at the sudden tails on the phase space plot; France, Italy, United Kingdom, Sweden, Spain, and Denmark are all seeing upticks that are likely omicron.

South Africa is showing a near stagnation of cases; don’t know that I believe it.  Their positivity is through the roof, so data must be suspect.

What's set the UK apart from other nations on these plots over the last 2 months is the very good level of vaccine uptake in older people here and the quantity of post-vaccination spread of delta in younger adults - vaccinated and unvaccinated.  That mans we have a stand-out low number of people with no prior immunity, and we have a lot more people with some immunity against viral proteins other than the spike used by the vaccines.  Seems to me like that's all going to confer advantages in terms of moderating severity of a variant that escapes immunity-against-infection like Omicron. Going to be a lot of tightening of control measures ahead I suspect.  

The EMA were showing unusual signs of haste over the Pfizer anti-viral; if that works as well in the trial world as trial data suggests, it's as close to a Deus-Ex as anyone could reasonably hope for.  Some interesting opinions in the news split over the UK ordering too little of it and over the risk of using too-much potentially driving anti-viral mutations.  I'm hoping for more news on all this soon as it's got big possibilities it seems.

Guess the lake.

Bolam?

Phew!

Thanks. About 9C here in Brum so impressed  there’s ice up where you are but it’s a long way North!

‘Latest’ SAGE minutes out. Meeting on 8 Dec, which feels like a lifetime ago. Why can’t they publish them within a couple of days…

https://www.gov.uk/government/publications/spi-m-o-consensus-statement-on-c...

As Misha says, your weather is very different to ours up on Skye at the moment. 10 degrees and dry is almost like a summer without midges. 

Similar size and same side of the pennies…

In reply to jonny taylor:

When you get the right conditions in that part of the world it’s something else. Above the fog it was wall-to-wall blue skies here, no wind and just above freezing.  As close to perfect as possible.  Fog looks gunkier today, if I had a little sounding balloon with a camera I’d send it up to get an idea which way we should go to escape it…

Shaping up for a white Christmas which is very exciting.

So for all the political self inflicted stupid wobbles at Downing Street we might just get through this ok. 

Might even be as good a 3.0 days?  Blessed be.

No worse than some university outbreaks then.  Which makes me wonder how they’re going to avoid a tidal wave of students having to isolate when they return next term.  Halls of residence are an awful place to have to self isolate in their tiny dorm rooms.

In reply to neilh:

Anything can happen in the next half hour.  We’ve got a chance I think…

i'll just drop this in: 

https://www.theguardian.com/theobserver/commentisfree/2021/dec/19/exponenti...

"David Spiegelhalter is chair of the Winton Centre for Risk and Evidence Communication at Cambridge."

I think he needs to work on his communication skills. That was a terrible explanation of exponential growth and how to estimate doubling time.

I know a few people are looking more to the ONS survey as we expect PCR testing to saturate.

I've plotted the ONS estimates for the percentage of the population with Covid and the daily new case count. These aren't the same measure, the ONS survey will detect a live infection some time (weeks) after a new infection.  The cases data detects some fraction of new infections a few days after infection occurs.

The striking thing is that the ratio between live infections (ONS, red) and new infections detected as cases (dashboard data, black) has halved since the summer low; how much is due to a change in cases remaining live for longer (change of variants) and how much is down to changing attitudes and demographic shifts I'd not like to proclaim.

It does suggest that we're probably close to 3.5% of the population having live Covid infections by today (given lags in all the data, especially the ONS data which obviously lags a bit as it has to be acquired, processed, QCd and released on a cadence).  

• Which is interesting because the demographic cases data suggests infections are strongly concentrated in to younger adults and in to the south east in particular. 
• Which is really interesting because that suggests perhaps those limited bins could be running incredibly hot (circa 40% with Covid in ages 25-30 in London???) - slamming through to locking in a recovered/immune status within days to a week maximum.  Which then goes to remove pressure driving infections in older people, just as we saw with the school ages over the last couple of months.

Those tentative conclusions over London seem absolutely bonkers.  If they're right, there's going to be a. collapse in new daily cases by this time next week for ages 25-30 in London.

• Remember, people are infectious for some time after they become new infections, and things are changing faster than the mean lifetime of infections at the moment, so even if cases are plummeting in that cohort, they still represent an incredible transmission risk for another week or two...

Edit: To give an idea of my estimate:

New daily cases are running ~11x higher in the 25-30 bracket in London than at the top level for England

• Ages 25-30 new daily case rate is ~1700 cases/100,000 people/day (dashboard data)
• England top level new daily case rate is ~150 cases/100,000 people/day (dashboard)

New daily cases at the England (including those likely in provisional data) have roughly doubled on the plot from 40k / day to 80k / day in the last week or so, and the ONS data hasn't caught up yet...  

• 40k / day was corresponding to 1.75% of people with live infections, so 80k/day is about 3.5%
• 11x hotter cases in the specific London band would correspond to 11 x 3.5% ≈ 40% with live infections

A lot of estimates, suppositions and other shonky methodology lumped in there, but it does suggest to me that - assuming recovery from omicron confers immunity against catching omicron again in the short term - things are moving very quickly, and - topically to Šljiva's link above - exponential growth in that band can't go on indefinitely...  Or Indeed for much longer at all...    Plenty of other ages where the process can't be allowed to happen this quickly due to elevated risk to healthcare.  Thankfully the rise is slower in those ages, and if the growth in young adults does collapse down shortly, that'l remove one source of "forcing" for growth in older bands.

They should have got Alan AtKisson to pen the piece -  youtube.com/watch?v=bghbxemp4kQ&

The article mentions the infection survey. The next one should be revealing. I think they’re out each Friday and take into account infections up to the previous Sunday. The last one didn’t show that much change. Not surprising as absolute numbers of Omicron were still relatively low and the growth will have been diluted by the c. 1m ongoing infections. The next one should show a fair bit of change though, given the strong growth this week. If it doesn’t, something won’t be adding up…

Here is some anecdotal information about the current state of Covid in London.

I work for a small company that employs 8 full time staff and use lots of freelancers.

Last week 3 of my colleagues had to self isolate because of close family members testing positve. Also a freelancer had to be replaced on a job last minute after testing positive.

Then on thursday this week my wifes Booster side effects actually proved to be an infection!

Im now hiding out at home waitting for the results of a PCR ( currently showing no symptoms and negative on LFT)

I am due to fly out for work on the 27th of this month, so I phoned a couple of freelances in case I need replacing, the first 2 I called are positive with Covid!

Ive been in London for the entire pandemic, This level of infections in people I know. Is far worse than anything ive seen before.

Everyone has Covid!  Almost literally for under 40s in London, perhaps.

If I had my way, the only person crossing the London Orbital in the next week would be Snake Plissken.

Let’s hope those younger people do a lateral flow test before they go to see their parents / grandparents for Xmas… I think most people will be sensible and not visit if I’ll / symptomatic. However asymptomatic and presymptomatic spread is going to be the real issue (and has been all along, I suspect). As you say, it can’t keep doubling at a crazy rate within a given demographic. But it can spill over…

Edit - I’ve heard of a few people in our London office (out of around 100 in my immediate team) off with Covid but certainly nowhere near 40%. Might have been that if the Xmas party didn’t get cancelled!

So easy! (I won't say which lake.... I don't want to spoil it for others)

Just sticking my head above the lurkers parapet to thank Wintertree and to add that 3 Names experience mirrors my own; London feels insane at the moment. The uncertainty and anxiety is palpable although perhaps magnified by peoples (diminishing) hopes of travelling / gathering for Christmas. Anecdata - I'm sure I'm projecting my own insecurities here but I'm equally sure I'm not alone.

I thought he was dead.......

Not dead, but I thought he’d be taller.

The wave of isolation orders over Christmas is going to be horrid I think; and of dubious necessity for younger adults given the prevalence; allowing meeting other younger adults at private abodes, and allowing outside exercise feels more proportionate.  Different rules for different people though, divisive stuff; don’t want everyone suddenly thinking they’re invited to the party at No 10…

If it follows the normal pattern we can expect infected Royals to start arriving at Balmoral.

I'm thinking that what really, really needs to happen is restrictions to prevent the mass gatherings on New Year's Eve in London, and associated crush on trains and tube. Pretty ghastly to contemplate, but it will certainly happen if not made illegal.

Daughter lives in London, supposed to meet her this weekend and then back home for Xmas. She tested positive Saturday. She reckons nearly all her mates have had it. The thought of leaving her in London for ten days, and the total mess that it creates for our other plans, is a real pain in the arse. Not the first time it’s screwed our plans. Weighing everything up (who she lives with etc), the lest worst option is to bring her back up north. But that would be breaking the law. 

In the absence of restrictions in the meantime, I'm not sure that'll make much difference. Cases are so high in the age groups most likely to be out and about, and increasing so quickly that it may well have peaked and be on its way out by then

For example in Lambeth 4% of all 25-29 year olds have tested positive in the week up to the 14th December, half of that growth was in the last 2 days of that period. 

By new year, I think for the cohort that will be hitting the town hard, this might mostly be in their rear view mirrors by then.  Be interesting to see what Paris does for New Year's Eve.  The one I spent there was a night I'll never forget...  One worry here is the non-Covid workload NYE creates for A&Es...

In reply to mik82:

At this rate I'll be surprised if it's not peaked before Christmas Eve in that age band...

In reply to Bottom Clinger:

That sucks; I think there's going to be a lot of that going about at the moment.  

youtube.com/watch?v=84ebGxPONXU&

In reply to thread:

The doubling time for London has stopped rising by the week-on-week method

• Perhaps we’ve hit the buffers in terms of growth rate...
• Or, it could be provisionality, or it could be running in to capacity limits on testing
  • For this weeks update I already took another data point off the end of these plots as the leading edge was getting more provisional…

The demographic data stops a day sooner and has some hints about what might be behind the levelling off of the all-ages doubling time for London...

• The doubling time for the fastest growing age band (25-29) looks to be levelling off at between 3 and 4 days; this doesn’t look like a saturation in testing as it’s levelling off faster than the other ages on the plot.  So, I think this probably tallies with my noddy take on the ONS data and London demographic data up-thread - infection has spread so much I this age band that it’s now self-limiting and decay isn’t far away.

Beyond the young age bands, the plot D1.c update doesn’t look like great news.

• There’s a band of dark orange down the far right - more positive rate constants equivalent to shorter doubling times - at all ages. 
• The time-slice plot shows this in more detail; cases In particular, ages 40-65 are doubling every 6 days or so now, with those doubling times still getting shorter.
  • This isn't great as it's where there’s an overlap between people who are likely to go to hospital if they are unvaccinated and catch Covid, and people who have not engaged with vaccination.  London Mayor Sadiq Khan was commenting on this yesterday [1].
  • I don't see the doubling times coming much below 4 days, as there's a lot more 3rd doses fully bedded in in these ages now than there was in the younger ones a week ago (things are moving quickly on the vaccine front - next post) and as people are more precautionary.  Still, slamming through to immunity thresholds in this age group is a lot more risky than in the younger adults, especially with some very low immunity and zero vaccination people still about.  

[1] https://www.bbc.co.uk/news/uk-england-london-59710649

Ran out of time to do these with the main posts at the start of the thread.

Vaccine 3rd doses day-of-week plot (UK)

• Just astounding really, over 900,000 third doses yesterday. 
  • Come on 1,000,000....
• Not shown here, but the first dose rate is rising a bit as well, up perhaps 50% over the last week or so.  Every little helps.  Time is rapidly running out for people to get their first antibodies through vaccination.

Vaccine 3rd dose progress plot (England)

• The left plot shows the total number of 3rd doses given for each age bracket in England, and compares it to the number of people who received a 2nd dose at least 90 days before, as a proxy for who is eligible for a 3rd dose.
• The middle plot shows that 3rd dose number as a percentage of the eligible proxy. 
  • This over-estimates the size of the eligible cohort a bit as some people have died;  the eligible number is over-estimated by on the order of a few % by the far right of this plot. 
  • The black trails show the change in this measure over the last two weeks (using today’s eligible number, not that of two weeks ago)
• The right plot shows how long it would take in days to reach 100% if the average dosing rate over the last two weeks continues.   
  • The most recent week is faster, so if that is sustained these times may come down faster than 1 day/day.
  • I’d expect the time estimates for younger adults to drop as they’ve only just started being able to book, let alone receive, third doses so the new dosing rate isn’t fully embodied in the data used to measure vaccination rate. 
  • Some amount of the remaining time on the oldest ages will be down to the over-estimates in the eligible criteria, and some may be down to the difficulties of getting vaccination to some in care homes.  In general the last bits of each band will likely be slower.

Just astounding progress, it really is.  More-or-less completing third doses by min-January seems likely; and I don't think we've seen an end to the rising dose rates just yet.

 “Be interesting to see what Paris does for New Year's Eve.  The one I spent there was a night I'll never forget...  One worry here is the non-Covid workload NYE creates for A&Es..”

Official Paris is cancelled, I’m guessing official London will be by the end of the week (ish) 

My sister took my niece in to get her booster, and then stayed on to help inject, as they were a bit shorthanded that day; she wasn't supposed to be on shift...

If they still aren't reporting a second positive test as a case then the official case numbers are going to undercount Omicron cases to the point there's hardly any point in calculating R numbers or exponential constants based on them.

Peston brought this up a few months ago and is was argued it wasn't that relevant because reinfections were low but Omicron changes that.

https://fullfact.org/health/robert-peston-reinfections/

https://coronavirus.data.gov.uk/details/about-data#daily-and-cumulative-num...

"COVID-19 cases are identified by taking specimens from people and testing them for the presence of the SARS-CoV-2 virus. If the test is positive (except for rapid lateral flow tests which have negative confirmatory lab-based polymerase chain reaction (PCR) tests taken within 72 hours), this is referred to as a case. If a person has more than one positive test, they are only counted as one case for all nations with the exception of Wales."

Plus last weeks news was there was no PCR availability in parts of the UK due to hgh demand.

https://www.theguardian.com/world/2021/dec/14/uk-public-hit-by-second-day-o...

Yes, discussed a few times in recent weeks.

Abouf 16% of people have been a detected case in the UK so at most it reduces cases by 16%.   Or a day’s growth at current rates.

Having thought it through, it’s not bothering me.  I’ve never paid to much heed to the absolute value of cases as they’re most use as a barometer of change, and all sorts of biassing factors drop out when they’re used to measure change.

Just another reason why it’s a fool’s errand to compare case rates internationally.

PCR testing is clearly going to struggle to scale to this.  It was interesting to read Alan McNallys thoughts on Twitter - a suggestion that LFTs run by trained staff would be better; I can see the attraction in a data source with a a shorter delay despite the higher false negative rate.

It’s on his Twitter feed.

All going well we’ll have to week ourselves off the PCR testing at some point soon enough (the end of cases data as we know it!); perhaps some of the resource from that can go to scaling up the ONS random sampling and moving it to a twice-weekly cadence.  

It's a bit scary though, as the best analysis relies on PCR data and we are 'going dark' at exactly the wrong time.

One of the pandemics unsung heros, Chris Hopson, Chief Exec of NHS providers (I say this as he seemingly just ignores the government bad news bans but never exaggerates) is pointing out on the news this am that staff off sick in London doubled from 1900 at the beginning of last week to 4000 by last Thursday and numbers are still growing.

Off sick or off with isolation orders due to a positive result?

Hopefully the worst of the isolation orders lands up front with the younger adults, and as they return to work the rate of infection slows as it shifts in to older adults with more 3rd doses and more precaution; with the bulk of hospitalisations coming as spread moves to older adults the timing should work out…

Absolutely insane way to run a ship so close to the wire.  

Not sure what else we could do except allow positive but asymptomatic people back to work, as happens with all the other circulating respiratory viruses.  Given the concentration of at-risk people in healthcare it should probably go the other way - screening and “don’t come to work” orders on pre existing respiratory viruses.  I suspect that would not prove viable for staffing…. Not sure how all this gets squared off in the end; a much less immediate problem than the next few weeks.

Sorry; I didn’t really spell this out.  It’s good to spell it out sometimes.

I think this makes bugger all difference to the rate constants.  

To a first order, in the face of a dominant variant that evades prior immunity against infection,  it’s a multiplicative effect such that reported cases are fractionally smaller than detected cases.  Have you jotted it all down on paper with some maths and some variables for the conversion ratios?  Infections > detected cases > reported cases, and then the maths of a rate constant fit?  See which variables drop out and grok why?  The one you hilight here is a much smaller effect than the conversion ratio of infections to cases, and we’ve lived with that.  That ratio changes over time, but it seems to do so more slowly than changes in the spread of infection which is part of why I did the passband stuff, although I appreciate that’s a bit esoteric for a climbing forum, and not in many people’s skill sets to follow.

Multiplicative factors drop right out of rate constant measurements.  The rate constants remain what they always have been - a scale free way of understanding how much the virus is spreading or decaying, and one that is equally applicable to hospitalisations and deaths.

So, pray tell, why do you think this means there’s hardly any point in calculating the rate constants?  

They are what they are and as far as I can tell, they’re as useful as they’ve ever been.  I’m far more concerned about saturation of PCR testing; this does *not* drop out of the measurements of rate constants or indeed of anything else.

How do you work that out?

16% of people is 10.56 million.  The jobs/lifestyle factors which were correlated with catching it the first time are probably also correlated with catching it again when there's a variant which avoids immunity.

We are still only on about 80k cases reported a day.

You could have another 80k cases among people who have already had it and not shift the daily case number at all.  Testing capacity running out would have the same effect.

If a lot of the new cases are not counted it makes the growth look lower than it is.

The latest SAGE minutes state "...much higher levels of reinfections are being seen with Omicron (8% to 9%) than Delta (around 1%)..."

So it's significant, but not enough to change the big picture. It's certainly not another 80k cases. There are much bigger missing numbers due to people who are asymptomatic or don't get tested for other reasons.

As WT, rate constants aren't affected by it - except where comparing how the rate constant is affected by the change in variant. My feeling is (and confirmed by SAGE who know better) that we will currently be missing a lot of infections. Hence why the next couple of ONS surveys will also be interesting.

https://www.gov.uk/government/publications/sage-99-minutes-coronavirus-covi...

But if one variant is rising whilst another is falling, and the new variant causes more reinfection, then the multiplicative factors are not constant; we hide the growth of the new variant. Like what happened with delta.

See comments from myself and mik82 up thread.   Around half the people in ages 25-29 have caught omicron most likely.  At that point your worst case assumption that re-infection risk front loads in new cases can be set aside for reasons I hope are obvious (50% > 16% and all that).  In other ages where growth is slower it could have more of an effect, but…

Even considering your dubious - but worst case - assumption that all reinfections are front loaded in this wave, if you look at the ONS reinfection survey data on Ct values and symptoms it’s clear that pre omicron, reinfections were much less likely to be symptomatic and so to become a detected case.  We’ll see how that holds up against omicron, but the immunology side gives some hints…

Yes.  Also, the sky is blue and rain is wet.

It would be good to get some data giving an indication of the scale of this bias, but I maintain it more or less drops out of the rate constants - you’ve been keen to dismiss those in favour of cases before but to my mind cases have far more issues as a basis for comparison and for understanding, many of which drop out of rate constants.

You haven’t explained why you think this issue makes the rate constant measurements pointless however.  Can you explain why you think this?  As explained the you mention “drops out” of the rate contestants.

But reinfections (let alone reinfections by being detected twice as a case) are a minor effect unless you assume reinfections are all front loaded which doesn’t fit with other data.  Even ignoring reinfections being less severe and so less likely to be detected as cases, only around a half to a quarter would be detected as a case this time around given likely infection > case detection ratios.  Chain it all together and we’re getting down to a few % bias per day in case numbers.

Did it?  Falling numbers of one variant absolutely masked the rising of the other, but that’s very different to claiming differences in reinfection rates contributed to the masking - I’ve seen no analysis suggesting that for delta?

Edit: also, the transition period from delta to omicron is going to be pretty brief…. Reinfections are a really thorny issue going forwards with community testing in the data as we move towards endemic status.  I think a wider, more frequently updated random sampling study that collects Ct and symptoms data could be much more useful, along with hospital admissions data including if it’s incidental or causal.

I'm happy to accept that it's a small bias. But the multiplicative factor is not constant as you argued.

The rate of growth of delta was masked by the fall in alpha. So once delta became dominant, the rate constant increased, to its 'true' value. Slightly different point, but related.

I didn't argue it was constant, I prefaced my comment with the opening words "To a first order" with all that that implies.

Yes - there are different rate constants for each bin in a granular analysis - age, location, variant, sex etc - and opposing changes in different bins often hide from the top level data, until the moment the shark jumps out of the water anyhow.  A drum I've banged many times over the last few months with my take that level cases did not imply a steady state (far from it).

Different re-infection rates for different variants will change the rate constants for each variant which will change the tipping point where omicron takes over driving top levels cases from delta.  It'll change it by about 12 hours or so I reckon given the running boots omicron is wearing...

Allow me to pounce on this as it backs up what I’ve been saying for ages and I now have the data to prove it. Natural immunity after being jabbed is the best way out of the pandemic.

- Delta infection: 91-92% effective 

- Boosted: 75% effective 

- double jabbed: 25% effective

Yes breakthroughs aren’t pretty in vulnerable groups, but in the face of an ever mutating virus they’re our best defence against future variants. 

I'm not sure you can pull those numbers together and present them as speaking to the same thing without more context than is given in the document.  Nonetheless, if they did happen to be fairly comparable...

The data as you interpret it shows (mostly vaccine moderated) delta infection to be the best of a limited serious of options.  There are other options not on the table that could well be better yet.   

The way I continue to see this is that vaccines take an order of magnitude of risk out of a person's first live infection.  

There remains a role for a high efficacy, multi-protein vaccine in hopefully taking another big lump out of that risk - a lump that is by many measures required to retire the risk to the level of truly endemic viruses.  One of two reasons I'm very unhappy about the Valneva decision - the other being the pre-empting of the independent regulator, publicly, during multiple phase 3 clinical trials.  I'm more interested in this with regards to the risks of serious illness than of transmission.

To look at it another way, my take:

Natural immunity after being vaccinated is the only way out of the pandemic.

What matters is how we use vaccination to best effect to lower individual and societal risk as fast and as low as practicably possible.

At some point we will have to* transition to endemicity, and that is essentially what that is likely to look like in my view - use vaccines (potentially 6 monthly boosters on an ongoing basis) and medications to smooth the bumpy path to our immune systems just getting used to it like it has with other coronaviruses.

So in principle I agree, but we do need to do that without the NHS going well past capacity, so we may well indeed need one or two more "lockdowns" along the way.

* The other option, eradication, is going to be very difficult with a virus with a potential R0 of 5+, unless you are going to go for a lockdown well in excess of the original one, e.g. no outdoor exercise and Army provided food deliveries, and for a long period, too.  And as other countries won't do it, near-permanent border closure without exceptions, too.  Would be easier for us than say Germany, but probably not viable.

The simplest solution would be to count all the infections rather than try and compensate for unnecessary bias after the fact.  Catching double counted people and excluding them is more work than not catching them, it should be trivial to tweak the code that does the excluding so there's a time limit after which reinfections are not excluded.

So why haven't they done it? 

It's more a case of it being inevitable.  The emphasis has to be on doing it safely though, which means (a) actually getting triple vaccinated and (b) even then, not all getting our 'natural immunity' at the same time. 

And this is what will probably lead to some sort of lockdown being announced today.  If we all get it at once, that means society pretty much breaking down (because everyone is ill) and potentially many avoidable deaths, including of other things as the ambulance service ceases, for example.

Were it much milder, then that might not be a bad thing (we all get the sniffles at once, big deal) but it's hospital occupancy and NHS staffing (plus fire, police, power supply etc) that is the danger.

Re: SAGE document.  They're really spelling it out as clearly as they can.

If I could use double-size bold red text to post point 8, I would.

8. The earlier interventions happen the greater the effect they will have (high confidence). This may also mean that they can be kept in place for a shorter duration. 

Given the worsening doubling times now showing in older ages, its hard to see the implementation of signifiant additional control measures can be delayed much longer.

Point 11 is really interesting as well - early suggestions that the apparent reduced lethality of omicron is down to increased population level immunity and not decreased intrinsic lethality.  Whodathunkit?  An awful lot of confused reporting around this pre-empting even preliminary controlled analysis.  Just the wrong kind of noise to have when the MPs are balkanising along the divide between believing what they prefer and listening to the scientists.

So in essence SA were seeing less severe cases because they were people who were entirely immune to previous strains?  That makes some sense.

Regarding interventions I think I'd go with 27th December at 0001 for 3 weeks minimum including schools starting back late (take 2 weeks off next summer so no learning time is lost, with holidays granted if you can prove you have a prebooked holiday as of the announcement), but that if you are staying with someone you can remain there if already there and return home as you see fit.  The reason I say that is that people cramming onto trains and coaches on the morning of 27th would be a superspreader event, better that they spread it out, and those people are already close contacts with the host family anyway.

I was reacting more to this comment:

"Multiplicative factors drop right out of rate constant measurements."

You asked people to pick up errors in your analysis. It seemed to me you were arguing multiplicative factors were constant. I'm glad you have clarified they are not.

Can you explain why you think this is biassing the rate constants?

I'm not trying to compensate for it, I'm letting it drop out.  16% of people are a detected case, 8% risk of reinfection following a recent delta infection, do the maths.   We're coming out with single-digit percentages here which then drop out of the rate constants anyway.

A suddenly change to the data would break all the analysis using it - they're going to need to give a view of the data broken down by past infection status - but that's becoming increasingly meaningless as the immunological picture fragments in to every possible permutation of infection and vaccine doses...

PCR data is a mess already with an eclectic mix of symptomatically gated, non symptomatically gated and hospitalisation gated samples flowing in to the labs.  A binary approach to past infection detected as a case  is almost just noise in many ways. 

Far better I think to move towards a higher powered, more frequently updated random sampling survey with a faster cadence than the current ONS one, with a more detailed analysis of symptoms (more like ZOE), and to pick some more rigid criteria for P1/P2 PCR testing and for categorising the results.  

I assume you're referring to the last sentence? Is that not simply referring to multiple tests for the same infection i.e. person goes for PCR and tests positive, then goes to hospital a week later and is tested multiple times during their stay, each time testing positive.  We want all those positive tests to be counted as one.

If the argument is the immune system needs to see the full virus then we can use a deactivated virus in the vaccine.  We don't need exposure to the live virus.  It's not just Valneva taking this approach, the Sinovac vaccine does it too.  No problem at all with buying 66 million doses of an inactivated virus vaccine.

You want tests within a few days/weeks for the same illness to count as one but tests for a new illness to count again.  Easy enough to put a time criterion in e.g. after two months it is a new case.

It certainly seems that mixing vaccines produces the best overall immunity.  AZ+Pfizer seems to do pretty well from the stats I've seen, even though the former *isn't* inactivated-virus.

I think you've misinterpreted the data. As I understand it, it's not "natural infection provides (100-X)% protection against Omicron", it's "approximately X% of Omicron cases are reinfections" (edit, known to be reinfections due to a previous positive case, of course). Which is something completely different. It's ambiguous though, I looked for the source briefly in the other documents on the page but couldn't find it.

(X being 8-9)

Regardless, we already know that a combination of vaccination and infection gives you better immunity than vaccination alone. That's not really up for discussion. The difficult bit is the other effects of the infection.

Like every previous Covid infection was a recent Delta infection.  Plenty of Covid infections were earlier and not Delta.

I appreciate the comments and I'm happy to clarify.

Multiplicative factors drop out of the rate constant calculations.

I'm only claiming the re-infection biassing factor is multiplicative to a first order.

• It will change with the changing demographics within a variant, and it will change with the shift from one variant to another, but these are limited duration and limited scale changes.

But it's a small factor in daily case data then it's constant to a first order approximation, with constants dropping out of the rate constant analysis.  The details beyond that approximation don't, but they're small effects applied to a small cohort.  Well within the noise of all the other problems with cases/day data...

So I'm not clear why TiE thinks "there's hardly any point in calculating [...] exponential constants" from the data.

I didn't claim they were.

Yes, and I can do the maths on that and I came out with a single-digit percentage.

That was what a lot of people thought, but it turns out to be as TiE describes.  There's been some back-and-forth on Twitter with the lead developer of the dashboard over it.

Although I'm a fan of this as part of the vaccination protocol, there's a problem in that some key viral antigens aren't present as proteins in the virion particle but are expressed from the viral genome immediately after entry into the cell.  An inactivated virus would (presumably) contain no viable viral RNA and so wouldn't provide the same immunological challenge as a natural infection.   

For example, one promising target is the nsp12 polymerase, which would wouldn't be present in an inactivated virus vaccine.  A more sophisticated approach might be an mRNA vaccine encoding all or just the important bits of spike protein, plus conserved epitopes from nsp12 and other non-structural proteins (not functional), combined with viral structural proteins in protein form (either together or in separate shots).

That's the way I see it.  People who had so much immunity they didn't catch delta on exposure to it are catching omicron - and going on to have a lot of mild symptoms.  So, as mild as it is, exposure to omicron is worse for those people than delta.  

If you then take that lens and apply it to the question "What happens when someone who would have gone to intensive care after being exposed to delta is now exposed to omicron instead?"  and combined it with the question "how much more rapidly are those people going to be exposed to omicron as it spreads through everyone" you get the pertinent and rate limiting answer.

Moving some school weeks from winter to summer was an achievable goal that was not planned in advanced.   Likewise laster winter.  We can't keep hitting parents with school closures and home schooling...  It's not good for anyone involved.

Ours came home with a QR code for logging in to their distance learning system on the last day of term...

This is the whole situation in a concise nutshell IMO.

The main issue is whether we think Covid infections are spread randomly/evenly through the 66 million population or concentrated in groups whose life/work involves more contact.

If it isn't random then reinfection has a more serious impact on case numbers than you think.

I really don't think everyone is on the same page here.

Indeed.  

Also of the use of isolation orders when the plan apparently is to sit back and watch as all younger adults get infected within a couple of weeks in the leading regions.  People make things work.  We need things to work.  At this point, having London function appears to pivot on the significant fraction of infections not being detected as cases...  It doesn't feel like a very integrated approach to policy...

They do. But only if they are constant multiplicative factors.

Agreed; it's 'constant enough not to make a big difference'. There are other factors changing more rapidly (and probably more markedly) than the shift in re-infection rates due to change in variant.

I disagree.

• It's at most about 4% on case numbers if we take it as mooted that only about 25% of infections are now detected as cases...
• Even if I agreed it was a serious impact on case numbers (I don't), I would  disagree that this translates to a comparable impact on the rate constants for reasons already given.
• Even if I did think it translated through to a comparable change in rate constants, that can only be achieved by your "front loading" hypothesis, which can only lead to bias over a short period of time - and a bias that is small compared to what we're seeing happen in the data right now.  

So I don't just disagree, I triple disagree.

It just doesn't make a material difference to interpreting the stuff I'm presenting.

I appreciate the irony of you digging down in to all these details when a few weeks ago you were angrily defending your cherry picking of a single-day weekend sampling low from Germany and a Monday sampling high from the UK to make the (false) case things were 15x worse here than in Germany. 

There's something rather unbalanced about all this, I just can't quite put my finger on it...

Interesting, to me "multiplicative factor" implies constant but I'm happy to tighten up language there!

Also cases are perhaps going to hit a wall real soon now - depends on how fast cases in the peak ages in London start collapsing and if that opens up space for those to follow.  

I've long maintained cases aren't a useful measure for comparison between different times or places, and this underscores that.  The rate constants - judged empirically - seem to still be as meaningful as usual.

Upadte to LSRH's sweepstakes from last week's thread.  I took the liberty of adding a few people based on their comments.  Those not in bold have seen their date passed.  Perhaps they were thinking about when "should" rather than "will".

December:
15:
16:
17:
18: LSRH
19: MM
20: NW
21:
22:
23:
24:
25
26: JK
27:
28: WT (±  1 day - “or so”)
29:
30:
31:
January:
01:
02:
03: BC
04: RaJ (ironically?)

"Primary diagnosis supplement" are the Google magic words 

They're doing it. WT linked this on the 10th in another thread: https://twitter.com/pouriaaa/status/1468534853117304836?s=21

You are right - I did misinterpret it.

The point is valid and I think as Wintertree has pointed out, people are definitely not on the same page.

Theres still a lot of people hoping to get all of their immunity from vaccines and to never catch it. 

I'll just refer to doubling time as I think it's more plain English than rate constant.

There are many true doubling times and they're not all the same.

For example, the true NHS staffing doubling time of interest to Sadiq Khan is about 3 days from numbers below. That is a true doubling time of the relevant demographic for gauging how quickly a location and sector specific crisis is looming.

"One of the pandemics unsung heros, Chris Hopson, Chief Exec of NHS providers (I say this as he seemingly just ignores the government bad news bans but never exaggerates) is pointing out on the news this am that staff off sick in London doubled from 1900 at the beginning of last week to 4000 by last Thursday and numbers are still growing."
https://www.ukclimbing.com/forums/off_belay/friday_night_covid_plotting_57-...

The calculated doubling times (plural) don't matter to the extent that:

• you know it's going to change as behaviour changes due to caution or fear
• you know it's going to change as behaviour changes due government guidance, regulations or law
• they might not be the doubling time specific to staffing at NHS London, Sainsburys or train drivers
• if they're not based on hospital admissions (or something else) it's not the most immediately relevant to hospital admissions (or something else)
• they're not the doubling time of government political courage or the doubling time for IQ of Covid Recovery Group members

Where the calculated doubling times (plural) do matter is that they are short compared to

• the period of growth so the growth is compounded
• the time taken for behavioural changes to feed through to reduced deaths (a couple of weeks?), reduced staff absences (longer?) or some other measure
• time to complete the booster programme
• time taken to find evidence Omicron is milder (still none from UK or Denmark)
• political dithering

TLDNR - where the calculated doubling times (plural) do matter is that they are short compared to the period of growth so the growth is compounded.

My view things are so badly f*cked at the moment about the only thing you can say with confidence is:

a. things are f*cked. and

b. the Tories are responsible.

c. if we get 10% of infections turning into long covid as in earlier waves it will be a disaster even if deaths are low.

Can I have 23rd, please?

(I'm not sure abou this +/- business either.  Sounds like hedging bets to me)

You may be able to say them with confidence but not with accuracy

LSRH take note.

But I only get 1/3rd of the kudos if I get it right.  Conserved quantity, you see.

Is it "announced" or "coming into force", in terms of restrictions?

If it was announced I could go for tomorrow; but probably just pleading before Christmas with anything with legal weight coming in on the 27th. I think Boris' ego is too fragile to take having his face photoshopped onto the Grinch 2 years running.

Agree with all of that. Especially

Announcement - https://www.ukclimbing.com/forums/off_belay/friday_night_covid_plotting_56-...

You still haven't made a compelling case for why it makes it pointless to look at the rate constants. 

(a)  it's looking that way, and if we actually wait for hard data on omicron hospitalisations in ages 40+ before implementing more restrictions, and the increased precaution seen recently doesn't rapidly translate in to lower rate constants really soon, sadly you might well be right.  I'll regard it as more of a broken clock moment mind you, as you've been saying the same thing for the duration.

(b) Have you tried screaming this it in to a hollow tree?

(c) In terms of disease (not causal agent), most people are now catching a cold, not Covid, it seems.  We don't get many "long-colds" - post viral fatigue in other words.   So, applying that 10% to all infections is a bit dunderheaded.  I wouldn't want to be out in the world without at least two doses of vaccine in me, preferably three, right now mind you...

Unfortunately if you live in Tory Britain you are governed by a bunch of psychopaths who are setting out to make it impossible to avoid catching it and will almost certainly be successful.

Latest wheeze: ask retired teachers to help keep schools open because too many of the younger ones have caught Covid.   What could go wrong?

These f*ckers sit about in Downing Street drinking wine and asking other people to risk their health for peanuts.

Thing is Tom, even if you’re ruled by the evil English Tories, the Chinese communist party or Jacinda Ardern in New Zealand the end result will always be the same. You’re going to get a dose of immunity the good old fashioned way in the end.

I won’t dwell on the path we’ve taken to get here, as it’s still controversial, but we’re probably one of the most immune nations on the planet right now with high vaccine uptake in vulnerable groups and high levels of prior infection.

Despite all of our advantages we’re still contemplating a lockdown. This alone should prove to you that a vaccine only strategy has its limitations.

This, hospitalisations and deaths still not increasing despite rising cases since November 7th, widespread public immunity, 3rd jabs soon to exceed 50% uptake.

Why are we expecting a lockdown again?

TiE seems to be grasping at vague notions of prevalent long covid and having a right never to catch covid at all (eg. delusions of elimination). Despite clearly being totally flawed arguments, I can't see any better reasons being put forth.

To be fair so far cases have been confined to the young. Who cares if under 40s have it?

It will be the impact as we start to get breakthroughs in 65+s that would start to put pressure on the NHS.

Theres been some promising anecdotal stories about it being less serve, but we don’t know yet.

Having said that, to jump straight to a lockdown seems premature to me.

I’d be horrified if we lockdown without first advising the elderly to shield, suggesting over 65s don’t mix socially, freeing up hospital capacity by shunting people into care homes, banning care home visits ect.

I think you misunderstand.  Detected cases are perhaps going to hit the wall of limited PCR capacity.  True infections have got a lot further to go...

True infections might be about to peak in one limited age band in one geographic region, but that's a band not very susceptible to hospitalisation.   There's an awful lot more to come.

Factually untrue.

London hospitalisations are clearly in to a significant rising phase as a result of the rising spread of infection.  This has been happening before cases started to rise rapidly in older adults - that latest rise is yet to play out in hospitalisations.  It doesn't look promising, does it?

https://coronavirus.data.gov.uk/details/healthcare?areaType=nhsRegion&a...

Where London goes, the regions are bound to follow - all-be-it with more 3rd dose uptake and hopefully more precautions, school closures, festive workplace closures and more WFH.

Going off continued comments (most recently from Sadiq Khan) in no small part due to the totally avoidable pressures being put on intensive care by the number of unvaccinated people requiring intensive care to try and stop them dying.   The daily rate at which people in susceptible ages are going to be catching Covid is skyrocketing as most of the rest of the population moves on to a more endemic phase.

We're contemplating locking the whole of society down so that those who declined vaccination can continue to receive high quality medical care when they suffer the life threatening and life limiting consequences of their free choices.

Thankfully that process is rapidly self liming as the number of people without any notable immunity decreases rapidly.  

Then it comes down to a question of if we have enough healthcare capacity to handle the demands of the winter viral pantheon on our oldest and most susceptible.  Difficult one that as Covid expands the pantheon and a failure to suitably control peak hospitalisation rates in the past combined with staff losses and - right now - a large number of isolation orders on staff means something may be needed. 

https://coronavirus.data.gov.uk/details/healthcare?areaType=nhsRegion&a...

I'll take the 22nd.

Unexpected (/unscheduled) cabinet meeting called at 2, the guardian reports, with Whitty and Valance presenting...

Could be moment of truth time for the sweepstakes...

The options for ministers are apparently:

1. Urging the public to limit social mixing, without legal enforcement

2. Mandating curbs on household mixing, the return of social distancing and forcing pubs and restaurants to close at 20:00 GMT

3. A return to full lockdown. The Telegraph says it has been told one cabinet minister would resign if that happened

I suspect we’ll see a fudge between 1 and 2. I can’t see an appetite for 3 before Xmas.

I feel like either 2 or 3 is the right answer in terms of covid spread (I'm just not sure many people are willing to sacrifice Christmas plans voluntarily at this stage, and whatever moral authority Boris has has been diminished by the various #lockdownpartygate stories), but I certainly wouldn't bet against you (hence my slightly hedging reply to the sweepstake - I can see 1 now and then 3 after Christmas). 

At the moment the majority of rapid Omicron growth is in younger people who appear to be less susceptible to severe illness etc.

What we don't want is for this to spread rapidly into older people before the vast majority of 3rd vaccinations have taken effect; otherwise lots of people with severe illness etc.

When are younger people most likely to mix with older people - anything coming up on the calendar to give us a clue - oh yes, Christmas with all the family get togethers.

So of course any lockdown (they'll say "it's not a lockdown, just a firebreak") will come into force just after Christmas day when the opportunity to have the greatest effect (in avoiding spread) will have been lost.

I think this will turn out to truly be a case of shutting the stable door after the horse has bolted

🐎...........................................................................🚪

27th please if it's not already gone, otherwise 29th

It's trending up slightly in London, a high density urban sprawl and travel/business hub, with myriad other contributing factors to consider before extrapolating to the rest of the UK.

We're in a good place in terms of uptake and I don't think we'd really expect much higher without serious intrusion into, as you called it, free choice. I think you understand the net detriment of forced vaccination or health status segregation on society as much as I do, as such there's not much positive to be derived from trotting out scapegoats repeatedly, as far as I can see.

Well exactly.

You addressed isolation orders above. Even if it isn't yet, at a certain point the isolation requirements are going to be causing more issues than the virus if most cases are asymptomatic/mild, especially in critical roles.

I think there are a number of intermediate steps that we can take before jumping into nationwide lockdown if buying a bit more time for some jabs and gathering/analyzing more data is the goal. You don't seem fully sold on the idea yourself this time around. Certainly I don't think people are putting up with a new lockdown looking anything like the previous two in terms of length and severity.

You do know that in the shropshire byelection there was a candidate standing on a no restrictions platform etc. He got 500 votes, suggesting that your view on the " people" maybe a bit out of sync.

This is the point I’m pretty sceptical of. The 65+s got boosted ages ago. When I went for my booster 2 weeks ago everyone was under 50.

If we do lockdown it’s to prevent the NHS from being overwhelmed this winter, not to give time for the booster program to finish. 

And if it’s NHS capacity that’s on the line it’s going to be a LONG lockdown. I’ll see you on the other side next May.

It seems lot of people are looking for a relatively small scale and quiet Christmas this year, an earnest urging from government would go a long way to limit contact, especially if threats of harsher measures are left out for once (treated like fellow humans rather than scolded children).

I think social distancing is out the window really, I don't foresee us going back to the pantomime days of giving everyone a wide berth in supermarkets, one way aisles etc in a vaccinated population. Lip service will be paid where it absolutely has to. I guess you can slow spread somewhat with pub/restaurant closures but is it going to do more good than harm? Are these businesses to be compensated for the loss of earnings?

A full lockdown should require similar circumstances and levels of threat as the first two, is that the case? 

I don't see appetite increasing after Christmas unless you have very clear cut indicators that things are getting significantly worse.

It's doubled in the last month, and the doubling time is currently getting worse.  These infections would have been happening during the early rise of omicron in younger adults, where-as now grow is rapidly increasing in cases in ages actually susceptible to much hospitalisation.  It seems reasonable to expect this to keep rising quite a lot.

If someone hasn't understood the behaviour of exponentials and the time lags from infection > cases > hospitalisation > leaving hospital by now, it's hard to treat this as genuine naivety. 

Yes, hopefully the inevitable will be slower in the regions, but it's still inevitable - noting that the scale of the inevitable is still mutable through 3rd doses.

Much freer choice than is remaining in some of the countries in Europe, thankfully.  I am so incredibly grateful that we've got this far despite the significant noise against vaccination that has been so all-pervasive throughout. 

Segregation will flow naturally if we get to a triage situation, and I will feel gutted on behalf of the medical staff carrying the emotional brunt of this.

Pushing too hard on vaccination is a concern IMO in terms of undermining future engagement with vaccinations.

They're not scapegoats, they're (rather graphically and literally in many cases) the choke point in healthcare.  If my child's education has to be disrupted again because of people who chose to not engage with vaccination I will accept it, but I will be clear about my feelings on the issue, too.

Once the pool of unvaccinated hospitalisations-to-be are drained down.  That was happening quietly in the background before Omicron.

Problem is 1/4th of the Tory MPs are now pretty set against this sort of thing, and the PM is on thin ice for wider support; the politics - and those who have sought to influence them over the last 21 months - are toxic to taking sufficient intermediary steps, meaning we end up reaching for the big red panic button.  We've just had one minister resign over some of this level of precautionary, pre-lockdown control measures.

If we need it, we need it, but it will be a failure of state this time around more than an inevitable consequence of a brand-new pandemic IMO.  A lot of people who are better informed than me are making a lot of very concerning noises right now. 

In reply to neilh:

Hints at a shocking revelation that the single-issue cohort of Brexit MPs now confronted with Covid aren't very in-sync with the people.  

It's ok, there's a new rapid test on the way:
https://nitter.42l.fr/_BillieBelieves/status/1472886369609469954

No I didn't know that. If they were a one issue candidate against the established big parties, which everyone saw as a time to tactical vote to give the conservatives what for for the Christmas party debacle, in a time where there were basically only mild restrictions, I can see why most people weren't that worried.

I don't think I claimed anywhere my views are reflective of those of the people of Shropshire.

When were you last in a supermarket. Seen plenty of social distancing already over the last week. Do you not observe that restaurent bookings and foot fall in shops are already down.

I was in London on Saturday, the amount of face mask wearing in the streets was unreal.

I reckon you must walk round with your eyes and ears not functioning or you live remotely.

The dashboard update we didn't want to see?

20 December 2021

From 21 December 2021, an additional category will be added to the cases map to show 7-day rates of newly reported cases that are greater than 1,600 per 100,000 people.

I bagsied 26th first, but  I don't mind sharing with JK

I appreciate they are your views, most people's actions out in the real world would suggest that you are a small minority.

Ah I’m not so sure he is…

UKC is abnormally puritanical about Covid restrictions. Let’s put it another way - if people took UKC as representive of the wider population Corbyn would be the PM now and Labour would win by a landslide every time.

Most people are mask wearing in my area, but would they comply with not mixing with households. I suspect not. Compliance isn’t black and white and I don’t think we’ll get the same drop in social interaction this time.

On the sweepstakes the Times are now reporting new measure will come in on the 28th.

I though they leaked 27th as very likely an hour or so ago?

Times: A government source said: “The data isn’t there yet to justify further restrictions. There’s still so much that we don’t know about the severity of the variant and the extent to which vaccination and boosters sever the link between cases and hospitalisations.” However, ministers are still considering plans for a circuit breaker after Christmas. This would ban people from different households from mixing indoors and introduce possible restrictions on hospitality. December 28 has been pencilled in by officials as a possible starting point for new curbs — taking into account the 48 hours needed for recall.

It was interesting reading editorials in the newspapers yesterday pointing out that people are already making their own choices. So I am not sure your view stands upto scrutiny irrespective of UKC.

The times yesterday pointed this out.

Might have been more efficient to add a 3,200 per 100,000 people at the same time...?

In reply to thread:

A good read from Nick Triggle .  He's also suggesting the UK may be better placed than much of Europe for Omicron given both the vaccine uptake and the level of spread of delta that it enabled.   Sometimes I almost wonder if he lurks on here, but I think it's more likely that he talks with a lot of experts in the epidemiology community and front line medicine, and has really learnt his beans over the course of the pandemic.  Nice to read a news story saying "don't panic (yet)".  Hopefully they're on to something... 

https://www.bbc.co.uk/news/health-59658486

Last December Nick Triggle said don't panic too. https://www.bbc.co.uk/news/uk-55379220

Yesterday I commented hat there slowing down of growth in the demographic rate constants for London could be due to running out of people to infect in those age bins, or that it could be down to limits on testing capacity.

I think it's probably some mix of the two; today's update hints much more firmly about testing saturation.  The doubling times for three different bins have all started to back off at once in London, where-as we'd expect different bins to reach saturation at different times with 25 to 29 getting there first, given its stand-out higher growth.

Likewise, the doubling times for all the English regions have started to back off.

I can interpret this in two opposite ways:

• One big argument for this being a saturation of capacity is that it's very synchronised across all the English regions, suggesting an artefact rather than an epidemiological cause as the progression of omicron is very desynchronised across the regions.
• One big argument against this being a saturation of capacity is that it's very synchronised across all the English regions, where-as I thought samples were processed regionally, and because the decease is very desynchronised, different labs will saturate at different times.

The last day of data used in the rate constant plots is now five days in to the past (moved back to trim back rising provisionality in the most recent few days of data) so Wednesday Dec 15th; that feels a bit too soon to be in response to the WFH mandate (active from Monday 12th) but late enough to respond to the returning mask mandates in England.

I was wondering what to do about potential PCR capacity saturation.  I think at this point not using it for community samples from adults aged under 35 might be a pragmatic solution - better to retain high quality data on the ages where hospitalisation remains a concern with a buffer zone down the ages.  Clinically administered LFTs where possible for under 35s and more responsive random sampling using a small sub-set of the liberated capacity.

Increased Uncertainty!   Great.

Regardless, cases are definitely rising in adults aged over 40 in London now which is worrying news for hospitalisations.

I don't read his inset there quite as that.   

His wrap up doesn't seem far off to me compared to what happened - things did feel very different 3 months later, and things did get harder.  They got a lot harder and a lot worse mind you, so he was on the optimistic side.

That could see two million people a week being vaccinated. Within a matter of months all the over-65s could have been offered a jab. This could then start to feel very different.

But for now, the slog of the pandemic continues - and for many it just got harder.

I just remember reading quite a bit by Nick last autumn and he always seemed over-optimistic to me compared to how things panned out.

Maybe just my take on it.

I’ll have to go back and do some reading; I agree with you that article turned out to be optimistic.  He was - in retrospect - also optimistic with the line there was no evidence alpha was no more lethal.  I was pleased to see the clarity SAGE spoke with on omicron and lethality.  An increasingly complex issue…

So the guardian reckons the primary driver of restrictions right now will be the London hospitalisation rates, currently ~200/day. A month ago they were half that. Applying a 10 day lag, cases were around 10k, a month earlier, 5k. So on this raw data analysis (eyeball smoothing applied), and assuming the ‘pre-spike’ growth in London cases over the stable 5k baseline over dec is all O (reasonable?), hospitalisation is similar to delta..? how does that check out with the WT analysis?

Is there data to suggest hitting limits on testing capacity or running out of people to infect (is that likely when the triple jabbed can be infected?) rather than people reducing social contacts because they see soaring infection rates or because they want to reduce risk in advance of meeting vulnerable relatives over Christmas?

The dashboard data would suggest there is still plenty of PCR capacity. The graph below is for the whole UK. I'm not sure how capacity is measured though (eg is it just theoretical lab capacity or does it recognise the limits of test sir, test kit provision etc.)

Another question in my mind though is why PCR test numbers haven't gone up faster than they have. Positivity is heading northwards but the last date for which the dashboard presents data (and only for England) is 15/12, as a seven day average. I'm not sure if more recent positivity data is available in the downloads (WT?) It got to almost 20% at the peak last winter and I suspect is probably already in the same ballpark. A really high positivity would probably suggest the number of tests taking place is insufficient.

Good point.

Bear in mind vaxxed contacts are being asked now to LFD rather than PCR and that testing is liable to reduce when testing positive means you wipe out your Xmas… 

It feels like this applies to at least some of the cabinet, to be honest.

Is anyone else getting a horrible feeling of deja vu?

Arg, lost my post. In brief:

Positivity is suddenly rising rapidly - see screenshot below - even though it's still priming the 7-day moving average.  Red flag that the ratio infections : detected cases is getting worse, even though we're not hitting the dashboard capacity limits (likely ensemble logistics not on-paper capacity?)  

• I think so in the demographic in London that's the leading part of current growth.  See [1] for my thoughts.
• See also mik82's comment [2].
• Todays dashboard heatmap for London [3] shows 2884 / 100,000 as the rate for cases in ages 25-29; that's 2.9% per being detected as a covid case.  With only a third to a quarter of true infections (including asymptomatic, which many young omicron ones are likely be) getting detected as cases, perhaps 10% of young adults in London are catching covid every day.  In the last 7 days, 11% of this band have tested positive, perhaps between 33% and 44% of all people in that band.

Absolutely staggering situation.  

[1] https://www.ukclimbing.com/forums/off_belay/friday_night_covid_plotting_57-...

[2] https://www.ukclimbing.com/forums/off_belay/friday_night_covid_plotting_57-...

[3] https://coronavirus.data.gov.uk/details/cases?areaType=region&areaName=...

I've avoided doing case hospitalisation rate analyses as

1. As with CFRs, it really has to be demographic to have any utility and not to be badly open to misunderstanding/misrepresentation
2. The very different age bins in case data and hospitalisation data make a demographic analysis with public data impossible, and the attempts I made to try and back out a proper demographic curve for hospitalisation crossed the line in to modelling territory that I can't hope to do competently. 

Crude analysis -London hospitalisations not rocketing, wouldn't expect them to for another 5-6 days; lots of data holes over Christmas so it'll probably be next Thursday before we know if it's going properly south.

Rant alert - using different demographic bins for English cases and deaths, Scottish cases and deaths and hospital admissions is just bone headedly stupid.  I get the impression it frustrates some of the actual scientists feeding stuff in to SAGE as well (could be wrong) and is just an unnecessary hassle.    It seems like such a simple thing to fix (although that's probably because I don't have a cell wall covered in scribblings depicting the NHS IT systems) and low hanging fruit for improving the fluency of the data pipelines feeding in to cabinet.  A bit like adding more lanes to the A1(m) until it gets to Morpeth....

In reply to Ramblin dave:

I had high hopes for some resignations over a tightening of control measures today, silver linings and all that.

Sort of; I think there is a lot more hope this time around, but if it works out it will involve a large dollop of luck; there was no way it was ever going to work out last year.  But any interview with anyone from healthcare points in the same direction - exhaustion, stress, nothing good; far worse than this time last year.  

Also, Johnson has said he's going to be reviewing the data hour-by-hour.  I'm not quite sure which data sources actually update that often, but good on him........

It's all in the downloads, but handily if you access it on a desktop device you get a proper graph - used for my screenshot above.  Up and up it goes, where it stops, nobody knows.

I'm just remembering the Newsthump headline from last year: "Government insists there will be no U-turn on houses mixing over Christmas until everyone has bought a massive turkey." I'd call it prophetic, but it doesn't really seem like prophesy when the thing you're talking about is so predictable.

It's also not particularly comforting to watch friends who work in hospitality (including supply chain) having to gamble on whether they'll be able to sell any stock that they order.

I can get the same graph on mobile, but it stops at 15/12 - I was wondering if any download contained the more recent data so we could see the trajectory better. I guess maybe they only calculate it once the case and test numbers data have passed their 5 day cutoff.

There are still lots of LFTs happening. Over a million a couple of days ago, now dropping slightly, probably as schools have started to break up. Remember that even as a vaxxed contact, no-one actually checks you have done an LFT, it's just strong guidance so it won't force someone who really doesn't want to. Tbh I suspect more and more people will do LFTs for non work/school reasons as Xmas day approaches but some (hopefully very small) % of the positives won't get registered.

Bloody hell! 

Positivity reported in Scotland was 15.2 % today, equalling the previous record high from January). 

date      %positive      n tests

20/12       15.2           49,000

19/12       13.9           47,000

18/12       12.2           53,000

17/12       10.1           49,000

16/12       10.8           60,000

15/12        9.1            61,000

https://www.gov.scot/publications/coronavirus-covid-19-trends-in-daily-data...

(I am currently self isolating).

It also does not follow that everyone who gets a positive and doesn't register it won't use it to at least influence their behaviour, e.g. not go and see Nan but still go out for walks alone where you basically have no chance of spreading it.

Ah right - I thought you meant it cut off at that date last year on the other side.  Got it, now.  A lot of API downloads have conservative cut-offs in them.

In reply to Ramblin dave:

Indeed; for all the "but the economy" cries, what many industries need is certainty and predictability (not just hospitality...).  Would be nice.

In reply to thread:

Updated demographic case fatality rate plot; when I was looking at the CFR for different lags (7 days to 28 days), I was generating the bounds for shading by case date; engaging my brain I realised it makes more sense to do this by death date; it has't changed the trends but it has tightened the shading. This is almost entirely pre-omicron deaths data, so it's a baseline to see what happens over the next month.

Good work on the table formatting; also useful to see the numbers for Scotland; thanks.  Hope you've got some comfy walls to bounce off in isolation.  Hopefully you'll be out in time for the next cold spell...

In reply to Neil Williams:

Indeed; at some point the rules become a perverse incentive in a system driven entirely by voluntary engagement.  I'm both surprised and disappointed that lone outdoors exercise isn't permitted.  It was one thing in early 2020 when many things weren't understood and when healthcare was on a real precipice, but over recent months it's seemed pretty hard to justify.

I know many people are doing them as instructed but not bothering to register the result

yup.  vimeo.com/455043781

What data, though?

Torygraph readers' comments?

Backbencher tweets?

Number of rioters?

At times I imagine Cummings talking to WOPR as it ingest’s the world’s social media feeds and emits popularity rankings.  I suspect there’s an image search set up for a grinch with Johnson’s features…

Very good video. If only government messaging was that clear.

My self-isolation is nearly at an end, thank you (and thank god!). Which means I missed all the inversion weather - I'd rather have missed Christmas.

I have this vague memory of saying "let's hope it's all got so boring that #52 can be the last". Mmm.

What a great video - especially...

Go Hard Act Early - Handling Pandemics Day 1 Lesson 1

I'm a bit unclear what you're saying here:

• Are you saying that you think all the oldies (I'm >60 so I'm probably one) have already been boosted - many/most areas yes, other areas not yet (I believe).
• If you accept that oldies even with 3 jabs are more likely to go to hospital, then surely you can see that rapid spread into oldies is worse for the NHS than slow spread into oldies.
• If "all" oldies have had 3rd jab (over a week ago) then I'd agree that the continuing booster program is not particularly relevant to this part of the pandemic demographic except that it might slow the spread into oldies.

> If we do lockdown it’s to prevent the NHS from being overwhelmed this winter, not to give time for the booster program to finish. 

More boosters means less serious illness which means less chance of NHS being overwhelmed. But you're right that the primary reason is to protect the NHS, anything else is a secondary benefit (although on an individual level, I think most people would see not being dead as a primary benefit).

My cynical side reckons that towards the end of a 2 week firebreak, it will be necessary to extend (and deepen?) it into a longer lockdown - hopefully not, but the thing I'm most certain of with Omicron is that the government are unlikely to make the right decision at the right time.

I opted for the 26th last week. 

Interesting news of yet another vaccine, the more the merrier!

"First doses of protein-based Nuvaxovid are expected to be used in new year after European Medicines Agency gives go-ahead"

https://www.theguardian.com/world/2021/dec/20/approval-new-covid-jab-german... 

Yesterday evening. People are being sensible, despite being quite crowded with the stocking up for Christmas, but there's no maintaining a strict 2 meter, one way system, comically exaggerated dodging of others and tape all over the floors. As I say I think that's a bit panto and unlikely to return.

You can't be suggesting people are capable of making their own informed decisions on where they go and who with? That's the governments job!

Was it just co-incidence that the Lib Dems are strictly anti-vaccine passport and got a shock win over the conservatives who are very much split on the matter? Must have been.

London New  Year event (Trafalgar Sq, not fireworks this year) now cancelled. 

The actual truth is that England is abnormally lax about Covid restrictions. 

Yes, obviously! Do you honestly believe that the Tories are detested by every reasonable person because of vaccine passports, not because they're disgusting lying scum who treat every ordinary person with contempt by behaving as though they are above the rules they make for the rest of us?

Did you see what they campaigned on? Was it vaccine passports or government lies about parties? What are people angry about? And while the LDs voted against vaccine passports, they're not freedom-twats going for the anti-vax vote!

Are you saying that Scotland is that much better. Slightly better in that its legislation is usually slightly more prudent, but is the Scottish observence of the regulations that different from the English?

I think it's a notable and relevant factor in the context of Neil's points, I'm not saying it was a main factor. Certainly if the people of Shropshire were massively in favour of bringing in covid-passports they wouldn't be voting for the party with 100% opposition to them, would they?

You rage and attack points I'm not making, as usual.

Of course! He sees people breaking the rules sometimes but he knows they're English (and probably Tories).

Nice description, would look good in the Wikipedia entry for Conservative Parliamentary Party.

And would be ok (ish) if they were in opposition, unfortunately...

Yes they would if they didn't know or care about vaccine passports.

So are you asserting this is the case? Or just giving a what if?

It's just obvious to me that since the LDs haven't publicly pushed their opposition to VPs, rather they voted against them for pragmatic reasons, this was not a significant factor in the by-election. The point you make is not credible, it's snatching at obviously irrelevant facts to try to support the impression that your opposition to VPs is some kind of consensus.

Perhaps it is, but the by-election isn't evidence for that.

What I said was England was abnormally lax on Covid restrictions.  I define normal with respect to advanced democracies such as the EU, Canada, Australia, New Zeeland, Japan, South Korea etc and now the Democrats are in charge the US.

For the level of infection being experienced the lack of action is way outside the international norm.

Scotland is a special case because it does not have the powers of a nation state.  If we were independent we'd be far closer to Ireland or Germany than England.

England was also “abnormally low” in terms of ability to sustain exponentially growing spread of delta cases, hospitalisations and deaths as winter landed.  We’ll see how that translates to omicron, but spoiler alert: more immunity across the population is a good thing at this point going forwards.

Your continued pretence that this is a one sided issue is nonsense.  Beating a broken drum.

In terms of control measures post vaccination things are far, far from the simple minded view you cling to.

You're being a silly goose at this point

https://www.bbc.co.uk/news/uk-politics-58594038

Go and google 'lib dem covid passport', read some results. Maybe we have a different definition of what 'publicly' means.

I didn't even claim it was a significant factor, let alone a consensus. If your going to strawman me at least do it with a bit more subtlety. Or just rage out and start waffling and name calling as is your signature.

Farrer's interview on R4 was interesting and as usual very informative.

Agree with you on that.

The 500 votes I referred to related to Laurence Fox's party which is vehemently anti any covid restrictions.

You also need to ask yourself if vaccine passports for night clubs  in what is a rural area is really a critical issue.I would suggest not ( maybe city centre issue).

The drum is only broken in England.  In plenty of other countries it is working just fine.  By coincidence England has far more deaths than those countries over the full course of the pandemic.

There's absolutely nothing good about catching Omicron on Dec 20 before you get your third jag when in a couple of weeks you would have been jagged and loaded up with antibodies.

I'm got g to try a more concise way of replying rather than derail the thread with points you're not going to read before you go back to square 1.

Whatever.

Updated plots for London.  

PCR positivity continues to rise in England (17.5% on Dec 15th, rising about 2% a day) which suggests some of the slackening off of the growth rates could be down to increasing pressure on testing.

The time-series of some of the rapid growth ages from D1.c is interesting; growth rates started to slack of in ages 25-29 two days before the other ages - and before the slackening of growth seen nationwide.  

• I think this is probably the mid-point of the sigmoidal curve for cumulative infections in this age bad, because up to 10% of people in this band in London are catching Covid each day, going off the reported rates and spitballing over infection:testing ratios.   If I'm right, it'll become impossible to sustain such high daily cases in this band in another 5 days or so, at which point they have to go in to decay...

Cases in 65+ are really taking off in London now which is the bad news; we're seeing doubling times of between 4 and 6 days in all ages; this is nowhere near the apocalypse-level 2 day times mooted for the variant but still it means the situation is changing rapidly; any doubling time under about 10 days gives insufficient time for the outcomes > data > analysis > cabinet > decision cycle we've been used to.  

We await news on what the hospitalisation rates are shaping up like; we might reasonably expect them to fall both from ongoing boosters and from the exhaustion of the pool of people with no prior immunity (infection or vaccination).

Unless things change it's hard to see how signifiant additional control measures for London at least can be delayed much beyond Dec 27th.  As for the rest of England...

FTFY

The same set of plots as above, but for all of England except for London.  Gives an idea what's coming...

To me it looks very much like the situation in London about 5 days ago, except that the doubling times the peak growth age of 25-29 is more moderate than seen in London. at around 4.5 days; we expect this from cases taking off later after more precautionary behaviour got started.  The other subtle difference is that the growth rate in this age bin starts slacking off at the same time as others, not in advance of them.  This fits with the idea that it's passing 50% infected in London, as that would obviously not be the case for the rest of the nation yet, and so that wouldn't start tipping the rates over yet.  All a bit speculative on my part...

Growth in the not-London cohort was led by young adults, just as with London, and now it's starting to dial up in older adults, just as with London.

Plot 18 is interesting - the fall in hospital admissions continues in most of the non-London regions; more precautionary behaviour and more 3rd doses kicking in before Omicron takes over?   It's important to keep in mind that the growth in London hospitalisations is mostly from the early period where Omicron was only driven cases in younger adults - who don't go to hospital much.  So we're really waiting to see what happens with the older cases.  

East of England is probably the region to watch next in terms of hospitalisations.

Having checked, there's no recent figures released for NHS staff absence (think the most recent figures cover the first quarter of 2021).

Thinking this number is as important to the situation as cases and hospitalisation as if it gets too high due to staff having to isolate it reduces the NHS capacity to provide care. (Feel free to tell me if I'm missing something really obvious here or just being daft).

Hopefully it won't end up being a problem but it's definitely something that's been nagging away at the back of my mind.

Rose from 1900 to 4700 within a week, in London:

https://www.bbc.co.uk/news/uk-england-london-59741331.amp

Thanks, just been looking for figures myself but think the NHS only releases them quarterly. 

Hoping it doesn't get too bad, they're under enough pressure as it is.

No data to add but it’s clearly a big concern.

I think the rate it’s slammed through young adults in London isn’t going to be repeated in other ages and/or other regions.  So hopefully the rate of isolation orders is going to back off.

I think the early cohort of young Londoner cases will be released about when older hospitalisation start to land in their currently unknown numbers.  Not I think what anyone is going to want after isolating through Christmas; it really is a raw deal to cap off a year of demoralising events.

Lots of news stories about different services being impacted today.

In other news, I went for a walk in the woods today to get away from it all.  The bloody virus was there as well…

Disagreeing with you isn't derailing anything because your opinions are getting more and more flaky.  England is the crazy outlier in terms of opinion on how to deal with Covid.

That’s me, the flaky one.  Perhaps I’ll move on to 99s next week.  Less messy.

I said I wasn’t going to derail it, not you.  I was going to try and give a detailed, balanced reply but then I remembered you make an active point of telling me you don’t read what I write before going back to square 1.

https://www.politico.eu/article/jens-spahn-germany-biontech-pfizer-covid-19... 

Germans will be ‘vaccinated, cured or dead’ by spring, warns health minister

Doesn’t seem so different to me, except they squandered the autumn and are going to struggle to manage this in winter.