UKC

I’m not leading with a look at cases this week, as it all boils down to “cases are going up, lots”.  There is much debate to be had about if this is the right thing to do, and if so if this is the right time to do it - a very academic debate suggested another poster on the previous thread’s continuation given the pace at which things are developing. There's lots of room for  different opinions on this - room expanded all the more by the absence of more clear communication and policy decisions from government to address the issues this raises, and to evidence that it is the least worst option.

As always I'm having a quick look at the data - I did want to mention that I recognise that there's a lot more to this than the data itself - the support and burden on people being required to isolate as cases are apparently deliberately allowed to rise to high numbers, the uncertainty over longer term health consequences of infection in the young and healthy.  Lots for everyone to think about, and not the clarity or messaging I would like to see the public being respected with.

IMO, the critical bit now is what happens with hospitals.  The early outbreak of the “Delta” or Indian variant in the North West was quite alarming in that regards, but the wave seems to be rapidly moving away from that as the centre of infection moves in to populations that for various reasons are less vulnerable.

As well as issues around where the outbreaks first landed and relative vaccine uptake, we are seeing cases concentrate more in the young.  Plot C shows this clearly, with a massive change from the winter (blue) to now (red), with the shift on-going as can be seen by looking back a few weeks ago (orange) vs now.  The difference to hospitalisations is much more than the difference in overlap of the graphs, as the hospitalisation risk increases near-exponentially with age.

Plot P1.e is another way of seeing this shift.  Both show the relative distribution of cases at a particular time point, not the absolute number.  In the period before P1.e starts there's an even more dramatic shift in the centre of infection from older to younger people as vaccination started, but what I'm focusing one here is how that shift is still ongoing in recent weeks.

This shift is I presume the main driving factor in separating cases from all other healthcare measures as the next post goes in to.

https://www.ukclimbing.com/forums/off_belay/friday_night_covid_plotting_31-....

The Lissajous plots:

• These plots shows cases on the x-axis vs other measures on the y-axis  Other measures are hospital admissions, hospital and ITU occupancy and deaths.
• Time is not explicitly shown, but I've coloured the last rising wave purple and this one blue; both start in the bottom left with low numbers and move diagonally up and right as both measures on a plot rise.
• In a pure-exponential situation, both blue and purple curves would have a gradient of 1:1 on these plots, and the lower the conversion ratio from cases to the other measures (or the larger the lag between them...) the further right the blue curve would be from the purple one.

Interpretation:

• Most blue curves are bending round to the right, indicating that the conversion ratio from cases to other measures is getting lower - things are getting better from a healthcare and not dying perspective.  
• On some (e.g. deaths in England), the blue curve is moving right and down, meaning that cases are increased but the other measure is decreasing.  This is no mean feat.
• ITU occupancy in the North West - where this variant first took hold perhaps in a more vulnerable sub population with lower vaccine uptake -  hit the same level as in the last wave.  It now looks to be separation.
• It's perhaps what quite a few of us expected to happen - that the exponential growth in the North West wouldn't be sustained for long, and that the high hospital rates wouldn't track cases for long, but it's very reassuring to see that start to play out.
• Cases in Scotland have grown at such a bonkers rate suddenly that the blue curves have shot far off to the right; I expect them to curve around somewhat as the lag from detection of a case to the other measures plays out.  They've left themselves no room to manoeuvre if conversion rates turn out worse than expected - this sudden rise of cases from such a high number feels positively reckless.   It'll probably all be fine, but passing the point of no return before you find out for sure...  

In terms of the recent fast growth in Scotland, I was casting my mind back to late last summer, when various outbreaks started appearing in the news and locally, and these gradually grew in frequency until they coalesced in to what was the exponential phase of the wave beginning.   Now, the data looks to me like outbreaks are still dominating in many places, but that they're running to much larger numbers but not coalescing in to a global exponential phase.  A result of vaccination being inhomogeneous across the population, leaving many islands of transmission-risk that fire off stochastically but can't coalesce?  Or I'm just stringing words together that sound clever but don't really have a clue - interpretation of everything is getting a lot harder IMO as the situation gets ever more fractured and fractal.

The plots for England.

• You can see in these the same trends as in the plots above - cases are rising fast and it won't be long before we perhaps exceed the peak rates of the last wave, but hospital admissions have stalled or perhaps even gone in to decay; it's a very stark effect.
  • Perhaps there are other areas of vulnerability as with the early hospitalisations of this wave, perhaps ongoing vaccination has moved us beyond that.  
• Cases can't keep rising indefinitely without hospitalisations turning back to rise, but the susceptible population will be dropping rapidly with vaccination and infection happening at pace.
• I have a nagging suspicion that engagement with testing is going to start dropping off rapidly as attitudes change, so perhaps the scale of real infection is rising faster.  ONS data will give us some insight in to that in a few weeks perhaps.
• Perhaps deaths are turning to decay; the numbers are very low so the statistical noise is high.

I don't think I'm the only person to have looked at the cases data for Scotland in disbelief in the last week.  They lost a thousand cases or so down the back of the sofa as well [1], I'm not sure how much that's resolved yet.

Anyhow, hospitalisations aren't currently rising as drastically as cases, but it'll be another couple of weeks to see how much the recent meteoric rises end up translating.  A slight oddity in that the admissions rate is going down, but the occupancy rate is going up - with very low deaths, this implies to me that the discharge rate is going down, so perhaps fewer but more serious cases are being admitted; this perhaps aligns with observations from Si dH on a past thread that ITU occupancy is proportionally higher compared to general occupancy this wave, it rather suggests to me that there's a correlation between who is less well protected by vaccination and who is at worse risk of bad consequences.

Anyhow, I imagine a lot of people in England and other parts of the world will be watching very closely over the next two weeks as Scotland leads the charge.

[1] https://twitter.com/scotgov/status/1408409705513422854

The four nations plots for cases.  

• The same basic failure of control measures is happening everywhere, but at different times.  I haven't looked but I imagine its not unrelated to arrival of significant numbers of  India variant infections.  
• Where Scotland leads, I think everyone else will follow.  England is likely to exceed it's previous peak case rate in another week or so.
• I am very grateful that the last round of unlocking was delayed; whilst it looks like the healthcare measures are detaching well from cases, if we'd gone any faster there would've been scant little evidence of this until well past the point of no return.
• I'd be more grateful still for a few more weeks of vaccination progress before this had happened. 

Plot A shows a similar set of UTLAs hitting my arbitrary "outbreak" criteria this week.  Most of the remaining areas continue to rise.  None of the long running outbreak areas have sustained exponential growth for long, but nor are they going in to decay.   

Plot E shows that the outbreak areas are currently very close to zero growth, and almost everywhere else is growing, but with a somewhat moderate doubling time (for total cases including LFDs) of ~ 18 days.  That still translates in to a lot of daily growth in absolute numbers however.

Plot 18 shows how hospitalisations are going in to decay in some regions (e.g. North West & London) despite cases continuing to rise.  Presumably the demographic shift, ongoing vaccination and perhaps infection moving beyond the places it first landed.

The PCR cases doubling time plot shows doubling times backing off for England, so despite concerns we're still not seeing unconstrained exponential growth; if the orange areas on plots A & E show similar behaviour to the earlier outbreak red areas, then we could only be a couple of weeks away from the end of growth - but somehow I think the strongly targeted "boots on the ground" public health interventions that helped curtail growth in the early outbreak areas are less likely to be replicated for all the others, unless any worrying regional or sub-regional hospitalisation signals emerge. 

The vaccine plots shows the second dose rate dropping - but it's dropping more slowly than the rate at which second doses are coming due, hence the notional delay estimated between doses (red curve) is coming down.  First doses seem to be slowing down as well - as with many hard jobs in life, the first 90% is often the easiest, and perhaps we're running in to that with vaccination; there are a lot of walk-in centres being opened up now and the rising cases certainly gives impetus to people.  

My thoughts now turn to sequencing - with this scale of rapidly rising infection, a Bad News Variant could hide in plain site for far too long, and as mind-bendingly large as the UK's daily sequencing capacity is, it's not going to keep up with this.  Hopefully priority is given to sequencing samples from every hospitalisation in all parts of the UK, driven by science and not targets.

One anecdotal observation on testing - a lot of the folk I work with are slipping into using LF as tests to use when they feel unwell/are symptomatic rather than going for a PCR and self isolating. 

This is pretty depressing as they are a medically literate population.

assuming this is replicated to the rest of the population, then we may get less PCR positives - and more missed infection due to the lower sensitivity of LF, fuelling spread.

I think in the event of Scexit your proposed new name for Scotland should be adopted. 

Is it possible to tell this for Scotland? Without a map of infection rates down to MSOA level I don't think one can really tell whether the outbreaks are well joined up.

It's also worth noting I think that the demographic shift will make cases concentrate higher in cities because of the higher proportion of younger people living there. It doesn't mean the countryside or towns in between aren't effectively caught up in the same large outbreak, just that their doubling time is longer because they have fewer unvaccinated or/and young adults.

The hospital data looks encouraging. But the Scottish case data seems to have broken from the trend we have seen on the outbreak areas in England so far so I'm now wondering whether case rates in the north west (as the leading area in England and in which doubling times have reduced) could jump again with some sort of minor trigger event. Previously, I had thought it seemed like the case rates were going to max out in different built-up areas around the UK somewhere between what we saw in Bolton / Blackburn and what we saw in Bedford, before flattening and then eventually dropping with further vaccination progress. The Scottish data would appear to throw that in to significant doubt.

Sorry; could have explained myself in more detail.  I imagine outbreaks as being bounded by an inner scale (smallest detectible size) and an outer scale (largest containable size).  The outer scale seems to be moving upwards a lot recently with rising immunity in the population. Scrabbling very much in the dark - as you say without MSOA level data - perhaps the recent spike is still within the outer scale for an outbreak; we'll know in a few days based on if case numbers level off/fall or continue rising.

A few months ago, the next couple of weeks of data could be predicted pretty well by applying momentum to the rate constants, tweaked based on changed control measures, and following that through in to cases etc.  Now it's all over the place; I think this is down to the outer scale of outbreaks rising significantly; in terms of "can we have significant extra growth in the north west" - that depends if the most significant reservoirs with outbreak potential have been triggered yet or not under my noddy mental model.  The longer this goes on without any more sustained growth, the less likely it is that such reservoirs exists I suppose.

Good observation re: urban areas and demographics. 

I haven't crunched any sort of numbers on this, but is it plausible that the roll-off of increases in some areas that had initial spikes of the delta variant are down to immunity due to infection adding to immunity due to vaccination to actually hit the herd immunity level (given current controls), and that happening at different points depending on how much immunity already exists in the unvaccinated part of the population from previous waves? Scotland having been basically less worse during previous waves explaining why they're getting hit quite so hard now...

Edit: or if the model of "hitting the herd immunity threshold" is too simplistic, could immunity from previous waves at least be a factor?

From what I've seen locally there have been a ton of cases in S5/S6 school leaver year groups in the last couple of weeks as they have their end of school events.  That will likely also hit the parent age groups and siblings.  Football probably hasn't helped either.

The short answer I think is yes, it is definitely plausible. I don't think it's likely we have reached a threshold where immunity is high enough across the whole population of these areas that the virus could no longer grow if everyone relaxed tomorrow. However I do think the simple model of herd immunity is overly simplistic, and I think population immunity is probably having a big effect. Obviously some areas have higher pre existing immunity than others as they go in to this wave, especially those areas hit worst in December/January.

ONS estimate that well over 86% of adults in the UK now have antibodies to covid*. This is from a combination of vaccines and prior infection. We don't have data for kids, which will presumably be lower due to lack of vaccination. (A quick Google tells me that 22% of the population is under 18. If we guess that 50% of them have antibodies, which seems plausible to me if they are retained for a long time, then that would be 78% of the overall population with antibodies, illustratively. If only a third of under 18s have antibodies, it drops to 74%.)  Vaccine efficacy data would suggest that having antibodies does not make you completely immune to catching and passing on the disease but it will certainly help a lot. Let's keep things simple here and assume that individual immunity is something that is a binary on/off.

The simple model says that if virus r0 = 8, then greater than 7 in 8 people (ie > 87.5%) need to be immune to reach herd immunity, because each infected person, who would transfer enough virus particles to infect 8 susceptible people, will then only infect less than one person on average. If each infected person infects less than one other, the pandemic will shrink. (I have seen people guesstimating that Delta would have a theoretical r0 of 8, but have no idea how good the guesstimate is.)

In practice r is not fixed either in time or across the population. Firstly, the average person infects fewer people when restrictions are in place, regardless of any pre existing immunity. Secondly, some groups of people will infect far more than others, ie if r0 is 8 overall, it might actually be 0.5 for recluses and 20 for bus drivers or hospitality workers.

The consequence of the first point is that reaching herd immunity with the current remaining restrictions is easier than when they are completely relaxed. Let's guess that with the current limited restrictions in place, if there was no pre existing immunity at all, r would be 4 instead of 8. That means you could reach a temporary state of herd immunity if more than 3 in 4 (> 75%) had immunity, so the level of cases would slowly drop. If we aren't there yet when <18s are included, we must surely be much closer to 75% than we are to 87.5%. Temporary, because when restrictions were relaxed it would no longer be enough.

The consequence of the second point is that each immune person in a group with a high r value will reduce transmission more than each immune person in a group with lower r value. So if more people in groups with high r value get infected in a particular wave (as seems likely) then it will make more progress towards overall population immunity than if the infections are concentrated in groups who are unlikely to transmit. In this scenario the idea of a single r0 and easily-calculated herd immunity value breaks down. However hopefully the effect can be easily understood as a concept. Basically, if you manage to get antibodies/vaccines into people who are in groups most likely to transmit the virus in a future wave, then the herd immunity threshold is likely to be bounded by (lower than) the simple estimate.

I think there is a high probability that infections in the early outbreak areas have reached temporary herd immunity thresholds as they stand for their particular populations in the current level of limited restrictions. To me, this is evidenced because it seems that Bolton and Bedford have been not just caught, but overtaken, in infection rates by surrounding areas. I see no plausible mechanism for this other than local population immunity. There might be a flaw somewhere in this logic though. I'd like to know if the half of Glasgow that was badly hit a few weeks ago but seemed to have peaked is now hitting new highs along with the rest of Scotland, or not.

Edit to add, the current vaccination programme in young adults should be helping an awful lot, at least theoretically, because that's where most of the transmission has recently been.

Edited for clarity.

* https://www.ons.gov.uk/peoplepopulationandcommunity/healthandsocialcare/con...

Bedford (and Bedfordshire) not quite so clear cut as Bolton in this respect actually, as surrounding areas still have lower rates than I thought. It's getting caught quickly though.

Is it possible that single vaccination immunity isn’t as long lasting as perhaps thought(as well as the reduced efficacy of one dose)? It wouldn’t be great is it tails off after say 4 or 5 months if a second jab isn’t taken up and you have significant numbers thinking they are protected when they aren’t.

Regarding Scottish cases I’ve only really been following Dundee and Renfrewshire. From what I’ve been able to see case numbers are very concentrated into certain areas.

The point made above re use of LF tests is a good one. You can imagine wishful thinking being applied by many around the quality of the results.

The other thought I have around all of this is when do we rip the proverbial plaster off?  We have to do it as some point and there will always be an element of risk/leap of faith. Perhaps we’ve all become cowed(myself included) and need to break out of that mindset!

With the easy availability of LFT, and the fact that you get an immediate (if less accurate) result, I think it is inevitable that people will use LFT.

I suspect covid is also becoming 'normalised'

There’s nothing wrong with that as a test to find positives. False positives are rare. But the issue comes in taking a negative as meaningful. The issue is partly the LFT itself but also the self administration. You can get you a negative on PCR with absolute certainty if you don’t swab them properly..

I've not followed the discussions of LFTs very much, but I thought they were reasonably good with symptomatic covid, and less good with asymptomatic covid? I.e. if you are using it because you're already symptomatic it isn't a terrible "poor man's PCR", especially if you do several, but that without the filtering of being symptomatic the accuracy of the asymptomatic testing was less good. Is that not right? (Like I say, I've not been following it much, so genuine question)

I know...

All previous mailed tests were self-administered, too.

I suspect that this has been a problem at the school that one of my children go to. In their twice-weekly reminder email they've recently started prominently reminding parents and students that LFDs can give a false negative if you have symptoms.

An article in today's Graun quotes the founder of Independent SAGE as stating “by the end of July, you’d expect it to go up to 60,000 cases a day” with 2,400 (4%) daily hospital admissions.

https://www.theguardian.com/commentisfree/2021/jun/27/matt-hancock-hope-dre...

They seem to have a lot of youtube videos but I'm struggling to find hard details of any predictive modelling. Pointers, anyone?

Good question:

https://www.bmj.com/content/372/bmj.n823
 

suggests 58% sensitivity in symptomatic people.

So a fair few symptomatic infected people would slip through the net, and also miss out on a couple of days self isolation that would occur with a PCR.

SPI-M is what you need to Google.

If you want a well informed amateur take, look for James Ward

When we actually hit a herd immunity threshold, R will become less than 1 and daily cases will go in to permanent decrease; the curves don't look like that quite yet to my eye.  

But we've got to be getting close; antibody levels in the ONS surveys were pretty high in data from before these outbreaks, and both live infections and immunisation have carried on since - with live infection naturally falling more towards gaps in immunity.

But it's way complicated because the herd immunity  threshold isn't a clear, binary line in the sand. Susceptibility to infection depends on the strain in circulation, and the strain or vaccine that induced immunity.  (Susceptibly to serious illness being much less affected by variants to date than susceptibility to symptomatic infection).  But in terms of severity of consequences, infection with a variant other than the one that induced immunity is much less of a problem than infection by a variant when not endowed with any immunity.  It may be we don't hit herd immunity and R<1 in this wave due to weakened cross-immunity with what has gone before, but that circulation becomes no longer a healthcare quaking problem.

Won't be long till we know...

In reply to SFM:

No great sign of people declining the second dose in the data; I can't imagine withholding a second dose to measure relative speed of decay of immunity is going to get much study time.  

Proverbially speaking, I think it's happening.  I think any decision made at this point is subject to many intangibles meaning there's an aspect of faith to any direction from here.  The key thing is to not rush the process, so that the data can be continuously re-evaluated in case things take an unexpected turn.  I think if June 21st hand't been delayed, I'd have to have gone and moved to the forest for a month and turned off all communications, ostrich style.  As it is, when restrictions are likely dropped on July 19th, I don't think it will make such a difference as another month of both vaccination and infection will have raised immunity levels considerably more, and schools will be closed.

No idea where their models come from, I almost totally fail to engage with videos of talking heads, drives me batty.   No pointers (as LSRH says, look to SPI-M for models input to SAGE).  My take

60,000 cases/day by end July:

• This is about 2 doubling times from now, and in England a doubling time is currently ~18 days; 2x18 days from now is 2nd August - about the end of July - so that seems reasonable as an extrapolation of the present, but...
• Given a hunch that infections are ever more detached from cases, and historically infections were perhaps 2x cases, that's perhaps another 4 million infections between now and that daily case rate about.
• There's ~9 million unvaccinated adults in the UK - and that number is going to shrink between now and end July.  With some not insignificant fraction of those and of children having had Covid, I don't think it's right to assume exponential growth can continue at the present doubling time until late July; so unless growth is slowed significantly soon and then raised later, I'll be a bit surprised if we actually hit 60k/day.  (Optimistic mode is engaged).

4% case > hospital admission conversion rate

• I think it's likely to be below 3%, perhaps below 2%.  (Optimistic mode is engaged).
• This is from eyeballing the ratios on the Lissajous figures, looking at where they're perhaps heading, and thinking about more 1st and 2nd doses going ahead by then.  I think these estimates can firm up with another week of data - we're just shifting to massive infection growth in the young, and the latencies haven't swung around fully to admissions yet.  So it's unknown from this top level data where the conversion rate will fall.
• But....  Hospital admissions aren't the limiting factor; hospital occupancy is - especially ITU occupancy which seems to be running out of proportion (in a bad way) to general occupancy vs previous waves.  
  • It's still a lot better compared to cases than in previous waves, just not as much better as general occupancy. 
• Perverse factor alert - if engagement with testing goes down over time as people normalise this (as captain paranoia says), the hospitalisation rate goes up - the limiting case is where people don't bother to get tested, but where hospital admissions are tested.  Then, the rate is going to be 100%...

All spitballing to be taken accordingly.  

It's really hard to figure out how the next 6 weeks are going to run if cases are allowed to rise as they may, but if close contacts are still required to isolate under some pretence of controlling the spread (and to give the "trace" part of test and trace some purpose...)  It feels like a massive set of mixed messages, and one that's going to incentive individuals to think carefully about how much they want to engage with the systems, with not insignificant legal consequences a possibility if they break engagement.  Comments from the bailiwicks on their dropping of isolation requirements on double dosed vaccine recipients who are identified as a close contact seems like a logical step, but I can't see it being made until after school term is over and until after July 19th.  

Thank you for commenting on that 60,000 estimate, I was hoping you would! And for continuing to season with rationality the ill-flavoured soup of current events.

So whilst the first wave and the second wave peaked, and then decayed, because of lock-down, that isn't going to happen this time. An article in the Guardian the other day remarked that the side-effects of lockdown were so bad that we all know we mustn't do it, or anything like it, ever again. Interesting.

I'm out on a limb with guesses at late July - the devil would seem to be in the details of the intersection of vaccine uptake, where infection has landed in the waves to date, where the new infections are landing and the susceptibility of those not protected by full vaccination.  My sense is that any proper model going forwards is going to be hyper-sensitive to the details wrapped up in that list.

With cases rising here and with the travel industry having successfully lobbied for more travel opening, a variant with more immune evasion (in terms of infection, not necessarily serious illness), remains on the table.

The latest ONS dataset on measured/modelled immunity has 86.6% of the adult population testing positive for Covid antibodies as of 10th June 2021, with a tight CI.    Thats on the order of 10% to 15% higher that one would extrapolate from first dose + 14 days for immunity granted antibodies; trying to infer how much natural immunity is concentrated in the non-vaccinated from that rapidly has errorbars expanding beyond meaning.  

It's a shame the ONS survey doesn't break immunity down by protein/antigen so we can see what fraction of that are due to vaccination only.  

https://www.ons.gov.uk/peoplepopulationandcommunity/healthandsocialcare/con...

Rant alert - and before anyone piles on, I hold many of these points against governments of all colours going back several decades

I think knowing what we do now, the early predictions of death and healthcare overload looked rather optimistic.  If we are to avoid such a lockdown ever again, we need to up our game and be prepared in advance.  The meta point is that whilst we're going to be better prepared for the next pandemic, the next crisis could be one of many different things and the core problems underlying our lack of preparedness and making consequences far worse remain - a poor intersection of leading science and cabinet, a deeply unequal society, investment in blue skies research (not limited to the life sciences) becoming untenable, systematic underfunding of healthcare and social support. The United Kingdom is a  powerhouse of research in industry and academia with significance on the world stage, and is one of the wealthiest nations on the planet in per-capita terms.  The easy thing to do is to point at a country less fortunate than us and say "look, they didn't do any better".  The hard thing to do is to point at ourselves and say "We can do much better".  

The problem is that lots of people have the mindset that everything is fine once they've been jabbed when it might not be.  Jabbed people who have 'important' things they want to do (work, weddings, parties etc.) instead of self-isolation aren't going to test properly because they're not worried about their personal health and cases will rise 'in secret' as alluded to above.  We can only hope that enough people have been vaccinated because it's going to be tremendously difficult to put the genie back in the bottle now.

And last jan/Feb the messaging was ‘and a single jab is brilliant’ so lots of folk not getting the double jab. Hearing lots of interesting stuff here in Wigan.  Very high Covid rates in schools. One school I know: loads parents refusing to test their children so the council have intervened and said ‘isolate until you get tested - no test, no school.’  A school my niece works at, had at one point all staff with Covid apart from two non teaching staff and has had regular bubble closures. Think someone upthread (Si dH?) mentioned about areas of high Covid (who I refer to as the Usual Suspects) could hit herd immunity soon (think Bolton) and these areas could come out of Covid sooner. I have connections in Edinburgh - lots of Covid there, more than Glasgow, adds weight to this idea. And I know of a number of double jabbed people who have recently caught Covid. Twitchy bum time in many ways. Relative low vaccine take up as well (or vaccine supply is low), but good to see the govt saying ‘open house for jabbing’.  Hospitalisations and deaths going up locally and have been for a few weeks but, so far, not to worrying levels. 

40k is the middle of the road estimate from all the models I'm finding. That's about as much of a consensus as you're going to get. I think that's a realistic number for end of July. However it's critical to view that in the light of the (highlighted, but can't say it enough) likely disengagement with testing, so cases =/= infections by then. And as a result (also mentioned above but can't say it enough) the % hospitalisations will be misleading.

irt wintertree:

Still no appreciable funding going into GFBL research.....

I think I agree with your optimistic view that, here in the UK, we may be entering the herd immunity endgame.  However, I don't think this means restriction-free foreign travel any time soon without some international acceptance that fully-vaccinated (2 jabs plus three weeks) travellers pose a negligible risk of transmission. 

Most other countries are nowhere near the rates of vaccination required to be comfortable allowing visitors from anywhere with the levels of active infection that we are going to have for the next month at least.  If Angela Merkel gets her way there will soon be an EU-wide ban on travel from the UK - it remains to be seen whether a UK vaccine passport (such as you can currently access on the NHS app) would be acceptable to allow entry at all, never mind without quarantine and further testing.     

Considering Germany has just imposed a 14 quarantine period on fellow EU state member Portugal we have a long way to go before European foreign travel is really opened up.

Good point; we’re increasingly at odds to many other destinations on infection numbers, and they often have more vaccination to do.  Staling for time against this new variant is very much in their interests.

There was a notice on the dashboard yesterday saying some English cards data had been delayed but I still wasn’t expecting to see 23k reported today given it’s Monday. I imagine tomorrow will be in the low 20s as well. That’s broadly consistent with a doubling time of a couple of weeks but the absolute numbers are now getting chunky. Just imagine if we didn’t have the vaccines…

I don’t think your post earlier today was a rant, by the way. 

There have been various comments above about the absence of explanation to the public of what the government's plan is and what is the thinking behind letting cases increase as they are doing, and I agree this is bad.

I don't often get to hear Radio 4 between 5 and 6 but on my way to the wall this evening I listened to a very upbeat medic saying now is the best time, or the least worse time to release restrictions; and that the only way we can get herd immunity is for every child to get covid. Blimey I thought. Well well said the interviewer, logically that must mean not making schoolchildren isolate on contact with infection? Yes, maybe from September said the medic.

And then whaddya know, I look at the Gurdian when I get back from the wall and find that ministers are "set to end" isolation for schoolchildren.

Which is a lot of messaging in a short time, none of it directly from the government, none of it conveyed by anyone with responsibility for policy,  all couched in terms of opinion or of speculation ... 

Wowsers.  

Do you think the radio interviewee was someone genuinely independent giving their take on the situation, or another case of A/B testing of potential policy by careful (ab)use of the media?

The only time I listen to the radio is when driving, and I like to stay calm and collected when driving so I’ve stopped using the radio (I cracked when jeremy vine came on 6 months or so ago) and now have my antique 6-CD autochanger loaded up with:

• The Levellers - Levelling The Land
• Buffy the Vampire Slayer - The Album
• The Lost Boys Original Soundtrack
• A homemade soundtrack to 1967’s Casino Royal
• A home made rip of Wintertree Sr’s LP of Peter Ustinov’s “Grand Prix du Rock”
• The Cranberries - “Everybody Else is Doing It”

It’s better this way.

Edit: the soundtrack to Casino Royale (1967) was and still is astounding.  Can’t exactly say that about any other part of the movie…

Great to see the lissajous curves separating. If this continues, it will largely be a debate about whether letting cases get out of hand is a good idea or not. Well, it would be a debate here (with no clear answer) but it seems that government have made up their mind (in fact the number of case numbers was never a parameter for the official exit roadmap or whatever it’s called).

Totally agree on the messaging point.

One other point - Javid is considered less cautious than Hancock and this may well tilt the balance, though it probably didn’t need tilting anyway...

Anecdotal but on a work call today one person said they’d just recovered from Covid, someone else was coughing away but said it wasn’t Covid as they’d tested negative on LFD and a third person was mentioned as being off work with Covid. Definitely going round and if people were in the office it would be a lot worse. Also a good illustration of people not understanding what LFDs are for. 

LFDs as used here sit at a potent juncture of science and psychology IMO.

From my naive and data limited viewpoint I can’t form an opinion on if they’ve made the situation better or worse. (Not helped by the publicly available data not being sufficiently tagged by test type in any breakdown below national level, and then only for England).  Used appropriately they’re clearly a powerful tool, but a lot of my anecdotal experience suggests that, as used, they’re not well understood, with the potential for perverse consequences.

Wouldn't worry too much about getting to the bottom of it; it'll soon be moot when they're safe in their box gathering dust on the side of the bath while everyone's coughing their way down to the pub and getting their first taste of escalator handrail in 18 months.

I'm honestly surprised they've been used as much as they have given the disincentives for the individual. 

It seemed obvious to me yesterday that it was the latter, given the "set to relax" article in the Grauniad. And sure enough, early morning news on Radio 3 tells me this is Sajid Javid's new policy.

Rant coming: As for "no herd immunity till all children have had it" ... any hesitancy on the part of parents to get children vaccinated is now moot because *the plan* is for them all to get Covid. No need to worry about making choices. That's how I read it anyway.

    Restricting incoming travel from Delta hotspots for a country with insufficient levels of vaccination is a no-brainer. Every seeding event is potentially going to bring the date of restrictions forward by the doubling time, and seeding by travellers often occurs outside the normal geographic spread locations and straight into an area where people are being slack in their behaviour.  It's a race between vaccination and spread now, so reducing seeding is key.

Updated L-plots below.  

• To my reading they're all towards the less concerning range of possibilities.  
• A cautionary note however that the exponential rate is quickening over the last week or so in England, back towards a doubling time of ~8 days, so we expect the curves to bend in an optimistic direction temporarily due to the latency - it takes a while for the faster infection rate to translate in to faster rates for the other measures. 

Across the country, universities are reaching the end of their terms, mass outbreaks are happening with very few serious symptoms reported, and a wave of mass student travel back towards the households of older relatives is imminent.  This'll perhaps slow the growth rate in the university aged demographic and lead to a burst of transmission to single- and double- vaccinated parentals, as well as obvious consequences for relatives without any vaccination.  Previous terms haven't seen this to any clear standard, but things are a bit different this time around.  Some universities are offering to extend accommodation for free for isolation periods upon a positive test, so if you're concerned you might check this and then advise Jr to get a test before you pick them up.  

There will also be the 'meeting up with mates from home' once they're back too.

Vaccine uptake map launched on the dashboard. Now just need to have a toggle between that and the cases map... There’s certainly a correlation here in West Mids. 
 

https://coronavirus.data.gov.uk/details/interactive-map/vaccinations

Or, another way of looking at it, a "how vaguely do we know the size of the adult population" map.

Correlations between the maps are massively convoluted by age and also to some extent (linked to age) to population density. Younger adults are both more far more likely to get covid, more likely to live in cities and far less likely to have been vaccinated yet. So in itself the correlation doesn't really say anything about the effect of vaccination on spread I don't think... but you can get some of that from other data like demographic cases, as we've discussed in the past.

The data from London is a bit of a counter factual - given the residual cases that have been present there for a few weeks, the population density and the low vaccine take-up, it's amazing that growth hasn't been faster. Possibly an indication of residual immunity in young adults there from alpha's spread in December?

P.S. the guardian have been presenting a vaccination map like this for a couple of months, but they do it as a proportion of total local population including kids too.

If you made it a map of antibodies, London might look a little different.

I just edited my post in perfect time

Wow. That reveals my MSOA is 44.1/25.4% That's appalling. I thought the LSOA figures were bad enough for me to write to my MP, but those figures are far worse than the LSOA.

My MSOA includes the University, though, so that is the most likely problem...

Can anyone do worse...?

60/36.8, but if we lived the other side of the main road it'd be 47.5/17.7. For the same reason as you, I think.

The 'low first does uptake' appears well correlated with both urban areas and student areas. Amusingly, the 'high second dose uptake' seems to be a pretty good filter for finding nice areas to retire to (probably also areas without any good employment opportunities)

Fair point but there’s a definite trend. I know some people have moved out of the large cities but it won’t be enough to explain the differences… 

Fair points, it’s hard to disaggregate them various factors. I should have said there is a correlation but it’s not simply due to the vaccinated %. 

Just had a interesting thought.... Apologies if it's been covered already, but.... What happens to the ratio of hospital admissions to cases as we vaccinate younger people? It's going to start going up, right?? Like, it should go up, shouldn't it? If I've got things straight in my head.

Might not be an easy graph to explain, that one. Especially to people who don't want to understand it.

Good point. But the key thing is absolute numbers. The ratio is mostly relevant for modelling and you’d think the modellers would consider how the ratio changes over time and what that means. Also with fewer cases among younger people there would be less bleed through into the vulnerable population and hence fewer hospitalisations. What that does to the ratio is beyond me…

There is going to be some difficult-to-decipher data over the next couple of months I suspect.

Increased vaccination of young people along with building natural immunity levels should partially reduce cases in those age groups, but it'll be offset by increased cases due to further relaxations. That will also affect those age groups the most because they do the most mixing in places that are currently subject to social distancing rules. 

We are also soon reaching the point where immunity could potentially start to wane in the earliest groups to be vaccinated. So need to be looking out for a signal of that.

If I had to guess, I'd expect that all hospitalisations will rise with cases, the % in younger age groups will gradually drop back, but overall occupancy will stay manageable and deaths stay low, if the current hospitalisation trends roughly continue. But things could change and it'll be hard to understand and explain the causes.

Yes and yes. Thing is, a lot of people have been pointing at that ratio coming down and saying "this means vaccines are working". When it starts to go up it's not gonna be trivial to explain in an accessible way why that still means vaccines are working.

I went to a concert in London yesterday evening, preceded by dinner in a restaurant near King's Cross. It is all very bizarre - mostly there is just no sign of a pandemic, apart from masks and distancing on the train and at the concert. 

Some of the rules are starting to look silly: singers on a platform, no masks ... then two or three of them put masks on, step two or three yards to change places and take their masks off again to sing the next item. The conductor doesn't wear a mask but is, as conductors do, constantly moving about.

(This was music that isn't heard at climbing walls and that even I have to admit would not work at a climbing wall).

https://www.theguardian.com/world/2021/jul/02/we-are-a-petri-dish-world-wat...   

I was reading that this morning. Lots of resonance with discussion here.

Yes, leading the charge once again.  Sadly I don't think attention is going to remain on us for very much longer.

In reply to Longsufferingropeholder:

I think it's incredibly hard to predict how the behaviour of the data is going to proceed, especially considering the oft discussed possibility of many people disengaging from testing. It's going to be an absolute minefield subject to accidental misinterpretation and deliberate misrepresentation.  Misrepresenting the situation here could become key to people playing silly buggers in the USA, for example.  We've seen it happen before with emotive issues around healthcare and I except we'll see it again.

In reply to BusyLizzie:

It really is, and getting weirder by the day.  Perhaps the guidance issues after the end of school term will normalise things a bit.

A rather meta plot this, it shows the number of views each plotting thread has seen (black) and the average daily cases for that week (red).

Some whimsical interpretation:

• These threads started after the rise in November was well under way. Interest rose and then fell with that wave.
• Interest then rose again around May 2021 when cases bottomed out and it became apparent that we were returning to a rising phase.  
  • As this became inevitable and clear in the data, interest faded.  
• Interest seems to be returning now, and these are indeed interesting times.  
  • Assuming the wheels don't come off in the next couple of weeks, interest is probably going to fade.  

(The last black data point is provisional)

I’m not a human SPSS, but I suspect your 2 sample Kolmogorov Smirnov is insignificant..

Which of the two aspects do you deny?  Enquiring minds want to know…

I’m torn between two retorts:

1. Stop making words up.
2. Lots of hidden parameters you see.

Less flippantly, it’s staggering in how many disconnected professional areas  the work of Kolmogorov pops up.

(excess flippancy may be due to after effects of my second dose)

I’m several human SPSSs.

I’m also on my second dose .. of laphroaig.. mmmm

‘Lots of hidden parameters you see’ initially reads as tautology, but with my cryptic hat on..

the bells, the bells?

Quantum virology.  It’s the future.