Coronavirus in the UK - news, lockdown and discussion
- Thread starter Sasaferrato
- Start date
Lockdown 1 came in early enough to stop it getting a toe hold and Lockdown 2 was too late for that?
elbows
Well-Known Member
Hve you managed to get your head round death increase curves more than when you said this around a month ago?
The UK deaths by date of death number is somewhat stable for the date in question now, October 26th, so now is the time for me to bring this subject up for the last time. Currently the figure for that date is 274.
So if that models number had been for the whole of the UK, it would have been within the range of that particular run of their model. But as I've mentioned previously, their numbers were actually for England only, and the very next time they shared their results they had downgraded their prediction. The current figure for that date for England is 220. So their older model runs range of 241-686 for that date was not met. The real figure was a better fit for their next published model run, which had a range of 178-344 for England for that date.
Anyway I'm not planning to talk about subsequent runs of that model as I'd rather use real data at this point, but they did do another update today if anyone is interested https://joshuablake.github.io/public-RTM-reports/iframe.html
littlebabyjesus
one of Maxwell's demons
It's certainly got bad rather quickly, as it has in Spain and France. I actually thought the model was for the UK so assumed they had just sneaked in right at the bottom. But I still maintain that they got something badly wrong to miss the number while giving themselves such an enormous range. I don't think I said what my guess would be just on the 'tracking Spain/France' criterion, but it would have been around 160-200 (for the whole UK), so I would have been significantly under.
One thing that I hadn't factored in then was the difference in hospitalisation figures in different countries. That probably led me to think that the UK wasn't quite as bad as it was as both France and Spain seem to measure a bit differently and get higher figures.
elbows
Well-Known Member
I think the fundamental problem in most of these models is how simple they are. And I think the number of deaths that they know already happened is the main thing they feed in, and this makes their prediction ranges especially broad and lacking in confidence when the period they have data for doesnt include a very large number of daily deaths yet. And when we comment on such models we normally have a good number of days additional data compared to what they used. Every subsequent run of that model that has been published gave results that should be much less hard for you to believe plausible than the earlier one we looked at, an extra week or so's data made a big difference to their results.
I suspect one of the differences between our hospital data and some other countries may be that they have admissions policies (and discharge policies/policies on when someone stops counting as a covid patient) that differ to ours. I know I said that Frances was especially high relative to some others when we discussed this before, but when I added intensive care cases to the figures from Italy, which it turns out were not a part of the hospital numbers for Italy I was previously using, they hit much more similar levels to that of France.
Its tricky trying to predict by sight and feel. Because these things are curves but with periods of growth that look more like straight lines than curves, and its hard to say exactly when trajectories will change so that the overall trend ends up a curve rather than a straight line. Your 150 prediction was reasonable if imagining a straight line of growth upwards from the period when you made the comment, but the trajectory got steeper, leading to a large gap between your prediction and what actually happened by October 26th. France at the time in earlier October wasnt the best guide because things were only just really getting going there (in terms of death), and Spains numbers were complicated by additional data issues and lag, and maybe also because the timing of Madrid was ahead of the other regions and started to plateau in a manner that affected the overall national figures. The hospital figures for Italy have really exploded, up to similar levels as France, over 30,000, so I really should start looking at their death figures. I should also compare intensive care numbers this time to last time, for countries where I have enough data. Generally expect to see that where hospitalisation levels are matching or exceeding thsoe of the first wave, intensive care numbers are not close to those seen the first time round.
Last edited:
elbows
Well-Known Member
I wouldnt rely too much on number of cases testing positive in the area when trying to compare this wave to the first wave in any particular location. Primarily because the testing was very limited the first time around, the figures from the first wave are a tiny fraction of the full picture of that time.
I would rely quite heavily on daily death figures and hospital data to judge. I havent looked at Hulls hospital levels yet, but I have looked into deaths there....
Unfortunately the media cant be relied upon to do that bit properly, leading to stories like this one:
Hull's deadliest day so far as coronavirus ravages city
'These aren’t just statistics. The numbers represent real people who are seriously ill'
That reporting is a mess, focussing on totals for individual days rather than the broader picture. And focussing on deaths by day they were announced. And using the NHS England death data per hospital trust instead of the broader measure of deaths per area that can be found on the official dashboard.
So what they are actually describing is the spike in this picture via the NHS data, for Hull University Teaching Hospitals NHS Trust:
And that does offer some guide as to how to think about this wave in Hull so far compared to the last one.
The one I would favour using from the dashboard is this one, because it deals with where people live rather than hospital trusts. Very similar pattern but slightly different peaks for newspapers to make stories out of.
I wouldnt rely too much on number of cases testing positive in the area when trying to compare this wave to the first wave in any particular location. Primarily because the testing was very limited the first time around, the figures from the first wave are a tiny fraction of the full picture of that time.
I would rely quite heavily on daily death figures and hospital data to judge. I havent looked at Hulls hospital levels yet, but I have looked into deaths there....
Unfortunately the media cant be relied upon to do that bit properly, leading to stories like this one:
Hull's deadliest day so far as coronavirus ravages city
'These aren’t just statistics. The numbers represent real people who are seriously ill'www.hulldailymail.co.uk
That reporting is a mess, focussing on totals for individual days rather than the broader picture. And focussing on deaths by day they were announced. And using the NHS England death data per hospital trust instead of the broader measure of deaths per area that can be found on the official dashboard.
So what they are actually describing is the spike in this picture via the NHS data, for Hull University Teaching Hospitals NHS Trust:
View attachment 238454
And that does offer some guide as to how to think about this wave in Hull so far compared to the last one.
The one I would favour using from the dashboard is this one, because it deals with where people live rather than hospital trusts. Very similar pattern but slightly different peaks for newspapers to make stories out of.
View attachment 238455
You don't need to tell me the HDM is shite!
Tbf I've been looking mainly at case rather than death figures and obviously they're not really comparable now with the spring. Nevertheless, watching the figures rise inexorably in a city that's had low numbers for a long time has been alarming, and anecdotally I know of a lot more people who've either had symptoms or been told to self-isolate than before.
elbows
Well-Known Member
Alternatively, the very difference pace so far in London means that admissions standards and public attitudes towards seeking care are so far different in this wave and more milder cases are being seen in the healthcare system than the first time around.
Since I posted those charts for Hull I suppose I mayas well post a few others from the government dashboard to show a few more examples. Partly because I dont know as people have really been looking at this stuff much.
An obvious example to choose, Liverpool:
An example of an area experiencing a deadly 2nd wave that probably hasnt had much attention because its in a region that is usually considered to not be at big risk, and numbers overall have been low in the first wave and the second wave so far. Bournmouth, Christchurch and Poole. But I post it because it could help bust some myths and impressions people hold about the second waves spread so far:
A different sort of example, Birmingham:
As usual with me these are deaths by date of death and so the figures shown for more recent days are subject to future increases.
elbows
Well-Known Member
Now then, when it comes to studying the two waves, it seems to me that in addition to drowning in graphs, I should really look at totals for each wave in order to be able to correctly emphasise that its the area under the graph that counts, and how lower peaks can still produce more deaths over time if they carry on for too long.
So the question is, what is a reasonable date with which to seperate the first wave and second wave death statistics? July? 1st August? 1st September? I want a few opinions on that before I start doing it for national, regional and local death numbers.
So the question is, what is a reasonable date with which to seperate the first wave and second wave death statistics? July? 1st August? 1st September? I want a few opinions on that before I start doing it for national, regional and local death numbers.
StoneRoad
heckling from the back!
1st August looks a good break point between the two waves. as I think July had the last of the first wave deaths.
elbows
Well-Known Member
Cheers for the feedback. Having poked around with the data a bit it doesnt make a huge difference so I may as well be conservative about it and start from September 1st. In which case using the 28-days positive test daily deaths by date of death UK data from the dashboard, I get 41,544 deaths before 1st September and 8,817 since. I would normally rather use ONS data as it was more complete in the first wave, but its lagging behind a lot when trying to measure the 2nd wave so far in a reasonably timely manner, so I probably wont. But just to give a one off indication of ONS levels in the 2 waves, by date of death it has 57,546 before September 1st and 5,459 after, with no November deaths included yet and likely more late October ones to add.
Sunray
Its sunny somewhere.
I think the takeaway message from the two graphs is:-
If there is one thing this government has done (generally they are doughnuts) is to put down massively expensive punts on possible vaccines. If a gov is going to gamble big on anything, surely this is one thing worth gambling on. Some might fail, but it appears to have paid off big time. We have 30 million doses of the Pfizer 90% effective (will be revised down) vaccination on order. Double dose, so 15 million people.
- We are testing a lot more people now, so I suspect the 1st wave and the second are similar in size, the 1st perhaps being bigger than the 2nd.
- Health professionals have been talking and are much better at keeping the very sick alive.
If there is one thing this government has done (generally they are doughnuts) is to put down massively expensive punts on possible vaccines. If a gov is going to gamble big on anything, surely this is one thing worth gambling on. Some might fail, but it appears to have paid off big time. We have 30 million doses of the Pfizer 90% effective (will be revised down) vaccination on order. Double dose, so 15 million people.
Last edited:
elbows
Well-Known Member
Somebody tell Nick Triggle that London has gone missing from one of the graphics in his article.
Covid: How the UK reached 50,000 virus deaths
The death toll has hit a grim milestone. But behind the figure, there's a complex picture.
www.bbc.co.uk
mx wcfc
Well-Known Member
There's a big difference between ordering 30 (I thought it was 40) million doses and getting them.
If Pfizer is manufacturing in the US, I can't see us getting any whilst Trump is in charge. Maybe not even under Biden.
Supine
Newt Member
It is also being manufactured in Germany and Belgium for EU supply. Remember it's a BioNTech development, so not just a Pfizer US product.
And if we leave the EU without a deal, maybe getting it from the EU might be a problem too
Sunray
Its sunny somewhere.
Let's consider anti vaxx people, 1st. If you are one of those, die collectively. Be a marter. You can even cough your very last cough in my face if you like. * shrug *
Ordering from a big company like Pfizer* is a reasonable gamble, IMO. Life is a risk. Pfizer started manufacturing the vaccination way before they knew it even worked on the basis if it was even slightly effective, it would be approved.
We have 15 million doses.
*Pfizer Market Cap: 213.94B for Nov. 11, 2020
Ordering from a big company like Pfizer* is a reasonable gamble, IMO. Life is a risk. Pfizer started manufacturing the vaccination way before they knew it even worked on the basis if it was even slightly effective, it would be approved.
We have 15 million doses.
*Pfizer Market Cap: 213.94B for Nov. 11, 2020
SpookyFrank
A cheap source of teeth for aquarium gravel
All this talk of a vaccine being rolled out tomorrow morning is unlikely to help with lockdown compliance levels IMO.
weepiper
I fix the machines that fight climate change
Scotland's reputation as the sick man of Europe might have to be revised.Somebody tell Nick Triggle that London has gone missing from one of the graphics in his article.
Covid: How the UK reached 50,000 virus deaths
The death toll has hit a grim milestone. But behind the figure, there's a complex picture.www.bbc.co.uk
View attachment 238466
mauvais
on reddit or something
Good spot - raised and fixedSomebody tell Nick Triggle that London has gone missing from one of the graphics in his article.
Covid: How the UK reached 50,000 virus deaths
The death toll has hit a grim milestone. But behind the figure, there's a complex picture.www.bbc.co.uk
View attachment 238466
Thought people might like to know some figures I've just got back from a traffic volume survey I did last week, which straddled the new restrictions coming in on Thursday - one site was in Malvern, the other in Knowsley, and both showed a drop in vehicle movements of almost exactly 10%
I included it within the worldwide stats thread, but this includes this table:
| Table 3: Mobility and Reproduction numbers | ||||||
| Decline in mobility in first wave (%-pt) | Decline in R in first wave | Impact on R of a 10%-pt decline in mobility | Latest peak in R | Recent decline in mobility (%-pt) | Implied latest reading for R | |
| Germany | -63 | -2.2 | -0.35 | 1.49 | -27 | 0.55 |
| Spain | -92 | -2.9 | -0.32 | 1.15 | -25 | 0.35 |
| UK | -81 | -3.4 | -0.42 | 1.36 | -20 | 0.52 |
| US | -49 | -1.7 | -0.35 | 1.30 | -4 | 1.16 |
| Source: J.P. Morgan. 'R' refers to the reproduction number and is based on hospitalization data. | ||||||
The mobility rate comes from the Google Mobility Index, which is the graph below.
I have to say that I'm not sure based on that graph where they are getting a 20% recent decline in mobility from. It looks to me that it has dropped from approximately -25% of baseline to -35%, or .65/.75, which is more like a 15% decline. Hard to tell from just inspecting the graph, though. Either way, your 10% seems like a reasonable ballpark that agrees with the mobility index.
Just walked past the McDonalds in Angel. Very, very busy with teens ordering take-away. An absolute free for all and obvs no masks or distancing. Oh hum.
I really think it would have been better if we as a country at least tried to get the kids on board with it all. For a long time now they've suspected this virus is nothing to do with them and the decision to shut vast swathes of our economy just to keep their plague pits open just confirmed it to them. Not about blame because I would have been the same and its their future we're currently mortgaging to the skies.
I really think it would have been better if we as a country at least tried to get the kids on board with it all. For a long time now they've suspected this virus is nothing to do with them and the decision to shut vast swathes of our economy just to keep their plague pits open just confirmed it to them. Not about blame because I would have been the same and its their future we're currently mortgaging to the skies.
not a trot
Well-Known Member
Todays numbers looking grim.
cupid_stunt
Chief seagull hater & farmerbarleymow's nemesis.
Fucking hell, 33,470 new cases, lockdown v2 is working well.
And, another
EDIT - corrected figure, I hadn't spotted the error on the government's dashboard.
Last edited:
The death rate is going to be bad for a while, we knew that. They key is whether infections are slowing as has been hoped. That r thing again.
emanymton
A cat politely sat on the flaming gardener.
There was never going to be much impact from it. People's work patterns have basically settled into whatever they are going to be.Fucking hell, 33,470 new cases, lockdown v2 is working well.
And, another 595 deaths, up from last Thursday's 378.
Schools are still open
And rules on no household mixing are routinely ignored.
Frankley I didn't see much point is restricting pubs past what was in place for level 3 and closing a few shops. It might slow the growth a bit but not much more.
Similar threads
