Donate to Remove ads

Got a credit card? use our Credit Card & Finance Calculators

Thanks to johnstevens77,Bhoddhisatva,scotia,Anonymous,Cornytiv34, for Donating to support the site

Coronavirus - Modelling Aspects Only

The home for all non-political Coronavirus (Covid-19) discussions on The Lemon Fool
Forum rules
This is the home for all non-political Coronavirus (Covid-19) discussions on The Lemon Fool
scotia
Lemon Quarter
Posts: 3561
Joined: November 4th, 2016, 8:43 pm
Has thanked: 2371 times
Been thanked: 1943 times

Re: Coronavirus - Modelling Aspects Only

#450891

Postby scotia » October 17th, 2021, 9:02 pm

Another 14 days of data - with the same model parameters as in the preceding submission.
First we have England - projected deaths 0.14 of admissions 11 days previous.

Image

The reported deaths (blue) are now running slightly above the projected deaths (red) - but not by significantly more than the standard deviations (vertical bars). Both the deaths and admissions are now increasing slowly. Probably to be expected by the reduction of constraints.

And now for Scotland - projected deaths 0.14 of admissions 12 days previous.

Image

The statistics are considerably poorer - but even so, it looks like the reported deaths (blue) are now levelling off significantly higher than the projected deaths (red). And once again, the actual Scottish death rate significantly exceeds the English death rate. From the above, the current weekly ratio of reported Covid-19 deaths in Scotland are around 1/5 of those in England, whereas from population sizes you would expect 1/10.
I know that the age demographics are different - but are they so different as to create such a difference in Covid-19 death rates?

Hallucigenia
Lemon Quarter
Posts: 2618
Joined: November 5th, 2016, 3:03 am
Has thanked: 166 times
Been thanked: 1718 times

Re: Coronavirus - Modelling Aspects Only

#450908

Postby Hallucigenia » October 18th, 2021, 12:13 am

scotia wrote:I know that the age demographics are different - but are they so different as to create such a difference in Covid-19 death rates?


This talks a bit about the differences and may be of interest :
https://twitter.com/jburnmurdoch/status ... 2207239176

scotia
Lemon Quarter
Posts: 3561
Joined: November 4th, 2016, 8:43 pm
Has thanked: 2371 times
Been thanked: 1943 times

Re: Coronavirus - Modelling Aspects Only

#450961

Postby scotia » October 18th, 2021, 11:22 am

Hallucigenia wrote:
scotia wrote:I know that the age demographics are different - but are they so different as to create such a difference in Covid-19 death rates?


This talks a bit about the differences and may be of interest :
https://twitter.com/jburnmurdoch/status ... 2207239176

Thanks for the link. What is doubly puzzling is that mask wearing in public places is still mandatory in Scotland - and is observed. Its possible that the cooler weather in Scotland leads to more indoor contact - with possibly a heavier transmitted dose of the infective virus. As one who is approaching the "very elderly" state, I'll be happy to get my booster at the end of this week.

jfgw
Lemon Quarter
Posts: 2539
Joined: November 4th, 2016, 3:36 pm
Has thanked: 1097 times
Been thanked: 1146 times

Re: Coronavirus - Modelling Aspects Only

#452087

Postby jfgw » October 21st, 2021, 6:55 pm

New cases in the South West (teal with crosses) are rocketing,
Image
My graph. Contains public sector information licensed under the Open Government Licence v3.0. : https://www.nationalarchives.gov.uk/doc/open-government-licence/version/3/

Mostly schoolchildren, possibly followed by parents. Higher age groups are showing signs of an upcoming peak, but I fear that cases may rise faster again as these people catch it from the lower age groups.
Image
My graph. Contains public sector information licensed under the Open Government Licence v3.0. : https://www.nationalarchives.gov.uk/doc/open-government-licence/version/3/

Compare this with England as a whole,
Image
My graph. Contains public sector information licensed under the Open Government Licence v3.0. : https://www.nationalarchives.gov.uk/doc/open-government-licence/version/3/


Julian F. G. W.

zico
Lemon Quarter
Posts: 2139
Joined: November 4th, 2016, 12:12 pm
Has thanked: 1074 times
Been thanked: 1086 times

Re: Coronavirus - Modelling Aspects Only

#452561

Postby zico » October 23rd, 2021, 9:26 pm

scotia wrote:
Hallucigenia wrote:
scotia wrote:I know that the age demographics are different - but are they so different as to create such a difference in Covid-19 death rates?


This talks a bit about the differences and may be of interest :
https://twitter.com/jburnmurdoch/status ... 2207239176

Thanks for the link. What is doubly puzzling is that mask wearing in public places is still mandatory in Scotland - and is observed. Its possible that the cooler weather in Scotland leads to more indoor contact - with possibly a heavier transmitted dose of the infective virus. As one who is approaching the "very elderly" state, I'll be happy to get my booster at the end of this week.


I think the linked piece explains it by saying that 'risky' indoor events are far more common in UK than Europe, so even wearing masks doesn't help that much when lots of people are in close proximity for hours.

scotia
Lemon Quarter
Posts: 3561
Joined: November 4th, 2016, 8:43 pm
Has thanked: 2371 times
Been thanked: 1943 times

Re: Coronavirus - Modelling Aspects Only

#454685

Postby scotia » November 1st, 2021, 12:14 am

Another 14 days of data - with the same model parameters as in the preceding submission.
First we have England - projected deaths 0.14 of admissions 11 days previous.

Image

The reported deaths are still reasonably following the predictions based on hospitalised admissions of the preceding 11 days - although the multiplier of 0.14 is now on the low side for recent data - and 0.15 provides a better fit. It appears that the admissions are levelling out - and hopefully the deaths will follow. But there is no evidence of a falling rate.

And now to Scotland (with much poorer statistical accuracy) -projected deaths 0.14 of admissions 12 days previous.

Image

The published death rate remains above the falling projections, although an optimist may say that they are now coming back into line. On a deaths per head of population basis Scotland still remains above England - but the gap has substantially narrowed to around 30% - with Scotland down, and England up. Could this effect be related to the earlier return to school in Scotland?

I have been looking at the recent (12th and 13th October) model results presented to the Sage committee by Warwick University and Imperial College. These cover Autumn 2021 and Winter 2022, with the Warwick predictions extending to June 2022. Under optimistic assumptions, Imperial College predict that levels will remain similar to, or possibly lower than, current. However they stress that in more pessimistic assumptions they foresee a substantial wave of infections, hospitalisations and deaths. Warwick University comes to a similar conclusion - and they consider that a decline in vaccine efficacy would be a substantial driving force in a new wave. They suggest extending boosters to lower age groups.

Both reports can be accessed from the Sage site - https://www.gov.uk/government/organisations/scientific-advisory-group-for-emergencies

scotia
Lemon Quarter
Posts: 3561
Joined: November 4th, 2016, 8:43 pm
Has thanked: 2371 times
Been thanked: 1943 times

Re: Coronavirus - Modelling Aspects Only

#458049

Postby scotia » November 14th, 2021, 11:02 pm

Another 14 days of data - with the model parameters slightly adjusted from those in the preceding submission.
First we have England - projected deaths 0.15 of admissions 11 days previous. The 0.15 is an increase from 0.14 - and gives a better fit to the more recent data.

Image

It looks like the admissions may now be on a modest downward trend, with the deaths following. I'm hoping that at the end of the next fortnight the booster vaccinations for the elderly should start to show in further reductions, in both the number and the ratio

And back to Scotland - where I have also increased the ratio to 0.15.

Image

As expected, the statistical accuracy of the data is poor, but we are also seeing a continuing downward trend.

Comparing the Scottish and English data as displayed over the 60 day period, the English numbers have increased, whereas the Scottish numbers have substantially decreased. The Scottish deaths are now back down to a level (96 for the current week) which is reasonably compatible with those in England (897 for the current week), given the 1 to 10 (approx) population ratio. Living in Scotland, I'm relieved to see our reduction - since it was worrying to see it higher than the (population adjusted) English number by almost a factor of 2 in mid October.

scotia
Lemon Quarter
Posts: 3561
Joined: November 4th, 2016, 8:43 pm
Has thanked: 2371 times
Been thanked: 1943 times

Re: Coronavirus - Modelling Aspects Only

#461667

Postby scotia » November 28th, 2021, 6:33 pm

Another 14 days of data - with the model parameters remaining unchanged from those in the preceding submission.
First we have England - projected deaths 0.15 of admissions 11 days previous.

Image

Both admissions (projected deaths) and actual reported deaths are on a downward trend. And the actual deaths now seem to be falling below the projected deaths - possibly due to the Boosters? Encouraging.

And in Scotland, with projected deaths 0.15 of admissions of 12 days previous we have

Image

Again encouraging - with a downward trend, and the actual deaths now seem to be falling from above to below the projected deaths.

scotia
Lemon Quarter
Posts: 3561
Joined: November 4th, 2016, 8:43 pm
Has thanked: 2371 times
Been thanked: 1943 times

Re: Coronavirus - Modelling Aspects Only

#465332

Postby scotia » December 12th, 2021, 6:38 pm

Another 14 days of data - with the model parameters remaining unchanged from those in the preceding submission.
First we have England - projected deaths 0.15 of admissions 11 days previous.

Image

The actual deaths have levelled out, and the projected deaths have started a modest climb.
I suppose as Christmas approaches, with shopping, and group lunches, it is to be expected.

And Scotland - projected deaths 0.15 of admissions 12 days previous.

Image

The statistics are poorer, but it looks like the projected deaths are continuing to fall, with the actual deaths wobbling around the projections. Possibly slightly better than the English performance? Although the actual ratio of English to Scottish deaths is a reasonable match to the population ratios.

I don't expect the Omicron infections have yet had any significant effect on these results, although I suspect that in 14 days time (Boxing day) the number of Omicron infections will be significant - but what effect this will have on admissions and deaths is as yet unknown. The reporting of data will also be subject to significant fluctuations over the holiday period - so it may be difficult to get a clearer picture until well into January.

funduffer
Lemon Quarter
Posts: 1328
Joined: November 4th, 2016, 12:11 pm
Has thanked: 122 times
Been thanked: 831 times

Re: Coronavirus - Modelling Aspects Only

#465634

Postby funduffer » December 14th, 2021, 8:14 am

scotia wrote:I don't expect the Omicron infections have yet had any significant effect on these results, although I suspect that in 14 days time (Boxing day) the number of Omicron infections will be significant - but what effect this will have on admissions and deaths is as yet unknown. The reporting of data will also be subject to significant fluctuations over the holiday period - so it may be difficult to get a clearer picture until well into January.


Indeed.

If Omicron is less lethal than Delta, then we might expect to see your 0.15 factor reduce as Omicron becomes more prevalent.

On the other hand, if it is no different, then we are potentially in big trouble.

FD

Hallucigenia
Lemon Quarter
Posts: 2618
Joined: November 5th, 2016, 3:03 am
Has thanked: 166 times
Been thanked: 1718 times

Re: Coronavirus - Modelling Aspects Only

#465647

Postby Hallucigenia » December 14th, 2021, 9:26 am

scotia wrote:I don't expect the Omicron infections have yet had any significant effect on these results, although I suspect that in 14 days time (Boxing day) the number of Omicron infections will be significant - but what effect this will have on admissions and deaths is as yet unknown.


Admissions are already taking off in London, which is where omicron first took hold :
Image

Also worth noting the Danish data which has the age profile of omicron infectees, it's mostly in 20-somethings at present, which may affect apparent severity :
https://covid19.ssi.dk/virusvarianter/omikron

Data fans should also be aware of the South African dashboard, which is a pain to link to as it's all clever and dynamic, but if you full-screen it and select Gauteng it's quite interesting :
https://www.nicd.ac.za/diseases-a-z-ind ... ov-report/

Cases in Gauteng are levelling off, albeit at a level about 50x where they were a month ago, this site tries to model R across South Africa :
https://unsupervised.online/static/covi ... 54_Gauteng
Image

servodude
Lemon Half
Posts: 8271
Joined: November 8th, 2016, 5:56 am
Has thanked: 4435 times
Been thanked: 3564 times

Re: Coronavirus - Modelling Aspects Only

#465689

Postby servodude » December 14th, 2021, 11:19 am

Hallucigenia wrote:Cases in Gauteng are levelling off, albeit at a level about 50x where they were a month ago


That's one of those rollercoaster sentences ;)

-sd

zico
Lemon Quarter
Posts: 2139
Joined: November 4th, 2016, 12:12 pm
Has thanked: 1074 times
Been thanked: 1086 times

Re: Coronavirus - Modelling Aspects Only

#466473

Postby zico » December 16th, 2021, 6:42 pm

Some odd figures on hospitalisation by regions. I don't have a link to the graph, but it shows a very steep increase in London hospital admissions (about 60%) between 24th Nov - 8th Dec. Other regions are fairly flat in a similar timeframe. I can't find corresponding figures for infections at a regional level.

Another point to bear in mind with daily infections figures is that pretty soon the UK is going to run out of test capacity, probably both testing kits and people able to test them, so the exponential growth will appear to be declining, when really it'll just be that not everyone can be counted.

On the exponential growth itself, I'd expect this to start to slow down after Whitty's comments yesterday, because many people will start to be more cautious so they can be clear of infection in time for Xmas family gatherings.

On the plus side, it appears that infections are growing massively amongst younger people, who are far less at risk. The vulnerable people (unhealthy and/or older) have mostly got boosted, so there will be a massive number of younger generation infections, hopefully dashing against a strong protective wall of highly protected older people.

jfgw
Lemon Quarter
Posts: 2539
Joined: November 4th, 2016, 3:36 pm
Has thanked: 1097 times
Been thanked: 1146 times

Re: Coronavirus - Modelling Aspects Only

#466500

Postby jfgw » December 16th, 2021, 8:09 pm

zico wrote:Some odd figures on hospitalisation by regions...


Admissions data taken from the NHS website, https://www.england.nhs.uk/statistics/statistical-work-areas/covid-19-hospital-activity/, Daily Admissions and Beds spreadsheet, https://www.england.nhs.uk/statistics/wp-content/uploads/sites/2/2021/12/COVID-19-daily-admissions-and-beds-20211216.xlsx. Regions with the same name do not cover the same areas.
Image
My graph. Contains public sector information licensed under the Open Government Licence v3.0. : https://www.nationalarchives.gov.uk/doc/open-government-licence/version/3/

Cases data downloaded from the government website, https://coronavirus.data.gov.uk/details/cases
Image
My graph. Contains public sector information licensed under the Open Government Licence v3.0. : https://www.nationalarchives.gov.uk/doc/open-government-licence/version/3/

Age data from the government website (heatmap data):

Compare England with London, Those 20—39 Londoners have been going a bit wild.
Image
My graph. Contains public sector information licensed under the Open Government Licence v3.0. : https://www.nationalarchives.gov.uk/doc/open-government-licence/version/3/

Image
My graph. Contains public sector information licensed under the Open Government Licence v3.0. : https://www.nationalarchives.gov.uk/doc/open-government-licence/version/3/


Julian F. G. W.

zico
Lemon Quarter
Posts: 2139
Joined: November 4th, 2016, 12:12 pm
Has thanked: 1074 times
Been thanked: 1086 times

Re: Coronavirus - Modelling Aspects Only

#466508

Postby zico » December 16th, 2021, 8:17 pm

Thanks for finding the link to the charts I mentioned.
It almost looks as if London has zero lag time between infections and hospitalisation, which surely can't be right. Maybe for Omicron, people don't think they have Covid initially, and only get tested when symptoms start - but that might simply be trying to find a theory to explain the odd data.

scotia
Lemon Quarter
Posts: 3561
Joined: November 4th, 2016, 8:43 pm
Has thanked: 2371 times
Been thanked: 1943 times

Re: Coronavirus - Modelling Aspects Only

#467456

Postby scotia » December 20th, 2021, 1:35 am

Another 7 days of data - with the model parameters remaining unchanged from those in the preceding submission.
My two excuses for not waiting the usual 14 days are (1) curiosity and (2) next week's data will be skewed by holidays
First we have England - projected deaths 0.15 of admissions 11 days previous.

Image

The projected deaths (based on admissions) continue to climb at a modest rate, however the actual deaths seem to be following a slow downward trend. Encouraging. But why this divergence? A guess might be that the boosters to the most vulnerable (us oldies ) are having an effect. Or are we seeing some reduced death rate due to Omicron? I think its too early to say. I'll stick by my earlier statement that it may be well into January before we get a clearer view.

And now for Scotland - projected deaths 0.15 of admissions 12 days previous.

Image

Given the poor statistical accuracy I can only say that both the actual and projected deaths seem to be on a wobbling downward trend.

Although Omicron does not appear to feature strongly in the above (English and Scottish) admissions and deaths data, It is becoming obvious that Omicron is resulting in fast increases of infections, leading to substantial increases in quarantined NHS staff. This may prove to be a serious factor in future data.

servodude
Lemon Half
Posts: 8271
Joined: November 8th, 2016, 5:56 am
Has thanked: 4435 times
Been thanked: 3564 times

Re: Coronavirus - Modelling Aspects Only

#467458

Postby servodude » December 20th, 2021, 1:43 am

scotia wrote:Although Omicron does not appear to feature strongly in the above (English and Scottish) admissions and deaths data, It is becoming obvious that Omicron is resulting in fast increases of infections, leading to substantial increases in quarantined NHS staff. This may prove to be a serious factor in future data.


Indeed!
It feels like the secondary effects from this strain passing through will be material

I concur that we're probably not going to have a decent picture of what's going on until January (and probably towards the end of it - unless there's an obvious overwhelming trend)
That will be down to the disruption to data due to the holidays, the changes that they make on behaviour and that fact that we still need more data to pick out the effects of Omicron

- sd

zico
Lemon Quarter
Posts: 2139
Joined: November 4th, 2016, 12:12 pm
Has thanked: 1074 times
Been thanked: 1086 times

Re: Coronavirus - Modelling Aspects Only

#467937

Postby zico » December 21st, 2021, 10:58 pm

People are talking about Omicron having exponential growth, and that it'll soon run out of people to infect, but it's surely a lot more complicated than that.
Seems to me there are lots of sub-populations.
1. Club/pub crowd. People who aren't taking precautions -maybe 20million (including their families). Omicron spreads very easily, so will infect all these people, via exponential growth, until most of them have been infected.
2. Careful people, but who need to use public trans and/or have public-facing jobs. Infections in this group should rise more slowly, probably over a much longer period of time.
3. Careful people who can avoid the public (Probably most people on Lemonfool). It'll take ages to infect this group, and it may never happen if the Omicron wave has passed through and died out.
Problem is, here immunity % is much higher if transmitability is much higher.

Seems to me likeliest scenario may be big wave of 'careless' infections, followed by gradual low level cases as virtuality everyone gradually gets infected over a year or two.
How does this above logic sound?

scotia
Lemon Quarter
Posts: 3561
Joined: November 4th, 2016, 8:43 pm
Has thanked: 2371 times
Been thanked: 1943 times

Re: Coronavirus - Modelling Aspects Only

#467998

Postby scotia » December 22nd, 2021, 11:08 am

zico wrote:People are talking about Omicron having exponential growth, and that it'll soon run out of people to infect, but it's surely a lot more complicated than that.
Seems to me there are lots of sub-populations.
1. Club/pub crowd. People who aren't taking precautions -maybe 20million (including their families). Omicron spreads very easily, so will infect all these people, via exponential growth, until most of them have been infected.
2. Careful people, but who need to use public trans and/or have public-facing jobs. Infections in this group should rise more slowly, probably over a much longer period of time.
3. Careful people who can avoid the public (Probably most people on Lemonfool). It'll take ages to infect this group, and it may never happen if the Omicron wave has passed through and died out.
Problem is, here immunity % is much higher if transmitability is much higher.

Seems to me likeliest scenario may be big wave of 'careless' infections, followed by gradual low level cases as virtuality everyone gradually gets infected over a year or two.
How does this above logic sound?

Add another group to category 1. Parents with young (in the un-vaccinated age group) children in school. They are little virus factories - and although it will only have a minor affect on them, it will be passed on to their parents. And short of closing the schools, I don't see a way round that. There is also a group of grandparents who normally pick up the kids from school, until the parents return from work - although working from home may reduce that cohort.

vrdiver
Lemon Quarter
Posts: 2574
Joined: November 5th, 2016, 2:22 am
Has thanked: 552 times
Been thanked: 1212 times

Re: Coronavirus - Modelling Aspects Only

#468014

Postby vrdiver » December 22nd, 2021, 11:49 am

zico wrote:People are talking about Omicron having exponential growth, and that it'll soon run out of people to infect,

<snip>

Seems to me likeliest scenario may be big wave of 'careless' infections, followed by gradual low level cases as virtuality everyone gradually gets infected over a year or two.
How does this above logic sound?

The assumption in that logic is that people only get infected the once. How does it play out if Omicron (or one of its successors) can be caught multiple times?

As a basis to this scenario: https://www.imperial.ac.uk/news/232698/modelling-suggests-rapid-spread-omicron-england/

VRD


Return to “Coronavirus Discussions”

Who is online

Users browsing this forum: No registered users and 7 guests