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

#436097

Postby scotia » August 19th, 2021, 10:38 pm

Another two weeks of data - But its not particularly good news, and there also seems to be a break from the simple model of the deaths following the admissions with a time slip

Image

The dip following the peak in admissions which we saw two weeks ago was short lived - after a modest dip it has started climbing again. And more seriously the death rate has continued climbing without any dip.

As I have previously mentioned, the Scottish statistics are poor, but I have plotted them with the same parameters

Image

With such poor statistics, only broad brush comments can be made. The dip following the peak in the hospital admissions has continued - although it now seems to be levelling out. There s some evidence of a double peak in the deaths , but these now seem to be falling.

funduffer
Lemon Quarter
Posts: 1327
Joined: November 4th, 2016, 12:11 pm
Has thanked: 122 times
Been thanked: 831 times

Re: Coronavirus - Modelling Aspects Only

#436574

Postby funduffer » August 22nd, 2021, 8:05 am

Although the England and Scotland data are different in shape, the deaths points break away from the predicted deaths points at around the same point in time - day 35, which I think is around the beginning of August.

So we are seeing relatively more people dying from this point onwards.

I cannot see an obvious reason for this, but it may be because only sicker people are being admitted to hospital, as the system is under so much stress.

FD

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

#436727

Postby jfgw » August 22nd, 2021, 8:55 pm

For England, case rates are highest in the 15 to 29 age groups. There is a gradual increase in cases in older 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/

Death rates, however, are highest in the older 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/

Admission rates are a lot higher in 85+ age groups than in lower groups,
https://coronavirus.data.gov.uk/details ... tes_by_age


Julian F. G. W.

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

#438083

Postby jfgw » August 28th, 2021, 5:23 pm

Comparison graphs for different variants seem thin on the ground. This one was derived from "Variants of concern: technical briefing 21 – underlying data", Fig. 7: Sequenced Cases, https://assets.publishing.service.gov.u ... roved.xlsx and cases data, https://coronavirus.data.gov.uk/details ... me=England .
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.

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

#438273

Postby jfgw » August 29th, 2021, 8:45 pm

I have combined data used for the above graph with data from the Milton Keynes Lighthouse Lab from last year to show the progress of the Alpha and Delta variants. The Milton Keynes data may not be representative of England as a whole.

The (total all variants) new cases data are taken from the gov.uk coronavirus website, https://coronavirus.data.gov.uk/details ... me=England .

Other data up to 5th January are taken from https://twitter.com/The_Soup_Dragon/sta ... 88/photo/1 . There is a more recent relevant graph here but I have not had time to read off the data and include them in my graph,
https://twitter.com/The_Soup_Dragon/sta ... 8557363201 .

I have used the weekly data published here, https://assets.publishing.service.gov.u ... roved.xlsx (Fig. 7) to determine the ratios of the Alpha, Delta and total other variants since 30th January. Daily ratios have been interpolated and applied to the daily new cases figures.

You will observe that there is a gap in some of the data.

ENGLAND NEW CASES PER DAY, 7 DAY AVERAGE:

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.

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

#438497

Postby jfgw » August 30th, 2021, 8:36 pm

A comparison of UK countries / England regions,
Northern Ireland (shamrock green), high but levelled off;
Wales (dragon red), levelling off;
Scotland (St. Andrew blue), climbing fast;
England: Noticeable peak in the South West not evident elsewhere (except for a similar, smaller peak in the South East). Some other areas are still rising but others appear to have peaked.

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/

The same graph except over a longer period:

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/


England cases by age,
Under 15s, and 40 and over: cases rising;
35—39: fairly level;
15—34: cases falling;
Hump unique to 15 to 19 year-olds, cases falling rapidly.

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.

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

#439530

Postby jfgw » September 3rd, 2021, 7:12 pm

Vaccine by age data now include 16—17 year-olds,

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/

Uptake for this age group has peaked,

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/

The data are downloaded from here (heatmap chart), https://coronavirus.data.gov.uk/details ... me=England



Julian F. G. W.

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

#440271

Postby scotia » September 7th, 2021, 11:56 am

I have been away from base, so my fortnightly graphs are a few days late. I have retained the previous "model" parameters - although its now clear that an adjustment is due. The data is as reported on 6/9/20.

First the English data

Image

Most notable is the effect on Death registrations of a Bank Holiday! Looking at the Hospital admissions, it seems to be levelling out - and possibly the deaths is also moderating - albeit at a significantly higher level than the 0.09 multiplier of the "model". Why?

And now the Scottish data

Image

The trend in both admissions and deaths is upwards, and once again the 0.09 multiplier is now too low.

Scottish schools re-opening dates are around the middle of August, while English schools re-open around two weeks later.

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

#441060

Postby zico » September 9th, 2021, 7:01 pm

Any thoughts as to what's happening with the daily statistics?

Infections - big drop in mid-July, and even though they've risen since, still below the mid-July level.

Hospitalisations - small drop end-July, and now higher than mid-July.

Deaths - no drop, just a levelling out for a week or so in early-August, and now resuming a steep upward trend.

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

#441152

Postby servodude » September 10th, 2021, 5:06 am

zico wrote:Any thoughts as to what's happening with the daily statistics?

Infections - big drop in mid-July, and even though they've risen since, still below the mid-July level.
Hospitalisations - small drop end-July, and now higher than mid-July.
Deaths - no drop, just a levelling out for a week or so in early-August, and now resuming a steep upward trend.


The observations you've listed could imply that there's been a shift towards only getting tested when you have symptoms
- or conversely.. a shift away from trying to comprehensively catch cases as you would in the "trying to prevent spread" phase
- if we had a breakdown of symptomatic/asymptomatic in the cases we might be able to lend some credence to this

Considering alternatives:
vaccine efficiency waning would case the proportions of cases that convert to hospitalisations and deaths to increase
- but the relative swings in these data sets with respect to the infections seems too abrupt for that to be the case here
- as the vaccines were administered over a fair period of time and you'd expect waning to be similarly spread

So I'm going to suggest that "freedom day" (as a concept more than a specific point in time) brought about a paradigm shift in approach

- sd

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

#441360

Postby jfgw » September 10th, 2021, 8:36 pm

There have been a few changes since my last "cases by age" graph.

15—19: New cases rose sharply, rate of increase easing off. I don't know how much term dates vary across England but the minimum looks too early for school infections to have caused the rise. More testing could be the major factor here.
5—9 and 10—14 age groups were rising anyway but have seen a more rapid increase since just before the new school term, again, suggesting an effect of more testing.
20—24 and 25—29 age groups continued to fall but have now levelled off.

Image


Julian F. G. W.

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

#443726

Postby scotia » September 20th, 2021, 12:50 am

I previously mentioned that the model of a correlation between deaths and time slipped hospital admissions had clearly broken down, with the ratio climbing significantly from the previous value of 0.09. With another two weeks of data, the ratio seems to have stabilised around 0.14, and here is the result for England

Image

It would seem that the English Admissions have levelled and are possibly starting to fall. The deaths have also levelled - and are around 750 to 800 per week. Although this is disappointing compared to the 30 to 50 in the second half of May, it is a factor of 10 below the peak in the last week of January, before the vaccines took effect. I can understand that with the removal of restrictions, it was bound to rise - but I'm puzzled why, over the past two months, the ratio of deaths to admissions has increased substantially. Is there a new mutation, are the vaccines waning, or is it simply that older persons are being less careful? Hopefully the government statisticians are looking at this.

Now for Scotland - using the same parameters as for England.

Image

The most obvious difference from the English graph is that in Scotland the Admissions and Deaths are on a significant upward trend. And given that the population of Scotland is approximately one tenth of that in England, the pro-rata death rate is about 75% higher. Rather worrying.

Itsallaguess
Lemon Half
Posts: 9129
Joined: November 4th, 2016, 1:16 pm
Has thanked: 4140 times
Been thanked: 10023 times

Re: Coronavirus - Modelling Aspects Only

#443733

Postby Itsallaguess » September 20th, 2021, 6:23 am

scotia wrote:
It would seem that the English Admissions have levelled and are possibly starting to fall. The deaths have also levelled - and are around 750 to 800 per week. Although this is disappointing compared to the 30 to 50 in the second half of May, it is a factor of 10 below the peak in the last week of January, before the vaccines took effect.

I can understand that with the removal of restrictions, it was bound to rise - but I'm puzzled why, over the past two months, the ratio of deaths to admissions has increased substantially.

Is there a new mutation, are the vaccines waning, or is it simply that older persons are being less careful? Hopefully the government statisticians are looking at this.


I think recent studies have clearly recognised that even double-jabbed immunity starts to wane after around 6 months, so given that the England vaccination programme started in early December 2020, and given that the early invites to that initial roll-out were to the oldest and most vulnerable, then we've now got that cohort well into the period (in the absence of widespread boosters...) where any drop in immunity might well be hitting the most susceptible people to being admitted to hospital the hardest...

The UK’s Covid vaccine rollout has been a huge success by any measure, with more than three quarters of adults receiving both doses in less than eight months. However, the latest research shows that the protection provided by the vaccines may start to wear off within six months, especially against the now-dominant Delta variant.

Earlier this week our ZOE Covid Study team published an analysis of data from our app contributors showing that the protection provided by two doses of either the Pfizer or AstraZeneca vaccine wanes over a number of months.

We found that initial protection against infection a month after the second dose of the Pfizer vaccine was 88 per cent, falling gradually to 74 per cent within five to six months. For the AstraZeneca vaccine, there was around 77 per cent protection a month after the second dose, falling slightly less rapidly to 67 per cent after four to five months.


https://inews.co.uk/opinion/covid-19-immunity-waning-booster-jabs-delta-1171114

Cheers,

Itsallaguess

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

#443929

Postby jfgw » September 20th, 2021, 8:23 pm

scotia wrote: I can understand that with the removal of restrictions, it was bound to rise - but I'm puzzled why, over the past two months, the ratio of deaths to admissions has increased substantially. Is there a new mutation, are the vaccines waning, or is it simply that older persons are being less careful? Hopefully the government statisticians are looking at this.

Itsallaguess wrote:I think recent studies have clearly recognised that even double-jabbed immunity starts to wane after around 6 months, so given that the England vaccination programme started in early December 2020, and given that the early invites to that initial roll-out were to the oldest and most vulnerable, then we've now got that cohort well into the period (in the absence of widespread boosters...) where any drop in immunity might well be hitting the most susceptible people to being admitted to hospital the hardest...


I would expect a new mutation to show up first in one region or in adjacent regions but I cannot see anything in regional data that suggests this. However, while the new case rate seems to have peaked in almost all age groups, the numbers of deaths in higher age groups are both high and (if you filter out the noise), probably rising.

New case data do not include people who have tested positive previously. Older people are more likely to be reinfected so actual cases, for the higher age groups, will be higher than "new cases" figures. How significant this is, I do not know.

It would be good to have detailed "admissions by age" data. I am aware that data are published for broad age groups by the NHS but I have not plotted these.

All graphs: 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/
Sources:
https://coronavirus.data.gov.uk/
https://www.england.nhs.uk/statistics/w ... 10920.xlsx


Image

Image

Image

Image


New Cases: 10—14 highest and rising. 5—9 also rising. 15—19 falling after a couple of peaks. Other age groups peaked.
Image

Image


Julian F. G. W.

9873210
Lemon Slice
Posts: 984
Joined: December 9th, 2016, 6:44 am
Has thanked: 226 times
Been thanked: 296 times

Re: Coronavirus - Modelling Aspects Only

#443974

Postby 9873210 » September 20th, 2021, 11:18 pm

jfgw wrote:I would expect a new mutation to show up first in one region or in adjacent regions but I cannot see anything in regional data that suggests this. However, while the new case rate seems to have peaked in almost all age groups, the numbers of deaths in higher age groups are both high and (if you filter out the noise), probably rising.

I don't think you have to speculate about new mutations.

Over 10% of cases are being gene sequenced. If a mutation is responsible for a significant increase in cases there would be direct evidence for it. It takes a short time to do the sequencing and notice that a particular mutation deserves a Greek letter, but not that long and then stored sequences can be search retroactively.

Even a far lower rate of gene sequencing would catch an important mutation before it shows up in increased cases. If you apply the usual sampling statistics a few hundred random gene sequences a week should catch mutations long before they show up in case rates. (unless it's the Omega strain that goes from 0 to 100% in a few weeks, but we can rule that out on other grounds)

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

#444254

Postby jfgw » September 21st, 2021, 8:02 pm

jfgw wrote:It would be good to have detailed "admissions by age" data. I am aware that data are published for broad age groups by the NHS but I have not plotted these.


I have now created a graph using published NHS data. Populations used are 2019 ONS data.
NHS data available up to 1st September 2021: https://www.england.nhs.uk/statistics/w ... -2021.xlsx . The next publication is due 7th October.

You will see that admissions are steady or falling for the under-65 age groups.
65—84s: admissions are rising or peaked (wait for next month's data to find out which).
85+ age group: admissions comparatively high but this has always been the case. Still rising.

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.

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

#447571

Postby scotia » October 3rd, 2021, 8:02 pm

Another 14 days of data - and it looks like there is some stability returning to the model parameters.
First we have England - projected deaths 0.14 of admissions 11 days previous.

Image

We now have a few weeks of data following the model parameters. The deaths have been falling, but the projection suggests that they will level out - following the admissions.

And now for Scotland - projected deaths 0.14 of admissions 12 days previous. (in England deaths must be registered within 5 days - whereas its 8 days in Scotland)

Image

Its encouraging that the deaths and projected deaths are now falling

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

#448603

Postby jfgw » October 7th, 2021, 6:59 pm

Vaccine data now include 12—15 year-olds,
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.

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

#450333

Postby jfgw » October 15th, 2021, 12:34 pm

The NHS October monthly publication was published yesterday. This includes admissions by age data, which I have plotted.

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/

Source of data: https://www.england.nhs.uk/statistics/w ... -2021.xlsx
Populations used are 2019 ONS data.


Julian F. G. W.

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

#450738

Postby vrdiver » October 17th, 2021, 11:25 am

jfgw wrote:The NHS October monthly publication was published yesterday. This includes admissions by age data, which I have plotted.

Thanks for this.

I am a bit bemused why they have lumped the 18-64 data into one curve, as I'd expect the cases/million to be very different for the 55-64 versus the 18-24s.

I'd still like to see the hospitalisation numbers broken down by vaccination status within age group as well, which seems a very obvious bit of data to be publishing...

VRD


Return to “Coronavirus Discussions”

Who is online

Users browsing this forum: No registered users and 7 guests