Page 60 of 68
Re: Coronavirus - Modelling Aspects Only
Posted: June 25th, 2021, 1:03 pm
by redsturgeon
Moderator Message:
We are drifting off "modelling aspects" here. Back on topic please or post elsewhere
Re: Coronavirus - Modelling Aspects Only
Posted: June 25th, 2021, 1:13 pm
by 1nvest
Re: Coronavirus - Modelling Aspects Only
Posted: June 25th, 2021, 1:23 pm
by murraypaul
1nvest wrote:Nowadays I'm more inclined to just look at the death rates and in seeing relatively low numbers it seems more like that Covid is just another cold. 1% of the population will die on average each year, so 700,000 type number, near 2000/day. I suspect there are nowadays more suicides/day perhaps as a function of lockdowns/controls than Covid related deaths.
https://www.bmj.com/content/372/bmj.n834From the earliest days of the pandemic there was concern that suicide would increase. It wasn’t hard to see where the risks might come from: anxiety about infection, isolation, disrupted care, domestic violence, alcohol, recession. Actual figures, though, took months to appear. Now we have reports from several countries, based on national or state level suicide data. They come from Australia, Canada, New Zealand, Norway, Peru, Sweden, and the US—high income countries in most cases—and they carry a consistent message. Suicide rates have not risen.
In England we have had to set up new data collection because our coroner system is not quick enough for the close monitoring we need—the median time from death to inquest is 166 days. We now have “real time surveillance,” recording deaths by suicide as they happen, providing figures for a population of around 9 million, one sixth of the country. Here, too, we have found no increase in the months post-lockdown. The same appears to be true of self-harm.
Re: Coronavirus - Modelling Aspects Only
Posted: June 26th, 2021, 2:21 pm
by jfgw
All graphs are mine. All graphs: Contains public sector information licensed under the Open Government Licence v3.0. :
https://www.nationalarchives.gov.uk/doc/open-government-licence/version/3/New cases rising, North East rate increasing rapidly,
A disproportionate number of new cases are arising in the 15-19 and 20-24 age groups,
Admissions hopefully peaking (or peaked) - one to keep an eye on,
Julian F. G. W.
Re: Coronavirus - Modelling Aspects Only
Posted: June 26th, 2021, 7:14 pm
by Sorcery
Thanks for doing this Julian. Never mind the NE being on the rise, the NW is ballistic.
Re: Coronavirus - Modelling Aspects Only
Posted: June 29th, 2021, 9:06 pm
by jfgw
Vaccine uptake seems to be slowing down at well below 100% for lower age groups,
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/Second jabs are following first jabs as one would expect,
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.
Re: Coronavirus - Modelling Aspects Only
Posted: July 2nd, 2021, 12:17 pm
by murraypaul
Covid deaths by age:
Shows a reduction in deaths among older age groups, but they still make up the majority of covid related deaths.
The downward trend in older covid deaths has reversed since the beginning of June.
Covid deaths by age as a %age of all death by age:
Shows that this is due to the higher total number of deaths among those age ranges, not to a higher rate of deaths from covid.
The percentage of deaths related to covid has actually fallen to less than 2% for all age ranges 50 and above.
Simplified version of above:
Data source for all graphs:
https://www.ons.gov.uk/peoplepopulation ... ndandwales
Re: Coronavirus - Modelling Aspects Only
Posted: July 8th, 2021, 10:45 pm
by scotia
Two weeks ago I noted that the climb in hospital admissions was levelling off, and I said
Looking at the admissions curve, it is not climbing exponentially - rather an optimist would suggest it is reaching a peak - although possibly we should wait for more data.Regrettably - the new data shows that it was merely a lull, followed by a steeper climb.
The deaths curve is also now showing a clear correlation with the admissions - with a much reduced time difference between them. This looks like the delta variant, imported from the Indian sub-continent, and is it possible that the communities to which it was initially returned have had a lower uptake in vaccinations.
Where next? I'll wait for more data.
Re: Coronavirus - Modelling Aspects Only
Posted: July 9th, 2021, 5:03 pm
by zico
scotia wrote:
The deaths curve is also now showing a clear correlation with the admissions - with a much reduced time difference between them. This looks like the delta variant, imported from the Indian sub-continent, and is it possible that the communities to which it was initially returned have had a lower uptake in vaccinations.
Where next? I'll wait for more data.
Wouldn't you always expect the deaths curve to have a clear correlation with the admissions, once the deaths increased to a level where there wasn't "lumpiness" in the data? If the hospitalisations and deaths were purely within the Indian community (with their slightly lower vac rate) then surely the increases should have slowed by now because the Delta variant has been widespread across all of the UK for several weeks now.
Re: Coronavirus - Modelling Aspects Only
Posted: July 9th, 2021, 6:50 pm
by scotia
zico wrote:scotia wrote:
The deaths curve is also now showing a clear correlation with the admissions - with a much reduced time difference between them. This looks like the delta variant, imported from the Indian sub-continent, and is it possible that the communities to which it was initially returned have had a lower uptake in vaccinations.
Where next? I'll wait for more data.
Wouldn't you always expect the deaths curve to have a clear correlation with the admissions, once the deaths increased to a level where there wasn't "lumpiness" in the data? If the hospitalisations and deaths were purely within the Indian community (with their slightly lower vac rate) then surely the increases should have slowed by now because the Delta variant has been widespread across all of the UK for several weeks now.
I would expect a clear correlation if the circumstances remain constant. So while there will inevitably be lumpiness at low numbers, the most significant change which created the low numbers and the broken correlation was the ongoing vaccination campaign. But now we are back to a clear correlation and an upward trend. Yet the vaccination program is ongoing - so why the change? We imported the Delta variant, initially into the Indian community, where the uptake of vaccinations has been reported as being significantly lower than the UK average. This seems a likely candidate for the sudden change in direction. Of course, there are other possibilities - is the Delta variant breaking the vaccine protection? Or have we become more careless in our behaviour? But there is little doubt that there has been a significant change - with the weekly deaths at 32 (on 24/5/21) and now 148 on 8/7/21.
Re: Coronavirus - Modelling Aspects Only
Posted: July 9th, 2021, 7:20 pm
by zico
scotia wrote:I would expect a clear correlation if the circumstances remain constant. So while there will inevitably be lumpiness at low numbers, the most significant change which created the low numbers and the broken correlation was the ongoing vaccination campaign. But now we are back to a clear correlation and an upward trend. Yet the vaccination program is ongoing - so why the change?
What I meant was I'd expect there to be a clear correlation between hospitalisations and deaths, but hopefully lower.
To give simple made-up numbers for an example.
Pre-vac (say)
100,000 infections leading to 1,000 hospitalisations leading to 100 deaths.
Post-vac
100,000 infections leading to 200 hospitalisations leading to 10 deaths.
(So in both cases, there's a correlation, just at different levels. We know that post-vac, there are fewer hospitalisations per infection, and hopefully fewer deaths per hospitalisation.)
Re: Coronavirus - Modelling Aspects Only
Posted: July 9th, 2021, 11:34 pm
by servodude
zico wrote:scotia wrote:I would expect a clear correlation if the circumstances remain constant. So while there will inevitably be lumpiness at low numbers, the most significant change which created the low numbers and the broken correlation was the ongoing vaccination campaign. But now we are back to a clear correlation and an upward trend. Yet the vaccination program is ongoing - so why the change?
What I meant was I'd expect there to be a clear correlation between hospitalisations and deaths, but hopefully lower.
To give simple made-up numbers for an example.
Pre-vac (say)
100,000 infections leading to 1,000 hospitalisations leading to 100 deaths.
Post-vac
100,000 infections leading to 200 hospitalisations leading to 10 deaths.
(So in both cases, there's a correlation, just at different levels. We know that post-vac, there are fewer hospitalisations per infection, and hopefully fewer deaths per hospitalisation.)
Doesn't that depend on whether the vaccine projects you from "hospitalisation and death", or hospitalisation, and death?
I suspect that the vaccine effects are more seen in the early part of the immune system (as it increases the chance of being asymptomatic etc)
- but once you're in need of serious intervention it's got passed that point
-sd
Re: Coronavirus - Modelling Aspects Only
Posted: July 9th, 2021, 11:41 pm
by zico
servodude wrote:
Doesn't that depend on whether the vaccine projects you from "hospitalisation and death", or hospitalisation, and death?
I suspect that the vaccine effects are more seen in the early part of the immune system (as it increases the chance of being asymptomatic etc)
- but once you're in need of serious intervention it's got passed that point
-sd
The stats I've seen on vaccine protection seem to indicate a threefold protection at all stages -
70% less likely to get infected, 90% less likely to be hospitalised, and 98% less likely to die.
Re: Coronavirus - Modelling Aspects Only
Posted: July 9th, 2021, 11:54 pm
by servodude
zico wrote:servodude wrote:
Doesn't that depend on whether the vaccine projects you from "hospitalisation and death", or hospitalisation, and death?
I suspect that the vaccine effects are more seen in the early part of the immune system (as it increases the chance of being asymptomatic etc)
- but once you're in need of serious intervention it's got passed that point
-sd
The stats I've seen on vaccine protection seem to indicate a threefold protection at all stages -
70% less likely to get infected, 90% less likely to be hospitalised, and 98% less likely to die.
That would be nice.
The only results I've really paid attention to were the clinical trials ones where no-one died so they were presented as "hospitalisation and death"
- but again that was with original covid and the proper timing regards doses
-sd
Re: Coronavirus - Modelling Aspects Only
Posted: July 10th, 2021, 7:15 am
by Itsallaguess
Whilst there's clearly some concern over the rising UK case numbers, it's the
rolling 7-day percentage change in new cases that's going to help to indicate where any existing
vaccine or
prior-infection-driven 'herd-immunity' benefits start to push back on what would otherwise be a much larger wave of infections this time around.
Looking at the current 7-day case charts below, there's clearly opportunities for 'false-dawns' in the short to medium-term data, as highlighted on the recent re-growths on the right hand chart below, but given that there's likely to have been lots of indoor football-based gatherings in recent weeks, it'll be interesting to see if what
might be seen as the beginnings of a' plateauing' of the rolling 7-day case-rate, shown on the left-hand chart below, continues over the coming weeks and months -
Source -
https://coronavirus.data.gov.uk/details/cases#card-percentage_change_in_recent_7-day_case_rates_by_specimen_dateCheers,
Itsallaguess
Re: Coronavirus - Modelling Aspects Only
Posted: July 10th, 2021, 8:05 am
by servodude
Itsallaguess wrote:Whilst there's clearly some concern over the rising UK case numbers, it's the
rolling 7-day percentage change in new cases that's going to help to indicate where any existing
vaccine or
prior-infection-driven 'herd-immunity' benefits start to push back on what would otherwise be a much larger wave of infections this time around.
Looking at the current 7-day case charts below, there's clearly opportunities for 'false-dawns' in the short to medium-term data, as highlighted on the recent re-growths on the right hand chart below, but given that there's likely to have been lots of indoor football-based gatherings in recent weeks, it'll be interesting to see if what
might be seen as the beginnings of a' plateauing' of the rolling 7-day case-rate, shown on the left-hand chart below, continues over the coming weeks and months -
Source -
https://coronavirus.data.gov.uk/details/cases#card-percentage_change_in_recent_7-day_case_rates_by_specimen_dateCheers,
Itsallaguess
Is positivity tracked similarly?
-sd
Re: Coronavirus - Modelling Aspects Only
Posted: July 10th, 2021, 8:22 am
by Itsallaguess
servodude wrote:
Is positivity tracked similarly?
I'm not sure.
Cheers,
Itsallaguess
Re: Coronavirus - Modelling Aspects Only
Posted: July 10th, 2021, 8:28 am
by servodude
Itsallaguess wrote:servodude wrote:
Is positivity tracked similarly?
I'm not sure.
Cheers,
Itsallaguess
No problem.
I've been trying to find a source for it for the data in
https://coronavirus.data.gov.uk/- in my mental model it relates to confidence in the cases figure; which I expect might take a bit of a hit with load
-sd
Re: Coronavirus - Modelling Aspects Only
Posted: July 10th, 2021, 8:55 am
by murraypaul
scotia wrote:We imported the Delta variant, initially into the Indian community, where the uptake of vaccinations has been reported as being significantly lower than the UK average.
Just to clarify this, the ONS surveys show that while vaccination rates were lower in the Asian/Asian British community, this is most likely because of age demographics meaning they hadn't been offered yet, not 'vaccine hesitancy'. The survey covering the period just before vaccinations were opened up to all adults shows:
Re: Coronavirus - Modelling Aspects Only
Posted: July 10th, 2021, 11:48 am
by scotia
murraypaul wrote:scotia wrote:We imported the Delta variant, initially into the Indian community, where the uptake of vaccinations has been reported as being significantly lower than the UK average.
Just to clarify this, the ONS surveys show that while vaccination rates were lower in the Asian/Asian British community, this is most likely because of age demographics meaning they hadn't been offered yet, not 'vaccine hesitancy'. The survey covering the period just before vaccinations were opened up to all adults shows:
Yes - that's the ONS data from 26/5/21 to 20/6/21.
However if you look at the ONS data earlier this year (e.g. 31/3/21 to 25/4/21) you will get
And if you go further back to the ONS report of 13/1/21 to 7/2/21 you will get
I'm delighted to see that the vaccine hesitancy has significantly reduced in all three of the above categories. The reduction in the Asian community is particularly impressive.
Sources from ONS "Coronavirus and vaccine hesitancy, Great Britain"