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

Re: Coronavirus - Numbers and Statistics

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

Re: Coronavirus - Numbers and Statistics

#499451

Postby Hallucigenia » May 9th, 2022, 3:40 pm

Itsallaguess wrote:Reading these Covid-related boards over the past couple of years, no-one could be blamed for coming away thinking that the UK had a uniquely bad outcome from the pandemic, and the above WHO excess-deaths table shows that to not be the case when compared to the US and some of our large European neighbours...

I very much suspect that we'd have heard a lot more about the above ranking on these boards if the UK had been higher up on it...


I'd take these kinds of thing with a bit of a pinch of salt - if you look at the error bars on these kinds of things they're rather bigger than some of the effects you are looking for. And they've made a rather arbitrary decision to model the baseline with a cyclic cubic spline which seems to have had the effect of making Germany look worse and Sweden quite a lot better. I look forward to the Telegraph discussion of whether spline fitting is the right choice before they make any leaps to simplistic political conclusions...

I don't think anyone is saying the UK has been uniquely bad, but it's still possible for the UK to have been quite crap, particularly in 2020, and for other European countries to have made a lot of mistakes as well. Indeed, the outcome in some of the Pacific countries suggests that in general the response of most of Europe has been sub-optimal.

An interesting one that people don't talk much about here is Canada, which has seen everything from Australia-style tight exclusion in the Atlantic provinces to "let it rip" in the western provinces, and an incompetent populist reacting a month late to everything in Ontario. I leave you to decide which is closest to the UK... Their death rates have ranged from 5-18% of UK levels in the Maritimes, up to 50% in Manitoba. However the "leader" at 60% is Quebec, which has done a lot of things right and wasn't too bad for much of the pandemic, but has been really slammed by omicron - more deaths in 4 months this year than in the whole of 2021. I've not really seen an analysis of what's been going on there, it could be that even if you do most things right, if things are bad enough in your neighbours then they will spill over, particularly when you have a variant as infectious as omicron and everywhere is opening up again "because it's over".

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

Re: Coronavirus - Numbers and Statistics

#499537

Postby Hallucigenia » May 9th, 2022, 10:03 pm

Ah, the grownups in the form of the COVID-19 Actuaries Response Group have responded in similar vein :

https://covidactuaries.org/2022/05/09/w ... ss-deaths/

The choice of measure for expected deaths makes a huge difference to the comparison between Sweden and Germany. Under the WHO approach, excess mortality in Germany was close to twice that in Sweden. But under the linear trend, excess mortality in Sweden is three times that in Germany.

No doubt we'll see a headline to that effect in tomorrow's Telegraph.

It's almost like this stuff is complicated....

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

Re: Coronavirus - Numbers and Statistics

#499546

Postby scotia » May 9th, 2022, 11:48 pm

Hallucigenia wrote:Ah, the grownups in the form of the COVID-19 Actuaries Response Group have responded in similar vein :

https://covidactuaries.org/2022/05/09/w ... ss-deaths/

The choice of measure for expected deaths makes a huge difference to the comparison between Sweden and Germany. Under the WHO approach, excess mortality in Germany was close to twice that in Sweden. But under the linear trend, excess mortality in Sweden is three times that in Germany.

No doubt we'll see a headline to that effect in tomorrow's Telegraph.

It's almost like this stuff is complicated....

Intriguing! I have no great faith in the accuracy of excess deaths in Europe since the vaccines took effect, and the death rate fell dramatically. It was clearly by far the most accurate statistic at the beginning of the outbreak when Covid deaths were large, and were being substantially underreported. But now with the Covid death rate at low levels, I prefer the reported Covid deaths as a better measure. But I hadn't realised how badly some excess death estimations had become. In response to the Daily Telegraph graph, I calculated the reported deaths per head of population - and made it clear that Sweden's performance was far from optimum when compared to the other Nordic countries. I checked out the French reported deaths, and confirmed that it should lie just above Sweden - as in the Daily Telegraph graph - so I assumed that the rank order of the excess deaths was similar to the rank order of the reported deaths. But if only I had looked at Germany!
Who Covid Cumulative Reported Deaths per million population (to two decimal places)
Sweden 1800; France 2200; Germany 1600
Quite a different ranking order!

Now getting back to the Daily Telegraph graph. The numbers are excess deaths per 100000 in 2020 and 2021. But per 100000 what? No units - a hanging offence in my far off days as a scientist. It would be reasonable to assume per 100000 population, but I noted that there was a factor of about 3 difference in my numbers for Sweden and France compared to the Daily Telegraph numbers (although the rank was the same), so I wondered if it was per annum, compared to the reported deaths which I believe are summed from the start of the outbreak. Any clarification? Surely the excess deaths are not running at around one third of the reported deaths. It was certainly the opposite way round at the beginning of the epidemic. Or am I making some schoolboy error? All help gratefully accepted. :)

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

Re: Coronavirus - Numbers and Statistics

#499547

Postby servodude » May 10th, 2022, 12:04 am

scotia wrote: Or am I making some schoolboy error? All help gratefully accepted.


I think you might be making the mistake of starting with the figures rather than the answer you want :roll:

Anyways if you go to the source (a workbook entitled WHO_COVID_Excess_Deaths_EstimatesByCountry.xlsx
And find the "country rates by year" sheet it does seem to confirm the 100,000 is their denominator for population

Code: Select all

pop.e5   Population (in 100,000)
excess.mean   Excess deaths associated with COVID-19 pandemic from all-causes per 100,000 (mean)


- sd

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

Re: Coronavirus - Numbers and Statistics

#499549

Postby scotia » May 10th, 2022, 12:15 am

servodude wrote:
scotia wrote: Or am I making some schoolboy error? All help gratefully accepted.


I think you might be making the mistake of starting with the figures rather than the answer you want :roll:

Anyways if you go to the source (a workbook entitled WHO_COVID_Excess_Deaths_EstimatesByCountry.xlsx
And find the "country rates by year" sheet it does seem to confirm the 100,000 is their denominator for population

Code: Select all

pop.e5   Population (in 100,000)
excess.mean   Excess deaths associated with COVID-19 pandemic from all-causes per 100,000 (mean)


- sd

Thanks - I'm even more confused :)

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

Re: Coronavirus - Numbers and Statistics

#499802

Postby Hallucigenia » May 11th, 2022, 11:33 am

Interesting chart from Paul Mainwood on the difference between BA1 and BA2, which is hitting the elderly much harder :
Image

Julian
Lemon Quarter
Posts: 1385
Joined: November 4th, 2016, 9:58 am
Has thanked: 532 times
Been thanked: 676 times

Re: Coronavirus - Numbers and Statistics

#499904

Postby Julian » May 11th, 2022, 5:15 pm

Also more hospitalisations in the elderly as shown in a tweet that Paul Mainwood was replying to when he tweeted the graph Hallucigenia just posted...

Image

Obviously there is an implicit assumption here that the first wave was BA.1-driven and the second wave BA.2-driven.

For the benefit of others like me who had never heard of Paul Mainwood until Hallucigenia's last tweet Google pretty quickly found this for me - https://twitter.com/PaulMainwood?ref_sr ... r%5Eauthor - and then on to this - https://quoranoftheweek.quora.com/Smoot ... -marketing - which I assume is the correct Paul Mainwood.

- Julian

XFool
The full Lemon
Posts: 12636
Joined: November 8th, 2016, 7:21 pm
Been thanked: 2608 times

Re: Coronavirus - Numbers and Statistics

#500047

Postby XFool » May 12th, 2022, 11:15 am

United States passes one million Covid deaths

BBC News

The US has passed more than one million Covid-related deaths, says the White House.

"The US has also recorded more than 80 million Covid cases, out of a 330 million population."

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

Re: Coronavirus - Numbers and Statistics

#506407

Postby scotia » June 11th, 2022, 12:48 am

Following on from the comments on the No Statistics board, where it has been mentioned that Covid cases may be once again on the rise, I have plotted the latest 7-day totals of Covid hospital admissions in England. The vertical bars represent the statistical standard deviations.
Image

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

Re: Coronavirus - Numbers and Statistics

#507177

Postby Hallucigenia » June 14th, 2022, 2:36 pm

Hallucigenia wrote:Ah, the grownups in the form of the COVID-19 Actuaries Response Group have responded in similar vein :

https://covidactuaries.org/2022/05/09/w ... ss-deaths/

The choice of measure for expected deaths makes a huge difference to the comparison between Sweden and Germany. Under the WHO approach, excess mortality in Germany was close to twice that in Sweden. But under the linear trend, excess mortality in Sweden is three times that in Germany.

No doubt we'll see a headline to that effect in tomorrow's Telegraph.

It's almost like this stuff is complicated....


The WHO did subsequently tweak their numbers, albeit by less that the above - Germany down 37% (so now less than the UK) and Sweden up 19% :
https://www.nature.com/articles/d41586-022-01526-0

I still don't remember seeing anything in the Telegraph about it.

As for what's happening now, a week after the Jubilee try

https://twitter.com/COVID19actuary/stat ... 9034253312
Hospital admissions with COVID are rising sharply across England, up 33% week-on-week. Admissions rose by between 23% and 50% in all English regions. Bed occupancy is up by 12%.

Hospital deaths with COVID are no longer falling. Our estimate suggests deaths rose by 8% this week....

the current rate of increase in hospital admissions is faster than we saw during the BA.2 Omicron wave in March.

This chart shows what that means as an R estimate, currently approaching 1.2.


Adam Kucharski, one of the co-directors of the epidemics centre at LSHTM, reckons we're now at the stage where the virus is so inherently infectious, any further increase won't make much difference, the main driver of infection levels is more the rate at which immunity fades. The current cycles and infection levels suggest it fades after around 3-4 months. So don't believe the people who say "one more push and it will all be over".

https://twitter.com/adamjkucharski/stat ... 8010201088
Image

And for another view, here's a presentation from the Professor of Biosecurity at the University of NSW :
https://exchange.isid.org/isid/2022/isi ... demic.html
COVID-19 eradication is unlikely. During the pandemic, as vaccination programs began nearly everywhere, herd immunity was suggested as the outcome to control SARS CoV-2 replication and dissemination. However today, as we are facing waning vaccine efficacy, new variants of the virus with different pathogenic patterns, scientists and public health authorities are suggesting the pandemic is over and that we are in an endemic phase. Raina MacIntyre explains why this is not so, and that SARS-CoV-2 will continue to cause epidemics and likely a substantial burden of chronic disease. Better control will improve these outcomes.
To achieve this, a vaccine-PLUS strategy is needed, with layered protections including testing, tracing, safe indoor air and masks. Vaccine equity globally must also be addressed.

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

Re: Coronavirus - Numbers and Statistics

#508003

Postby scotia » June 17th, 2022, 10:25 pm

Another week of 7-day totals of Covid Hospital admissions in England

Image

The numbers are modest, but I find the continuing upward trend somewhat worrying

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

Re: Coronavirus - Numbers and Statistics

#509644

Postby scotia » June 25th, 2022, 10:41 pm

Another week of 7-day totals of Covid Hospital admissions in England

Image

Worryingly the curve continues to rise

Now the ONS survey of Covid infections for the period 11th to 18th June estimates 2.36% to 2.64% for England, whereas the Scotland estimate is 4.16% to 5.4% - roughly double. So I have also plotted the 7-day totals of Covid Hospital admissions in Scotland - with a 10:1 scaling as for England, which roughly matches the population ratios.

Image

The Scottish totals ( per unit of population) may be slightly higher than the English but the difference is not nearly as large as the ONS infection rate estimates. Why?

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

Re: Coronavirus - Numbers and Statistics

#511217

Postby scotia » July 1st, 2022, 6:52 pm

Another week of 7-day totals of Covid Hospital admissions in England

Image

Not good news.

I have also plotted the 7-day totals of Covid Hospital admissions in Scotland - with a 10:1 scaling as for England, which roughly matches the population ratios:-

Image

Now the latest ONS survey of Covid infections (for the week ending 24th June) estimates 3.35% for England, and 5.47% for Scotland. Last week (ending 17/18th June) it was 2.5% for England and 4.76% for Scotland

There is no sign of the Scottish hospital admissions being any higher (per unit of population) than the English, which seems a bit contrary to the ONS infection data.
However the Scottish hospital admissions seem to be levelling off, rather than climbing as in England. I suppose if I take a cavalier attitude to statistical accuracy, I could possibly state that this trend is possibly in agreement with the trend of the ONS data.

ReformedCharacter
Lemon Quarter
Posts: 3120
Joined: November 4th, 2016, 11:12 am
Has thanked: 3591 times
Been thanked: 1509 times

Re: Coronavirus - Numbers and Statistics

#511236

Postby ReformedCharacter » July 1st, 2022, 9:17 pm

According to the ONS the figures they produce are:

Percentage of people testing positive for coronavirus (COVID-19) in private residential households in England, Wales, Northern Ireland and Scotland...

Is it unreasonable to suggest that fewer infected people are now testing for C19 because tests are now no longer free for most people? If that is true then surely the present figures cannot be accurately compared with previous 'waves' when more people were testing? I hope these aren't stupid questions :)

RC

daveh
Lemon Quarter
Posts: 2191
Joined: November 4th, 2016, 11:06 am
Has thanked: 409 times
Been thanked: 807 times

Re: Coronavirus - Numbers and Statistics

#511241

Postby daveh » July 1st, 2022, 9:47 pm

ReformedCharacter wrote:According to the ONS the figures they produce are:

Percentage of people testing positive for coronavirus (COVID-19) in private residential households in England, Wales, Northern Ireland and Scotland...

Is it unreasonable to suggest that fewer infected people are now testing for C19 because tests are now no longer free for most people? If that is true then surely the present figures cannot be accurately compared with previous 'waves' when more people were testing? I hope these aren't stupid questions :)

RC

The ONS figures are produced by random testing of a sample of people in each country (a bit like a covid opinion poll) so the data is unaffected by the stopping of free tests for the general public.

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

Re: Coronavirus - Numbers and Statistics

#512855

Postby scotia » July 9th, 2022, 12:20 am

Another update from the most recently published data (extracted at 11pm 8/7/22).
First we have the 7-day totals of Covid Hospital admissions in England.
Image

Still not good news - the steep climb continues.

I have also plotted the 7-day totals of Covid Hospital admissions in Scotland - with a 10:1 scaling as for England, which roughly matches the population ratios:-

Image

Optimistically, it appears that Scotland may be past the peak.

Now the latest ONS survey of Covid infections (for the week ending 29th June (England) and 30th June (Scotland)) estimates 3.95% for England, and 5.94% for Scotland. Last week (ending 24th June) it was 3.35% for England, and 5.47% for Scotland. And the previous week (ending 17/18th June) it was 2.5% for England and 4.76% for Scotland.

I still continue to be surprised at the higher number of (ONS reported) infections (per head of population) in Scotland when compared to England -which doesn't seem to match the hospital admissions.

Sorcery
Lemon Quarter
Posts: 1229
Joined: November 4th, 2016, 6:38 pm
Has thanked: 147 times
Been thanked: 366 times

Re: Coronavirus - Numbers and Statistics

#513550

Postby Sorcery » July 11th, 2022, 8:59 pm

On a topic which has been discussed before Ferguson's modelling but brought bang up to date :
https://dailysceptic.org/2022/07/10/its ... ish-model/

Grim reading, belief in his models causes government scientists to rush us into lockdowns, furlough and ultimately the highest taxation for 70 years. The bill for which is somewhere round £300-410 billion https://commonslibrary.parliament.uk/re ... /cbp-9309/

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

Re: Coronavirus - Numbers and Statistics

#514584

Postby Itsallaguess » July 15th, 2022, 12:51 pm

Some hopefully encouraging data from the ONS this morning, which highlights Scotland being slightly ahead of the rest of the UK in terms of the latest wave of rising infections, but where they are currently seeing something of a slow-down of infection rates -


Image

Source - https://tinyurl.com/2d9rq5sz

From the above BBC news link -

Are we approaching the peak?

It feels like an odd question to ask after another chunky rise in the number of people infected going from 2.7 million to 3.5 million.

However, there are some postive signs if you read between the lines.

Scotland has been ahead of the rest of the UK in this wave, however, study the charts and it looks like the number infected is slowing down and approaching the peak.

Meanwhile, the number being admitted to hospital with Covid in England is also slowing down - another sign we could be approaching the peak.

It's too early to know for sure, but the figures may start to look more positive in the coming weeks.


Cheers,

Itsallaguess

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

Re: Coronavirus - Numbers and Statistics

#514603

Postby Itsallaguess » July 15th, 2022, 1:36 pm

Some more commentary on the latest ONS figures regarding the current UK wave -

This wave is so different to early in the pandemic -

The Covid waves we've had this year have infected incredible numbers of people.

Go back to January 2021 and the highest number of people infected in the ONS weekly surveys was 1.2 million.

Already this year we've had a wave cap out at 4.9 million at the end of March and now one at 3.5 million and climbing.

And yet we're seeing nowhere near the levels of severe illness.

The number needing ventilation in intensive care went above 3,700 in England alone in January 2021.

The equivalent figure now - in a much bigger wave - is just 274.


https://tinyurl.com/3sxxs7wn

Cheers,

Itsallaguess

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

Re: Coronavirus - Numbers and Statistics

#514724

Postby scotia » July 16th, 2022, 12:18 am

Another update from the most recently published data (extracted at around 11pm 15/7/22).
First we have the 7-day totals of Covid Hospital admissions in England.
Image
Since it appears that a peak may be near, I have extended the time range to compare it with the previous peak around the end of March

And I have given the same treatment to the Scottish Data, with the vertical multiplier a factor of 10 larger - to reflect the approximate 10:1 England to Scotland population.

Image

Conclusions? - it looks like the current Scottish Covid Hospital Admissions are now well over their recent peak, and this peak is significantly lower than the level back in March.
And I'll be an optimist and suggest that the English Covid Hospital Admissions also appears to be reaching a peak - which is lower, but not significantly, than its peak back in March.


Return to “Coronavirus Discussions”

Who is online

Users browsing this forum: No registered users and 6 guests