Scrutable

This is all qualitative and anecdotal. We must stick to hard data to make rational decisions.

sheldrake wrote: Mon Oct 18, 2021 5:39 pm This is all qualitative and anecdotal. We must stick to hard data to make rational decisions.

Serious question - I notice you weren't involved in discussions in the "Pandemic Arena" part of the forum until about a month ago.

What suddenly got you interested in this idea that COVID-19 hasn't caused substantial excess mortality? If it was a particularly persuasive article on this issue - please do share it.

Of course you could have been reading all the "Pandemic Arena" discussions for the past 18 months or so. But right now we seem to be rehashing discussions that were had on many previous occasions.

Sheldrake's a troll. Ignore the stupid f.cker.

bob sterman wrote: Mon Oct 18, 2021 5:45 pm
Serious question - I notice you weren't involved in discussions in the "Pandemic Arena" part of the forum until about a month ago.

What suddenly got you interested in this idea that COVID-19 hasn't caused substantial excess mortality? If it was a particularly persuasive article on this issue - please do share it.

edited:-
I wasn't posting at all on the forum until about a month ago because I was angry with the moderation policy at the beginning of 2020. I began to notice bizarre statistical untruths being cited by government spokespeople last summer eta: in particular a lack of mature discussion about how Covid risks compared to other risks we live with and a very single-dimensional focus on increases in Covid case numbers or deaths. My own father died in December partly as a result of lockdown policy which impacted his access to GP care. My impression is that Covid is definitely more dangerous than average seasonal flu and people with immune disorders or over the age of 70 with other health problems require special protection, but it's still less dangerous than many other things we live with and 'behavioural science' (i.e. propaganda techniques) have been used to terrify the ordinary population out of proportion with the real risk. This is being exploited by authoritarian opportunists.

Of course you could have been reading all the "Pandemic Arena" discussions for the past 18 months or so. But right now we seem to be rehashing discussions that were had on many previous occasions.

People often said that about brexit and other topics, and when I looked back I found some perspectives (and evidence) hadn't really been surfaced, or had been dismissed in an emotive way due to groupthink. If you feel something I'm raising has definitely already been covered by a specific post, please link me to it.

sheldrake wrote: Mon Oct 18, 2021 4:44 pm

jimbob wrote: Mon Oct 18, 2021 1:31 pm

No but I have data for all cause deaths for England and Wales going back to wk31 1999 showing how it was completely abnormal to then.

I'd love to understand this plot more deeply. I have some questions about it

1) Is it showing all deaths, or all hospital deaths?

2) Is the bulge for week 18 of 2019-20 April/May of 2019. or does it represent some other date

1) Raw weekly all cause deaths in England and Wales - historic data from figure 4 of https://www.ons.gov.uk/peoplepopulation ... o2016final

2015 data onwards from the ONS weekly all cause deaths reports. There were not (Looks at the radius of the circle and the scale) around 520,000 deaths (roughly 10,000 per week x 52) in hospitals in England and Wales.

2) It is implicit in the graph. It runs from wk31-wk30 so 2019-2020 would have wks 31-52 as 2019, and wks 01-30 as 2020.

Also it's pretty obvious that the only time this century when there were more than twice the usual weekly deaths was not in April 2019.

jimbob wrote: Mon Oct 18, 2021 6:17 pm

1) Raw weekly all cause deaths in England and Wales - historic data from figure 4 of https://www.ons.gov.uk/peoplepopulation ... o2016final

Thank you.

2015 data onwards from the ONS weekly all cause deaths reports. There were not (Looks at the radius of the circle and the scale) around 520,000 deaths (roughly 10,000 per week x 52) in hospitals in England and Wales.

2) It is implicit in the graph. It runs from wk31-wk30 so 2019-2020 would have wks 31-52 as 2019, and wks 01-30 as 2020.

Also it's pretty obvious that the only time this century when there were more than twice the usual weekly deaths was not in April 2019.

Thanks, it wasn't obvious to me that the later weeks of something labelled 2019-2020 would actually correspond to 2019. It's also not clear to me that this would be the case because our age-adjusted mortality in 2020 was lower than 2009 and every year back to 1990.

sheldrake wrote: Wed Oct 13, 2021 10:33 pm

1) 'Excess deaths' depend on what you take as a 'normal' level of deaths. Previous 5 years we had unusually low age-adjusted mortality compared to the norms going back to 1990. You can't assume that all, or even most, deaths above the previous 5-year average were really the result of untested Covid.

2) Deaths that occur even when somebody tests positive to Covid aren't necessarily primarily caused by Covid. I think the cause of admission to hospital (e.g. admitted for Covid, or unknown respiratory) would be the best thing to look at, don't know if it's possible to dig that out.

1) yes you can. When you dig into the numbers. Which it's perfectly possible for you to do. I'm using London as that's the first of my graphs that I found.






or England and Wales



notice how the non-Covid excess peaks and then falls with the Covid deaths at about 2000 deaths per week for the first five weeks of the peak then vanishes, which coincides with clarification of guidance on writing death certificates.

https://assets.publishing.service.gov.u ... vid-19.pdf



2) The death certificate looks at that. The ONS analyses it. The overwhelming majority of deaths where Covid was a contributory factor, it was the underlying factor. This is quite different to influenza.

sheldrake wrote: Mon Oct 18, 2021 6:20 pm

Thanks, it wasn't obvious to me that the later weeks of something labelled 2019-2020 would actually correspond to 2019. It's also not clear to me that this would be the case because our age-adjusted mortality in 2020 was lower than 2009 and every year back to 1990.

The thing about age-adjusted mortality is that it's a read herring. For the last 20 years, the fall in that has been pretty much offset by the rise in the population, so the weekly number of deaths has been remarkably constant.


I am not an actuary, so I don't try to reinvent the wheel. Putting in corrections where you don't understand them is silly. If you want an actuarial analysis - look at what actuaries are saying.

But you don't need to.

Because it is so massive that you need to do really stupid fudges to pretend it's not a huge problem. I have seen quite a few of those.

jimbob wrote: Mon Oct 18, 2021 6:35 pm

1) yes you can. When you dig into the numbers. Which it's perfectly possible for you to do. I'm using London as that's the first of my graphs that I found.

But again, you've posted a comparison that doesn't look back far enough. How does 2003 look, for example?

sheldrake wrote: Mon Oct 18, 2021 5:39 pm This is all qualitative and anecdotal. We must stick to hard data to make rational decisions.

After the weekend it usually takes a day or two for the hospital admissions and occupancy figures to get caught up, and after last weekend also I will need a day or two, then we can see if things are trending up or down (although for sure hospital figures are only trending slowly compared to cases, whichever way they're going).

The extra rapid tests being taken by Italians who for some f.cking reason decided not to get vaccinated* but now suddenly realize they need to go to work hasn't spiked the case rate; the positivity rate down below 1%.

(* - or maybe decided at the last minute but didn't realize it takes two weeks for the vaccination Green Pass to become valid)

sheldrake wrote: Mon Oct 18, 2021 5:39 pm This is all qualitative and anecdotal. We must stick to hard data to make rational decisions.

What, hard data like a photo of an outdoors bar in July 2021?

jimbob wrote: Mon Oct 18, 2021 6:40 pm

The thing about age-adjusted mortality is that it's a read herring. For the last 20 years, the fall in that has been pretty much offset by the rise in the population, so the weekly number of deaths has been remarkably constant.

Woah, this is really weird reasoning. You're suggesting that we shouldn't look at normal number of deaths per 100,000 adjusted by age cohort? Why would you do this?

I am not an actuary, so I don't try to reinvent the wheel. Putting in corrections where you don't understand them is silly.

I haven't put these corrections, the ONS has. Ignoring age-adjustment in mortality is a really odd thing to do when you know your population is, on balance, getting older. It's an odd thing to just look at absolute numbers when you know your population is getting larger.

If you want an actuarial analysis - look at what actuaries are saying.

But you don't need to.

Link me to an actuarial analysis if you believe that this profession has a different view, please.

Because it is so massive that you need to do really stupid fudges to pretend it's not a huge problem. I have seen quite a few of those.

It seems like you don't believe age-adjustment matters when comparing mortality across years in an aging population? I just find that weird if so.

lpm wrote: Mon Oct 18, 2021 6:50 pm

sheldrake wrote: Mon Oct 18, 2021 5:39 pm This is all qualitative and anecdotal. We must stick to hard data to make rational decisions.

What, hard data like a photo of an outdoors bar in July 2021?

Yes, mixed in with other data it's a pretty straightfoward way of highlighting how different the restrictions were in Sweden compared to the UK. You're not disputing this, are you ?

sheldrake wrote: Mon Oct 18, 2021 6:50 pm

Link me to an actuarial analysis if you believe that this profession has a different view, please.

https://twitter.com/NickStripe_ONS/stat ... 32544?s=20

https://twitter.com/ActuaryByDay/status ... 81696?s=20

jimbob wrote: Mon Oct 18, 2021 7:06 pm
https://twitter.com/NickStripe_ONS/stat ... 32544?s=20

In this first thread, the author starts using recent post-2015 data (already covered problems with this) then gets as far as population-adjusting the data This makes the increase in deaths look much smaller on a long scale like so: -



The author then considers also age-adjusting this data, and hand-waves it like so: -

Whilst the UK-born population is ageing, with post-war baby boomers now in their 50’s to 70’s, recent immigration will have predominantly been younger

This is wrong https://commonslibrary.parliament.uk/re ... n%20people.

If he'd have continued to do the analysis properly he would've gotten to the ONS data I already posted.

The second twitter thread uses data only as far back as 2011. Neither of your sources demonstrates a good counter argument.

sheldrake wrote: Mon Oct 18, 2021 7:16 pm If he'd have continued to do the analysis properly he would've gotten to the ONS data I already posted.

ONS dude wrote:We will produce age standardised rates soon

Sounds like he was planning to at the time of writing. Looks like he got there. For the ONS and everything.

monkey wrote: Mon Oct 18, 2021 7:33 pm

sheldrake wrote: Mon Oct 18, 2021 7:16 pm If he'd have continued to do the analysis properly he would've gotten to the ONS data I already posted.

Sounds like he was planning to at the time of writing. Looks like he got there. For the ONS and everything.

Then what was the point?

The data is there.

Tell me another April when there was a death rate anything like this in the last 100 years.

sheldrake wrote: Mon Oct 18, 2021 7:34 pm

monkey wrote: Mon Oct 18, 2021 7:33 pm

sheldrake wrote: Mon Oct 18, 2021 7:16 pm If he'd have continued to do the analysis properly he would've gotten to the ONS data I already posted.

Sounds like he was planning to at the time of writing. Looks like he got there. For the ONS and everything.

Then what was the point?

Because the age related data is not the only data of interest.

My point was that I don't think people should take you seriously.

monkey wrote: Mon Oct 18, 2021 7:41 pm
Because the age related data is not the only data of interest.

What is the interesting data?

Eta: You're just link dropping with no useful commentary

sheldrake wrote: Mon Oct 18, 2021 7:18 pm The second twitter thread uses data only as far back as 2011. Neither of your sources demonstrates a good counter argument.

I have plotted the raw numbers and given you the source of my data.

Others have given Euromomo data.


On one hand you are asking really basic questions and on the other coming back with questions that imply you think you know more than everyone else.


Anyway here's the cumulative all-cause deaths for England and wales by week starting at wk10 for every week from wk10,2000



Can you spot the odd year out?

jimbob wrote: Mon Oct 18, 2021 7:44 pm
I have plotted the raw numbers and given you the source of my data.

Others have given Euromomo data.

But you're segueing back and forth between age-adjusted data, population-only or unadjusted data. If we look at age-adjusted data, we returned to 2009 norms with a spike above the 5 year trend that looks simillar to 1951 flu.

At best, you're making a weak argument about acute risk without addressing acute vs chronic comparisons or historic norms.

Eta I ask these questions because you don't seem to realise that's what you're doing.

The other segue you keep making is going back to a comparison that only goes as far back as 2015, when I'm looking at age-adjusted data since 1990 or earlier.

Showing me a spike that's terrible compared to 2015-2019 when we know that 2010-2019 were lower than 2020, but 2020's mortality only represented a return to the mortality level of 2009 isn't really the comparison I'm trying to understand.

eta: the graph above drops age and population adjustment.