B.1.1.529 Omicron variant

Covid-19 discussion, bring your own statistics
User avatar
shpalman
Light of Blast
Posts: 6415
Joined: Mon Nov 11, 2019 12:53 pm
Location: One step beyond

Re: B.1.1.529 Omicron variant

Post by shpalman » Fri Dec 31, 2021 12:24 pm

shpalman wrote:
Fri Dec 31, 2021 7:31 am
South Africa reckons the fourth wave has passed thanks to... not selling alcohol after midnight and limiting gatherings to only 1000 people?

They're still wearing masks though.
... oh and of course it's summer for them. Can't assume our European winter Omicron wave will pass as easily as the self-limiting Delta waves which the UK had all summer.
molto tricky

User avatar
bob sterman
Catbabel
Posts: 772
Joined: Mon Nov 11, 2019 10:25 pm
Location: Location Location

Re: B.1.1.529 Omicron variant

Post by bob sterman » Fri Dec 31, 2021 5:15 pm

Little publicised fact - the daily UK updates (except for Wales) on new "cases" do not include reinfections (e.g. people testing positive for omicron who perhaps previously tested positive for alpha or delta earlier in the year).

It's the number of new people testing positive for the first time. A person can only be recorded as a new case once during the pandemic according to this metric....
COVID-19 cases are identified by taking specimens from people and testing them for the SARS-CoV-2 virus. If the test is positive this is referred to as a case. Some positive rapid lateral flow test results are confirmed with lab-based polymerase chain reaction (PCR) tests taken within 72 hours. If the PCR test results are negative, these are no longer reported as confirmed cases. If a person has more than one positive test, they are only counted as one case for all nations with the exception of Wales. Cases data includes all positive lab-confirmed polymerase chain reaction (PCR) test results plus, in England, positive rapid lateral flow tests that are not followed by a negative PCR test taken within 72 hours.
So the number of people testing positive each day in the UK is going to be way higher than the number showing in the daily updates.

This means you can't compute a test positivity rate by just taking new cases by specimen date / tests conducted on that date.

But it is good news when it comes to estimating the severity of omicron vs other variants (e.g. hospitalisations as a proportion of cases). Current simple estimates based on new hospitalisations / new cases are using a denominator that's too small.

Also, if the government stick with this metric - once the whole population has been infected and tested positive for either alpha, delta or omicron, in future when they are infected with Pi, Rho, Sigma etc the daily new case numbers will drop to zero! And on the way the number of new hospitalisations per day will climb past 100% of new cases!

User avatar
jimbob
Stummy Beige
Posts: 3144
Joined: Mon Nov 11, 2019 4:04 pm
Location: High Peak/Manchester

Re: B.1.1.529 Omicron variant

Post by jimbob » Fri Dec 31, 2021 5:53 pm

bob sterman wrote:
Fri Dec 31, 2021 5:15 pm
Little publicised fact - the daily UK updates (except for Wales) on new "cases" do not include reinfections (e.g. people testing positive for omicron who perhaps previously tested positive for alpha or delta earlier in the year).

It's the number of new people testing positive for the first time. A person can only be recorded as a new case once during the pandemic according to this metric....
COVID-19 cases are identified by taking specimens from people and testing them for the SARS-CoV-2 virus. If the test is positive this is referred to as a case. Some positive rapid lateral flow test results are confirmed with lab-based polymerase chain reaction (PCR) tests taken within 72 hours. If the PCR test results are negative, these are no longer reported as confirmed cases. If a person has more than one positive test, they are only counted as one case for all nations with the exception of Wales. Cases data includes all positive lab-confirmed polymerase chain reaction (PCR) test results plus, in England, positive rapid lateral flow tests that are not followed by a negative PCR test taken within 72 hours.
So the number of people testing positive each day in the UK is going to be way higher than the number showing in the daily updates.

This means you can't compute a test positivity rate by just taking new cases by specimen date / tests conducted on that date.

But it is good news when it comes to estimating the severity of omicron vs other variants (e.g. hospitalisations as a proportion of cases). Current simple estimates based on new hospitalisations / new cases are using a denominator that's too small.

Also, if the government stick with this metric - once the whole population has been infected and tested positive for either alpha, delta or omicron, in future when they are infected with Pi, Rho, Sigma etc the daily new case numbers will drop to zero! And on the way the number of new hospitalisations per day will climb past 100% of new cases!
Good point. There is somewhere data about possible reinfections on the gov.uk website, but can't remember where.
Have you considered stupidity as an explanation

User avatar
jimbob
Stummy Beige
Posts: 3144
Joined: Mon Nov 11, 2019 4:04 pm
Location: High Peak/Manchester

Re: B.1.1.529 Omicron variant

Post by jimbob » Fri Dec 31, 2021 5:56 pm

Found it

https://www.gov.uk/government/statistic ... 022-season

Week 51 data

Image

The dotted line is scary
Have you considered stupidity as an explanation

User avatar
shpalman
Light of Blast
Posts: 6415
Joined: Mon Nov 11, 2019 12:53 pm
Location: One step beyond

Re: B.1.1.529 Omicron variant

Post by shpalman » Fri Dec 31, 2021 6:09 pm

If they survived it once I don't think reinfection will kill them.
molto tricky

User avatar
lpm
Stummy Beige
Posts: 4031
Joined: Mon Nov 11, 2019 1:05 pm

Re: B.1.1.529 Omicron variant

Post by lpm » Fri Dec 31, 2021 6:22 pm

Volume of cases is massively understated. I think this will throw all shpalman equations out. There simply isn't the data.

Hospitalisations are rising as expected. Christmas Day infections will start to arrive in hospitals in about another week.
⭐ Awarded gold star 4 November 2021

User avatar
Martin_B
Dorkwood
Posts: 1192
Joined: Mon Nov 11, 2019 1:20 pm
Location: Perth, WA

Re: B.1.1.529 Omicron variant

Post by Martin_B » Sat Jan 01, 2022 12:44 am

shpalman wrote:
Fri Dec 31, 2021 6:09 pm
If they survived it once I don't think reinfection will kill them.
If it's a different variant, this isn't necessarily the case; omicron seems to be milder, but future variants may be more aggressive, or produce different symptoms which could kill them.
"My interest is in the future, because I'm going to spend the rest of my life there"

User avatar
bob sterman
Catbabel
Posts: 772
Joined: Mon Nov 11, 2019 10:25 pm
Location: Location Location

Re: B.1.1.529 Omicron variant

Post by bob sterman » Sat Jan 01, 2022 7:06 am

shpalman wrote:
Fri Dec 31, 2021 6:09 pm
If they survived it once I don't think reinfection will kill them.
Perhaps. But they can still turn up in the numerator (but not the denominator) of the new hospitalisations / new cases ratio, which messes up severity calculations.

User avatar
shpalman
Light of Blast
Posts: 6415
Joined: Mon Nov 11, 2019 12:53 pm
Location: One step beyond

Re: B.1.1.529 Omicron variant

Post by shpalman » Sat Jan 01, 2022 8:35 am

Martin_B wrote:
Sat Jan 01, 2022 12:44 am
shpalman wrote:
Fri Dec 31, 2021 6:09 pm
If they survived it once I don't think reinfection will kill them.
If it's a different variant, this isn't necessarily the case; omicron seems to be milder, but future variants may be more aggressive, or produce different symptoms which could kill them.
The vaccinated population might all be screwed then too.
bob sterman wrote:
Sat Jan 01, 2022 7:06 am
shpalman wrote:
Fri Dec 31, 2021 6:09 pm
If they survived it once I don't think reinfection will kill them.
Perhaps. But they can still turn up in the numerator (but not the denominator) of the new hospitalisations / new cases ratio, which messes up severity calculations.
Fewer official cases (because we don't count reinfections) for the same number of hospitalizations would make it look more severe than it is. And Johnson would look at the data and decide not to do anything anyway.

It's totally lame that the UK gave up right at the beginning with a protocol for declaring people recovered from covid though. In Italy, there's roughly half a million (including children) who have officially recovered from covid within the past 6 months such that they have a valid Green Pass from it, though, which is roughly the number we can vaccinate in a day.

Anyway based on UK daily hospital admissions being 1.8% of the daily cases with a week's lag, and that (official) cases are already 7-day averaging 150,000 per day, expect to reach

Image

lots of admissions soon.

Well, the case rates haven't quite baked in a peak which will surpass the 4500 a day from last winter's wave, but I expect it'll go past 3000 a day by mid January.
molto tricky

User avatar
lpm
Stummy Beige
Posts: 4031
Joined: Mon Nov 11, 2019 1:05 pm

Re: B.1.1.529 Omicron variant

Post by lpm » Sat Jan 01, 2022 9:44 am

Official cases aren't a usable metric, same as March 2020.

Around 5,000,000 actual infections over the 10 days holiday days? 0.5% hospitalised, 25,000, in the next 10 days? I think it looks pretty likely the UK will match the Jan 2021 peak of 40,000 in hospital.

But it's impossible to get any sense of how good the voluntary lockdown has been over the holidays.
⭐ Awarded gold star 4 November 2021

User avatar
lpm
Stummy Beige
Posts: 4031
Joined: Mon Nov 11, 2019 1:05 pm

Re: B.1.1.529 Omicron variant

Post by lpm » Sat Jan 01, 2022 9:59 am

SHOCK! Model looking accurate.

https://twitter.com/COVID19actuary/stat ... ZwL4g&s=19

Model a few days late, presumably due to the fast start with insane 1.6 day doubling. We'll see if the shape matches soon enough.

But why oh why does everybody work on daily admissions instead of in hospital? Makes no sense.
⭐ Awarded gold star 4 November 2021

User avatar
shpalman
Light of Blast
Posts: 6415
Joined: Mon Nov 11, 2019 12:53 pm
Location: One step beyond

Re: B.1.1.529 Omicron variant

Post by shpalman » Sat Jan 01, 2022 10:59 am

lpm wrote:
Sat Jan 01, 2022 9:59 am
SHOCK! Model looking accurate.

https://twitter.com/COVID19actuary/stat ... ZwL4g&s=19

Model a few days late, presumably due to the fast start with insane 1.6 day doubling. We'll see if the shape matches soon enough.

But why oh why does everybody work on daily admissions instead of in hospital? Makes no sense.
A fast rise in the case rate increases the percentage of them in hospital because they all get admitted at once, but a shorter stay in hospital decreases the percentage. So without modelling that properly (to see which effect "wins" or if they balance out) I wouldn't obviously be able to compare this wave to the previous one.
molto tricky

Millennie Al
Dorkwood
Posts: 1170
Joined: Mon Mar 16, 2020 4:02 am

Re: B.1.1.529 Omicron variant

Post by Millennie Al » Sun Jan 02, 2022 12:42 am

And a mild but highly infectious variant means more people in hospital with Covid but only in hospital because of something else, further complicating the picture.
Covid-19 - Don't catch it: don't spread it.

User avatar
jimbob
Stummy Beige
Posts: 3144
Joined: Mon Nov 11, 2019 4:04 pm
Location: High Peak/Manchester

Re: B.1.1.529 Omicron variant

Post by jimbob » Sun Jan 02, 2022 6:45 pm

lpm wrote:
Sat Jan 01, 2022 9:59 am
SHOCK! Model looking accurate.

https://twitter.com/COVID19actuary/stat ... ZwL4g&s=19

Model a few days late, presumably due to the fast start with insane 1.6 day doubling. We'll see if the shape matches soon enough.

But why oh why does everybody work on daily admissions instead of in hospital? Makes no sense.
Each admission is a single event, just as each new case is a single event.
Have you considered stupidity as an explanation

raven
Snowbonk
Posts: 539
Joined: Mon Nov 11, 2019 8:58 pm

Re: B.1.1.529 Omicron variant

Post by raven » Mon Jan 03, 2022 11:19 pm

bob sterman wrote:
Fri Dec 31, 2021 5:15 pm
Little publicised fact - the daily UK updates (except for Wales) on new "cases" do not include reinfections (e.g. people testing positive for omicron who perhaps previously tested positive for alpha or delta earlier in the year).

It's the number of new people testing positive for the first time. A person can only be recorded as a new case once during the pandemic according to this metric....
I thought there was a cut-off of 90 days. As in, a second positive PCR within 90 days is discounted because it could just be picking up viral fragments left over from the initial infection.

We're counting ones after 90 days though, surely?

User avatar
bob sterman
Catbabel
Posts: 772
Joined: Mon Nov 11, 2019 10:25 pm
Location: Location Location

Re: B.1.1.529 Omicron variant

Post by bob sterman » Tue Jan 04, 2022 1:24 pm

raven wrote:
Mon Jan 03, 2022 11:19 pm
bob sterman wrote:
Fri Dec 31, 2021 5:15 pm
Little publicised fact - the daily UK updates (except for Wales) on new "cases" do not include reinfections (e.g. people testing positive for omicron who perhaps previously tested positive for alpha or delta earlier in the year).

It's the number of new people testing positive for the first time. A person can only be recorded as a new case once during the pandemic according to this metric....
I thought there was a cut-off of 90 days. As in, a second positive PCR within 90 days is discounted because it could just be picking up viral fragments left over from the initial infection.

We're counting ones after 90 days though, surely?
Nope - according to the documentation an individual person can only become a new case once (except in Wales!).

Full details are given here...

https://coronavirus.data.gov.uk/metrics ... ecimenDate

https://coronavirus.data.gov.uk/metrics ... ublishDate
If a person has more than one positive test, they are only counted as one case for all nations with the exception of Wales
There's no 90 day limit in there.

Bottom right of this page...

https://coronavirus.data.gov.uk/details/cases

There's a table...
Cases by area (whole pandemic)
Total number of cases (people who have had at least one positive COVID-19 test result) since the start of the pandemic, and rate per 100,000 people.
Which today gives 11,408,560 as the total number of cases for England since the start of the pandemic.

And if you look at the Cases by specimen date at the top of the page - the cumulative total is 11,408,560 today.

So you can only be a new case once (rubbish title for next Bond film?)
Last edited by bob sterman on Tue Jan 04, 2022 1:26 pm, edited 1 time in total.

User avatar
wilsontown
Stargoon
Posts: 121
Joined: Tue Feb 23, 2021 11:51 am

Re: B.1.1.529 Omicron variant

Post by wilsontown » Tue Jan 04, 2022 1:42 pm

Press conference incoming at 5pm, with our esteemed leader alongside Whitty and Vallance. I assume this will be to announce nothing at all.
"All models are wrong but some are useful" - George Box

User avatar
lpm
Stummy Beige
Posts: 4031
Joined: Mon Nov 11, 2019 1:05 pm

Re: B.1.1.529 Omicron variant

Post by lpm » Tue Jan 04, 2022 2:39 pm

lpm wrote:
Thu Dec 09, 2021 6:52 pm
Outcome of my 9 December "model":

1) Slower rise of Omi by 31 December compared to most of the other speculation. I only see a rise from current 50,000 to 119,000 cases at 31 Dec.
Oddly enough, I was a lot closer with my spreadsheet "model" than most other speculators. Official cases on 31 Dec were 189,846 compared to my 119,000. Which is a lot in absolute terms but not much in terms of exponential growth. Certainly the "1,000,000 cases by the end of the year" crowd were well out.

Obvs the official case number is understated though.

Compared to my "model" boosters rolled out a lot quicker. The doubling speed of the initial Omi ramp up was a lot faster than assumed, but that speed faded away towards the end of the month (in official case stats anyway). Delta has declined faster, which I presume is due to the success of the voluntary lockdown taking Delta's R to well below 1.
4) Hospital numbers (two week lag for hospitalisations):
- Match Jan 2020 peak of 40,000 in hospital on 21 Jan
My spreadsheet gave in hospital numbers rising from 7,500 on 9 Dec to 12,657 on 31 Dec. This compares to the most recent figure of 11,918 on 29 Dec. My model underestimated cases but also underestimated "mildness" and those two things cancelled out.

40,000 on 21 Jan would still look reasonable as an outlook except it did not assume any change to the average length of stay. If that has indeed halved then it would be 20,000. Which is inevitable - there's a lot of growth in hospital numbers already locked in by the case load of a couple of weeks ago. The early stage of Omicron was 20-30 year olds. The inevitable spread up the heat map happened with the usual lags and hence the >65s cases are currently on the upward curve, feeding into hospitalisations now. Doesn't look good at all. It's just back to the old battle of vaccine protection against hospitalisation vs sheer volume. And volume would win because exponential always does.

But while R was clearly well above 1 for >65s over the Christmas period, maybe it will drop below 1 now? I suspect a lot of the elderly/vulnerable will have put themselves into lockdown on completion of holiday activities. Over the next couple of days we need to look for a bend in the >65s curve. We don't need to care too much about cases in the <50s which could easily resume their exponentials with the return to work and school - so long as more vulnerable people keep themselves at home.
⭐ Awarded gold star 4 November 2021

raven
Snowbonk
Posts: 539
Joined: Mon Nov 11, 2019 8:58 pm

Re: B.1.1.529 Omicron variant

Post by raven » Tue Jan 04, 2022 4:34 pm

@bob sterman

I'm not completely convinced. I mean, I know our statistics are a bit sh.t, but I didn't think they were that sh.t...Putting your quote in context:
COVID-19 cases are identified by taking specimens from people and testing them for the SARS-CoV-2 virus. If the test is positive this is referred to as a case. Some positive rapid lateral flow test results are confirmed with lab-based polymerase chain reaction (PCR) tests taken within 72 hours. If the PCR test results are negative, these are no longer reported as confirmed cases. If a person has more than one positive test, they are only counted as one case for all nations with the exception of Wales. Cases data includes all positive lab-confirmed polymerase chain reaction (PCR) test results plus, in England, positive rapid lateral flow tests that are not followed by a negative PCR test taken within 72 hours.
Reading that whole paragraph, particularly the bit I've bolded, it seems to refer to people having a positive LFT, then a confirmatory postive PCR within 72 hours and that only counts as one case. Which would make sense.

I got the 90 day thing from the reinfection bit of the Weekly Flu & Covid Surveillance Report -( for example, see pg 18 of this one from Dec 23rd.) They define possible reinfection as two positive tests at least 90 days apart, so I figured we are counting those as cases in their own right.

But if the right hand doesn't know what the left is up to, as is entirely possible, I could very well be wrong.

raven
Snowbonk
Posts: 539
Joined: Mon Nov 11, 2019 8:58 pm

Re: B.1.1.529 Omicron variant

Post by raven » Tue Jan 04, 2022 4:40 pm

Ah.

But now I've read the Welsh bit, which specifically mentions someone testing positive again after six weeks counts as a new case, maybe you are right

Although that doesn't make much sense either, not if it's a known thing that you can still be testing postive on PCR up to 90 days post-infection without being reinfected....

raven
Snowbonk
Posts: 539
Joined: Mon Nov 11, 2019 8:58 pm

Re: B.1.1.529 Omicron variant

Post by raven » Tue Jan 04, 2022 4:49 pm

bob sterman wrote:
Tue Jan 04, 2022 1:24 pm
Bottom right of this page...

https://coronavirus.data.gov.uk/details/cases

There's a table...
Cases by area (whole pandemic)
Total number of cases (people who have had at least one positive COVID-19 test result) since the start of the pandemic, and rate per 100,000 people.
Which today gives 11,408,560 as the total number of cases for England since the start of the pandemic.

And if you look at the Cases by specimen date at the top of the page - the cumulative total is 11,408,560 today.

So you can only be a new case once (rubbish title for next Bond film?)
The interesting thing about that table is that the figures are identical to the Cases by specimen date for England, for Northern Ireland and for Wales.

They're not for Scotland though. Cases by specimen date for Jan 3rd are 914,841; table at the bottom says 983,270. I dinnae ken why.

raven
Snowbonk
Posts: 539
Joined: Mon Nov 11, 2019 8:58 pm

Re: B.1.1.529 Omicron variant

Post by raven » Tue Jan 04, 2022 5:24 pm

@bob sterman

OK, it took a bit of digging but I think I've found an answer on this page here about how UKHSA & the NHS compile the pillar 1 & 2 data.
From 15 October 2020, the methodology for people tested and people testing positive has changed

Previously, the number of people newly tested and newly testing positive was reported, where the figures were deduplicated over the entirety of the pandemic so an individual would only appear once. This meant someone tested in March and again in September would only be counted in the March counts. This was progressively becoming less meaningful the longer the duration of the pandemic and meant that it was not appropriate to calculate a positivity rate from this data.

Figures are now reported as people tested and people testing positive at least once in the reporting week. People tested or testing positive are only counted once over the 7-day reporting period (Thursday to Wednesday), with a positive test being prioritised over a negative test. A person can be counted within more than one 7-day reporting period. If someone was tested more than once in different reporting weeks, they would be included in the count for all reporting weeks they were tested in.

For example, if a person was tested on Thursday and Friday of the same week, they would only be counted once in the reporting week. However, if someone was tested on Tuesday and Friday of the same week, that individual would be counted in 2 reporting periods, as the 2 tests fall into different 7-day reporting periods. If a person is tested under both pillar 1 and pillar 2 in the same reporting week, then only the pillar they were first tested under is counted, unless they were tested in both pillars on the same day, in which case, they are counted under pillar 2.

If someone had both a positive and a negative result in the same week, only the positive case would be counted, with the result being counted in pillar associated with the positive result. All previous reporting weeks have been backdated and revised with the new methodology.
So it seems that a person couldn't be counted twice at all, but that changed 14 months ago and now you will be included in the cases if you get it again.

raven
Snowbonk
Posts: 539
Joined: Mon Nov 11, 2019 8:58 pm

Re: B.1.1.529 Omicron variant

Post by raven » Tue Jan 04, 2022 5:41 pm

Oh, ffs. Just ignore that last post. That's just for calculating the positivity rate.

I should've read 6 more paragraphs down to this bit:
The cumulative number of people newly testing positive (deduplicated over the entirety of the pandemic) is available on the COVID-19 in the UK dashboard.
Yep, bob is right & we really are only counting people once ever.

That means the kid my sister knows who had Covid 4 months ago and has just recaught it won't count in this weeks figures. There's 'only' been about 127,000 suspected reinfections to date, but not counting them in the case figures might make a big difference with Omicron which is likely to cause more of them.

User avatar
bob sterman
Catbabel
Posts: 772
Joined: Mon Nov 11, 2019 10:25 pm
Location: Location Location

Re: B.1.1.529 Omicron variant

Post by bob sterman » Tue Jan 04, 2022 5:55 pm

raven wrote:
Tue Jan 04, 2022 5:41 pm
Oh, ffs. Just ignore that last post. That's just for calculating the positivity rate.

I should've read 6 more paragraphs down to this bit:
The cumulative number of people newly testing positive (deduplicated over the entirety of the pandemic) is available on the COVID-19 in the UK dashboard.
Yep, bob is right & we really are only counting people once ever.

That means the kid my sister knows who had Covid 4 months ago and has just recaught it won't count in this weeks figures. There's 'only' been about 127,000 suspected reinfections to date, but not counting them in the case figures might make a big difference with Omicron which is likely to cause more of them.
I was astonished when I realised this was the method behind the daily figures.

I guess it's a method that made sense early in the pandemic - if you believed everyone would get it once, or be vaccinated, and then the pandemic would be over!

On the bright side - it does mean eventually we'll get close to zero COVID using this metric (aside from kids getting it for the first time!).

User avatar
shpalman
Light of Blast
Posts: 6415
Joined: Mon Nov 11, 2019 12:53 pm
Location: One step beyond

Re: B.1.1.529 Omicron variant

Post by shpalman » Tue Jan 04, 2022 6:11 pm

Scientists call for Covid reinfections in UK to be included in case figures
Reinfections should be included in Covid figures for the UK as soon as possible, scientists have said, amid warnings that up to 15% of Omicron cases could be people who have had coronavirus before.

At present, daily Covid case figures – which reflect the number of infections picked up by testing – do not include reinfections for all countries of the UK.

While figures for Wales do include multiple cases for the same person, provided the positive tests are more than 42 days apart, this is not the case for England, Scotland or Northern Ireland.
molto tricky

Post Reply