B.1.1.529 Omicron variant

Covid-19 discussion, bring your own statistics
Herainestold
After Pie
Posts: 1821
Joined: Mon Nov 25, 2019 1:23 pm

Re: B.1.1.529 Omicron variant

Post by Herainestold » Tue Jan 25, 2022 11:57 pm

Bird on a Fire wrote:
Tue Jan 25, 2022 11:09 pm
Image

That's right. I'll let you know when its over.
Masking forever
Putin is a monster, Russian people are not.

User avatar
Bird on a Fire
Princess POW
Posts: 9279
Joined: Fri Oct 11, 2019 5:05 pm
Location: Portugal

Re: B.1.1.529 Omicron variant

Post by Bird on a Fire » Wed Jan 26, 2022 12:48 am

Thanks. If you could hurry it along I'd be grateful. I've already had two conferences cancelled this year and I need to start finding work after my current contract ends in September.
Tree dwellers leaping out the boughs shouting «Get the paper»
Trunk hugging rebels chucking petals at a detonator

Herainestold
After Pie
Posts: 1821
Joined: Mon Nov 25, 2019 1:23 pm

Re: B.1.1.529 Omicron variant

Post by Herainestold » Wed Jan 26, 2022 4:25 am

Bird on a Fire wrote:
Wed Jan 26, 2022 12:48 am
Thanks. If you could hurry it along I'd be grateful. I've already had two conferences cancelled this year and I need to start finding work after my current contract ends in September.
There will be lots of job openings in health care. I'm sure they could use a bird scientist.(I think thats what you are. Apologies if I got it wrong)
Masking forever
Putin is a monster, Russian people are not.

User avatar
bob sterman
Catbabel
Posts: 901
Joined: Mon Nov 11, 2019 10:25 pm
Location: Location Location

Re: B.1.1.529 Omicron variant

Post by bob sterman » Wed Jan 26, 2022 6:47 am

raven wrote:
Tue Jan 25, 2022 9:42 pm
We're supposed to start adding reinfections to the case data soonish, aren't we.Cynically, I suspect they're waiting for things to die down enough that it won't cause a dramatic increase.
REACT suggests that most current positive cases are reinfections...

https://www.bbc.co.uk/news/health-60132096

So the daily new PCR positive case count is probably currently at least double the headline daily reported figure which excludes reinfections (although some reinfections would be in the current figures if the original infection wasn't confirmed by PCR at the time).

Of course this puts the current hospitalisation data into an even more favourable light.

User avatar
Woodchopper
Light of Blast
Posts: 5410
Joined: Sat Oct 12, 2019 9:05 am

Re: B.1.1.529 Omicron variant

Post by Woodchopper » Wed Jan 26, 2022 8:29 am

raven wrote:
Tue Jan 25, 2022 9:42 pm

We're supposed to start adding reinfections to the case data soonish, aren't we.
Here's the announcement: https://coronavirus.data.gov.uk/details ... 109f2cfbdd

User avatar
bob sterman
Catbabel
Posts: 901
Joined: Mon Nov 11, 2019 10:25 pm
Location: Location Location

Re: B.1.1.529 Omicron variant

Post by bob sterman » Wed Jan 26, 2022 9:01 am

Woodchopper wrote:
Wed Jan 26, 2022 8:29 am
raven wrote:
Tue Jan 25, 2022 9:42 pm

We're supposed to start adding reinfections to the case data soonish, aren't we.
Here's the announcement: https://coronavirus.data.gov.uk/details ... 109f2cfbdd
Thanks.

This line seems to suggest that the headline number of new cases (Cases by date reported), that the media focus on, is still not going to show reinfections...
new metrics will show first episodes of infection (equivalent to the current case definition) and episodes of reinfection, shown by specimen date only.

User avatar
shpalman
Princess POW
Posts: 7177
Joined: Mon Nov 11, 2019 12:53 pm
Location: One step beyond

Re: B.1.1.529 Omicron variant

Post by shpalman » Wed Jan 26, 2022 9:12 am

bob sterman wrote:
Wed Jan 26, 2022 9:01 am
Woodchopper wrote:
Wed Jan 26, 2022 8:29 am
raven wrote:
Tue Jan 25, 2022 9:42 pm

We're supposed to start adding reinfections to the case data soonish, aren't we.
Here's the announcement: https://coronavirus.data.gov.uk/details ... 109f2cfbdd
Thanks.

This line seems to suggest that the headline number of new cases (Cases by date reported), that the media focus on, is still not going to show reinfections...
new metrics will show first episodes of infection (equivalent to the current case definition) and episodes of reinfection, shown by specimen date only.
It suggests that they will define two new metrics, but that "cases in England by report date will change to the new definition of an episode of infection" and it's the cases reported each day which gives the headline number. i.e. people being infected a second time more than 90 days after their first infection will show up again in the daily case report.

The two new metrics will help to distinguish people being infected for the first time, and reinfections.

(The new definition of an "episode of infection" is a positive test more than 90 days after any other positive test, because a positive test within 90 days of a previous one counts as the same episode as before.)
molto tricky

User avatar
Martin Y
Stummy Beige
Posts: 2586
Joined: Mon Nov 11, 2019 1:08 pm

Re: B.1.1.529 Omicron variant

Post by Martin Y » Wed Jan 26, 2022 10:11 am

bob sterman wrote:
Wed Jan 26, 2022 6:47 am
raven wrote:
Tue Jan 25, 2022 9:42 pm
We're supposed to start adding reinfections to the case data soonish, aren't we.Cynically, I suspect they're waiting for things to die down enough that it won't cause a dramatic increase.
REACT suggests that most current positive cases are reinfections...

https://www.bbc.co.uk/news/health-60132096

So the daily new PCR positive case count is probably currently at least double the headline daily reported figure which excludes reinfections (although some reinfections would be in the current figures if the original infection wasn't confirmed by PCR at the time).

Of course this puts the current hospitalisation data into an even more favourable light.
The trend for Wales (where they already count reinfections) vs rest of the UK hints otherwise.

Just googling the UK covid 19 statistics for the last 3 months, the trend for England looks like there's a post-peak levelling off at about double the pre-peak level but the trend for Wales shows a post-peak level that's about the same as pre-peak. If reinfections were a significantly increasing element in the Wales figures, I'd expect those trends to look rather the other way around.

User avatar
bob sterman
Catbabel
Posts: 901
Joined: Mon Nov 11, 2019 10:25 pm
Location: Location Location

Re: B.1.1.529 Omicron variant

Post by bob sterman » Wed Jan 26, 2022 11:07 am

Martin Y wrote:
Wed Jan 26, 2022 10:11 am
bob sterman wrote:
Wed Jan 26, 2022 6:47 am
raven wrote:
Tue Jan 25, 2022 9:42 pm
We're supposed to start adding reinfections to the case data soonish, aren't we.Cynically, I suspect they're waiting for things to die down enough that it won't cause a dramatic increase.
REACT suggests that most current positive cases are reinfections...

https://www.bbc.co.uk/news/health-60132096

So the daily new PCR positive case count is probably currently at least double the headline daily reported figure which excludes reinfections (although some reinfections would be in the current figures if the original infection wasn't confirmed by PCR at the time).

Of course this puts the current hospitalisation data into an even more favourable light.
The trend for Wales (where they already count reinfections) vs rest of the UK hints otherwise.

Just googling the UK covid 19 statistics for the last 3 months, the trend for England looks like there's a post-peak levelling off at about double the pre-peak level but the trend for Wales shows a post-peak level that's about the same as pre-peak. If reinfections were a significantly increasing element in the Wales figures, I'd expect those trends to look rather the other way around.
From Dec 1st to Jan 22nd - England reported 4,745,555 new cases (excluding reinfections) and 72,855 hospitalisations (which can include reinfections). Using these figures gives a crude (incorrect) hospitalisation rate of 1.54%

During the same period Wales reported 246,224 new cases (including reinfections) and 1,788 hospitalisations (which can include reinfections). Using these figures gives a crude (but less incorrect) hospitalisation rate of 0.73%

So it does look like the calculated hospitalisation rate is lower when you include reinfections in the overall case denominator.

User avatar
shpalman
Princess POW
Posts: 7177
Joined: Mon Nov 11, 2019 12:53 pm
Location: One step beyond

Re: B.1.1.529 Omicron variant

Post by shpalman » Wed Jan 26, 2022 11:44 am

Martin Y wrote:
Wed Jan 26, 2022 10:11 am
bob sterman wrote:
Wed Jan 26, 2022 6:47 am
raven wrote:
Tue Jan 25, 2022 9:42 pm
We're supposed to start adding reinfections to the case data soonish, aren't we.Cynically, I suspect they're waiting for things to die down enough that it won't cause a dramatic increase.
REACT suggests that most current positive cases are reinfections...

https://www.bbc.co.uk/news/health-60132096

So the daily new PCR positive case count is probably currently at least double the headline daily reported figure which excludes reinfections (although some reinfections would be in the current figures if the original infection wasn't confirmed by PCR at the time).

Of course this puts the current hospitalisation data into an even more favourable light.
The trend for Wales (where they already count reinfections) vs rest of the UK hints otherwise.

Just googling the UK covid 19 statistics for the last 3 months, the trend for England looks like there's a post-peak levelling off at about double the pre-peak level but the trend for Wales shows a post-peak level that's about the same as pre-peak. If reinfections were a significantly increasing element in the Wales figures, I'd expect those trends to look rather the other way around.
What, you expect the addition of reinfections to the England data to make the overall case numbers go down?

Recent case rate data in cases/week/100,000, 7-day averaged:
UK-case-rates.png
UK-case-rates.png (48.46 KiB) Viewed 1489 times
molto tricky

User avatar
Martin Y
Stummy Beige
Posts: 2586
Joined: Mon Nov 11, 2019 1:08 pm

Re: B.1.1.529 Omicron variant

Post by Martin Y » Wed Jan 26, 2022 1:32 pm

shpalman wrote:
Wed Jan 26, 2022 11:44 am
What, you expect the addition of reinfections to the England data to make the overall case numbers go down?
No. The opposite. I'm noting the trend for England already looks worse than the trend for Wales despite England not counting reinfections. So if the Welsh figures were boosted by an increasing proportion of reinfections, I would have expected that trend to be the other way around, all other things being equal, which they aren't of course. What Bob said about hospitalisation rates is more compelling.

OffTheRock
Fuzzable
Posts: 271
Joined: Mon Dec 02, 2019 6:52 pm

Re: B.1.1.529 Omicron variant

Post by OffTheRock » Wed Jan 26, 2022 11:33 pm

raven wrote:
Tue Jan 25, 2022 9:42 pm
Post-Christmas peak has passed, cases plateau and then everybody relaxes slightly/goes back to work as they've been told so you get a slight rise perhaps? Or people've stopped testing so much because they're not visiting the elderly relatives like they were over Christmas, so we're not picking up so many asymptomatic cases, and the apparent fall/plateau wasn't totally accurate.

Speaking of which,sister who works in a primary school can't believe cases are falling because so many people she knows have it, including quite a few people who are getting it for the second time.

We're supposed to start adding reinfections to the case data soonish, aren't we.Cynically, I suspect they're waiting for things to die down enough that it won't cause a dramatic increase.

ETA: Just looked at the case data for where sister is, and yep, cases are rising in her area, plus the heat map is pretty scary:2022-01-25 Cases heatmap.png.
Most of the infection wave in schools before Christmas was delta. Primary schools have provided a lovely unmasked, unvaccinated population for omicron to spread through. And it’s had a ball. Secondaries have been affected but to a lesser extent. Until this week, from what I’m hearing. Friends in different parts of the country saying their cases have shot up over the last 24-48hours. Turns out that getting rid of masks in secondary from last Friday might not have been a brilliant idea.

raven
Snowbonk
Posts: 609
Joined: Mon Nov 11, 2019 8:58 pm

Re: B.1.1.529 Omicron variant

Post by raven » Thu Jan 27, 2022 11:11 pm

bob sterman wrote:
Wed Jan 26, 2022 6:47 am
raven wrote:
Tue Jan 25, 2022 9:42 pm
We're supposed to start adding reinfections to the case data soonish, aren't we.Cynically, I suspect they're waiting for things to die down enough that it won't cause a dramatic increase.
REACT suggests that most current positive cases are reinfections...

https://www.bbc.co.uk/news/health-60132096

So the daily new PCR positive case count is probably currently at least double the headline daily reported figure which excludes reinfections (although some reinfections would be in the current figures if the original infection wasn't confirmed by PCR at the time).

Of course this puts the current hospitalisation data into an even more favourable light.
Yes. When I said case rates looked scary where my sister is, I should've added that of all the many people she knows who've caught it last month and this, most of whom are kids/teachers/parents, none of them have ended up in hospital.

raven
Snowbonk
Posts: 609
Joined: Mon Nov 11, 2019 8:58 pm

Re: B.1.1.529 Omicron variant

Post by raven » Thu Jan 27, 2022 11:25 pm

Woodchopper wrote:
Wed Jan 26, 2022 8:29 am
raven wrote:
Tue Jan 25, 2022 9:42 pm

We're supposed to start adding reinfections to the case data soonish, aren't we.
Here's the announcement: https://coronavirus.data.gov.uk/details ... 109f2cfbdd
Cheers, Woodchopper.

I see they're using 90 days as a definition. I seem to remember something about PCRs picking up viral fragments for a while post-infection, meaning there's chance of a false positive, so you have to have some sort of cut-off to avoid that. (Do LFTS have the same problem or do they only detect active infection, I wonder?)

I guess evidence has suggested 90 days is the right time frame, but has Omicron changed that? A teacher my sister works with was positive in December, tested positive again last week, so that's more like 4-6 weeks.

raven
Snowbonk
Posts: 609
Joined: Mon Nov 11, 2019 8:58 pm

Re: B.1.1.529 Omicron variant

Post by raven » Thu Jan 27, 2022 11:45 pm

OffTheRock wrote:
Wed Jan 26, 2022 11:33 pm

Most of the infection wave in schools before Christmas was delta. Primary schools have provided a lovely unmasked, unvaccinated population for omicron to spread through. And it’s had a ball. Secondaries have been affected but to a lesser extent. Until this week, from what I’m hearing. Friends in different parts of the country saying their cases have shot up over the last 24-48hours. Turns out that getting rid of masks in secondary from last Friday might not have been a brilliant idea.
Just to add some n=1 to that:
Nephew tested positive mid-December, along with 2 kids in his class so maybe Delta?
Niece tested positve in January, along with almost half her classmates, so probably Omicron.

You can see signs of it moving from Primary (pale grey= 0-9 yr olds) to Secondary (blue =10-19) in this week's surveillance report:
UKHSA Influenza Surveillance graphs 27 January 2022.png
UKHSA Influenza Surveillance graphs 27 January 2022.png (172.2 KiB) Viewed 1332 times

User avatar
jimbob
Stummy Beige
Posts: 4016
Joined: Mon Nov 11, 2019 4:04 pm
Location: High Peak/Manchester

Re: B.1.1.529 Omicron variant

Post by jimbob » Fri Jan 28, 2022 5:15 pm

Image

You can see BA.2 rising now
Have you considered stupidity as an explanation

User avatar
bob sterman
Catbabel
Posts: 901
Joined: Mon Nov 11, 2019 10:25 pm
Location: Location Location

Re: B.1.1.529 Omicron variant

Post by bob sterman » Tue Feb 01, 2022 9:22 am

Well reinfections are now included on the UK dashboard - and numbers haven't gone up nearly as much as I was expecting.

Only about 830,000 cases of reinfection have been added to the cumulative total number of cases.

User avatar
shpalman
Princess POW
Posts: 7177
Joined: Mon Nov 11, 2019 12:53 pm
Location: One step beyond

Re: B.1.1.529 Omicron variant

Post by shpalman » Tue Feb 01, 2022 9:41 am

You mainly see the difference developing as the peak of the recent winter wave goes up.
uk-before-and-after.png
uk-before-and-after.png (36.21 KiB) Viewed 1192 times
molto tricky

User avatar
jimbob
Stummy Beige
Posts: 4016
Joined: Mon Nov 11, 2019 4:04 pm
Location: High Peak/Manchester

Re: B.1.1.529 Omicron variant

Post by jimbob » Tue Feb 01, 2022 10:50 am

bob sterman wrote:
Tue Feb 01, 2022 9:22 am
Well reinfections are now included on the UK dashboard - and numbers haven't gone up nearly as much as I was expecting.

Only about 830,000 cases of reinfection have been added to the cumulative total number of cases.
With a 90 day threshold, BA.2 reinfections of BA.1 won't be counted, for example - that's small, but many Omicron (either type) reinfections of Delta will also be missed.
Have you considered stupidity as an explanation

User avatar
Martin Y
Stummy Beige
Posts: 2586
Joined: Mon Nov 11, 2019 1:08 pm

Re: B.1.1.529 Omicron variant

Post by Martin Y » Tue Feb 01, 2022 11:00 am

bob sterman wrote:
Tue Feb 01, 2022 9:22 am
Well reinfections are now included on the UK dashboard - and numbers haven't gone up nearly as much as I was expecting.

Only about 830,000 cases of reinfection have been added to the cumulative total number of cases.
Well that answers the WTactualF question I arrived with after clicking my usual UK Covid 19 statistics link and seeing the sky-high spike today.

User avatar
shpalman
Princess POW
Posts: 7177
Joined: Mon Nov 11, 2019 12:53 pm
Location: One step beyond

Re: B.1.1.529 Omicron variant

Post by shpalman » Tue Feb 01, 2022 11:33 am

I originally thought nah 830,000 is too high you're an order of magnitude out and then realized it was correct and corresponds to only about 4-5 days' worth of infections at the peak...

So yeah I suppose they didn't just have to run the stats through with a new algorithm to count reinfections yesterday, but had to also run the previous days' stats through in the same way to get the difference for the headline daily reported case figure.
molto tricky

User avatar
shpalman
Princess POW
Posts: 7177
Joined: Mon Nov 11, 2019 12:53 pm
Location: One step beyond

Re: B.1.1.529 Omicron variant

Post by shpalman » Tue Feb 01, 2022 12:11 pm

The Guardian isn't really able to figure it out and was maybe hoping to see a spike in today's report coming from all the reinfections suddenly being included.

Of course yesterday's number was higher than the number from the day before, because yesterday was a Monday and the day before was Sunday. The previous two Mondays however only had about 13,500-13,600 more cases than the previous Sunday, but yesterday had about 30,000 more than the day before.

The post-wave vaguely constant bit starting in mid January has nearly 10,000 more cases per day now on average (was ~90,000 per day, now ~100,000 per day), so reinfections are about 10% of cases?
molto tricky

User avatar
shpalman
Princess POW
Posts: 7177
Joined: Mon Nov 11, 2019 12:53 pm
Location: One step beyond

Re: B.1.1.529 Omicron variant

Post by shpalman » Tue Feb 01, 2022 6:24 pm

It seems like the protocol which allows for reinfection also changes which deaths were within 28 days of a positive test, i.e. the one signifying reinfection not the first one.

Anyway the total deaths in England within 28 days of a positive test quietly went up by about 900 more than the headline number.
molto tricky

User avatar
shpalman
Princess POW
Posts: 7177
Joined: Mon Nov 11, 2019 12:53 pm
Location: One step beyond

Re: B.1.1.529 Omicron variant

Post by shpalman » Tue Feb 01, 2022 9:42 pm

https://twitter.com/Pouriaaa/status/1488560398618050572
Deaths — within 28 and 60 days of a positive test — have now been updated to reflect reinfections, which were added to the UK Coronavirus Dashboard yesterday.

Based on the new definition, an additional 1,087 deaths within 28 days of a positive test have been identified.
Image
molto tricky

raven
Snowbonk
Posts: 609
Joined: Mon Nov 11, 2019 8:58 pm

Re: B.1.1.529 Omicron variant

Post by raven » Wed Feb 02, 2022 11:13 pm

jimbob wrote:
Tue Feb 01, 2022 10:50 am
bob sterman wrote:
Tue Feb 01, 2022 9:22 am
Well reinfections are now included on the UK dashboard - and numbers haven't gone up nearly as much as I was expecting.

Only about 830,000 cases of reinfection have been added to the cumulative total number of cases.
With a 90 day threshold, BA.2 reinfections of BA.1 won't be counted, for example - that's small, but many Omicron (either type) reinfections of Delta will also be missed.
Fits with what's just happened in my nephew's class: a group of kids who got Covid mid-December have just got it again now, 6 weeks later. They won't get counted as reinfections.

Post Reply