Coronavirus (2021) thread

[Elbow beards]

Ideally one that compares the 2nd and 3rd waves in some form of stylised pyramid design would be best, i think.

Yes that's the one.

 

[blue b4 the moon]

Yes that's the one.

Dave Whelan has it on a black/white board but what with his broken leg he's not managed to get it here yet.

 

[M18CTID]

Just thought I’d get this one in first:

 

[Healdplace]

Zoe had a major tweak of its data yesterday as they said they would.

To try to match the UK data.

Now (as in Yesterday) their data has ended up far higher than the UK data and not far lower as was but might just hve been a trend they saw early!

Ongoing symptomatic cases 687, 666 (up from 478, 323)

And predicted daily cases now 60, 016 (up from 34, 637).


The term educated guessing is looking more and more appropriate to Zoe.

For what it is worth (not sure these days)

LONDON is now top region on 1160/1401 with NE & YORKSHIRE Next on 1006/1404

That at least is not far off the right order from two days ago in Gov UK data.


NW is now third on 849/1130 with N Ireland fourth 367/1008 and Midlands fifth 700/997

And the way they calculate placings is no more clear than it ever was.


Greater Manchester on the new look reads as:-

BOLTON 10,889

BURY 11, 248

MANCHESTER 13, 921

OLDHAM 15, 830

ROCHDALE 10, 829

SALFORD 14, 988

STOCKPORT 14, 708

TAMESIDE 5016

TRAFFORD 9477

WIGAN 14, 826


Not sure how much this has really clarified if anything - other than bumped everyone up high and left Tameside as by far the best and Trafford the only other one below 10K.

Will see what today looks like versus those numbers. Tameside's current cases are not justifying that rating at all, But Zoe and Tameside have been miles apart for ages and I am starting to think some residents are playing a joke on Zoe on purpose.

 

[roubaixtuesday]

The term educated guessing is looking more and more appropriate to Zoe.

It's a nice example of how self- selected surveys are very difficult to generate unbiased data from.

People who choose to download Zoe are likely to be more concerned than average about COVID, so more likely to get vaccinated and maybe more likely to behave responsibly. So they have to try and correct for these and other biases. As presumably the number and range of people using the app also changes all the time, that's really very difficult.

Testing data is also self selecting to an extent, but less biased probably.

The weekly ONS survey which attempts ac random sample of the population is probably most representative of the overall spread, but obviously lags by a week or two.

I think REACT uses the same panel every time?

All useful data, but all needs to be treated cautiously.

 

[Healdplace]

SCOTLAND DATA

22 deaths - was 19 last week

1825 cases - was 2086 last week

6.0% positivity - was 6.6% last week

488 patients - down 41 on yesterday (!) - was 543 last week

BUT 58 ventilated icu - up 7 on yesterday - was 47 last week


GOOD NEWS: Cases week to week still falling as is positivity so not obviously fewer tests related. Clearly now not a fluke. Though did not fall by a lot today.

Even better a BIG fall in patients and down week to week as well.


BAD NEWS: Highest death numbers in a good while and the icu numbers rising. But these are the last measure to change and so will be reflecting still the higher numbers . We need to watch to see them fall too next week or beyond if we are to really be convinced.

 

[Rammy Blue]

SCOTLAND DATA

22 deaths - was 19 last week

1825 cases - was 2086 last week

6.0% positivity - was 6.6% last week

488 patients - down 41 on yesterday (!) - was 543 last week

BUT 58 ventilated icu - up 7 on yesterday - was 47 last week


GOOD NEWS: Cases week to week still falling as is positivity so not obviously fewer tests related. Clearly now not a fluke. Though did not fall by a lot today.

Even better a BIG fall in patients and down week to week as well.


BAD NEWS: Highest death numbers in a good while and the icu numbers rising. But these are the last measure to change and so will be reflecting still the higher numbers . We need to watch to see them fall too next week or beyond if we are to really be convinced.

They didn’t have an extended Euros to cause issues, but they have had their schools shut for a few weeks now.

 

[Healdplace]

That Scotland and North West are still now seeming to be clearly tracking one another down in the order of data we would expect to see happening.

From cases to admissions to patients to icu and deaths.

AND with Scotland a week or so ahead of the North West in both but going the same direction.

That looks more than a coincidence given they were the two first first into this wave.

 

[Alan Harper's Tash]

That doctor has lots of time on her hands to tweet and write books

Do you prefer your doctor to work 24/7?

 

[Toney’s bookie]

It's a nice example of how self- selected surveys are very difficult to generate unbiased data from.

People who choose to download Zoe are likely to be more concerned than average about COVID, so more likely to get vaccinated and maybe more likely to behave responsibly. So they have to try and correct for these and other biases. As presumably the number and range of people using the app also changes all the time, that's really very difficult.

Testing data is also self selecting to an extent, but less biased probably.

The weekly ONS survey which attempts ac random sample of the population is probably most representative of the overall spread, but obviously lags by a week or two.

I think REACT uses the same panel every time?

All useful data, but all needs to be treated cautiously.

I got a letter inviting me to join the REACT study a few weeks ago having never been involved before. I assumed they have some turnover of contributors.