Coronavirus - Data Analysis Thread

[andy_s]

Elysium said:

Excess deaths update with todays ONS figures for the Wk ending 5th March:

Interesting; just a small point, the pool of 'vulnerable/elderly' has been diminished far beyond what would occur naturally [an acceleration/compression if you will]. Does this not now skew, to some degree, the average baseline of mortality - and our interpretation of comparative data sets over the long term [ie 1st wave vs 3rd wave].

Not sure whether these are factored in or not - or even whether it's a significant thing to factor in, but just 'gut' feeling...

[Elysium]

Update on todays data:

1. Tests and cases. The massive increase in testing has managed to squeeze out a few more cases dislodging us from the 18 day halving that we have followed since New Year.



2. Key metrics. When you correct for testing volume cases per 100k tests are still following that 18 day halving. Admissions are also bang on that trendline, but deaths are still falling more quickly. Although there is reporting lag, we now have 10 days with less than 100 deaths per day:







3. Three metrics overlaid. You can see how closely admissions and cases per 100k tests are tracking:

[Elysium]

A couple of further graphs exploring LFD and PCR tests in England:

1. Yesterdays massive test number is mainly LFD and presumably driven by schools doing their first round of home tests. 1.7 million LFD in England alone on the 21st March, but positivity is just 11 per 10000

In fact we have now performed a grand total of 35 million LFD tests, which have found just under 50k cases (14 per 10,000 overall).



2. LFD and PCR positivity are continuing to track each other, just at different levels:

[RSTurboPaul]

Thanks as always, Elysium.

Those figures are just.... mental...

Elysium said:

2. LFD and PCR positivity are continuing to track each other, just at different levels:

I'm still not sure I understand this.

I guess that whether or not they are 'wrong' / false-positive-tastic, they are at least consistent over time and reflect general infection trends within the population?



EDIT: Forgot to say - for total clarity, could be worth labelling the Y axes?

[Elysium]

RSTurboPaul said:

Thanks as always, Elysium.

Those figures are just.... mental...

Elysium said:

2. LFD and PCR positivity are continuing to track each other, just at different levels:

I'm still not sure I understand this.

I guess that whether or not they are 'wrong' / false-positive-tastic, they are at least consistent over time and reflect general infection trends within the population?



EDIT: Forgot to say - for total clarity, could be worth labelling the Y axes?

The left hand scale is PCR positivity and the right hand scale is LFD positivity

The graph shows that positivity changes for each proportionally, which suggests a link with prevalence. However PCR consistently returns a much higher %age of positive results. So when LFD briefly hit 1% PCR was above 20%

Some of this could be explained by differences in the number of people who take each test that are symptomatic. However, I think it also highlights significant differences in sensitivity between the tests.

[NoddyonNitrous]

Elysium said:

Some of this could be explained by differences in the number of people who take each test that are symptomatic. However, I think it also highlights significant differences in sensitivity between the tests.

They are indeed testing different populations: LFTs screen an asymptomatic group of people, PCR testing is targeted on those who are ill, symptomatic or who have tested positive / equivocal on LFT. No great surprise that it has a much higher positivity rate. I've done two LFTs per week since mid December (about 28 tests so far, all negative, and I'm to continue to do so). In general, people only get one or two PCR tests even if positive.

[Elysium]

Update on Excess Deaths following todays ONS release:

1. As of week 10, ending 12th March 2021, there are no excess deaths in England and Wales. In fact weekly deaths are only just above the 5 year minimum:



2. This graph compares the 1st and second waves of excess deaths to deaths 'due to' COVID:



Over the last 12 months (ie the entire pandemic period) there are 100k excess deaths, which represents a 19% increase in mortality.

[Otispunkmeyer]

Elysium said:

Update on Excess Deaths following todays ONS release:

1. As of week 10, ending 12th March 2021, there are no excess deaths in England and Wales. In fact weekly deaths are only just above the 5 year minimum:



2. This graph compares the 1st and second waves of excess deaths to deaths 'due to' COVID:



Over the last 12 months (ie the entire pandemic period) there are 100k excess deaths, which represents a 19% increase in mortality.

So we're currently in week 12

It will be interesting to see if that trend continues. Imagine, 3 weeks with deaths below the 5 year average, yet you still can't do anything. Strange times.

[Terminator X]

Otispunkmeyer said:

So we're currently in week 12

It will be interesting to see if that trend continues. Imagine, 3 weeks with deaths below the 5 year average, yet you still can't do anything. Strange times.

Plus masks and social distancing needed for "years" apparently ...

TX.

[havoc]

Terminator X said:

Plus masks and social distancing needed for "years" apparently ...
.

Masks I agree with in e.g. shops (if people can be trained to wear them properly...i.e. not under their fking nose)...social distancing feels like it might be OTT.

[Smollet]

havoc said:

Terminator X said:

Plus masks and social distancing needed for "years" apparently ...
.

Masks I agree with in e.g. shops (if people can be trained to wear them properly...i.e. not under their fking nose)...social distancing feels like it might be OTT.

I've twice not used a mask when visiting my local supermarket. Once completely forgot and only realised when I went to remove it, and it wasn't there. The second was deliberate to see what happened. No one said a thing. A few months ago I would've been hounded out of the shop and chased down the road.

[Terminator X]

Smollet said:

havoc said:

Terminator X said:

Plus masks and social distancing needed for "years" apparently ...
.

Masks I agree with in e.g. shops (if people can be trained to wear them properly...i.e. not under their fking nose)...social distancing feels like it might be OTT.

I've twice not used a mask when visiting my local supermarket. Once completely forgot and only realised when I went to remove it, and it wasn't there. The second was deliberate to see what happened. No one said a thing. A few months ago I would've been hounded out of the shop and chased down the road.

It shows how easily we can be socially conditioned though ... went out to visit a site last week, no need for a mask so I left mine in the car. We then decided to go to a local office of someone who was also on site. People started to put on their masks, I didn't have one so just went in bareback as it were. I felt really awkward tbf as the only person not wearing a mask!

TX.

[Elysium]

OK folks - quick update on todays numbers:

1. Tests and cases. Our crazy testing volumes are continuing with little sign that the kids are getting bored with it. The additional volume has arrested the fall in cases causing them to 'flatline':



2. Key metrics. Cases per 100k tests are still nicely on the 18 day halving line despite the erratic testing volumes. Admissions also seem nailed on to this trajectory. Deaths still falling faster. Although reporting lag remains an issue, we now have 2 weeks below 100 per day:







3. Metrics overlaid. All neatly following the same steady decline. And that is the one thing that worries me slightly. If this is really down to the vaccine, it seems to me that the 18 day halving should be accelerating?

[Elysium]

But whats going on with all the LFD tests (said no one ever)?

1. Firstly, PCR testing volumes are steady, but positivity is slowly declining. LFD volumes are now super spiky and there is a slight increase in positivity, which may or may not be meaningful:



2. This shows LFD and PCR positivity overlaid using different Y-axis so they can be compared. There is a gradual increase in LFD positivity since the school testing began. This started with a flatlining around the first week and is now turning upwards with the home tests. I have read a suggestion that this could be due to less consistent reporting for negative tests now they are being done at home.

[RSTurboPaul]

Elysium said:

3. Metrics overlaid. All neatly following the same steady decline. And that is the one thing that worries me slightly. If this is really down to the vaccine, it seems to me that the 18 day halving should be accelerating?

It seems that the Deaths curve is now / has been in steeper decline than the Hospital Admissions curve - could that indicate the reduction in serious illness and death that the vaccines promise?

I guess a graph with the rates split by age would be needed (like the Israel graphs) but I don't think the UK has that granularity?

[RSTurboPaul]

With thanks to BasherX for doing the work, here we have Government's graph from today's Press briefing - saying 'oooooh, look at those massive scary bars' - and that graph reworked onto the same X-axis:






Not quite so scary now...



The original should be on this page or the next couple, depending on your page settings:

https://www.pistonheads.com/gassing/topic.asp?h=0&...


with the Govt briefing (with thanks to Boringvolvodriver for posting the link) here:

https://assets.publishing.service.gov.uk/governmen...

Edited by RSTurboPaul on Tuesday 30th March 00:32

[Terminator X]

Infuriating me that they are still pushing cases scary cases. If deaths are not rising to suit then surely the vaccine has worked, all back to normal.

TX.

[panholio]

Terminator X said:

Infuriating me that they are still pushing cases scary cases. If deaths are not rising to suit then surely the vaccine has worked, all back to normal.

TX.

Deaths lag cases. Many said this last time "casedemic" etc. The difference this time is the vaccine, and all indicators suggest it works. I understand the caution though. This government have constantly been ciriticised for acting too late, so I get where they are at.

[havoc]

anonymous said:

[redacted]

I strongly suspect it's both...but we won't know for sure until we (don't) see the seriousness of the "dreaded 4th wave", and again as we head into winter when it and its brethren are more active.


As an aside, I know of a handful of +/- middle-aged people who are still suffering with long Covid, and 3x 40-somethings (Ok, 1 was 50) who ended up in ITU. So I wouldn't just get hung up on death rates...plenty of people still alive but who've had their lives turned upside-down by it...

By all accounts it's an odd disease, as there are a high number of 'outliers' in terms of medical outcomes.

[Murph7355]

anonymous said:

[redacted]

You quite clearly haven't listened to what he said.

The govt have been criticised heavily before for not acting quick enough/too late. So now they are taking a cautious approach with a 5wk firebreak between phases to ensure the data trends fully continue and no gotchas spring up.

In the circumstances of the last 12mths, that is a perfectly reasonable approach to take. When you consider how rapidly the vaccines have been developed and deployed, it is a perfectly reasonable approach to take.

There will be those at both ends of the spectrum that thoroughly disagree. Nothing at all would even get close to satisfying either end.