COVID19

[Doobs]

I know a couple of people in Singapore and their reporting is a little different to ours.

https://www.thestar.com.my/news/regional/2020/04/07/singapore-reports-66-new-covid-19-cases

But they are a smaller country (size and population), with less cases and very good procedures from the start.

They are not typical or perfect (there were scenes when ikea closed for example)

[kukushkin88]

Here’s a clear answer to kukushkin’s conspiracy theory (and confirmation of Jon MW’s proposition) on the slow reporting of deaths:

https://www.bbc.co.uk/news/health-52167016

Admittedly, it’s taken several weeks too long for this information to be confirmed.

It’s a bit soon to be calling it as that clear cut. Why aren’t the extra 300 included in the numbers yet? Why are we seemingly finding it so difficult to keep a tally compared to all the other nations with an equivalently large outbreak? (decimation of PHE and Local Authorities a factor I guess).

The tinfoil hat has been off for weeks on this but our poor capacity to maintain decent records has definitely been favourable for a government accused of being inexplicably slow to act.

What makes you think other countries are doing better?

I mentioned before part of the problem is that we know so much more about the UK's situation. Part of that is it is much more readily available because we're in the UK, part of it is transparency - we are just being given more information than some other countries.

But without the detail how do you know that a countries figures are accurate and don't have a lag? How do you know if their figures are only hospital deaths or all deaths? (for example)

The easiest indicator I guess is that our line is seriously wonky (should try and find a better word really but I like wonky). Other nations are declaring more transparently what they are recording.

France for example are declaring their total tally, as in the equivalent of our DHSC and ONS numbers combined. Arguably France are our closest equivalent if meaningful comparisons are to be made.

[Doobs]

Just for info, as I guess most aren't used to seeing Singapore reporting.  The infographics they produce linking cases are amazing.

https://www.channelnewsasia.com/news/singapore/singapore-reports-6th-covid-19-death-88-year-old-man-12609326

They really aren't typical of all countries.

Guess wave 2 could look more like this :/

[Jon MW]

Here’s a clear answer to kukushkin’s conspiracy theory (and confirmation of Jon MW’s proposition) on the slow reporting of deaths:

https://www.bbc.co.uk/news/health-52167016

Admittedly, it’s taken several weeks too long for this information to be confirmed.

It’s a bit soon to be calling it as that clear cut. Why aren’t the extra 300 included in the numbers yet? Why are we seemingly finding it so difficult to keep a tally compared to all the other nations with an equivalently large outbreak? (decimation of PHE and Local Authorities a factor I guess).

The tinfoil hat has been off for weeks on this but our poor capacity to maintain decent records has definitely been favourable for a government accused of being inexplicably slow to act.

What makes you think other countries are doing better?

I mentioned before part of the problem is that we know so much more about the UK's situation. Part of that is it is much more readily available because we're in the UK, part of it is transparency - we are just being given more information than some other countries.

But without the detail how do you know that a countries figures are accurate and don't have a lag? How do you know if their figures are only hospital deaths or all deaths? (for example)

The easiest indicator I guess is that our line is seriously wonky (should try and find a better word really but I like wonky). Other nations are declaring more transparently what they are recording.

France for example are declaring their total tally, as in the equivalent of our DHSC and ONS numbers combined. Arguably France are our closest equivalent if meaningful comparisons are to be made.

This is from 3 days ago

http://www.rfi.fr/en/france/20200404-france-reports-588-deaths-in-hospitals-but-numbers-are-much-higher

My impression was France was doing basically the same as us. Hospital deaths reported every day (without mentioning a lag) plus they are now getting on to filling in the gaps from other sources.

[kukushkin88]

This is good:

https://www.businessinsider.com/tool-shows-coronavirus-cases-rising-country-by-country-2020-3?r=US&IR=T
(you get to choose how to interrogate it)

France were very wonky as well until they made the decision to tally all numbers in as close to real time as they could. If we are serious about people obeying lockdown instructions being the most critical issue, we really need to avoid giving people the idea this is nearly over. Scotland for example, yesterday, probably would have been better saying we don't have the numbers for this period, rather than saying 2. Similarly the England data should be issued with a huge asterisk on a Monday. We also then get the reverse, it seems clear Wednesday/Thursday this week will be >1,000, this is a problem. I absolutely understand that we shouldn't be trying to draw conclusions from any period shorter than a week but this isn't how the public will be making their assessment and there are clearly a lot of people hanging on the daily numbers.

[kukushkin88]

Here’s a clear answer to kukushkin’s conspiracy theory (and confirmation of Jon MW’s proposition) on the slow reporting of deaths:

https://www.bbc.co.uk/news/health-52167016

Admittedly, it’s taken several weeks too long for this information to be confirmed.

It’s a bit soon to be calling it as that clear cut. Why aren’t the extra 300 included in the numbers yet? Why are we seemingly finding it so difficult to keep a tally compared to all the other nations with an equivalently large outbreak? (decimation of PHE and Local Authorities a factor I guess).

The tinfoil hat has been off for weeks on this but our poor capacity to maintain decent records has definitely been favourable for a government accused of being inexplicably slow to act.

What makes you think other countries are doing better?

I mentioned before part of the problem is that we know so much more about the UK's situation. Part of that is it is much more readily available because we're in the UK, part of it is transparency - we are just being given more information than some other countries.

But without the detail how do you know that a countries figures are accurate and don't have a lag? How do you know if their figures are only hospital deaths or all deaths? (for example)

The easiest indicator I guess is that our line is seriously wonky (should try and find a better word really but I like wonky). Other nations are declaring more transparently what they are recording.

France for example are declaring their total tally, as in the equivalent of our DHSC and ONS numbers combined. Arguably France are our closest equivalent if meaningful comparisons are to be made.

This is from 3 days ago

http://www.rfi.fr/en/france/20200404-france-reports-588-deaths-in-hospitals-but-numbers-are-much-higher

My impression was France was doing basically the same as us. Hospital deaths reported every day (without mentioning a lag) plus they are now getting on to filling in the gaps from other sources.

I'll look further in to it. A French ex-gf (small brag, less so since we broke up o/c) assured me they'd changed their recording method.

[AndrewT]

Raab and whatsherface both seemed nervous to me. Raab should be used to this sort of thing by now you'd have thought.

Tonight was garbage. Not the same without Hancock in charge.

Peston should be banned imo and Laura should be shuffled down the order occasionally. We're all having enough Groundhog Day without this being like the Q&A equivalent of Last Of The Summer Wine.

I read that only 14% of CV victims who go on to a ventilator make it out the other end. Scarily bad number if so.

I don't think that's right.   I saw that number and think it was an at the time number.   Others were still in intensive care and alive at the time.   Think it will be more like 50/50 in the end. 

Seriously bad for Boris, having watched it first hand last year, it is amazing what they can do.

Good luck Boris.

I had a reply to this with a link to a study but the thread pruning last night chopped it off - maybe the moderators could put back the non-controversial posts that were pruned(?)

This post is back in the thread now.

[kukushkin88]

This is a good detailed breakdown:

https://blog.ons.gov.uk/2020/03/31/counting-deaths-involving-the-coronavirus-covid-19/?WT.mc_id=2a19838a4a228a1d716f781f72dabf93&WT.sn_type=TWITTER&hoot.message=This%20is%20slightly%20lower%20than%20the%20figures%20reported%20by%20%40DHSCgovuk%20739%20as%20it%20takes%20time%20for%20deaths%20to%20be%20reported%20and%20included%20in%20ONS%20figures%20as%20last%20week’s%20blog%20post%20explains%20%5BLINK%5D&hoot.send_date=2020-04-07%2008%3A40%3A03&hoot.username=ONS&hoot.send_dayofweek=Tuesday&hoot.send_hour=08&hootPostID=655c9e01f437e72c1cf1344b239bc24c

[4KSuited]

This is good:

https://www.businessinsider.com/tool-shows-coronavirus-cases-rising-country-by-country-2020-3?r=US&IR=T
(you get to choose how to interrogate it)

France were very wonky as well until they made the decision to tally all numbers in as close to real time as they could. If we are serious about people obeying lockdown instructions being the most critical issue, we really need to avoid giving people the idea this is nearly over. Scotland for example, yesterday, probably would have been better saying we don't have the numbers for this period, rather than saying 2. Similarly the England data should be issued with a huge asterisk on a Monday. We also then get the reverse, it seems clear Wednesday/Thursday this week will be >1,000, this is a problem. I absolutely understand that we shouldn't be trying to draw conclusions from any period shorter than a week but this isn't how the public will be making their assessment and there are clearly a lot of people hanging on the daily numbers.

Do you realize that you’re advocating exactly what you’re criticizing the govt of doing? I.e. editing the data
I can only imagine your reaction if they followed your advice.
The only person I’ve noticed “hanging on the daily numbers” is you, tbh

[Doobs]

This is a good detailed breakdown:

https://blog.ons.gov.uk/2020/03/31/counting-deaths-involving-the-coronavirus-covid-19/

FYP

[kukushkin88]

This is a good detailed breakdown:

https://blog.ons.gov.uk/2020/03/31/counting-deaths-involving-the-coronavirus-covid-19/

FYP

ty

[kukushkin88]

This is good:

https://www.businessinsider.com/tool-shows-coronavirus-cases-rising-country-by-country-2020-3?r=US&IR=T
(you get to choose how to interrogate it)

France were very wonky as well until they made the decision to tally all numbers in as close to real time as they could. If we are serious about people obeying lockdown instructions being the most critical issue, we really need to avoid giving people the idea this is nearly over. Scotland for example, yesterday, probably would have been better saying we don't have the numbers for this period, rather than saying 2. Similarly the England data should be issued with a huge asterisk on a Monday. We also then get the reverse, it seems clear Wednesday/Thursday this week will be >1,000, this is a problem. I absolutely understand that we shouldn't be trying to draw conclusions from any period shorter than a week but this isn't how the public will be making their assessment and there are clearly a lot of people hanging on the daily numbers.

Do you realize that you’re advocating exactly what you’re criticizing the govt of doing? I.e. editing the data
I can only imagine your reaction if they followed your advice.
The only person I’ve noticed “hanging on the daily numbers” is you, tbh

I’ll try to waste as little time as possible on this. In what way am I suggesting manipulating anything? They clearly don’t have meaningful data for the relevant period in Scotland. They are saying they had 2 deaths in 24 hours, it’s not accurate and is potentially misleading, there’d be no harm in being upfront about that. You’re saying I’d disapprove if they were more accurate and more transparent? Are you sure?

[kukushkin88]

Some decent explanation of why it is a problem:

https://twitter.com/chrisgiles_/status/1247458186300456960?s=21

[Doobs]

Some decent explanation of why it is a problem:

https://twitter.com/chrisgiles_/status/1247458186300456960?s=21

Surely someone who has a job at the FT can do basic maths?

If a is reported as 100 when it should be 180 is not 80% understated but 44% understated.   80% understated would be reporting 36 deaths.

The real world is not ideal.  We know there is a lag; so we do the easy thing and adjust our model rather than pretending that there is a way of making the real world like our model.

Having said that, the 7,000 low estimate for deaths looks even more ridiculous now.  It is almost like they have never done this before.

[Doobs]

Some decent explanation of why it is a problem:

https://twitter.com/chrisgiles_/status/1247458186300456960?s=21

Surely someone who has a job at the FT can do basic maths?

If a is reported as 100 when it should be 180 is not 80% understated but 44% understated.   80% understated would be reporting 36 deaths.

The real world is not ideal.  We know there is a lag; so we do the easy thing and adjust our model rather than pretending that there is a way of making the real world like our model.

Having said that, the 7,000 low estimate for deaths looks even more ridiculous now.  It is almost like they have never done this before.

You can fit a curve to the data and show where they expect to be when the real numbers come out.   You can refit daily as new numbers appear.  You don't need to wait 3 days to model better.  I assume the Government have someone that can do this as they got enough skilled volunteers on the modelling side.