Covid-19 in statistics

[ursus arctos]

Interesting data here. Particularly the link to average household size.

https://twitter.com/albertsamaha/status/1258745087456169991

[Nocturnal Submission]

The ONS website has got some fairly detailed, granular statistics for the UK deaths from COVID-19, at the urban area and the Middle Layer Super Output Area (roughly equivalent to ward-sized area) levels. Quite fascinating.

[Fussbudget]

So, a question about the ONS data: are deaths registered for the area where someone lived, or the area where they died (i.e. if they died in a hospital, are they counted as a death for the hospital's local area)?

Because I keep seeing Middlesbrough really high up in the list of worst affected areas, but it has a death rate 4 or 5 times higher than any of the other local authorities on Teesside (one of the tables on the ONS website has it as 8 times higher than Stockton which would be utterly bizarre), and I'm now wondering if they're counting all James Cook deaths as Middlesbrough. Given that it's the biggest hospital in the region and serves all of Teesside plus chunks of North Yorkshire and County Durham, that would make the local authority level statistic fairly useless.

[ursus arctos]

They seem to record both "usual residence" and "place of occurrence". Not sure which is used for the maps.

See Section 5 here
https://www.ons.gov.uk/peoplepopulat...isticsjuly2017

[Fussbudget]

Thank you, I get a bit overwhelmed by the ONS website and the amount of information on there. Their maps and tables also seem to variously use data from local authorities, MLSOAs and urban built up areas which all have different boundaries so it's a little difficult to follow.

[ursus arctos]

Very much so.

Their summary of methodology changes in this respect goes into that to a certain degree, and England's penchant for changing boundaries definitely doesn't help.

It appears that the regional data is by "usual residence", though I am not sure if that applies to the more granular data as well. We have a similar issue here, in that Manhattan has an extremely high concentration of hospitals, some of which are primarily dedicated to terminal patients.

[Paul S]

A good article in the Daily Mail today (yeah I know, a rarity) but it is worth reading to see where we have done well with Covid19 as much as where as where we have gone wrong. You'll have to wade through the political bias - but that is true of every newspaper - but generally it isn't a bad analysis.

[ad hoc]

Paul S, the Dan Hodges of OTF

[anton pulisov]

A good article in the Daily Mail today (yeah I know, a rarity) but it is worth reading to see where we have done well with Covid19 as much as where as where we have gone wrong. You'll have to wade through the political bias - but that is true of every newspaper - but generally it isn't a bad analysis.

If you go to the euromomo website and compare the overall excess mortality z-scores of the different countries, the UK is by far the worst. It's not even close.

[Paul S]

An update from the Office for National Statistics. Total deaths registered in England and Wales up to 1st May v a five year rolling average.

2020: 247,261
5 years rolling average: 205,589

This means we are running 41, 672 over normal deaths.

[Paul S]

Here's a good one for the statisticians. According to the Centre for Evidence Based Medicine, the coronavirus outbreak is not an epidemic in the UK. The reason why is because:

The latest data from the Office for National Statistics (ONS) suggests that just 0.24 per cent of adults – approximately 136,000 people – have the virus. Separate surveillance by the Royal College of GPs indicates it may be even less.

Figures released last week showed just 0.037 per cent of people have the virus, although this is likely to be lower than the actual number because few people are visiting doctors with symptoms.

An epidemic is declared if the surveillance rate exceeds 40 per 10,000, but the new figures suggest it is between 24 and three in 10,000.

[Paul S]

Time for an update on coronavirus deaths per million head of population and the latest standings are as follows:

1. Belgium - 784
2. Spain - 588
3. Italy - 523
4. United Kingdom - 511
5. France - 410
6. Sweden - 358
7. Netherlands - 327
8. Ireland - 312

Belgium's lead at the top looks unassailable, but the UK is surging up the table and may yet finish with the World's second highest number of Covid-19 deaths per million head of population.

[Nefertiti2]

Deaths plotted against time of intervention

fromthe BMJ

https://twitter.com/globalhlthtwit/status/1261925658030202880?s=20

[ad hoc]

Time for an update on coronavirus deaths per million head of population and the latest standings are as follows:

1. Belgium - 784
2. Spain - 588
3. Italy - 523
4. United Kingdom - 511
5. France - 410
6. Sweden - 358
7. Netherlands - 327
8. Ireland - 312

Belgium's lead at the top looks unassailable, but the UK is surging up the table and may yet finish with the World's second highest number of Covid-19 deaths per million head of population.

As Nef's tweet makes clear this number for the UK is based on the 32000 ish deaths which doesn't include care homes or any of the other ones that didn't happen in hospitals (which the Belgian number does include). I don't know whether the UK has the highest number of deaths per million but it must be very close

[Paul S]

As of today the UK is now up to third place in the World rankings basing corona virus on deaths per million head of population. We've passed Italy, just France and Belgium left to beat:

1. Belgium - 801
2. Spain - 597
3. UK - 537
4. Italy - 535
5. France - 419

[Jimski]

I though San Marino was top? And Andorra near the top too?

[Fussbudget]

I take it Paul has ad hoc on Ignore Poster.

[Diable Rouge]

As Nef's tweet makes clear this number for the UK is based on the 32000 ish deaths which doesn't include care homes or any of the other ones that didn't happen in hospitals (which the Belgian number does include). I don't know whether the UK has the highest number of deaths per million but it must be very close

TBF, UK have been including care home deaths in recent daily updates, according to the Guardian's daily newsfeed.

[Diable Rouge]

My own county has seen daily cases virtually disappear - just 3 over the last week, and 7 over the last fortnight:

https://twitter.com/kerryman_ie/status/1263520938127249409

[Diable Rouge]

Last month, air and sea travel into Ireland fell by 99%, from 1.7m to 16,000, of whom 12,000 came by air:

https://www.cso.ie/en/releasesandpub...ticsapril2020/

[Paul S]

I thought I'd resurrect this thread so I could say something about the daily coronavirus totals in the UK which today hit 202, the lowest Friday figure since Friday March 20th. However, if you look at the NHS England website you will find that there were in fact just 10 deaths in English hospitals in the previous 24 hours which have been attributed to Covid19, the others are down to a time lag in reporting from many weeks ago. In fact daily deaths in England from covid19 for the last week or so have been:

11 June - 10
10 June - 35
9 June - 12
8 June - 3
7 June - 3
6 June - 1
5 June - 1
4 June - 2
3 June - 2
2 June - 1

Source:

[pebblethefish]

Yes, but who's to say that the total of 10 for yesterday will not be retrospectively added to three weeks down the line in similarly "time lag in reporting" figures?

[Paul S]

Yes, but who's to say that the total of 10 for yesterday will not be retrospectively added to three weeks down the line in similarly "time lag in reporting" figures?

That may well happen, I'm just looking for good news right now. There seems to be precious little of it at the moment.

[Fussbudget]

Yeah it's nonsense to look at the figures from the last week or so as some of the deaths have obviously not been reported yet. It literally says so in that NHS link:

Interpretation of the figures should take into account the fact that totals by date of death are likely to be updated in future releases for more recent dates.

[S. aureus]

That may well happen, I'm just looking for good news right now. There seems to be precious little of it at the moment.

You might want to look somewhere that isn't the UK's butchering of the pandemic, then.