Daily numbers, graphs, analysis thread 11

[BigChocFrenzy]

Welcome to thread 11 of the daily updates

Resource links:

Slides & data UK govt pressers
NHS England stats including breakdown by Hospital Trust
ONS UK statistics for CV related deaths outside hospitals, released weekly each Tuesday
Financial Times Daily updates and graphs
HSJ Coronavirus updates
Worldometer UK page
Covidly.com to filter graphs using selected data filters ONS statistics for CV related deaths outside hospitals, released weekly each Tuesday
Plot COVID Graphs Our World in Data

We welcome factual, data driven, and civil discussions from all contributors 💐

Methyr has an even higher than Leicester, but there’s no talk of a lockdown there unless I’ve missed something? I wonder why?

@Derbygerbil

Methyr has an even higher than Leicester, but there’s no talk of a lockdown there unless I’ve missed something? I wonder why?

Because the FT have messed up.

The local health board for Merthyr added 104 old cases in on 27 June, backdated for over the previous weeks. The normal run rate is 2 -4 cases a day.

however, the FT have treated that 104 as actual new cases for this past week, which make Merthyr look like The Pit Of Doom. Its not. It's real number should be more like 25 cases per 100,000.

To be fair, this continual adding in of cases they found under an old mattress makes it very difficult to keep up to date in all the areas

Thanks @PatriciaHolm - that would explain it.

@BigChocFrenzy the incessant adjustments for historical data does reduce my trust in the data. Yes, I can understand this occurring now and again, but it seems to be routine... What’s going on?

@BigChocFrenzy

To be fair, this continual adding in of cases they found under an old mattress makes it very difficult to keep up to date in all the areas

Ha! True. But I would expect the FT journalists to have a good inbuilt sense of stats bullshit - as in, when you look at a piece of data like that, your first thought really ought to be - "huh, that doesn't look right. I'll double check that."

I suspect with the masses of data they just copied the PHE and other spreadsheets and applied the calculation macros

Too much trust in data they were given - bad habit for journalists not to question

All countries find a few bits of data when they shake out the old mattress,
but it does seem to be a habit for PHE and we often have to dig deeply to find the local stats

imo, incompetence rather than deliberately trying to confuse us all
The Leicester MP says Hancock signed a contract with commercial firms without realising the text didnot require them to give data to local authorities

Germany, France & Italy have more accessible systems to access test data, probably several other developed countries have as well

In Germany
I can chose to examine new infections / active cases / available ICU beds / deaths

Then I just type in the first letters of the administrative district name
(there are 401 districts, each with about 200k people)

For "new infections" I get an overview map of all the districts with circle size indicating the scale of the new infections

The RED circle indicates the district that exceeded the 7-day case threshhold for locallockdown

For the chosen local district, I also get:

Map showing size of incidence, with RED for the district that exceeded the lockdown 7-day incidence threshhold

and for the specific district:

A graph of rolling 7-day average of cases

of cases in 7 days and / 100k pop

total # of cases and / 100k
total # of active cases and / 100k
total # of recovered cases and / 100k - btw, why can't the UK state "recovered" ?
total number of deaths and / 100k
number of free ICU beds

Without a district, it shows these stats for all Germany

Patricia do you have a source re Merthyr?

@Adesignforstrife The PH wales dashboard shows cases and says that the 104 was old cases.

Hm well they do seem to be related specifically to a food packing plant (which is shown on the ft chart) as a result of increased testing so although the dates of infection are in question the ft info is not outright wrong.

That information looks really useful @Bigchoc. The UK government really needs to copy Germany and publish rates by local authority, colour code then - green, amber, red - based primarily on rates per 10,000, and set out what should happen with each (eg amber = extra resources to try to identify why it's so high, red = local lockdown) If they don't do this the media and social media are going to do their own analysis of where's at risk of local shut downs, making errors and causing panic about 200% increases when cases go from 1 to 3 in a week.

Simply sharing rates with local authorities - as they propose currently - isn't going to be the answer as they will ask have different ideas what to do about it - people will be angry if they're forced into local lockdown and a neighbouring area with higher rates isn't.

@Adesignforstrife apologies yes that does seem to be the source for many. Dropping them all on one day skews the data, but of course they do need to be accounted for. Seems to be an isolated outbreak hopefully.

May have unfairly maligned the FT a bit there!

Does anyone know if the wk 27 PHE weekly surveillance report plus the excel data file is due to be released at some point today? Wk26 was released 25th June so presumably as its a week later wk 27 should be today.

I presume it would be located here once released ?

www.gov.uk/government/publications/national-covid-19-surveillance-reports

ONS usually have their statistical releases (any subject) published by 09:30 on expected day of release. The surveillance report is from PHE of course so perhaps different standards.

On a super geeky note I see that the PHE Leicester report was using the newly released by ONS last week mid year 2019 population estimates in its calculations of rates per 100,000. Very much on the ball there, however have council ward level population estimates for mid 2019 been released yet note to self to actually have a look ? I see the Leicester report gave case numbers at ward levels, several of which were within a handful of cases, but not the case ratio which would have been more useful.

On the ONS page it is listed as scheduled for release today.

It's up

Survey is out, but I’m struggling to make sense of it. Estimate for prevalence is 25,000 people with Covid on average over the two week period. But the survey also suggests the incidence rate is 25,000 per week (3,500 per day). Surely both can’t simultaneously be true unless virtually 100% of community infections recover within a week? In previous weeks prevalence has always been somewhat higher than incidence rate - which makes more sense.

I think the low number of positives makes the data very hard indeed to interpret with confidence.

I'm making the cardinal error of posting before reading properly, but looking at the new excel file for cases per 100,000 a lot of areas seems to heading in a positive direction. I do need to check the time periods that the new excel and the last one cover however areas of concern like Bradford, Rochdale etc , whilst still quite high, have come down quite markedly as compared to the previous figures. Bolton looks to be going the other way, but still not at a particularly high level of cases per 100,000. Doncaster and Sheffield also a bit up, but again not hugely.

Given the overall P1+P2 case numbers are clearly showing a downward trend I think cause for optimism remains. Quick sort suggests 115 / 150 have a lower weekly case rate than the last file, and of the remaining a majority of those were low before or whilst marginally higher (so or example still less than

The PHE Surveillance doc is indeed up, with spreadsheet.

Summary - everything still headed slowly down.

Overall infections detected pillar 1 and 2 In the spreadsheet and word doc is 4,108 (week 26, June 22-29. This is not the number in the infographic, as that appears to be looking at a slightly different "week" to the spreadsheet and word document - there is a day overlap in the infographic between last graphic and this, and it seems to be about 2 day adrift from the spreadsheet. !! why??? why do they do this ridiculous shit?!!). Anyhoo, down about 30% whatever "week" you take when you look at the previous equivalent week.

Outbreaks in workplaces up (43 v 22 last week) but educational settings very stable (40 v 49, with only 18 proving positive). Only 58 in care homes vs 112 last week which is great progress if sustained.

Case rates for Pillar 2 down for all regions.

Of the 150 ULTAs, weekly rates of positive tests for P1 & P2 are down for 114 of them. Of the 36 who have seen rises most are very small, nothing even remotely like Leicester (who are on 141/100,000; the next biggest is Bradford on 45.8!). There are a handful that will bear keeping an eye on though (see attached for areas of biggest absolute growth).

All the other community indicators continue trending slowly down.

I think the low number of positives makes the data very hard indeed to interpret with confidence

Yes - they had 12/23,203 test positive this time, versus 14/24,256 last time. Which is a decline of 0.006%. Impossible to draw any major conclusions from, other than it's not going up!

I wouldnt assume the incompetence is in PHE England - it may be but it's just as likely, if not more likely, to be in the private contractors supplying them with the data. And of course Merthyr is actually in Wales - so it would be PHW correcting the data.

The good news is that the outbreaks are, generally speaking, not yet leading to lots of hospital admissions so probably wont cause a big increase in deaths. However people sometimes deteriorate after a week and in multigenerational households grandad may end up being admitted in 2 weeks time. Wales is still restricting social contacts so the outbreaks cause by workplaces can be brought under control quickly. Leicester and Bradford are not out of the woods yet.

I wouldnt assume the incompetence is in PHE England
I would
they have been incompetent in all areas since they were set up

Daily data is late tonight

@hopefulhalf I was just thinking that. Nothing mentioned on daily briefing either.

Oh was there a breifing I missed that ?

and no slides since friday....did they say they were going to stop them ?