Daily numbers, graphs, analysis thread 12

[BigChocFrenzy]

Welcome to thread 12 of the daily updates

Resource links:

Slides & data UK govt pressers
UK dashboard sub-national data, local authorities
Beta Uk dashboard deaths, cases, hospitals, tests, partially sub-national
UK stats updated daily by PHE & DHSC
ONS UK statistics for CV related deaths, released weekly each Tuesday
PHE surveillance report infections & deaths released every Thursday with sep. infographic
NHS England stats including breakdown by Hospital Trust
FT Daily updates
HSJ Healthcare updates
Worldometer UK page
Plot FT graphs compare countries deaths, cases / million pop. / log / linear
Covidly.com filter graphs compare countries
Plot COVID Graphs Our World in Data

We welcome factual, data driven, and civil discussions from all contributors 📈📶👍

Interesting pre print about inaccuracy of R number

The reproduction number R for COVID-19 in England: Why hasn′t ″lockdown″ been more effective?

www.medrxiv.org/content/10.1101/2020.07.02.20144840v1

A very meaningful proportion of care home deaths will be deaths brought forward. We discussed it here several threads ago.
The average length of care home stay is estimated at 26 to 30 months depending on the source. It means that on average the residual life expectancy for an average care home resident is half of this, or 13 to 15 months.
The topic of care home deaths is highly emotional and gets a lot of attention but from the statistical perspective these are indeed deaths brought forward.
The conclusion for deaths in community (especially for middle aged population) will be very different

This - with knobs on! When the majority of deaths are in very elderly care home residents, to suggest that they have lost an average of 11-13 years of life is nonsense.

Holdingtight I’m curious about the mortality rates in the US compared to Europe as well. I’m thinking it likely has to do with the average age of covid patients being lower now in the US (haven’t seen stats but from reading news reports so possibly anecdote)and medics having learned how better to treat patients (I read about one hospital in Southern US pioneering what they called MATH+ and avoiding ventilators for example). I’d have thought the daily death curve would be starting to tick up in the US by now, given the surge in cases. It may just be a case of a longer lag of course but it’s good news (apart from giving POTUS reason to crow about being so brilliant at handling the crisis )

Presumably the care home deaths is why the MEAN is 11-13 years - they'll be the ones at one end of the curve. There will also be people at the other end of the curve, such as the young nurses who have died, who lost a lot more than 11-13 years - probably more like 30-40 years of life. Mean implies a range.

No-one is saying there have been no imminent deaths brought forward, just that it's not just people on the brink of death who have been dying.

This is the more or less which addresses this question www.bbc.co.uk/programmes/m000kdr6

This week's average cases is at 559 per day. Definitely lower than last week (although I can't quite remember what it was last week).

@PumpkinPie2016

This week's average cases is at 559 per day. Definitely lower than last week (although I can't quite remember what it was last week).

Average for 24th-30th was 867.

Seven day rolling average now 573, down from 892 a week ago.

Thanks @PatriciaHolm couldn't find my figures anywhere and I'm so tired today due to random cats squawking during the night!

Looking at today's (England only) csv file and splitting the content by weeks and taking the last three complete weeks - so w/e 21/6; 28/6 and 5/7 - recognising the latter may still have a few more cases to add, although today's published file shows another 265 cases added for Sunday 5th , and finally stripping out county councils and using their constituent districts:

4 LAs - Leicester, Bradford, Kirklees and Sheffield account between them for almost 25% of total confirmed cases for the most recent week, w/e 5th July.

24 LAs account for 50% of that week's cases

79 LAs for 75% of cases

165 LAs (so over half the total number of English LAs) account for just 10% of cases in that week.

Today's data all shows a positive picture almost across the board in those areas that might have been considered hot(ish) spots. Taking their numbers for w/e 21st June; w/e 28th June and w/e 5th July:

Barnsley : 135 - 96 - 42

Bedford : 81 - 36 - 34

Bolton 45- 68 - 43

Bradford 372 - 263 - 171

Central Beds 59 - 46 - 22

Doncaster 53 - 66 - 37

Hull 58 - 14 - 8

Leeds 156 - 98 - 44

Leicester 502 - 511 - 412

Oldham 91 - 74 - 45

Peterborough 56 - 46 - 34

Rochdale 118 - 83 - 72

Rotherham 91 - 74 - 57

Sheffield 123 - 148 - 86

Tameside 76 - 40 - 31

Wakefield 64 - 40 - 35

Wirral 62 - 28 - 8

Flatlining somewhat

Kirklees 131 - 121 - 131

and showing an upward trend, albeit from a relatively low base

Hillingdon 8 - 20 - 25

Liverpool 31 - 39 - 47

At the district local authority (so generally smaller populations than most of the unitaries and mets listed above

High Peak (derbyshire) 21 - 16 - 4

North East Derbyshire 14 - 11 - 8

Ashford (Kent) 54 - 37 - 27

Preston (Lancs) 53 - 24 - 12

Harrogate (N.Yorks) 44 - 18 - 6

Bassetlaw(Notts) 53 - 23 - 11

and some of the districts bordering Leicester

Blaby 27 - 19 - 21
Charnwood 29 - 25 - 25
Oadby & Wigston 23 - 31 - 14

@boys3
Can I just ask- is the above data positive if you lived in say, Bradford, and couldn’t face a local lockdown?
We have abided by every rule and have a UK break booked next week - self catering and planning rural walks and eating in/ takeaway!
Thank you

@Langismyhero

The reduction below the average in deaths could be due to other things as well as some covid-19 deaths bringing deaths forward e.g. less sporting activity, fewer people on the roads (I'd guess this is a big factor), fewer accidents on nights out etc etc.

There is this study which suggests mean years of life lost from covid deaths are 11 for women and 13 for men taking into account underlying conditions. wellcomeopenresearch.org/articles/5-75

I think the narrative - which is very convenient for the government - that all covid deaths are deaths brought forward is pretty much flat out wrong. There was a 'more or less' episode (radio 4 -excellent on statistics etc) which addressed this too and said the same thing.

surely a good proportion of care home deaths during March/ April/May in particular will be "deaths brought forward". Not all by any means but a proportion of them will be. This would be a fair point to make. To deny this could be the case is as much a political narrative as to claim the opposite. The truth is probably somewhere inbetween.

Very few people (relatively speaking) die each year in road deaths, sporting activities and accidents on nights out i would think.

@TheMammothHunters I'd view the downward trend in a lot of those areas as a positive, albeit not a cause for complacency. Bradford cases have more than halved compared with the position in that first week, so with a downward trajectory you'd hope that would continue and negate any need for any reversal of the easing of lockdown.

Certainly not complacent- I still wash my shopping 😂
But desperate for a change of scenery and some different walks and takeaways.
The lodge has a hot tub as well
Thank you

Not all by any means but a proportion of them will be. This would be a fair point to make

Yes, this is exactly what I was saying - hence the curve, the mean, etc. That some will be deaths likely to happen in the next months/year but definitely not all. The narrative from some quarters has very clearly been it's 'all' deaths that were about to happen. The more and less episode on it is very good and contains even more facts and figures. Their conclusion is that it is definitely not JUST people about to die.

@boys3 thanks for posting that. I agree that those figures do look encouraging.

I am in one of the hotish spots and I am pleased to see cases in my area have decreased a lot.

I will have a listen to more or less - thanks for recommendation
I would think you could look at the care home excess deaths for March/ April/May/ June and then the (presumably not necessarily) negative care home death rate for June/ July / August maybe September and the difference between these two figures will be those that could well have lived for longer.
That analysis makes sense to me - but I am not a statistician.

boys3 interesting stats!
what is going on in Kirklees,? there figures are both quite high and not going down!

I think Kirklees was a cluster around a food processing factory.

Holdingtight, sleepwhenidie

Death rate for USA - not as low as you may think,
depends how you measure

Deaths / million:

US has ⅔ deaths million of the UK
and has about 4 x deaths / million of Germany and Denmark

The USA deaths / millions should be lower than the average European country, because the USA has v low population density in several states

Low pop density is probably the main factor in the low death rate in Scandi / Nordic countries
(good govt, super-efficient public health & health services seem the reason for Germany's low deaths)

Currently, in the USA like several European countries,
the average age of becoming infected is much lower than before, e.g. than in NYC

The young are going out after lockdown was raised, then getting infected

• and they have much lower death rates

while the vulnerable elderly are staying home more and fewer of them are getting infected

Some of those in care homes would have died within a few months,
but iirc the average stay is 3-4 years

Elderly people outside care homes would be those with normally a much longer life expectancy

Overall, a woman in the UK who reaches age 80 has an
average further life expectancy of
10 years for a woman
9 years for a man

www.ons.gov.uk/peoplepopulationandcommunity/healthandsocialcare/healthandlifeexpectancies/articles/lifeexpectancycalculator/2019-06-07

btw - elderly does not mean feeble ,
but it does mean much greater risk if one catches COVID

I'm 64 and a gym rat who lifts 141 kg / 311 lb legs and 80 kg / 176 lb upper body
BUT
.... theoretically, I have 64 x higher chance of dying if I catch COVID than a fat 24-yr-old couch potato

My fitness would help to some extent - maybe my chance of dying is only 60 x hers
I would count as elderly, but my mum lived 30 years longer

64 is not elderly, surely?

My Dad died at 64 and he never got to be old.

Some of those in care homes would have died within a few months, but iirc the average stay is 3-4 years

I wonder what the stats are for care homes prolonging life?

My grandfather lived for 6 long, painful declining years in a care home. A nurse I know pointed out to me once that there was almost nothing that killed the patients once inside. The occasional outbreaks of flu or notovirus were how most of them died, other than the ones left to die from dementia. Which is a bloody awful death.

Outside of a home, I can't see that he would have lasted another year. So in his case the care home seemed to extend his life significantly.

If he had caught Covud in year 1 of his 6 year stay, arguably he would have been "deprived" of 5 years of his life. But that would be a strange way of looking at it.

This will be a bit outing, but analysing mortality and life expectancy is part of my day job. This blog: www.actuaries.org.uk/news-and-insights/news/what-difference-word-makes explains from an actuarial point of view why the “deaths just brought forward” concept is incorrect (it’s written by one of leading research actuaries in the UK in this area - the longer article referred to explained in more detail the modelling behind his conclusions. )

Part of the issue is that in general, people underestimate life expectancy, particularly of the elderly. At birth, your life expectancy is say 80 years. But that doesn’t mean, having got to 79, your life expectancy is 1 year. The fact that you have survived is relevant, and your life expectancy at that point is generally significantly more than a year, even if you have an underlying health condition (and most 79 year olds will have at least 1 health issue!)

@AprilLady

This will be a bit outing, but analysing mortality and life expectancy is part of my day job. This blog: www.actuaries.org.uk/news-and-insights/news/what-difference-word-makes explains from an actuarial point of view why the “deaths just brought forward” concept is incorrect (it’s written by one of leading research actuaries in the UK in this area - the longer article referred to explained in more detail the modelling behind his conclusions. )

Part of the issue is that in general, people underestimate life expectancy, particularly of the elderly. At birth, your life expectancy is say 80 years. But that doesn’t mean, having got to 79, your life expectancy is 1 year. The fact that you have survived is relevant, and your life expectancy at that point is generally significantly more than a year, even if you have an underlying health condition (and most 79 year olds will have at least 1 health issue!)

ONS have a good tool that illustrates this -

www.ons.gov.uk/peoplepopulationandcommunity/healthandsocialcare/healthandlifeexpectancies/articles/lifeexpectancycalculator/2019-06-07

If I say my age is 15, it tells me my life expectancy is 89.
if I tell it I am 80, say, it tells me my life expectancy is 90.

This has been suggested to indicate risk during second wave by LTA. Any comments?

covid19.demographicscience.ox.ac.uk/demrisk?fbclid=IwAR28Exqc_t65vtQgQpVRPT56kqY_ZoMj5muNgPXVS-P8d1nl5DFvl4YlqGU

really interesting link @whatsnext2

I'm not convinced however that the model theory would play out in practice, unless a second wave is going to be geographically very different to that which we've experienced during the first wave. So at a very, very simplistic level (that's me being simplistic) the impact colours on the map, say for age based hospitalisation and baseline bed capacity, look the opposite of what actual experience has been. It would be interesting if the model also showed a comparison with prediction based on model inputs and the actual reality of what has actually happened in the pandemic.

I'm a bit non-plussed by the use of ceremonial counties, but I do like the lower super output area level of mapping, although again I think the theoretical modelled results are likely to be spectacularly wide of the mark.