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 📈📶👍

SHoots Do you have a ballpark figure of how many would typically be added to total deaths for a week within a following period of say 2 months ?

I like this data source

www.ons.gov.uk/peoplepopulationandcommunity/healthandsocialcare/causesofdeath/datasets/deathregistrationsandoccurrencesbylocalauthorityandhealthboard

Delete the 'Health board' records (Wales only) in the 'all data' sheets, as they are a duplicate of 'Local Authority'

Then check the 'Occurrences - pivot table' , click on the pivot table and choose the 'Analyse' tab in the ribbon bar, then 'Field List'

Then remove Area name/Area code, and drag 'Week number' to the 'Rows' section, and clear the filter.

You can see there the six categories:

• Home [normal residence]
• Hospital
• Care home
• Hospice
• Other communal establishment [e.g., prison, children's home]
• Elsewhere [e.g., died in a car crash, friend's house]

It's noted in the 'Information tab'

"Figures by place of death may differ to previously published figures (week 15) due to improvements in the way we code place of death."

that they changed the codes of some places from 'other communal' to 'hospital' and similar things, to reflect what those places actually were,

So the first two spreadsheets published under 'previous data' are NOT comparable

If one compares week 16 to week 27 for week 16, then:

17% have been added to deaths at home
6% to hospital
6% to care homes
3% to hospice deaths (which is also a much smaller category)
6% to 'other communal (but only 8 deaths, so caution needed)
17% to 'elsewhere' (but this is a small number of deaths)

Also note that week 16 was high in the pandemic so it's not necessarily true that things would be the same today

Also by week 17 a lot of deaths had been added, so comparing week 16 @ +15 days (rather than + 8 days)
then the numbers aren't so big:

+7% home

• 3% hospital +2% care home +2% hospice -2 deaths other communal (so it looks like there has been some further re-coding of 'other communal' into their correct categories)

I don't think 'elsewhere' has been recoded, and if we look at the week 15 + 8 days spread sheet, there were then 180, 140, 152 deaths recorded 'elsewhere' in weeks 1, 2, 3, not to mention 247 in week 14. As of week 27 that was 215, 173, 175, and then 256 respectively.

So presumably those 30 or so weekly deaths added at more than 13 weeks after death, are the result of inquests that have now been concluded. Whereas we can see that between week 15 and week 27 there do not appear to have been many inquests concluded in respect of week 14.

So we can conclude that there is quite a backlog of deaths of inquests, since there is very little movement in the week 14 deaths 'eleswhere' between week 14 + 2 weeks and week 14 + 15 weeks.

The week 1, 2, 3 'elsewhere' deaths will likely be far from complete also, since they will include murders, drug overdoses, etc., where the inquests typically take 9 months.

This recent ONS release (17 July) purports to cover this issue

www.ons.gov.uk/peoplepopulationandcommunity/birthsdeathsandmarriages/deaths/bulletins/deathsinvolvingcovid19englandandwales/deathsoccurringinjune2020

And has supporting data:

www.ons.gov.uk/peoplepopulationandcommunity/birthsdeathsandmarriages/deaths/datasets/deathsinvolvingcovid19englandandwales

Do ignore the dates there, which are misleading and just download the 'June 2020' spreadsheet, which contains 'all causes of death' for March, April, May and June in separate columns, and realise that each spreadsheet is based on deaths to the first Saturday of the following month.

So if we look at 'June'

and then we cross-reference the 'annual deaths bulletin'

www.ons.gov.uk/peoplepopulationandcommunity/birthsdeathsandmarriages/deaths/datasets/deathsregisteredinenglandandwalesseriesdrreferencetables

And we look at the ICD-10 death codes U.509 (homicide) and V, X, Y, which are accidents/murder/suicide essentially, which we don't expect to be affected too much covid-19, then we can get an idea of what's going on.

For the latest (2019) release, that is for deaths registered in a calendar year, which for 2019 is just fine, as we don't have anything particularly unusual happening.

The final worksheet gives the outcome of deaths for 2019:

530,841 deaths:
65% with doctor's MCCD and no coroner referral
17% further sent to coroner but coroner decides not to proceed, so also registered on MCCD
0.5% uncertified
11% coroner-certified without inquest (post-mortem only)
6% coroner-certified with inquest

and Table 10 gives the U509, V-Y codes:

14,353 males
8,541 females

Unfortuntely this is not broken out in the same way as the monthly covid-period death counts, but we can see clearly enough that these are 1900/month, or 440 per week.

Meanwhile, the monthly death reports for the covid-period show for the violent/accidental death categories:

March - 377
April - 280
May - 138
June - 18

all by 4 July

Clearly a lot of uncontroversial deaths occurring in late-June would not have been registered by 4 July, so we'd best compare say the April data as of June with the April data as of July to get an idea if the other ICD codes are missing in significant numbers.

Looking at this, we find that:

covid-19 increased by 0.4% (124) for April between early June and early July
cancer/neoplasms by 0.2% (27)
diabetes by 1.7% (15)
dementia by 0.2% (15)

and in general non-violent/accidental deaths went from 74,605 to 75,042, or around 0.6%.

Violent/accidental deaths increased by around 118 (45%) at +2 months compared to +1 months.

Comparing March @ June to March @ July, meanwhile we can see that the non-violent/accidental deaths went from 46,538 to 46,812, which is a rise of 278, or 0.6%.

I am not aware of any data suggesting suicide, homicide and accidents have either risen or fallen, but it's possible that the emergency services have something.

If we assume that total violent deaths are constant, then we expect 1900 deaths for each of March, April, May and June, noting that we have so far registered (to 4 July)

47,189 total deaths in March
75,322 occurring in April
44,863 in May (of which 138 violent/accidental)
29,445 in June - but this latter number is aggressively truncated at July 4.

Violent/accidental deaths are around 4.3% of total deaths, and monthly average deaths are 44,000 with around 30% variance from winter to summer. So for example we can see that May probably has another 1750 violent/accidental deaths to be registered in due, along with a smaller number of non-external cause deaths, perhaps 700 or so.

For comparison 49,016 deaths have been registered as occurring in weeks 1 to 4 2020 (pre-covid), which is 53,100 on a monthly basis, and this number will be higher when all deaths for January 2020 have been registered, sometime in 2021.

Even European countries that did "well" like Germany, had considerable initial problems with e.g. insufficient PPE emergency stocks, superspreaders / clusters at carnivals etc,
because they had prepared for the "wrong" kind of pandemic: flu

Thanks for that excellent answer, shoots

I'd have expected the oldest staff to be protected
The BAME increased risk, once we normalise for occupation etc doesn't compare with the risk from age,
which increases about 10x for every 20 years of age

Whether you a man or a woman is a big issue that is being ignored in order to focus on ethnicity.

The ONS studies have found no increase in covid-19 death risk for female health care workers as opposed to female workers generally, while male health care workers did have such a risk.

www.ons.gov.uk/peoplepopulationandcommunity/healthandsocialcare/causesofdeath/bulletins/coronaviruscovid19relateddeathsbyoccupationenglandandwales/deathsregisteredbetween9marchand25may2020#women-and-deaths-involving-covid-19-by-occupation

In addition, the absolute size of the risk of death for working age people was more than double for men as women.

Focusing on protecting ethnic minority staff while ignoring the issue of sex, and age is not an evidence-based approach.

And we look at the ICD-10 death codes U.509 (homicide) and V, X, Y,

erm, that should of course say V, W, X, Y

I do no my abcs.

We should compare excess deaths for the same period in previous years

As shoots has explained, we'll have some additional deaths registered for this period later, so the full picture will only be in 2021

The COVID deaths were high during March - May, which is past the UK flu season,
but that was mild, so fewer of the very frail died last winter

We need to consider that some of those who dies from COVID would have died within the year anyway

• I have seen estimates of as many as ⅓ of COVID deaths, but not an estimate with detailled stats working shown

Also, this coming winter flu season is likely to be much reduced compared to the usual,
because of SD, masks and increased flu vaccination takeup

So getting the full picture in say late 2021 will be complicated and must consider many factors
I'm really looking forward to getting my teeth into ONS reports then

True, sex is a significant factor

Studying the German deaths, when I weight the older deaths wrt the larger % of women alive in those age groups,
the factor looks even greater,
more like 2.5 higher risk for men

"Focusing on protecting ethnic minority staff while ignoring the issue of sex, and age is not an evidence-based approach."

Absolutely
and I'm agreeing as someone of mixed race;
my age of 64 is what makes me cautious about some activities, knowing that my risk is approx 100 x that of a 24-year-old
I don't even consider the risk of my race, compared to that

I've mentioned before that employers here several weeks ago ramped up their standard schemes for early retirement from age 60.

I know this could lead to staff shortages and slightly reduced pensions,
but I hope the age 60+ teachers and HCPs are allowed early retirement, rather than being pressured to stay on.

Bringing doctors out of retirement, especially male BAME, looks a seriously bad decision for them personally, even though it was for the public good.
We've seen in several countries how this group suffered a heavy toll in the early days, when there was insufficient PPE

imo, retired staff - especially male, overweight, with health conditions etc - probably shouldn't be on the frontline,
at least unless full PPE stocks have been prepared in advance

Bradford choir did NOT have COVID back in January:

www.bbc.com/news/health-53447899

"After two weeks of intensive laboratory work, the results come back.
The choir members and the landlady of the Bulls Head in Baildon - whom the singers appear to have infected with their virus -
are all negative, for both antibodies and T-cell memory.

This is fairly persuasive evidence that they did not in fact have Covid-19 and
shows that there are other nasty viral illnesses out there, not to mention flu, that can sometimes make people ill enough to need hospital treatment.

Bringing doctors out of retirement, especially male BAME, looks a seriously bad decision for them personally, even though it was for the public good.
We've seen in several countries how this group suffered a heavy toll in the early days, when there was insufficient PPE

These doctors volunteered. That was probably a wrong decision, but when I read the biographies of the first doctors to die, a number were either retired and working voluntarily, or had chosen to stop doing their normal jobs and work in covid.

I think we should be clear about agency here.

We are talking about experienced senior doctors. Not fruit pickers working for £3/hour.

That's not to say it was necessarily a good idea to appeal for 'people to come back to the NHS' when those people were in high-risk age categories. But likely these doctors felt they were doing a good thing by this in a sort of 'is there a doctor on this plane' kind of way.

I certainly didn't wish to suggest they were pressured:

Retired doctors volunteered in many countries, despite the risk, out of a sense of public duty

However, I'm thinking that now the initial emergency is over, the authorities should thank them for their valiant service - and then gently ease at least those 65+ back into retirement.
Unless PPE can be guaranteed, imo they are at too high risk

Just a personal opinion though

Covid has certainly pushed certain people into retirement.

The current England v W Indies cricket series is interesting in that they have separate covid bubles for teams, media, and no spectators.

Despite this, Geoffrey Boycott aged 79, opted to retire rather than risk it as a commentator.

Also a player was suspended after he briefly came into contact with another person (not infected), reportedly a fellow player not involved with the series. It is reported that the whole series would have been cancelled if his breach had been uncovered a few days later.

I think the player in question did a runner to meet his girlfriend. It was completely off, given that all the other players were holed up and weren’t going anywhere or seeing family (or having some fun going out and about in the case of the visiting team). He was not a team player when it was most required!

NHS England have announced today's Hospital deaths number - 8, all from this past week.

Pretty sure that's the first single digit day, super low even for a Sunday (last Sunday was 15; previous low was Monday just gone, on 11).

The grey area with a time limit to a positive tests comes from people who have been discharged from hospital but might die afterwards from complications (which is potentially likely given the toll covid takes on some patients). If a time limit is imposed then these people would not be included in the daily figures and the "official" UK death toll which only counts people with a positive test. They would be included in the ons figures though. Ideally you'd have a time limit on a test result unless the death cert said covid. This would rule out people who've obviously died of other causes (such as car crashes) but ensure we capture most covid deaths.

The problem with the current system is not (as have been said) that we've been over counting overall. The potential small over count now in no way makes up for the large undercounting at the peak as evident by excess deaths. The issue is that any overcounting now makes our recovery seem slower than it is and feeds in fear about opening up the economy as lots of people still focus on the day to day figures rather than the rolling averages. I'm not a fan of extremists at either side of the argument, we shouldn't open up too quickly but we do need to open up if the data shows we can do safely.

I welcome the review into how the figures are being reported but I wish the dhsc hadn't paused the reporting in the meantime like they did for people tested (although at least the data is still relatively easy to find). It encourages tin foil hat brigade (both sides) that think the gov are trying to hid things.

The grey area with a time limit to a positive tests comes from people who have been discharged from hospital but might die afterwards from complications (which is potentially likely given the toll covid takes on some patients

If they have Covid on the death certificate at all, then they are currently counted and still would be under a 28 day limit - the issue is that we are currently counting people who die not in hospital, with a positive test more than 28 days ago, and who do not have Covid on the death certificate at all.

If someone dies of CV complications later than 28 days, CV is likely to be on the death certificate as a contributor to death and thus they will still be counted.

Are they counted in the daily figures though? My understanding is those are just people with positive tests. The ons figures released weekly include those with covid on the certificate but not the daily figures which count towards the "official toll"

www.gov.uk/guidance/coronavirus-covid-19-information-for-the-public here under deaths it states deaths with a lab confirmed postive test. Afaik only Begium count covid on certificate without positive tesr

The ons death figure is above 50k so higher than the 45k used by the uk government

www.ons.gov.uk/peoplepopulationandcommunity/birthsdeathsandmarriages/deaths/bulletins/deathsinvolvingcovid19englandandwales/deathsoccurringinjune2020

@Littlebelina

Are they counted in the daily figures though? My understanding is those are just people with positive tests. The ons figures released weekly include those with covid on the certificate but not the daily figures which count towards the "official toll"

www.gov.uk/guidance/coronavirus-covid-19-information-for-the-public here under deaths it states deaths with a lab confirmed postive test. Afaik only Begium count covid on certificate without positive tesr

Sorry, now slightly confused - our daily announced figures include those who have had a positive test at any point right now. Is that what you mean?

The ONS figures cover those with CV on the death cert but don't require a positive test.

@ShootsFruitsAndLeaves

> Bringing doctors out of retirement, especially male BAME, looks a seriously bad decision for them personally, even though it was for the public good. > We've seen in several countries how this group suffered a heavy toll in the early days, when there was insufficient PPE

These doctors volunteered. That was probably a wrong decision, but when I read the biographies of the first doctors to die, a number were either retired and working voluntarily, or had chosen to stop doing their normal jobs and work in covid.

I think we should be clear about agency here.

We are talking about experienced senior doctors. Not fruit pickers working for £3/hour.

That's not to say it was necessarily a good idea to appeal for 'people to come back to the NHS' when those people were in high-risk age categories. But likely these doctors felt they were doing a good thing by this in a sort of 'is there a doctor on this plane' kind of way.

Absolutely Shoots. Matt Hancock asked for help from them 20/3 which was just at the beginning of lockdown. Did we know about about the BAME Factor this early?

Are we getting better at treating Covid19?

Considerable improvement has been reported in the USA and other countries too, now in the UK
A big improvement for the UK has probably also been in realising the importance of early treatment !

Some improvement in stats in these countries will be because of:
. the noticeably lower age of those infected after lockdown is relaxed
. improved immune systems in summer
. warmer weather is less optimum for COVID

However, Australia is reporting even greater improvements - and it is their winter

actuarialeye.com/2020/07/19/covid19-reflections-43/

'At least in the UK, yes.
Actuaries in the UKK_ (bulletin 50 in the link) have analysed the data there and found a halving of the death rate for those admitted to hospital with Covid19.'

www.google.com/amp/s/www.bloomberg.com/amp/news/articles/2020-07-18/florida-s-desantis-says-virus-antibody-tests-show-16-positive

What do you all think of this?
16% antibodies in florida. Population of 21.5 million suggests 3.5 million infections.
Isnt that the same amount of infections predicted we have had?
Yet they have under 5000 deaths.
Whats gone so wrong?

"Did we know about about the BAME Factor this early?"

Coquohvan We knew about the massive dominance of age as a risk, from China
As upthread, BAME is a much smaller increased risk in comparison to age
(politics increases its prominance)

The issue was that retired doctors were asked to return, whatever their race

Not just in the UK, but in some other countries e.g. France

Yes, that's what I mean. The daily figures only include those with a positive test. These are the ones included in our death toll of just over 45k used by the uk government/worldometer. A time limit on positive risks missing some of those.

They would still be counted in the ons figures (which I believe is over 55k for covid on death cert for whole of uk as link above only includes England Wales) but these aren't the figures widely used in the press/by the government.

It will all come out in the wash eventually but we need to be careful of the optics if we change how we report things if it looks like we no longer reporting some deaths we were before.