Daily numbers, graphs, analysis thread 8

[Barracker]

Welcome to thread 8 of the daily updates.

Resource links:
Worldometer UK page
Financial Times Daily updates and graphs
HSJ Coronavirus updates
Johns Hopkins Coronavirus Resource Centre
NHS England stats, including breakdown by Hospital Trust
Covidly.com to filter graphs using selected data filters
ONS statistics for CV related deaths outside hospitals, released weekly each Tuesday

Thank you to all contributors for their factual, data driven, and civil discussions.

sorry, just to be clear there is no statistically significant death rate difference between cleaners and everyone else - the big risks (far higher than observed for say ethnicity) were:

security, processing plants, construction, most likely in that order. Cleaning and storage work was not necessarily higher risk.

Note that these are age-standardised, so for example if hod carriers are younger than average, then that would be taken into account if they had a risk of death higher than say clergymen (who will be older on average)

Security guards had a bigger risk than care home workers.

Bus drivers (2.5x) and taxi/cab drivers (3.5x) had much higher risk, which makes it unsurprising that given that if 1 in 5 Bangladeshi/Pakistani men are drivers, then that should result in a higher death rate for that group. HGV/van drivers did NOT have a higher risk.

Women in every kind of job in existence had a lower death rate than the average man, and whereas men in the most risky jobs had a risk MANY times the risk of the least risky jobs, there was a much smaller disparity between 'rich woman' and 'poor woman' than between 'rich man' and 'poor man'.

Female nurses (90% women) did NOT have a higher death rate than women generally, but male social care workers had a much higher rate than men generally.

It seems that despite all the 'clap for the NHS' stuff, female NHS staff have a death rate a small fraction of men working in menial low-paid jobs, and male NHS staff have a normal death rate compared to men generally.

It suggests that some of the PPE rhetoric may be misplaced - if doctors die from covid-19 then there will be questions asked, but if a minicab driver dies of covid-19 nobody gives a shit.

The thing at the moment is that people in all walks of life are dying of covid-19, so this is not something unique to the medical profession, and in fact not even elevated to the medical profession.

It reminds me of the police in that a lot of noise is made about police being killed in the line of duty, but it averages about one per year compared to 167 per year in agriculture, dozens in construction, etc.

Thank you place marking

Shoot's analysis of professions is a logical approach

• and the comparative risks may change once lockdown ends and millions go out to work again.

imo, we need to avoid lumping all BAME people together, when trying to analyse risks in more detail and understand the differences.

As a mixed race woman (English / Arab - lapsed Christian) I have little in common culturally and probably epidemiologically with someone African, Indian, Pakistani, Chinese etc

• other than the fact we have probably all suffered racism at some time, of course

We have different skin colours / Vit D levels, common blood types, average socio-economic levels, job prevalence, household sizes.....

If some or all of us are at higher risk than whites of the same age & sex, lumping us all together won't help in understanding why

fyi - Spiegelhalter from 2011 on risk at work:

www.bbc.com/future/article/20120330-death-in-the-workplace

the highest-risk occupation in the UK today is, believe it or not, commercial fishing.

< social distancing would be the least of their worries now ! >

A recent study found 160 deaths in commercial fishing in the UK between 1996 to 2005, which works out as 1,020 micromorts per year per fisherman.
This is staggeringly high – about the same as the risk British coal-miners faced in 1911.

Again some HCPs have been wearing PPE since early March. I don't think this has been corrected for in the ONS figures. So impossible to say if they are at greater risk or not.

Germany status:

The calm before the storm ... or before more calm ?
I don't know
^
The last 3 days, new cases have been below 800
and deaths [hospital + care home] were: 39,^ 20,^ 84^

R0 was 1.10 on 9 & 10 May, but fell slightly to 1.07 (95% conf, 0.88 - 1.29) on Monday
^
RKI on R0:

"the point estimate of R for a given day is estimated as the quotient of the number of incident cases on this day^
divided by the number of incident cases four days earlier.
....
Due to statistical fluctuations, which are amplified by the overall lower number of cases,
it is therefore still not possible to assess whether the decreasing trend in the number of incident cases observed over the past few weeks will continue
or whether case numbers will again increase."

Medriaswhip - the data on excess deaths by age was an analysis Shoots Fruits and Leaves did on the earlier version of this thread, after last Tuesday’s ONS data. If you look at the thread for last Tues or Wed you should find it there

Article in today’s times saying that research indicated that 4% of population had antibodies on average with 10% in London just before lockdown.

Question: if each person on average infects one other, does that mean now double?

www.thetimes.co.uk/article/lockdown-rules-for-work-report-employers-who-fail-to-keep-you-safe-staff-told-mk580p63p?shareToken=6ee87bf4707a15a0e75e5cffd856e99b

An R0 of 1 means a stable number of cases.

But even with a stable R0 the number with antibodies will have increased since lockdown.

A Chinese study showed that 95% of those infected developed antibodies- however there can be a time lag, the 95% figure was after 3 weeks.

Question: if each person on average infects one other, does that mean now double?

No, because the percentage of people with antibodies shows how many have been infected and produced antibodies since the beginning of the pandemic. Only a fraction of that population would still be infectious and able to transmit the virus.

It will be interesting to see the underlying data. I would have expected the London number to be closer to 15% (based on the back of the envelope calculations) but, of course, overestimating the level of immunity may be wishful thinking on my part.

none of the first 8 children to develop kawasaki like illness were white. It is unhelpful to lump BAME together when you have enough data to separate groups but that isnt always possible initially.

None of them were brown, either.

Sorry that's not right!

6 black
1 'Asian'
1 Middle Eastern

I read a few weeks ago that it was pointless to test for antibodies until four weeks after the infection.

@wintertravel1980 I got totally confused with my back of envelope calculations. Maybe wishful thinking too but if 10 % showing antibodies before lockdown then those 10% were infectious 2 weeks before lockdown on average ( or more depending on how long to produce antibodies) which means they would have been passing it on for a week or so, as would the people they infected?

@ajandjjmum article didn’t specify which antibody test. Some develop after 7 days and some up to 28.

ONS figures to May 1st are out. Slow improvement. Some of the headlines:

The number of deaths registered in England and Wales in the week ending 1 May 2020 (Week 18) was 17,953, a decrease for the second week running, but 8,012 more than the five-year average for Week 18.

Of the deaths registered in Week 18, 6,035 mentioned “novel coronavirus (COVID-19)”, which was 33.6% of all deaths; a decrease of 2,202 deaths compared with Week 17 (37.4% of all deaths).

The number of deaths involving COVID-19 decreased in most age groups, the exceptions being age groups 15 to 19 and 35 to 39 years (which increased by 1 death each).

The number of deaths in care homes (from all causes) for Week 18 decreased to 6,409; however, deaths involving COVID-19 as a percentage of all deaths in care homes rose to 37.8% compared with 35.3% in Week 17.

For the first time, all regions showed a decrease in the percentage of deaths involving COVID-19 in Week 18; the South East had the highest number of COVID-19 deaths, making this the first week that London did not have the highest count.

Care home deaths still more than double normal, but hospital deaths now only about 14% above normal.

Note that care home/hospital deaths should increase by about 3% from the numbers here by next week due to late registrations, and home deaths 6.5%.

Graph

I’m new to the thread, but enjoying the analysis - thanks.

Here’s an interesting piece on the maths of why an increase in overall R₀ might not necessarily be bad news (due to Simpson’s paradox):

unherd.com/2020/05/what-the-headline-covid-figures-dont-tell-you/

Some develop after 7 days and some up to 28.

IgG is initially higher and can be detected in the 1-3 week range, though it drops relatively fast. IgM rises to detectable levels about week 3, and seems to stay higher for several weeks but we don't have much data on longer term levels yet obvs.

COVID deaths to date in 0-19 age group = 10
Add the baby recently, to make 11

There are about 15 million children aged 0-19

Schools reopening may be a risk for some vulnerable teachers, parents and adult family members,
but it is not for children.

Someone's just directed me to this, which mentions over fifty thousand excess deaths in 2017/18 as a result of flu www.independent.co.uk/news/health/flu-vaccine-deaths-nhs-ineffective-crisis-bad-weather-illness-2017-a8660496.html

Am I missing something here? Because the obvious question is why current excess deaths are causing such a disproportionate panic if they've been as bad or worse in past years