Daily numbers, graphs, analysis thread 6

[Barracker]

Welcome to thread 6 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.

There's an interesting article in Project Syndicate on the Swedish approach

www.project-syndicate.org/commentary/swedish-coronavirus-no-lockdown-model-proves-lethal-by-hans-bergstrom-2020-04?utm_source=Project+Syndicate+Newsletter&utm_campaign=554d179774-covid_newsletter_23_04_2020&utm_medium=email&utm_term=0_73bad5b7d8-554d179774-105044917&mc_cid=554d179774&mc_eid=733176e37e

Here's a graph I made with a log scale showing the relationship between deaths per million population, based on ONS registrations to 10th April.

As you can see the risk of death grows perfectly exponentially with age (there is almost no risk of dying of covid-19 in the younger age groups, so a missing death or two, or an extra one will affect that much more, hence the slightly imperfect fit to the line there)

This is based on population estimates 2020 in 2018 for England by age, 2018 estimates for 2018 in Wales, versus the 10,000+ reported deaths reported by ONS to 10 April. We've got far more than that number now, probably 30k?

Note that the risk of death is around 96 per million for girls aged 5-14, which increases to 524 per million for men aged 15-24 because young men are violent and stupid.

Road deaths peak at 127 per million for males aged 10-14, and 27 per million overall. Suicide is around 100 per million, and homicide 11 per million.

It is unfortunate the young population (under 30) cannot really isolate itself from the rest of us, as they really don't have anything to rationally fear. No deaths under aged 15 at all.

514 deaths today

I would prefer less anecdotal friend of a friend told me, even less of I heard somewhere and absolutely no blah blah.
I would absolutely prefer more direct words from the front line and much more Numbers we can trust.

Where is that number from? UK or England?

I am one of those who see Sweden’s approach as potentially one to follow and I see nothing in the evidence to change my mind.

When Sweden followed a lighter touch approach, we were told that they faced disaster and an exponential death rate. That has not been the case. Despite keeping bars, restaurants, gyms and primary schools open, their death rate per million is much less than ours, and probably a lot less as they include care homes in their stats and we don’t.

Their deaths are also plateauing (85 today which is lower than yesterday though there is day to day reporting fluctuation) and look to have peaked with much less cost to their citizens in terms of lost futures, livelihoods and quality of life. It also looks sustainable to follow for a longer period whereas here a lot of people are so fed up with the lockdown that it is becoming unsustainable.

Yes, their death rate is higher than their neighbours and if your only objective was to have zero coronavirus deaths, whatever the cost, then you may think that is how to measure it.

But if you want to balance competing objectives - saving lives from all diseases not just covid, protecting people’s livelihoods, people’s quality of life, it looks like their trade-off has been proportionate. We could minimise deaths from the flu by having a lockdown every winter but we do not do so.

Even if you think Scandinavia is different for some unspecified reason (oily fish, people living alone), these people still go to supermarkets, hairdressers, restaurants, and bars and have kept these going, albeit at a lower rate, without seeing exponential growth. This is good news. It gives us some idea as to what are risky situations and what aren’t.

That is England, the U.K is 641.

That's a relief that the trend is continuing downwards, but obviously very sad that over 600 people have lost their lives.

589 sorry! Multi tasking
589 total
England 514
Wales 17
Scotland 58

There's an article about Sweden's approach in The Times today. Share token for anyone who's interested. www.thetimes.co.uk/article/f2c80252-84c5-11ea-b876-ef9d21d57c48?shareToken

I'm trying to keep an open mind about whether full lockdown or other social distancing measures are going to be the must effective going forward. We're really only at the beginning of this pandemic and we need to keep reassessing the evidence and data as it emerges and be prepared to change course if necessary. What seem to be the most effective measures now may not be so in the longer term.

Since it's such a perfect relationship, I would be interested to see what would happen to the death rates per country if they were normalised on a similar age-based basis. I.e. if you take the equation there, which is

Σ(population (in millions) in age band e^(0.105age of youngest in band)) and then divide that sum by the population of the country then you will end up with a normalised death divisor.

For England & Wales this comes to 39,000 (this does not represent a total of deaths, it's just a reference number with which to compare a different countries which takes into account the population size AND the age of the population). So if you calculate a similar figure for other countries using the same equation and the relevant population pyramid, you can then compare the deaths.

For Italy, that number based on www.populationpyramid.net/italy/2019/ and treating 90+ as 90 in the same way as I have done for E&W, you get a figure of 31,112

Hence if you divide the corona deaths in E&W by 39.0 and those in 31.1 then you will get comparable figures, which are I believe 17,000 official deaths in E&W, and 25,000 in Italy, which normalises that the age-adjusted per-capita death rate in England & Wales is 45% lower than Italy. Which is quite considerable really.

That assumes of course that the 17k and 25k figures are themselves comparable, which might not be the case..... But at least it's better than comparing on countries blindly.

For example, if we take the population pyramid for Indonesia, which has 270.6 million people

www.populationpyramid.net/indonesia/2019/

then using the equation above we find that the age-adjusted population/death divisor is 31.1, which is almost identical to Italy with only 60.6 million people, and considerably lower than England & Wales with also around 60 million people

In other words, if all other parameters were the same, we'd expect based on the relative populations for Italy and Indonesia to have the same toll, despite much larger populations.

Of course other parameters are NOT the same, but at least we can see that all countries don't have the same risk level, and that a country with a relatively young population will, by default, suffer perhaps just 1/5 of the deaths per capita of a country like the UK.

There's good reason to suppose that poverty, economic damage, lockdown measures, etc. will kill more than covid-19 itself in developing countries.

Sorry, correction to my last post.... I forgot to multiply the population sizes by the exponents.

Correct numbers:

E&W: 39.0
Italy: 51.9
Indonesia: 52.5

Extra number
USA: 182.1

I.e. divide the 'total deaths' by these numbers to compare deaths per million population, normalised by age.....

So England & Wales is actually not much better than Italy, the figures are 436 vs 482 (Italy). But Indonesia with 260 million+ people does indeed have the same age-adjusted total death rate predicted as Italy with 60 million, based on deaths by age from the ONS data.

The relevant figure for China is 485 btw, that is to say almost 10x Italy's death total on an age & population-adjusted basis

Today’s numbers tweet

Good way of calculating risk of death due to COVID for developing countries @ShootsFruitAndLeaves.

@Barracker I am interesting in the clotting factor as a symptom too. Is the data available to show if any of the unexplained excess deaths in the ONS data up to week 15 are from strokes or pulmonary embolisms?

Article from the Economist comparing average deaths by country, additional covid deaths, plus unidentified excess deaths for the first part of the year.

Also interested in the clotting thing. I have factor V Leiden so another tick in the box for me...

Btw the above equation implies a doubling of risk of death with every 2/(e*0.105) = roughly 7years, so risk at age 28 is double age 21, 35 is quadruple 21, 42 is eight times higher, 63 is 64 times higher, etc.

So it would make sense to for example use more younger people in frontline roles

Has anyone update the graph showing death by actual date recently?

Teabaseddiet please don’t jump to the wrong conclusions!

What I know is that, so far, there has been no direct link to increase in the chance of getting Covid19 or developing any more severe symptoms if you are already suffering with risk of blood clots caused by the leiden factor v. I’ve been searching for this regularly because a close family member has found out he has this in their mid 30s. It turned out his father had it as well but it never manifested. So there has been no observation of this or caution or anything at all from a reputable source and I have been looking.

However, these blood clots the medics are noting as regards Covid19 sufferers, are baffling. Nobody knows why it is happening. It appears that those in their 30s and 40s even with otherwise mild or no symptoms, develop blood clots which lead to sudden severe stroke. As with most current findings, this is a relatively small sample but they’ve only noticed recently so they’ll probably take more notice and record better observations.

The two are not linked, but I, personally, did wonder if there might be a link. It may not be.

You’re not allowed to tick any box yet

I am in favour of a sciency thread. I started one but nobody contributed. I started it because I’ve come across a lot of interesting observations and studies from different countries. My internet history is ridiculous and it would be good to share. All small scale but very interesting. Like: is there alink to the TB vaccine and lower (reported!) rates of infection in counties from the African continent which lead me to another small study suggesting that in Nordic countries TB vaccine can be effective for longer (30-40 years as opposed to the expected 15-20 years). A Belgian study looking at the link between a mutated receptor/protein cell within the lungs prevalent in Nordic, Baltic and Eastern European natives which might explain reported lower infection rates. China released quite a few interesting findings on small scale right at the start.
I remember reading in The Times about a month ago that the government was warned against collaborating with the tobacco industry who, apparently, offered to share their knowledge and research on respiratory issues.

London deaths down hugely, well over 50% from peak.

England deaths down at least a third from peak.

All good.

Checking in... thanks for the informative thread

@Branster we got don't you link your thread here so we can start using it?

Ooh yes ditto Bran!

Thanks @branster 🙂