What's the counterfactual? How many of these people would have died this year anyway?

[Guineapigbridge]

As an economist I am finding the statistics on covid-19 frustrating. 227,300 people have died with Covid-19 as of this morning. According to the statistics from New York that are reported on the Worldometer site, 75 percent of those who died will have had pre-existing conditions including Diabetes, Lung Disease, Cancer, Immunodeficiency, Heart Disease, Hypertension, Asthma, Kidney Disease, and GI/Liver Disease. A further 14 percent of them were over 75 years old when they died, even though they didn't have pre-existing conditions. That makes it 89 percent of people who are at quite high risk of death in a given year anyway, right? So has anyone run the numbers on how many would've died anyway? Without a counterfactual a pure count of deaths is meaningless from a policy perspective.

We'll have to agree to disagree.

I have heard them, ministers as advisors alike, say, repeatedly, what the measure is and why it is limited but that it is one consistent measure that can/could be collected daily and reflects the changes in infection and death rate.

Thinking about it, that's where I got the "You have to remember..." from. Said in a lovely Irish brogue.

Spiegelhalter is a renowned academic on risk assessment and statistics. Just because Grant Shapps quoted him doesn't mean is bunk science.

I said right at the beginning of this that the best way to do this would have been incredibly unscientific but relying on the medical professionals to check an extra box on the death certificate / reporting that said 'in your opinion, should this death be attributed to Covid 19?'

That way you are turning the grey into binary yes / no via the 'translation' of a qualified medical person who will be best placed to answer.

So although unscientific in that it relies on a human, it's data 'from source' if you like that can then be totalled easily and quickly in real time - instead of trying to work with massive data samples and extrapolate info out of them based on stats of likelihood of dying of x y z etc etc

Nope! I read that again and again and still can't get to grips with it. Could you explain it again, please?

Is that for me?

Basically, nurse caring for person looks at notes / history and uses their knowledge and experience to judge if this should be a COVID death or a normal death.

Eg if all of the below died 'with' Covid:
my grandad with advanced vascular dementia - not a Covid death
my mate with final stage cancer, not a Covid death
my husband with well managed asthma - yes a Covid death
Very active 78 year old who lives opposite - yes a Covid death
Very frail 78 year old with multiple hospital admissions this year and underlying conditions - much trickier which is why it would be a judgement call

Ticks box

Box gets reported that day - same way death stats are recorded.

But nurses don't have any part of death certificates.

And the from / with is understood by doctors, clinicians, statisticians.

And I still don't understand why?

And hospital deaths are reported daily.

And now care home deaths are too.

my grandad with advanced vascular dementia - not a Covid death

Why not attribute that death to Covid infection ? if that is what has killed them ? If the same person had developed a community aquired pneumonia and died, that would have been put down as cause of death, with the dementia as a contributory factor ?

Sorry massive bolding fail there

So on a MCCD it would be along the lines of

I (a) Disease or condition leading directly to death - institial pneumonitis

(b) Other disease or condition if any leading to I(a) - Covid-19

(c) other disease or condition if any leading to I(b) - vascular dementia

II Other significant conditions contributing to death but not related to the disease or condition causing it - diabetes mellitus

The thing that actually causes you to die it what is written first, so the pneumonia, this developed because of infection by Covid-19 and you developed the pneumonia as a result of being infected by Covid-19, because you had vascular dementia. I hope that makes some sort of sense and I have explained it clearly enough ? If not , I have a link I can post , that I got the information from

Try to remember that they also state that those numbers are for those tested, that they are trying to show the pattern of infection and death, not necessarily the absolute rate.

Useful insight; two different objectives for the data.

One thing that's interesting is that although ^^now the deaths in the uk are above average (apparently, I haven't looked at the stats), previously, even when the bbc et al were doing super dramatic headlines, they weren't. Of course, that's partly to do with a death time lag (ill people not died yet) and a reporting lag (non hospital deaths) but it also shows that even when the death rate was actually lower than the 5 year average, newspapers were reporting covid deaths in a v dramatic headline grabbing way. At that point the story should have been 'omg where are all the covid deaths?', but it wasnt.

My point is it's wise to take newspaper articles with a pinch of salt!

David Spiegelhalter has written some good stuff on this

This is the article he wrote in The Guardian the other day, Guineapigbridge. As Myotherprofile says, it has been quoted by the government. Article says the same as CuriousAboutSamphire, that is that it will be a while yet before we can understand the underlying factors behind excess deaths:

www.theguardian.com/commentisfree/2020/apr/30/coronavirus-deaths-how-does-britain-compare-with-other-countries