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.

"the action of ramping up is useless without it having some kind of purpose"

A system of contact tracking & tracing, with mandatory isolation
plus monitoring for compliance, early treatment and for sudden deterioration

Thnks, lostmyshit That has v good tables

1 "underlying illnesses" include Diabetes, Lung Disease, Cancer, Immunodeficiency, Heart Disease, Hypertension, Asthma, Kidney Disease, and GI/Liver Disease

Doesn't tell us what type. The two types are very different diseases. Type 1 is not lifestyle related at all, couldn't be prevented and can only be managed. How well it is managed can be down to the individual but as anyone who has it will know, it is a constantly changing thing that has a mind of it's own and can go off the rails for seemingly no reason. Unfortunately, the current lockdown situation means, at least in everyone I know with Type 1, running higher than normal and constant monitoring and insulin corrections are having only a minimal impact. This information is worrying for all Type 1s, I think, although risk is very low overall.

@RedToothBrush I’ve posted about this a couple of times, not directly about frontline staff but answers some of herd immunity stuff

www.medrxiv.org/content/10.1101/2020.05.06.20093336v2

As regards the early seeding of Covid I would suggest that ski trips etc were nothing compared to the 11 million that left Wuhan before lockdown on January 23rd. All over Christmas, New year etc internationally.

Does anyone have risk data for different age groups with conditions like diabetes?

The other issues is that risk factors can change with age and type of disease; some types of asthma seem to get better with age, others get worse.

The danger of combining Rs
[[https://plus.maths.org/content/problem-combining-r-rates
plus.maths.org/content/problem-combining-r-rates]]

You might think that to get the overall value of R, which applies to the community, hospitals and care homes taken together,
you simply take the average of the two values.

In reality this isn’t the case and leads to a dangerous underestimate of the overall value of R.

It can even lead you to assume that the disease is under control when really it is not.

The reason for this is that the populations, people in the community and people in hospitals, are not completely isolated from each other.^
...
Say the value of R in the community is 2, so on average an infected person in the community infects 2 others in the community,
and the value of R in hospitals is 3, so on average an infected person in a hospital infects 3 others in the hospital.

We also need to take into account that there will be contact between the two groups, so say that on average a person in the community also infects 1 person in a hospital (in addition to the 2 people they will infect in the community).

Similarly, say a person in a hospital infects 1 person in the community on average (in addition to the 3 people they infect in hospital).

Next generation infections
Then as you can see from the diagram above, a person in the community and a person in a hospital together infect 7 people.
This means there are 7/2=3.5 new infections per person on average for this first round of new infections.

As we continue this, we see that the ratio between new infections and infections at the previous step grows to 25/7=3.57 (see the diagram below).

Carrying on further will eventually lead us to an overall value of R of 3.62.
Crucially this is higher than any of the individual values of R, which were 2 and 3.

@BigChocFrenzy there are alarmingly large 'unknown' numbers in that NYC data.

In the UK doctors are instructed to list all causes contributing to death on the death certificate. So if someone aged 30 has diabetes or cancer then that really should go down.

We have around 20% with no underlying conditions in the youngest age groups,but many of those with underlying conditions are things like pneumonia which are ambiguous.

Also it's very unlikely for a person of say 90 to have no underlying conditions. It seems that underlying conditions might increase your risk by up to double , but this is nothing compared to the raw risk due to age.

@Shoots.
When my mother died in February, after contracting respiratory virus , with chronic heart failure, all that was listed on her certificate was the cancer, which she had had for over 12 years. It was the complications that were the immediate cause of death but similar to the Alzheimer’s discussion there was earlier in the thread, they aren’t always listed.

"there are alarmingly large 'unknown' numbers in that NYC data"

Yes, too much data from most sources is incomplete, or lumps too many factors together
but age is dominant over all else

Hence my idea of age-related risk assessments for the classification of vulnerable, shielded in the discussion of people returning to work

In Germany, news report said ⅙ of the working population has an underlying condition - that's just roughly the 18-65 group
Obviously the remaining 5/6 can't support all of them to stay home on 80% previous pay or whatever.
So there needs to be some acceptance of increased risk, especially in the under 45 group

I also get irritated by "sex unknown" - what have they been teaching in medical school ? - when there is such a sex difference in outcomes.

It still hasn't sunk in to so many people that COVID is overwhelmingly a danger to the elderly, not most others
Hence the over-estimation of risk in workplaces and particularly at schools

Someone on another thread was claiming that the UK's high death toll is because of the BAME population being a higher % than elsewhere in Europe
That might be a factor if they had lived in the UK for several generations, but currently they are only a small % of care home residents and the very elderly

The BAME minority are not driving the overall UK death rate

It is the - much much smaller number of - deaths of the working population that indicate higher risk levels specifically for black people,
not BAME as an amorphous vulnerable clump.

re herd immunity:

Scientists need to determine:

• is there significant immunity conferred by very mild cases, or cases without symptoms ?

• how long immunity lasts

currently most scientists estimate about 2 years immunity
It could be very costly in lives - mostly very elderly ones - if people have to go through all this every 2 years

Maybe the temptation for government is that it would help solve 2 other crises:
pensions and housing / generational wealth

@whatsnext2 yes this is true. If someone happens to have diabetes dies from covid-19 then that probably wouldn't be listed.

BUT the large-scale study from patient records from the NHS computer does consider this and there's nothing there that particularly suggests that ridiculous notion that some people have that all the covid-19 deaths were about to drop dead anyway. We do know that those with uncontrolled diabetes were 2x more likely to have died, but the fact of diabetes being uncontrolled of itself suggests other patient issues beyond the diabetes itself....

The risk of death is very quantifiable by now, what a lot of younger people are concerned about is the risk of post-COVID chronic ill health. These risks are only just beginning to be observed and studied, and are likely to impact the NHS for months if not years.

I read somewhere that the average number of years lost to a COVID victim was calculated as 10,
but can't now find that

Yes, we need more studies into the risks of longterm illness for age groups and maybe risk factors too, sparetime

www.theguardian.com/world/2020/may/15/weird-hell-professor-advent-calendar-covid-19-symptoms-paul-garner

According to the latest research, about one in 20 Covid patients experience long-term on-off symptoms. It’s unclear whether long-term means two months, or three or longer

ShootsFruitsAndLeaves a more nimble NHS There's an oxymoron for you. It did make me laugh. Insider knowledge and all that. If only. DH consults (not medical) to the NHS, they have so little knowledge of "how to" it is frankly quite mind-blowing.

By the way, I really, really, appreciate your and BigChocFrenzy and Barracker inputs and stats, thank you so much.

It is the - much much smaller number of - deaths of the working population that indicate higher risk levels specifically for black people,
not BAME as an amorphous vulnerable clump

Not as such

We know more or less exactly how many black Africans live in Slough and are aged 72, etc.

We compare the numbers dead by age with the total deaths of that age dead and the total population dead of that age.

In this case there might be say 0.1% of the population aged 60-70 that are black, and 0.2% of the population aged 40-50, but they make up say 0.3% and say 0.7% of the deaths in those age groups.

Most of the BAME deaths are still old, because age is far more important than anything else, but the ratio instead of being 10 to 1 between over 60 and under 60 it might be 3 to 1, a ratio which has little to do with the BAME risk (though it might have something, e.g. if 70 yo black men are retired and 50 yo black men are working in jobs that are more exposed to other people than the population as a whole, then the working age black population might have a greater relative (but lower absolute risk than the retired white or black population) risk.

But the risk is calculated across the whole population and only appears if you adjust out the younger age profile of the BAME population.

Very interesting video piece on BBC this morning from the Jewish community in London. Also a very high percentage for infection and death. So many different angles to our issues in the UK.

Carrying on further will eventually lead us to an overall value of R of 3.62 Crucially this is higher than any of the individual values of R, which were 2 and 3.

No it isn’t. The community person infected 3 and the hospital person infected 4. You don’t just ignore the other 2 infections because they are in the ‘wrong’ category. Your own diagram shows this.

On BBC Horizon last night they were clear that black and Pakistani/Bangladeshi had a very much increased risk, anyone know where they got their figures from?

"No it isn’t. The community person infected 3 and the hospital person infected 4. "

No, my understanding is that he says those nominal values of 2 and 3 are only calculated within community and hospital respectively,
i.e. within their own domain
They are not whole population R values.

He says that they would additionally infect 1 other outside the domain for which they were calculated

Baaaahhhh Could that be due to an older population? Jews seem to have many people who live into their nineties.

Study showing IR and seroconversion in Birmingham HT health workers

www.medrxiv.org/content/10.1101/2020.05.18.20105197v1

BigChocFrenzy, do we have any data that breaks down the "non-Covid" excess deaths by age, please?