Daily numbers, graphs, analysis thread 10

[BigChocFrenzy]

Welcome to thread 10 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

We welcome factual, data driven, and civil discussions from all contributors 💐

It is wrong that they keep saying 'of those infected' when really it is 'of those tested positive' isn't it. I mean it could be that women have less symptoms and therefore not tested, (which would tie in with other countries where men are more severely affected) or there are cultural reasons why men are tested more, creating bias in the figures. (on the BBC India report)

"South Asian mortality puzzle"

https://www.ft.com/content/aaa2c3cd-eea6-4cfa-a918-9eb7d1c230f4

South Asian hospital patients with Covid-19 are significantly more likely to die than patients from other ethnic groups, according to the largest study of its kind,
but researchers say they cannot point to a single reason for the higher mortality rate.

The study, by academics at the University of Edinburgh, found that
patients with Indian, Bangladeshi, Pakistani and other South Asian backgrounds were 20 per cent more likely to die from coronavirus than white, black or other ethnic patients,
and were on average 12 years younger.

It also found that black patients were no more likely to die than white people once admitted to hospital.

“It might have to do with viral load, or it might be something to do with genes,”
said Ewen Harrison, professor of surgery and data science at the University of Edinburgh and lead author on the study,
which is being peer-reviewed for publication in The Lancet medical journal.

About a fifth of the excess deaths in South Asians were linked to pre-existing diabetes,
which was much more prevalent in all ethnic minorities than in white patients.

The study, which examined the records of 35,000 patients, found
there was no statistically significant difference between patient outcomes of people from black or other ethnic minority backgrounds compared with white people after they entered hospital with Covid-19.

But it found that Bame patients were at higher risk of being in intensive care and requiring mechanical ventilation than white people.

PHE and census charts on health conditions & job types for ethnic minorities:

btw, chart shows unusually high % of Pakistani woman aged 55+ diagnosed with diabetes
doesn't specify type, but presumably means T2, or at least T2 predominates hugely.

iirc an ONS report also found this higher mortality of South Asian patients

==> Analysing ICNARC report:

stuart mcdonald@ActuaryByDay

Let’s look closer at ethnic differences in admissions.

Non-white ICU patients are:

• significantly younger; 5 years on average!
• 44% more likely to belong to the two most deprived socioeconomic groups.
• more likely to be a healthy weight; much less likely to be obese. /10

Looking at the medical indicators, non-white patients are;

• less likely to need assistance with daily living
• less likely to have severe comorbidities
• needing quicker admission to ICU
• much more likely to need mechanical ventilation within the first 24 hours. /11

Some worrying indicators when considered together.
On paper the non-white patients look like they should have better prospects.

They are younger, less-obese and with fewer health conditions.
They are however more likely to be deprived.
Are they seeking medical help later? /12
....
Table 10 shows gives a good summary of outcomes data by individual variables.
Survival is worse for:

• older patients
• males
• Asian, black and mixed ethnicity
• more deprived groups
• those who are NOT obese
• people with comorbidities. /14

Non-white survivors average 15 days in ICU, compared to 10 for whites
They are much more likely to need renal support in critical care:

These are really interesting stats! Thank you!!!!

I need to further analyse the stats for obese, as compared to merely overweight

And at least the researchers are looking for physical causes rather than a lazy 'must be racism, end of' rhetoric we hear elsewhere.

ICNARC report 19 Jun on COVID-19 critical care:

https://www.icnarc.org/DataServices/Attachments/Download/da19fd54-70b2-ea11-9127-00505601089b

So roughly speaking I'd expect an IFR of India of around 0.2%.

That’s still 2-3 million people!

I am ahem obese. This is partly because I have respiratory problems and potentially the outcome is worse if I lose weight (for bronchiectasis, the death rate is 10% PER YEAR for someone with a BMI of 18 and falls with higher BMI). My respiratory consultant suggested that a higher weight was no bad thing (pre covid)

With COVID, probably depends on whether the obesity is accompanied by T2, as we know that is a high risk factor
However, many obesse people - especially whites - do not have T2

That’s still 2-3 million people! sad

Only if 100% of the population is infected

Non-white ICU patients are:

• significantly younger; 5 years on average!

This is implied by the fact that the non-white-British/Irish population is MUCH younger on average, so it wouldn't make sense at all if (can't be bothered to look up the figures now) 60% of the white population is under 65% and 80% of the non-white population, that the average ICU age of both populations would be the same.

Clearly the average age at death is younger in India also....

What we expect is a lower IFR for ethnic minority populations, but we know %I is higher.

Lack of comorbidities is also implied from the population necessarily being younger.

Not sure why BMI is lower however.

Survival rates would need to be compared by hospital, as I imagine this varies regionally (it does with other illnesses), and age adjusted out as a factor (though the lower age should help survival).

The study, by academics at the University of Edinburgh, found that
patients with Indian, Bangladeshi, Pakistani and other South Asian backgrounds were 20 per cent more likely to die from coronavirus than white, black or other ethnic patients, and were on average 12 years younger.

More imprecise reporting.

The hazard ratio was 19%, but not 'and were'.

Rather, that sentence should end with

"same age, sex and location" and not mention the age

if the South Asian deaths were not younger, then that would be evidence that South Asian people were much less likely to die from covid-19. Rather, the 12 years younger is an inevitable function of the South Asian population being younger.

To be clear, they identify the hazard ratio associated with being South Asian when adjusting out diabetes as 14%. In other words, they found a 19% extra death risk, but when they considered the greater prevalence of diabetes, which is already known, this dropped to 14%.

It is not clear where the remaining 14% comes from from that study -
could be other comorbidities which differ by ethnicity, or some inherent ethnic prevalence, but not seeing any evidence for either there.

"patients .... and were on average 12 years younger."

I though it significant that patients are 12 years younger - that's patients sick enough to be in hospital
and 5 years younger for ICU

If that included black patients too, I'd have just thought it was more exposure of working age population due to job types
However, it looks specific to S Asian

Younger population, all things being equal (which they aren't) should just mean fewer people of that ethnicity in hospital,
not people coming in at a much younger age

e.g. one reason London IFR could be expected to be lower than rEngland

Similarly, fewer deaths, not deaths at a younger age

Is there a significant difference in job types typically held by S Asians vs black minorities ?

My iPad / wifi is refusing to download reports atm, so I can't check if more detailed ethnic stats are in ICNARC , or only in the Edinburgh paper,

but are S Asians dying several years younger than black minorities, not just whites ?

oops, London CFR - % infected of course much higher

I though it significant that patients are 12 years younger - that's patients sick enough to be in hospital
and 5 years younger for ICU

If that included black patients too, I'd have just thought it was more exposure of working age population due to job types
However, it looks specific to S Asian

As I stated, the age at hospitalisation being younger is a necessary function of the population being younger. This is true because if you consider hospitalisation as a risk doubling every six years of age, then if you have 100k people in each age bracket up to 90 then you would have for example:

32 people aged 84-90
16 people aged 78-84
8 people aged 72-78
4 people aged 66-72
2 people aged 60-66

Resulting in an average age of around 78.

Whereas if you consider an ethnic minority or developing country population pyramid then the % of people in each older age group will fall almost as fast as the risk of hospitalisation is rising . So it's absolutely impossible for the hospitalisation age not to be younger. Western countries have a fairly flat population pyramid, developing countries had a small fraction of their current population 70 years ago. The old people weren't alive yet to be old now, relative to the huge numbers of births.

And as I also stated, the journalist needs to be more precise with language.

The south Asian population was 12 years younger than the white population. It was 1.6 years younger than the black population.

As of 2011 the Asian population was 1 year younger than the black population . I imagine that will have increased with more Asian than black immigration.

So there is nothing untoward here .

That is the South Asian hospitalised population was 12 years younger

Only if 100% of the population is infected

Ok, but if allowed to propagate to herd immunity, that’s still 1-2 million.

Herd immunity doesn't seem to be a thing anywhere, and current deaths are far below the millions in India. We'd expect a vaccine before that.

I imagine India has bigger health issues than covid-19

Sorry, Shoots but I still don't get it wrt ages:

Hospitalisation is because of how ill someone is,
not with the aim of hospitalising similar numbers of people among ethnicities

e.g. if the UK govt last December had invited a migrant population of 1 million aged 21-25
then regardless of their ethnicity I would expect v v few of them to end up in hospital for COVID

• roughly the same % as of white people aged 21-25 doing similar jobs and in similar housing etc

People of S Asian origin having to be hospitalised at a younger age than people of other ethnicities indicates something particular to S Asians
just as having a younger average age of death from COVID does

Ah sorry, I see, they are looking at the entire population and the age group 80+ is v small for S Asians

I was considering the population of each age group, but that's not how they do their stats