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.

Good point oralengineer

It's all those known unknowns and unknown unknowns.

Was not paying much attention to the news but there was a scientist from SAGE asked if they would all be happy to see their advice made public and saying they all believed in being open. That is probably yes minister speak for Boris has been lying again.

Germany has about 3 times the hospital beds we have in Britain and I think spends something like 50% more on health care, much easier to be flexible when you run with spare capacity. Instead of bickering over international comparisons we should be celebrating what worked and looking at how we all improve. I am proud of the community efforts that donated lots of money, fabric and thread to make scrubs for the NHS and of the volunteers who made visors, scrub bags, headbands to save nurses ears. There have also been some quiet innovations like using snorkel masks to deliver oxygen - the valve for that was made here, the snorkel masks themselves are French. Porton Down has been managing faster ways to sterilise ambulances. The phone aps being developed could make a considerable difference to track and trace. There will be many small changes that we'll probably never hear about.

Forgot that I meant to say international travel direct to the uk is often largely to a few major cities, especially in winter. However there have been cases across the UK. Tracking stopped after the half term ski trips because there were too many widespread outbreaks.

NHS England: Type 1 and Type 2 diabetes and COVID-19 related mortality in England: a whole population study

https://www.england.nhs.uk/2020/05/nhs-expands-offer-of-help-to-people-with-diabetes-during-coronavirus-outbreak/

Summary:

people living with type 1 diabetes are at three and a half times the risk, and
people living with type 2 are at double the risk
of dying in hospital with the virus, compared to people without diabetes.

However, by far the strongest risk factor for dying with the virus is age,
and people with type 1 diabetes are on average younger than people with type 2 diabetes.

Overall, 7,466 of those who died in hospitals in England had type 2
and 365 who died had type 1 diabetes

and the research suggests that the threat for those under 40 with type 1 or type 2 diabetes is very low, with no recorded deaths in those under 20.

The study also shows that in people with both type 1 and type 2 diabetes,
even when all other known factors are taken into account,
higher blood glucose levels and obesity are linked to higher risk.
.....
⏺ The overall death rate for people with diabetes doubled during the early stage of the pandemic.

⏺ In both type 1 and type 2 diabetes, men, people of black or Asian ethnicity, and people living in more deprived communities, were at higher risk.

⏺ In both type 1 and type 2 diabetes, those with pre-existing kidney disease, heart failure and previous stroke, were also at higher risk.

Professor Jonathan Valabhji, national clinical director for diabetes and obesity and lead author of the study said:

“This research shows the extent of the risk of coronavirus for people with diabetes
and the different risks for those with type 1 and type 2 diabetes.

“Importantly, it also shows that higher blood glucose levels and obesity further increase the risk in both types of diabetes.

I think I would describe Pakistani/Bangladeshi as ethnically very different from Indian

Erm I meant racially perhaps? Obviously Indian richer, different religion etc. But it's basically one country split down religious lines.

There seems to be a bit of an implication that there is something inherent to having black or brown skin and risk. That might be the case, but clearly a big part of this is the cultural aspects, which might be found in individuals of any race, such as working as a bus driver, or eating fried food , or whatever

Unfortunately many of the younger black nurses/healthcare workers who died were noticeably overweight

How many are we talking about? There's weak evidence for non-morbid obesity contributing to covid-19 death risk. It would not at all explain vastly higher death risk. We are talking perhaps 25% extra risk from obesity, which doesn't add up to much.

For people who are extremely obese such that they waddle, if you will, then there is going to be a bigger risk effect.

Taxi drivers are often older

This is adjusted for in the models. 50 year old taxi drivers are compared to the average 50 year old. They have a much higher death risk

perhaps Asian cab drivers similarly meet in close contact, or perhaps they meet in the local mosque during the day.

Forgive me if we have different understandings of this illness, but as I see it if 1 in 20 of the population are infected then if you take an infected person in your cab then you are basically guaranteed to catch covid-19 after a given period, since you are in a confined space.

I'm sure mosques and taxi huts and such like have been closed for weeks and weeks.

I.e. environmental factors, such as being in a covid-19 infested care home, or carrying covid-19 infected passengers, or working in a factory with covid-19 infected workmates, is likely to be far more important than any obesity, diabetes, etc.

Yep, much higher risk of catching it
if someone is in close proximity to other people in an enclosed space for an extended period of time,
especially if singing, shouting, heavy exertion etc

Institutions of all kinds where people live 24 / 7 are obvious higher risk areas:
care homes, prisons, homeless shelters etc
and care homes combines this with the highest risk age category

Much higher risk of dying
if you're old - dominates all other factors
but evidently T1 and T2 multiply the risk by x 3.5 and x 2 respectively

The increased risk wrt other (single) comorbities is much smaller in tables I've seen, more like 5-20%
Or have some of these been updated to be much higher too ?

Risk
Age
Sex
Job
Health issue
Public transport
Rural/city
Shielding or vulnerable
Other people in house (more -more risk)
Children

Diabetics have been vulnerable all along. So suggests (care home/hcp/not allowed to wfh).

That Birmingham study finding over 2% of those hcp infected.
How have 75% not caught it in that environment

Horizon broke it down as Black, Indian, Pakistani/Bangladesh and so on, and yes Black and Pakistani were higher than Indian. Risk of death. And they also said when all other factors were accounted for. So I’m afraid different conclusions to you, which is why I wondered where they got info from.

I don't know as you've not provided a link.

However the ONS are the most reputable source of information on this. Other sources are likely to be less so.

That said, it's unlikely that they said 'Pakistani people are at higher risk', since what we know is that 'Pakistani people are more likely to have died', except that's not exactly true either, because Pakistani people are much less likely to be old, which is far more important, so it's 'Pakistani people are more likely to have died compared to white people of the same age'.

That doesn't mean Pakistani people are more likely to die, because for example the ONS study cuts off in early April, when everyone dying would have been infected in March.

If we are talking about risk today, then it's not going to be the same, because for some time most of the deaths were in hospital, and now they are in mostly care homes, which are 96% white.

So we can't even say that disproportionately many Pakistani people will die of covid-19, only that that was true in March & early April.

And they very obviously haven't adjusted for all factors, since if they adjusted for all factors, then the risk would be the same. So it would be that they have not identified which risk factors are more important.

It would make sense, for example, to look at minicab drivers, and see of the 70-odd deaths, how many have been white and how many South Asian, and the age profiles of the underlying populations.

That would not be that easy, as for example a minicab driver in/around London in March is likely to have been infected, and my impression is most are non-white, whereas a minicab driver in say, Hull is much less likely to be non-white, and the risk would have been many times lower to be infected in Hull in March, so it really is NOT an easy task to work this out.

We know that certain non-white populations are more likely to have died. But we don't know exactly why. But once we have identified all the risk factors, then the risk would be the same. For example, it might be that simply having black skin doubles your risk of dying once infected. But we don't know that.

We do know for absolute certain that South Asian & black populations are overwhelmingly concentrated in places that were hard hit by covid-19 in March, when people were acting like nothing was happening, and would have disproportionately been doing jobs involving contact with others.

@itsgettingweird assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/878121/coronavirus-covid-19-testing-strategy.pdf

There's weak evidence for non-morbid obesity contributing to covid-19 death risk.

Some photographs of those who have sadly died here;

www.telegraph.co.uk/news/0/nhs-workers-died-coronavirus-frontline-victims/

Unfortunately I would describe some of them as morbidly obese - bearing in mind that families may have chosen more flattering photographs taken a few years ago as well.

I'm sure mosques and taxi huts and such like have been closed for weeks and weeks.

I did post that religious worship must have played a part in the early spread of the disease - i.e. through the local community - the mosque I linked to is one of the largest in Europe - it was still open a few days after the Cheltenham Festival had finished; "On a normal day, some 10,000 worshipers from all over East London attend the Whitechapel site across five daily prayers." We don't know how many cab drivers caught the infection from a member of their own family, a friend, a passenger... but yes, you are right about enclosed spaces.

I think this may be asking the impossible but... if, for example, there was an infection rate of 10% amongst the population and 2% death rate (totally made up numbers), aside from not accounting for a skew towards hospitals and care homes, it also doesn't account for how many people have been shielding or working from home and not gone out even to shops. If that was 50%, then you could argue a natural infection rate would be 20% and deaths 4% (I know the reality is more complex, but trying to isolate to one variable).

Has anyone seen data like this? I.e. we're seeing infection and deaths rates of x, but actually they aren't of the 'whole' population? Sorry I'm not being very clear but hopefully you can tell what I mean! I'm poking it how bad will it get when schools go back and if I should DS back to nursery!

BBC journalists failing on accuracy and basic understanding of statistics again.

www.bbc.co.uk/news/education-52662338

"Newham in east London has the highest proportion of deaths from coronavirus in England and Wales. BBC News looks at why, and what it means for the community."

"Covid-19 has preyed on Newham like nowhere else.

Data released by the Office for National Statistics shows the east London borough has suffered the highest proportion of deaths from the disease in England and Wales."

That's NOT what the ONS say.

What they actually said

www.ons.gov.uk/peoplepopulationandcommunity/birthsdeathsandmarriages/deaths/bulletins/deathsinvolvingcovid19bylocalareasanddeprivation/deathsoccurringbetween1marchand17april

The local authorities with the highest age-standardised mortality rates for deaths involving COVID-19 were all London Boroughs; Newham had the highest age-standardised rate with 144.3 deaths per 100,000 population followed by Brent with a rate of 141.5 deaths per 100,000 population and Hackney with a rate of 127.4 deaths per 100,000 population.

So first thing would be to note that it's clearly fucking bollocks to say that it's preyed on Newham like nowhere else, because 144.3 and 141.5 are essentially the same value, so it would actually be

'It's preyed on Newham like much of London'

It's even in the header of the report

"The 11 local authorities with the highest mortality rates were all London boroughs, with Newham, Brent and Hackney suffering the highest rates of COVID-19 related deaths."

Next point would be that it's not 'the highest proportion'. It doesn't even begin to claim that. It says 'highest age-adjusted death rate'. Which is to say that far more people might have died as a proportion of the population in other areas, but because they are older they won't be counted the same.

It's barely literate, but if they mean 'covid-19 deaths as a % of total deaths', then that was then Brent, where it was 304 out of 550, which is obviously more than the 208 out of 421 in Newham.

And thirdly, the release they are using is a fucking failure at this point because it was released by the ONS on 1 May, which is absolutely ages ago,

for "deaths occurring between 1 March 2020 and 17 April 2020. Figures only include deaths that were registered by 18 April 2020."

whereas the ONS published on Tuesday this week figures for

"Deaths (numbers) by local authority and cause of death, for deaths that occurred up to 8th May but were registered up to 16th May, England and Wales"

so we have a full month's more data, and it would take about 10 minutes to update the figure before publishing the report.

Anyway, Newham is no longer top by any measure, that is now Brent. And there have been 50% more total deaths in Brent, despite a slightly smaller population, it's only Newham's population being significantly younger that makes it get even close to Brent.

Currently there have been at least 41 covid-19 deaths at home, 362 in hospital, 37 in care homes in Brent, and 30, 206, and 41 in Newham.

And that's not the full story. There have been, for example, 331 deaths at home/hospital in Croydon, versus 236 in Newham. The 87 care home deaths in Croydon might be less of a surprise, but it's not obvious exactly that Newham is harder hit simply because it has a younger population while having 1/3 fewer deaths.

Nowhere in the BBC's report does the word 'age-standardised' appear. This is deceitful, in that you'd have to be a bit of a psychopath to imagine that the stacked bodies of mostly old people in northern Italy were not traumatic for the children, grandchildren, etc. of the deceased.

The BBC's chosen metric (though they quote a different measure) counts each death in Tower Hamlets 3 times, Newham 2.45x, while Hertsmere only counts 1x, because it has the same age profile as the country as a whole.

Thus the worst hit in terms of covid-19 deaths, which seems to me like the most reasonable measure for 'worst hit', is in fact Hertsmere, with 164 deaths from just 105,000 people. And these aren't just grannies in nursing homes - 38 out of those 164 were in care homes, and 126 were not.

In fact in terms of 'covid-19 deaths as a % of the population', Newham is only 75th in the list

Is it any wonder there are posters claiming we have 50,000 deaths as a result of a (supposedly) high ethnic minority population when we have bollocks stories like this?

In fact, of the places with >100 actual covid-19 deaths per 100,000

very few of them fit the BBC profile of BAME people being struck down in large numbers.

These are:

1. Hertsmere
2. Harrow
3. Brent
4. Epping Forest
5. South Lakeland
6. Middlesbrough
7. Croydon
8. Ealing
9. East Staffordshire

10. Walsall 11. Salford 12. Enfield 13. Barnet 14. Stratford-on-Avon 15. Bromsgrove 16. Solihull 17. Reigate & Banstead 18. Watford 19. Sunderland 20. Barrow-in-Furness 21. Wirral 22. Tewkesbury

Well, given they've only just worked out today there's a mathematical issue with the school plans (how do you fit double the amount of classes into a school without a rota and more rooms and staff magically appearing in a puff of smoke - I think the report is a rehash of a very good new statesman article), I don't hold up much hope for the bbc analysis.

I'd worked out some of the areas that had had the higher numbers of deaths per 100,000 simply from laboriously typing it into their own postcode checker. And had spotted that figures were higher in for eg Middlesbrough than places in London. (And I'm shit at maths.)

Convinced hertsmere is worst because Harry Potter studio tour is up the road

Many employed there live in hertsmere and Harrow

7000 visitors travelling through a day has to have an impact.

Nearly all the foreign visitors travel through Watford station and use the buses to get to the tour They also stay in local hotels.

That's interesting.

Middlesbrough have both v high numbers of key risk factors and was the most non compliant area at the start of lockdown.

Neuro. I worked there and I never felt more vulnerable in feb/early March. So many foreign visitors coughing all over the place and we have a lot of enclosed spaces. Tightly packed Waiting areas with 150 people in etc.

We also have a lot of shared touch and explore exhibits. It's a no brainer.

Ah, it's Voldemort's fault then!

Doubt it makes that much difference. It's not in that borough and visitors will stay all over London.

It's interesting that Hertsmere has the second highest % of Jews in the country.

The Board of Deputies of British Jews has been reporting on disproportionate deaths among Jews for weeks, but it doesn't seem to be a particular interest more widely.

The total appears to be close to 200 deaths per 100,000, which is approximately FIVE TIMES higher than the rate among Bangladeshi and Pakistanis.

Of course the Jewish population is much older.

But even so it does seem that there has been a disproportionate focus on reporting the deaths of non-white people, who experience a death rate below average because of their age, as compared to the Jewish community which has a higher death rate than any local authority in the UK.

I'm sure it's a contributing factor. Leavesdon is literally a stones throw away and all the visitors will be staying over using the petrol stations, fast food places, buses and trains to get back into London-through hertsmere

Strawberry 😂

I think in the end there will be many, many factors for key hot spots.

I found it interesting that York, which had the first confirmed cases, was then very slow thereafter; there will be other factors too (the actual area the first case was in was university but also an affluent area otherwise, older generation who would have been extremely careful immediately) but undoubtedly simply knowing there was a case locally would have made everyone panic and be more careful.

I can't remember the specifics but Cumbria's outbreak has been linked to a cluster of cases in a factory / docks? Or similar early on, where measures were slow to be taken.

Harry Potter world is prime half term hol visit material.

I had a chest infection but had the kids over half term and dragged them around various indoor attractions locally as staying at home with them was more daunting painful. I remember feeling quite faint in one museum. Feel extremely bad now! Definitely wasn't COVID though. Youngest was sporting a weird rash on that day too that turned out to be scarlet fever a few days later... we were a walking germ factory.

I don't think the Harry Potter hypothesis can be valid.

Harry Potter is in Three Rivers District, and close to Watford District.

It's in Three Rivers 002, which had 3 covid-19 deaths out of 7, to early/mid April, and borders TR 001 (1 out of 8), TR 003 (0 out of 15), Watford 001 (7 out of 25), Watford 002 (3 out of 5), Watford 003 (1 out of 9), Watford 004 (3 out of 12), Waford 007 (22 out of 45), Watford 008 (4 of 13).

Arguably the entire Watford district, where by early/mid April it was 57 deaths out of 159.

Meanwhile Hertsmere is

001, 002, 003 - Potters Bar 22/66
004 - South Mimms & Shenley 3/8
005 - Radlett 8/19
006, 007, 009 , 011 - Borehamwood 28/66
010 - Elstree, Letchmore Heath, Aldenham 10/18
008, 012, 013 - Bushey Heath - Bushey 29/84

So there doesn't seem to be a particularly obvious Harry Potter effect. Also we'd expect Watford deaths in particular (as it's Watford that is the relevant district, not Three Rivers or Hertsmere) to have peaked in March/April. In fact, Watford has gone from 59 deaths per 100k to 104, Three Rivers from 36 to 79, while Hertsmere has gone from 95 to 156.

So Hertsmere, which should be far behind Watford in suffering deaths in March/early April peaked early, while Watford had suffered much less as of that time.

@ShootsFruitAndLeaves BBC Horizon episode 2. It’s on bbc I player.