Daily numbers, graphs, analysis thread 11

[BigChocFrenzy]

Welcome to thread 11 of the daily updates

Resource links:

Slides & data UK govt pressers
NHS England stats including breakdown by Hospital Trust
ONS UK statistics for CV related deaths outside hospitals, released weekly each Tuesday
Financial Times Daily updates and graphs
HSJ Coronavirus updates
Worldometer UK page
Covidly.com to filter graphs using selected data filters ONS statistics for CV related deaths outside hospitals, released weekly each Tuesday
Plot COVID Graphs Our World in Data

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

Schools things pisses me off as a send parent.

They don't even stick to law re timescales and naming schools. Send children go months without an education.

Ds college has said it'll do blended learning all academic year - whatever. His EHCP says he needs a laptop because he had a scribe in school. But college day he can only have it when in college - but they won't allow him to attend full time.
Someone needs to get a grip of the situation.

For hospital admissions, I only look at the numbers for England because both Wales and NI seem to count suspected cases (i.e. all patients displaying potential C19 symptoms) and therefore report inflated numbers.

England reports both actual hospital admissions and transmissions happening in hospitals:

England data captures people admitted to hospital who already had a confirmed COVID-19 status at point of admission, and adds those who tested positive in the previous 24 hours whilst in hospital. Inpatients diagnosed with COVID-19 after admission are assumed to have been admitted on the day prior to their diagnosis.

[quote Derbygerbil]@PatriciaHolm

I’m feeling rather disconcerted that there were 391 hospital admissions and 921 reported cases on 23 June... implying 42% of those who tested positive are hospitalised... I realise some of those who are hospitalised would potentially have tested positive before, but nonetheless, this is a very high figure, and doesn’t seem right... especially as I thought the estimated new infections per day was around the 3-4,000 mark. Am I missing something?[/quote]
hmmmmm I hadn't looked at it that way before.

I think there are a few things going on; firstly, the data doesn't match as nicely was we would like it to, though trends are all down so it's not so important that it doesn't now.

I think it comes down to there being 2 sets of testing. Pillar 1, for those in clinical needs and healthcare workers etc (so those in hospital) and Pillar 2, community tests.

Looking at the Pillar One tests for, say, 23 June, we get 210 positive tests in England and 275 hospital admissions. Not a problem given we know tests are just those that come back that day, not those given, but indicative. Pretty much all the hospital admissions are coming from Pillar One.

Pillar 2 is community testing. Which makes up a much bigger percentage of tests and positives now. But maybe not hospitalisations?

Say...for example...we are testing 50,000 people a day, Say 1,000 are positive, but 4,000 in the community per day as a total are. 400 are hospitalised - 10% of infections, but 40% of positives. That suggests that a large amount of the infected are not actually getting tested, probably because they are asymptomatic?

So infections as a whole are low, with a large number of those being asymptomatic cases; those who bother to get tested are more likely to be ill, and require hospitalisation (which is a lower threshold than it used to be).

Does that make sense.....

@PatriciaHolm

Thanks. Back in late April, a study indicated that 1/3 of hospitalised patients died...

www.google.co.uk/amp/s/www.bloomberg.com/amp/news/articles/2020-04-30/a-third-of-u-k-covid-19-hospital-patients-die-study-suggests

Apply that to the 391 and c.130 would die. If the IFR is 0.5%, that implies there are 26,000 infections every day - or c.400,000 over a fortnight.... If the IFR is lower, it’s higher still! This is far higher number than latest ONS estimates which estimates that “51,000 people (95% confidence interval: 21,000 to 105,000) within the community in England had COVID-19 between 8 June and 21 June 2020.”

Even taking account of the likelihood that hospital deaths are better than 1/3 now... unless it’s different by a order of magnitude, there’s still a significant disparity.

I must be missing something here... I’ll have a think.

It would be interesting to compare with other countries’ hospitalisations.

was it really as high at 1 in 3 dying?

"The risk of dying from coronavirus after being admitted to hospital has fallen dramatically since the beginning of April, analysis by Oxford University has shown.

Around the peak of Britain’s epidemic, six per cent of people hospitalised with Covid-19 died from the virus but figures show that by June 15, just 1.5 per cent died."

www.telegraph.co.uk/news/2020/06/25/oxford-university-analysis-coronavirus-death-rate-hospitals/

"The proportion of people dying with Covid-19 in hospitals had been declining considerably since the end of March.

Of the 10, 387 people in hospital in England with Covid-19 on 2 April, 644 died, whereas on 15 June it was 50 out of 3,270, which shows that the death rate in April was four-fold higher than in June."

So it is as I thought that the "admissions" includes those already in hospital. They really shouldn't refer to these as "admissions" should they? It would be useful to know how many people are actually admitted as a result of catching it in the community. As opposed to going into hospital for some other reason and catching it in there.

I don't think you can figure this out from the pillar 1 stats because you don't know how many of these are patients and how many are staff. These stats are crap!

Another positive day in terms of overall cases.

In terms of the P1+P2 combined cases map as kindly posted a little up thread by patricia it is a shame the local authority data is not published along with it. However using the map every UTLA can be identified - hopefully accurately - and although the cumulative P1+P2 cases per 100,000 is shown as a range given the daily dashboard publishes the p1 actual cumulative total and rate per 100,000 the range of P2 cases in most areas can calculated.

This highlights some interesting areas.

Lincolnshire has one of the lower P1 cases per 100,000 at 152, and just under 1150 actual cases, being in the lowest 20 of the 150 UTLAs. But in terms of P1+P2 it is in the 313 to 354 cases per 100,000 suggesting, given its population that there have been somewhere between 1200 and 1500 P2 cases, more than double the P1 cases.

Peterborough P1 cases per 100,000 233, but combined rate in the 520 to 602 band.

Derbyshire and Northants P1 228 and 226 per 100,000 respectively and both in the 404 to 442 band.

Leicester P1 P1 296 cases per 100,000 and, not surprisingly in the over 600 combined cases per 100,000

Conversely a number of London Boroughs show as having really quite limited p2 cases

Does anyone know if the new dashboard is likely to have the P2 cases at council level?

I read in a paper (sorry, but I can't remember where) that people who catch Covid-19 in hospital are counted as an 'admission' on the following day. For around 20% it is a hospital acquired infection,

Those would be Pillar one.

@torydeathdrug

Yes, 1/3 did seem very high!... 1.5% similarly feels very low.

I’m not sure I understand “whereas on 15 June it was 50 (dead) out of 3,270“.... That suggests a daily death rate, as presumably, unfortunately some of the remaining 3,220 will die? Also, the April figure is consistent for daily deaths at this time.

yeah I agree - it's unclear. It says "Prof Carl Heneghan, Professor of Evidence Based Medicine, who carried out the analysis with Dr Jason Oke, of Oxford University, and Dr Jason Mahon of the University of York" but I can't find the actual analysis! More from that article

"The proportion of people dying with Covid-19 in hospitals had been declining considerably since the end of March.

Of the 10, 387 people in hospital in England with Covid-19 on 2 April, 644 died, whereas on 15 June it was 50 out of 3,270, which shows that the death rate in April was four-fold higher than in June.

“You’d expect 190 deaths if there was the same relationship,” added Prof Heneghan. “It is important to understand the reasons.”

The total number of people dying with Covid-19 in hospitals in England each week has fallen from a peak of 899 on the 8th April to 50 on the 15 June.

The number of people in hospital with coronavirus has also fallen from a peak of 15,702 on the 10th of April to just 2,891 on the 19th of June."

That's a good report.

I think there's so much to remember as well.

Some of the deaths we are recording now will be for people who've spent months on a ventilator since the peak.

think its from here

www.cebm.net/covid-19/declining-death-rate-from-covid-19-in-hospitals-in-england/

Yes, these are daily rates and the authors of the research hypothesised that some of the patients may be staying in hospitals for longer. However they did not think it could be the only driver.

Here is the initial analysis:

www.cebm.net/covid-19/declining-death-rate-from-covid-19-in-hospitals-in-england/

And here are comments from a doctor who confirms that that Covid patients in hospitals seem to be doing better than they were 2-3 months ago:

There are no Covid-19 patients in intensive care at England’s largest hospital trust for the first time since the pandemic began, its chief executive has said.

Dr David Rosser, who heads the University Hospitals Birmingham NHS Foundation Trust (UHB), also said there were signs coronavirus-infected patients “don’t seem as sick, on average, as they were”.

“I don’t think we understand or have the insights into what that is about, but it’s interesting how current patients don’t seem as sick, on average, as they were.

“That may be just a coincidence – it’s all very small numbers.”

www.expressandstar.com/news/uk-news/2020/06/19/no-itu-covid-19-patients-in-englands-biggest-hospital-trust-says-health-chief/

Can anyone point me in the direction of how to see the figures for Leicester? Is there public information about a clear spike in cases?

Looking at the spreadsheet of admissions and deaths, some 25% of admissions have died, but of course that ratio will be getting better over time as the article above suggested.
We are averaging 55 deaths in day in hospital in June, but of course that relates to infections 3/4 weeks ago, which at a IFR suggests 10,000 infections a day in May which doesn't seem unreasonable.

this is similar - declining ICU admissions, increasing ICU survival

www.cebm.net/covid-19/covid-19-declining-admissions-to-intensive-care-units/

@MarcelineMissouri not sure but it may be given somewhere on Leicester lockdown thread which has been running for a few days?

[quote StrawberryJam200]@MarcelineMissouri not sure but it may be given somewhere on Leicester lockdown thread which has been running for a few days?[/quote]
This was from the thread earlier - from the Leicester Mercury, though I can't find it online.

torydeathdrug The 33% death rate came from this earlier review of what was then all UK hospitalised COVID cases:

Features of 16,749 hospitalised UK patients with COVID-19 using the ISARIC WHO Clinical Characterisation Protocol

https://www.medrxiv.org/content/10.1101/2020.04.23.20076042v1

The median age was 72 years
....
47% had no documented reported comorbidity.

Overall, 49% of patients were discharged alive,
33% have died** and
17% continued to receive care at date of reporting.

"Covid patients in hospitals seem to be doing better than they were 2-3 months ago"

wintertravel Several countries are seeing this - and it seems to be because
the age of patients is significantly lower,
as more young and middle-aged people are going out

e.g. charts for Germany:

Thank you - and actually on the BBC just now they reported that 10% of our cases in the last week were in Leicester

the Heneghan analysis considers the age of patients

"The reasons for this steep and continual decline in the deaths per day in the hospital of patients with COVID-19 are unknown and should be explored. Potential reasons could include:

• Patients overtime being admitted are becoming younger with fewer comorbidities, although there is no evidence of this in the daily hospital death data which, if anything, suggests a greater proportion of deaths in hospital are over the age of 60 than at the peak of deaths in early April."

The most plausible explanation for the decline in hospital death rates has to be that they are admitting people who are not as sick as before. All the other possible explanations that that article offers don't fit with the data - by their own admission. And it fits with what we read generally in the press - at the height of the epidemic hospitals were overwhelmed and there were numerous reports of very sick people being told to stay at home, or dying at home after paramedics had been called and gone away telling them to stay home. I've not heard those kind of reports for many weeks now, so seems very likely that they're admitting people who are not so very sick.

The 1.5% death rate that they quote isn't 1.5% of patients dying btw - It's 1.5% of all hospital patients dying per day So if the average stay was two weeks, that would be 21% dying overall. I think the latest data indicate it is about that. The per day death rate would also fall if they have more slow-recovering patients lingering around for many weeks - which is also what I hear is the case (and obviously wouldn't have been the case in the early days)