Daily numbers, graphs, analysis thread 6

[Barracker]

Welcome to thread 6 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.

There is no data for a very important part of the picture (infection rates).

Or if there is, I'd like to read it.

Is the ONS calculation the additional risk of dying within 1 year ?

and does it include predictions about how the COVID death rate will continue over the year ?

Of course, if the infected % of the population is 4%, as Imperial calculate atm,
then any individual's risk of dying is very low

it would be difficult to calculate the alternative if the UK hadn't gone into lockdown
and it is difficult to calculate the deaths if all measures were removed prematurely, instead of in the proposed gradual stages

Most countries seem to have a strategy of stretching out the period over which the population become infected
as ntbo the deaths if 70 or 80% become infected would be non-trivial at all ages
and disastrous if this happened within a short period instead over many years

Chic The Imperial estimates are only for entire countries, not regions of the UK, if you are asking about those:

mrc-ide.github.io/covid19estimates/#/total-infected

Otherwise, there have just been a few studies within some countries of a particular region, e
.g. NYC, Gangelt

Interesting analysis of the NYC seroprevalence data from a US scientist:

Thoughts on seroprevalence in NYC. I'm not at all surprised by an estimate of 21% seropositive in NYC as discussed by @NYGovCuomo today.

I had previously been estimating a case-to-infection reporting ratio of 10-20X. I can't figure out when this 21% seropositive estimate refers to, but we can do some extremely rough calculations assuming 21% today. As of today, NYC has had 145k confirmed cases reported. Assuming seroprevalence of 21% gives 1.7M infections in a city population of 8.4M. Dividing 1.7M by 145k gives a reporting ratio of ~12X.

This reporting ratio of 12X seems entirely within the realm of expectation. If we then take deaths as of today as 17,200 based on excess deaths we'd get an infection-to-fatality ratio of ~1%.

This estimate deserves better statistics as there are active infections among the 1.7M that will resolve to deaths in the coming weeks (increasing the numerator) and the true seroprevalence may be greater today than when the study was conducted (increasing the denominator). Keep in mind that this infection-to-fatality ratio is heavily dependent on population demographics as well as health system capacity. IFR of ~1% (or more) in NYC may differ from IFR in other locations.

A completely non-statistical throw away comment for which I apologise. I think calculating "lost years" as a purely numerical loss, is missing the point which we should all consider. Are those intangible, potential years, healthy and fulfilling, or are they now just numbers to be attained at any cost? Sorry, obvious agenda and all that.

I know for sure that "Healthy Longevity" is now a key goal of government, rather than pure numerical extra years of life.

Baaah There is the same sort of moral / financial calculation when treating, operating on and keeping alive a severely disabled newborn,
who will never have any independence or much quality of life, but could live for years, maybe decades

The initial cost I read sometimes exceeds ½ million quid
then there would be large annual costs to the state for the rest of their life - and also massive effects on the lives of parents, siblings etc

If they have a chance of surviving, whatever godawful quality of life, then this is normally done,
because society makes this choice

These suggestions of there being a justifiable cost in terms of people close to the end of their lives seem to assume that the rest of us get the virus and mostly recover with a small percentage of unfortunate deaths.

Does anyone on the thread have any statistics about how many people make a full recovery and how many are left with long-term adverse health consequences?

I would particularly be interested in the % of those hospitalised in the different age ranges
and those assessed to have longterm (> 3 months) effects

If COVID progresses to the lungs, then reports are that at any age this requires a long recovery, especially to regain previous level of health & fitness,
with reports too of some longterm conditions wrt lungs, chronic fatigue, possibly CVD

We originally read 20% from the China data
German data gives 16% of confirmed cases

None of these have age divisions or longterm prognoses

UK & German data both have a mean / median age for ICU patients of about 60, so about half would be aged below 60
However, even requiring oxygen in a non-ICU ward likely has longer term consequences

Assessing the economic consequences needs to estimate this too, especially for the population under 65

Ah, the estimate about remaining years may have come from Prof David Spiegelhalter, Uni Cambridge ?

www.bbc.com/news/health-51979654

. Nearly 10% of people aged over 80 will die in the next year

. The risk of them dying if infected with coronavirus is almost exactly the same.

. He says there would be "a substantial overlap" in deaths, not just extra ones

. Ferguson at Imperial has suggested it could be up to ⅔

. Risk could increase each month, depending on the course of the epidemic and the capacity of the health service

@Eyewhisker, the mean covid-19 death (based on ONS stats) is aged 77.7 for males, 79.8 for females. For UK all causes the numbers are respectively 79.3 and 82.9.

If we consider for example breast cancer then the average age at death appears to be 68, and for prostate cancer 80.

Annual flu reports are here

www.gov.uk/government/statistics/annual-flu-reports

If we look at the most recent bad flu season, 2014-15, then for weeks 1-4 2015, we can see that 57,323 deaths were reported.

www.ons.gov.uk/peoplepopulationandcommunity/birthsdeathsandmarriages/deaths/datasets/weeklyprovisionalfiguresondeathsregisteredinenglandandwales

Of these 41,937 were aged 75+, that is 73.16%

In 2020 the week 1-4 number was 51,158, of which 36,158 were 75+, which is 70.68%.

If we look at say July 2019, (weeks 27-30), then of 36,433 deaths, only 24,396 were aged 75+, that is 66.96%.

So we can see that we have summer, where 67.0% of deaths are aged 75+, a mild winter, when that rises to 70.7%, and then a bad winter (flu) season, when it goes up to 73.2%.

What's happening now? Now we are at 70.6%

So it's extremely clear that while, yes covid-19 does affect the elderly more, it trends much lower than seasonal flu, in that while a winter such as 2015, with around 5500 more deaths per week compared to summer weekly deaths, resulted in 73.2% of deaths being 75+, we now have more than 9000 extra weakly deaths compared to summer, and yet our 75+ death proportion is still only 70.6%

So conclusions:

• 20,000 people dying of seasonal flu in a season is not as bad as 20,000 people dying of covid-19 because the 20,000 flu deaths had fewer years to live (healthy or otherwise)
• some causes of death tend lower than average, hence why breast cancer is a bigger health concern than prostate cancer, but even something that is only average in terms of age is still very bad if it increases the all-cause death rate by 80% as with covid-19, because within that dead population there will be some people who have lost many years of life
• flu really does kill those who likely have a short time left to live and so comparisons with covid-19 are a bit stupid even if we ignore the fact that we live with an manage flu every year - it's just not the same thing!

I.e. it might be that flu and covid-19 are similarly contagious (R0) and that (if you want to be someone trying to be clever about who covid-19 is no big deal on social media) both can be coronaviruses (though ethanol and methanol are both alcohols so this argument is only even slightly persuasive if you are particularly thick), and therefore we should ignore covid-19 as we do flu. However this would only be true if covid-19 and flu had the same risk of death by age. And they don't.

So there.

Someone asked up thread whether clinical diagnoses are being counted in the government's figures. The answer is no, tests only.
Doctors working in covid areas are also saying that the 'covid cough' has such a distinct sound compared to other coughs that they know as soon as someone coughs whether they have it even if they don't have access to xrays/CTs. In my area access to xray machines for covid patients is being limited anyway as it means a deep clean after each xray. If this is a countrywide policy it will mean we have less clinical diagnoses from xray than other countries even if they were to be added to the figures.

Thanks quarantinequeen. I am also curious about the criteria for testing patients in hospital, are they testing all new admissions in order that positives can be isolated, or only those with suspected CV19?

QuarantineQueen I hope you are feeling much better now 💐

Yes, German reports mention the mandatory deep clean after each X ray,
but this hasn't so far prevented doctors here ordering X rays when they want the info this brings.

Obviously depends on the number of machines available, the number of patients coming in, the turnround of the deep clean process

• so presumably X rays would be rationed if our current relaxation goes badly awry

I've been digging around for reports and stats on XRays or longterm prognoses, but can't find any published here so far

pocketem

Good analysis. The things that appear to be missing (and very hard to measure) is the proportion who:

a) have ”innate” immunity (ie are exposed but don’t get infected due to some protective characteristic they innately possess, and

b) those that were infected, but did not produce sufficient antibodies to be tested.

This would reduce the virus’s mortality rate...

However, those who are currently infected and alive, but will die in the coming days and weeks, will tip the balance back the other way... We only need to look at Italy to see how long and large this tail is. 60%+ of Italy’s death’s have occurred since its death rate peaked a month ago, and daily deaths have still not dropped below 400 per day...

“...those that were infected, but did not produce sufficient antibodies to test positive”

Also, the NY study was based on grocery store customers. I wonder how representative that would be. It potentially under-represents children, those in care settings, and those shielding or wealthy enough to have their shopping delivered.

Thanks @BigChocFrenzy I am recovering slowly at last.
@sleepwhenidie I gather the protocol for who is tested changes frequently but I can't see how testing is helping them isolate if a) the tests are taking 48-72 hours to come back and b) they are only 75% accurate anyway. There must be a lot of cross infection happening in hospitals.

587 today in England

761 excluding N.Ireland according to Twitter big jump for Wales (110)

I believe that those Wales deaths include 84 that weren’t previously counted.

Quarantinequeen, that does not fit with BMA or ONS guidance or reports from colleagues. :

Medical practitioners are required to certify causes of death “to the best of their knowledge and belief”. Without diagnostic proof, if appropriate and to avoid delay, medical practitioners can circle ‘2’ in the MCCD (“information from post-mortem may be available later”) or tick Box B on the reverse of the MCCD for ante-mortem investigations. For example, if before death the patient had symptoms typical of COVID- 19 infection, but the test result has not been received, it would be satisfactory to give ‘COVID-19’ as the cause of death, tick Box B and then share the test result when it
becomes available. In the circumstances of there being no swab, it is satisfactory to apply clinical judgement.

assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/877302/guidance-for-doctors-completing-medical-certificates-of-cause-of-death-covid-19.pdf

So sorry misread that as deaths not cases. Many apologies QuarantineQueen, we actually agree on what's counted!

Latest German stats

As Derbygerbil said, long tail of recovering / dying:

German deaths still 200 - 300 per day,
even though daily cases have reduced from a peak of nearly 7,000 around end March to not much over 2,000 atm

Worldometers and RKI:

www.rki.de/DE/Content/InfAZ/N/Neuartiges_Coronavirus/Situationsberichte/2020-04-23-de.pdf?

153,000 cases (confirmed by test)
41,000 currently infected
104,000 recovered

BUT not defined whether recovered just means no longer infected - RKI use an algorithm to calculate the number of recoveries

21% of confirmed cases made no report of symptoms, so presumably none or v mild

Of the remaining 79%:

50% have cough
42% fever
21% rhinorrhoea

3% pneumonia
17% hospitalised (this was 16% for a long time)
Note: extensive treatment at home, including with O2, by "Coronavirus taxis" of medical teams making regular home visits

ICU
2,776 Currently
7,900 Discharged, of whom 30% died

==> 10,676 total in ICU at some time
which is about 7% of all confirmed cases
74% of those in ICU were ventilated

Tables for ICU and also of
age breakdown over calendar weeks 12-15 for the 17% of confirmed cases who were hospitalised

(Apolgies for all the underlines that appeared)