Daily numbers, graphs, analysis thread 9

[Barracker]

Welcome to thread 9 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.

According to the BBC the 77 is in all settings where there has been a positive test so the usual measure.

Have been lurking on this thread for a while. Thank you all for your insight.

It includes community deaths. As quite often happens, the PHE number is actually negative, because of data cleaning - they often double count, so have a periodic clean up. So today's number is the normal UK, all settings.

Ah - thanks, all. I worry about what it will look like when PHE decide a "catch up" is due again, but at least it's going in the right direction

Im surprised infection is goung down mainly as if they were accurately testing and tracking they should be finding some of the 8k infections a day.
I want a breakdown on who has obly been tested because of contact tracing. Or are they only telling them to isolate?

Why did it take brazil so long to get going is it as they move towards winter?

"It would suggest the IFR is potentially miniscule though. "

That is overly optimistic, a hope without much evidence

e.g. New York IFR is significant

30,000 dead in New York
1,563 deaths per million population
Even if every single person was infected - ntbo impossible - that would be 0.1563 %
Assuming more likely that 30% pop infected, would give

IFR = 0.521%

To compare, when calculated for the estimated number of cases, as we are doing here, not just those confirmed,
scientists estimate

IFR flu = 0.044%

Most scientists estimate COVID IFR is about 10 x flu IFR

@Keepdistance

Im surprised infection is goung down mainly as if they were accurately testing and tracking they should be finding some of the 8k infections a day. I want a breakdown on who has obly been tested because of contact tracing. Or are they only telling them to isolate?

Why did it take brazil so long to get going is it as they move towards winter?

The ons estimate of weekly cases went down to 39,000 in the week ending 30th May which means the estimated daily infections had fallen from 8000 to 5600. New figures will come out this week I assume so will be interesting to see if it's fallen further but 8000 is now out of date.

As far as I'm aware if you are contacted by a contact tracer to say you've been in contact with a covid case you'll be asked to self isolate but you will only get tested if you develop symptoms. This is because it can take up to 14 days to develop so if you get tested on day 1 (when you don't have symptoms) and it's negative you'll still have to isolate for the rest of the period anyway.

The estimate of 8,000 new cases a day must have a very high confidence interval though. I think they said that out of the sample, only about 22 people in 15 households tested positive. With such low figures testing positive, it is easy for the sample to be widely out.

"if you go on the antibody test thread on here there are lots of people who had positive swabs but a negative antibody test. "

The fact that they did not produce antibodies, or too few to be detected, does not necessarily mean that they have immunity via T cells either
For that, we need to test for T cells

Also possible that the test gave a false positive

Let's calculate the effect of false positives:

Say we have a test which is 99% specific, i.e. only 1 in 100 tests give a false positive.
< and maybe their tests were not that accurate >

And imagine we’re testing 100 people
If say 5% are infected, then we should have 5 positives
However, we would also expect 1 false positive, given the error rate of our test.

==> 1 out of the 6 positives is a false positive

@BigChocFrenzy yes i see what you mean when looking at new york data. It perhaps wouldnt impact the overall IFR. It does support your argument about working age people returning to some kind of normality though. If you look at that PHE study the IFR for under 60s is very small, and thats based on antibody screening. If alot of people dont produce antibodies and more people have had it than have been detectes, this would lower the IFR even more and further strengthen your argument. Like I say, obviously anacdotel, but over on the several antibody threads, its quite common for people who have swabbed positive to have a negative antibody test

"about 22 people in 15 households tested positive"

Much too small numbers to be reliable
Different sampling could significantly change results
Large standard deviation likely - what SD did they give for that ?

BigChoc thanks so much for the “Non-COVID” excess deaths thread - really interesting! Especially to see that from 1 May there are hardly any excess deaths recorded as non-covid.

T cells are a mechanism that the young would use, but much less likely in the 60+ age group where about 90% of the deaths occur

So we have to look at antibodies there - however, I read that age 60+ are not very efficient at producing antibodies either

Germany has all along advised those aged 60+ to take extra care to avoid infection, as death rate starts to accelerate significantly from 60:

389 have died aged 0-59 out of a population of 83 million, then 815 died aged 60-69
(I'm nearly 64, btw)

Many firms here have rolled out plans for retirement at 60 or 61;
also allows them to bring in new blood

Even with the higher death rates in the UK, most of those up to say age 55 and not shielding seem at low risk when returning to work

However, imo the govt should offer UBI to all those aged 60+ who wish to give up work
This would particularly help lowpaid workers who statistically have poorer health
it would also help the young unemployed get work

Germany, with sharp increase of risk from age 60:
(even if still low)

Thanks for that table @BigChocFrenzy the lowers numbers don't make my poor historian's brain hurt as much as our bigger ones!

Also noticeable is quite how much lower the F to M ratios are, until (I presume) you get to the age groups where there's a higher proportion of females still around to fall victim.

BigChoc
And I - being of that age group - am very willing to take resposibility for my life and take the precautions I deem necessary so that my childeren and others can get on with their lives.
I am not prepared to go without parties, church services, concerts, choirs, theatre, football etc. just to exist and go on existing.

(to quote Erich Kästner: Leben ist immer lebensgefährlich.)

@ShootsFruitAndLeaves Bit off topic but may I ask if the lha rise is set to be permanent?

@BigChocFrenzy i see your point about false positives. I could be totally wrong and correct me if i am wrong.. But if there was the same risk of a false negative would they cancel eachother out? Not great for the individual getting an incorrect result, but in terms of data. If in 100 tests there was 1 false positive and 1 false negative

Prokup I go to the gym and hairdresser, but none of the others you mention - but being an Aspie hermit, I didn't before Coronavirus either !

I have moved reading this thread but what is an aspie hermit???

BigChocFrenzy
Well, the AS members of my family are not all "hermit-y" - only DS is, but even he welcomes some hours at school, now. His DM (i. e. me) is not quite like him.
But I want others to restart their choirs (DSis), their acting (DNiece), their literary meetings (DH's job), their uni courses (DD), I want my neighbour to open his music pup - he worked so hard to make it a success and so on.

jrob My comment on false positives was re your post on people testing negative but later being found not to have antibodies.

False negatives are far more common in swab tests than false positives, for several reasons:
difficulty of getting swab back far enough, virus moving to lungs etc

So the % of infected from swab tests announced daily is probably underestimated somewhat.

However, more reliable would be the latest blood tests - but only when carried out by professionals -
which should show when antibodies have been produced, or T cells activated for COVID.

We need results of representative national sampling, with data on age, sex, race

What we don't know is what % of mild or asymptomatic cases actually result in immunity

• whetehr via antibodies or T cells -

and whether this differs wrt age, or even sex.

re false negatives & positives:

Tests are very useful to give a reasonable estimation of infection rates or antibodies / T cells over a population,
because we can estimate % false results and calculate standard deviations.

imo less useful for an individual, because they will expect a binary "answer"
So I'm still dubious about "immunity passports"

Of course, the tested population must be large enough for a few false results not to skew results

Bigchoc - thanks for the table. I was struck by how many more women in their 90s died than men which reflects the widow generation after WWII. Sobering.