Daily numbers, graphs, analysis thread 15

[BigChocFrenzy]

Welcome to thread 15 of the daily updates

Resource links:

Uk dashboard deaths, cases, hospitals, tests - 4 nations, LAs, English regions
Slides & data UK govt pressers
UK added daily by PHE & DHSC
PHE Surveillance report infections & watchlists every Thursday with sep. infographic
ONS England infection surveillance report
ONS UK death stats each Tuesday
ECDC rolling 14-day incidence EEA & UK
Daily ECDC country detail UK
Worldometer UK page
Plot FT graphs compare countries deaths, cases, raw / million pop
Covidly.com world summary & graphs
Plot COVID Graphs Our World in Data test positivity etc

We welcome factual, data driven, and civil discussions from all contributors 📈 📉 📊 👍

ONS figures are estimations of a very small %:

www.ons.gov.uk/peoplepopulationandcommunity/healthandsocialcare/conditionsanddiseases/bulletins/coronaviruscovid19infectionsurveypilot/englandandwales14august2020

"During the most recent week (3 to 9 August 2020),
we estimate there were around 0.69 (95% credible interval: 0.42 to 1.08) new COVID-19 infections for every 10,000 people in the community population in England,
equating to around 3,800 new cases per day (95% credible interval: 2,300 to 5,900)."

I haven't seen anything that says they are adjusting for false positivity. The latest release did have the numbers - 17 positive out of 30,000.

Having delved a bit further they do acknowledge it as an issue, but think that the flase positive rate must be very low, for reasons similar to what I've said.

In an article on the BBC I read that Germany quarry times for two weeks OR tests on arrival from a high incidence country. I understand UK says two weeks regardless as a test could be negative whilst COVID was incubating. Do the Germans have better tests available?

I liked this analogy from Drosten:

< Drosten is Merkel's chief COVID adviser, who broadcasts frequently (without Merkel)
He is the one who developed Germany's reliable test back in mid-January, which gave them a major boost >

https://www.dw.com/en/coronavirus-what-you-should-know-about-tests/a-53349363

the infectious disease specialist Christian Drosten from Berlin's Charite Hospital compared this with

an attempt to catch a goldfish in a tank with a hand net
If you pull the net out of the water and there is no fish in it, it does not mean that there are no fish in the tank.

@Bigchoc That diagram explains the potential impact of false positives and negatives very nicely, but you really need to have actual real world figures on it for the calculations to work. If the rate of false positives was anywhere near 1% that would suggest that nearly all the cases currently being detected via pillar 1 and pillar 2 testing are in fact false positives. So I suspect it's much lower. But even 0.1% would be contributing around 10% of "cases" (assuming a positivity rate of around 1% which I think I'd what we have currently)

Is it actually that there are zero false positives if labs are careful and avoid cross contamination? I'd expect to have heard a bit more about it if not - and actual data not mathematical modelling!

[quote Firefliess]@Bigchoc That diagram explains the potential impact of false positives and negatives very nicely, but you really need to have actual real world figures on it for the calculations to work. If the rate of false positives was anywhere near 1% that would suggest that nearly all the cases currently being detected via pillar 1 and pillar 2 testing are in fact false positives. So I suspect it's much lower. But even 0.1% would be contributing around 10% of "cases" (assuming a positivity rate of around 1% which I think I'd what we have currently)

Is it actually that there are zero false positives if labs are careful and avoid cross contamination? I'd expect to have heard a bit more about it if not - and actual data not mathematical modelling! [/quote]
The additional complication here is of course that those numbers are based on random sampling - if we took a random 1,000 of the population and sampled them. We aren't doing that; we are testing people who feel the need for a test/have been told to test, so our (real) positivity rate should be higher than that of the general population.

That would make the false positives a smaller proportion of the whole positives.

I don't think false positives are generally a fault of the lab - they are a problem with the test itself, and different test types will be more reliable than others. For example, the FDA in the US has flagged one type of test as having up to a 3% false positive -

www.fox6now.com/news/fda-warns-labs-about-high-risk-of-false-positives-from-medical-companys-covid-19-test

713 positives in UK today, 3 deaths. (down from 816 last Monday).

https://www.theguardian.com/politics/live/2020/aug/17/uk-coronavirus-calls-boris-johnson-intervene-exams-gcse-a-levels-crisis-live-news?page=with:block-5f3a9e518f08f8e2a854b007#block-5f3a9e518f08f8e2a854b007

Birmingham’s public health director has said the city could “very easily” face additional lockdown restrictions imminently following a large rise in coronavirus cases.

Dr Justin Varney said the city’s infection rate had more than doubled since the start of August and that
there had been 321 new Covid-19 cases in the past week.

@AnyFucker

Wrong thread, sorry

Dunno.

Think that comment also relates to most covid stuff

Birmingham's graph coronavirus.data.gov.uk/cases?areaType=ltla&areaName=Birmingham

Coventry is looking worse again. coronavirus.data.gov.uk/cases?areaType=ltla&areaName=Coventry

If Blackburn goes on being stupid - 100 person wedding reception broken up by police - they can expect to stay high.

^The additional complication here is of course that those numbers are based on random sampling - if we took a random 1,000 of the population and sampled them. We aren't doing that; we are testing people who feel the need for a test/have been told to test, so our (real) positivity rate should be higher than that of the general population.

That would make the false positives a smaller proportion of the whole positives.^

Yes, it can be a very different thing whether the sampling is considered to be independent or not. That's why I was considering the ONS survey results, which are supposedly a random sample - if you assume that all of them are false positives then at the current rate of testing that means that you would expect 80-90 false positives a day, which is about 10% of the total positive tests. So, for the daily tests it shouldn't be making a massive difference. But it could still be making a significant difference to the prevalence survey, we just don't know how much of an affect it is having.

The other big assumption I am making is that the tests being carried out all have the same false positive rate. It could well be that the testing done for the ONS have a different rate (and that pillar 1 and pillar 2 tests also have different rates).

So they've started publishing the death numbers again which is good. Although shame they are now not breaking it down into regions it seems....

Map showing 14-day incidence rate in the UK & critical EEA countries, with changes (red) since previous 14 day period,
data from ECDC:

India being hard hit, may have a high death toll among the rural elderly:

www.theguardian.com/global-development/2020/aug/17/indias-invisible-catastrophe-fears-over-spread-of-covid-19-into-poor-rural-areas

With limited resources, their families will not rush an elderly person to hospital if they have a fever,”
said Muliyil.
“They will be allowed to die.
That is the reality in rural India where life expectancy is 65.”

The UK is doing relatively well. Who'da thought it.

That's what the figures now say.

I know this is a thread specifically about figures. But does anyone here really believe the figures?

I had read that in India in some poorer areas - there are not as many older people, but a lack of social distancing. In Mumbai slums, there is a report that 57% are have Covid19 antibodies, and may have reached herd immunity. So may now be relatively safe for a limited period. However this is controversial. And this will have come at a cost, who knows how many people are now chronically ill because of Covid in the slums. And nearly 7,000 people died in Mumbai but it’s unclear how many were from the slums.

Rural India is different and much may be unreported. However there may be a lower mortality because of a younger population. Survival rates generally for older people in poorer classes in India is low.

I think the UK is testing widely now, so the case figures are as good (if not better) as elsewhere. The NHS admission and death figures are reliable and the ONS death certificate figures are good now, although they weren't early in the pandemic.

After 3 weeks of my town showing no new cases it's been updated to having 3 this week.

2 on one day and 1 2 days later.

Shows the transmission of this virus is so high. It just appears out of nowhere in communities.

Ok realistic explanation is people returning from holidays with in U.K. or abroad or people meeting friends/family from high transmission area.

Just a reminder though how quickly it can grow exponentially and get out of control again.

Does anyone else feel that the public messaging about this is really non existent poor?

I've spoken to loads of people this week who can't understand why we are still having all the measures when cases are low and I find it quite frustrating trying to explain because if we went back to lives in February we'd very soon return to live as it was in March and April.

Our area is similarly frustrating; very low cases for ages and none anywhere near us, now little spotty blue patches are popping up all over the map, right across the London boroughs. Holidays are a likely source.

I live about 5 mins from Moulton Park and frequently see staff car sharing and piling into buses without masks over their noses and mouths. A lot of the staff live in HMOs.
I'm also pretty sure there was an outbreak at nationwide right at the beginning of the pandemic.

So, after going looking for it, there are some warnings about Birmingham cases rising, however I would imagine most people here are completely oblivious - why isn't there more out there to warn and advise people of this?! ,

Are students going in to schools to pick up GCSE results? Never mind schools opening, look for number of infections going up everywhere approx 2 weeks after that.

Results are being sent by email and schools are phoning students. In our case, they are able to go in to school in cases where they need assistance from teachers.

DS is going to school to meet the same kids he will be bubbled with in 10 days time !