Daily numbers, graphs, analysis thread 14

[BigChocFrenzy]

Welcome to thread 14 of the daily updates

Resource links:

Uk dashboard deaths, cases, hospitals, tests - 4 nations, LAs, English regions
Slides & data UK govt pressers
[[https://www.gov.uk/guidance/coronavi
rus-covid-19-information-for-the-public UK stats]] list of reports added daily by PHE & DHSC
PHE Surveillance report infections & deaths released every Thursday with sep. infographic
ONS England infection surveillance report ONS UK statistics for CV related deaths, released weekly each Tuesday
Daily ECDC report UK & EEA
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 additional data

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

on a slightly more up beat note looking at the cases per 100,000 at each decile, plus 95th and 5th decile possibly some cause for cautious optimism for w/e 2nd August. Whilst cases up again as compared to the previous week, and with the caveat that a few hundred more may still to be added, this is not reflected across every decile range (as with the previous week). This is based on UTLA cases per 100,000.

Thanks came looking for this as the Facebook Covid info section said there were 5 new cases today, which seemed unlikely

if they stop the current contract and redeploy the contract tracers to work for local authorities maybe track and trace would work. It's ridiculous paying people to contact members of the same family, or to sit around doing nothing.

I assume the track and trace on Isle of Anglesey was done by local people, it's the one place that has really knocked an outbreak on the head.

Could do with the changeson the graph.
Kids in school
Pubs open
Quarantine removed...
Kids breaking up from school.
Quarantine back for Spain
Etc

It's quite hard to understand what happened.
I think schools and non essential shops at the same time and pubs later?
But more kids did go back as time went on.

@PatriciaHolm et al

Sorry all but I'm struggling to understand if the % positive on the testing is on an apples for apples basis. And so need some help. :) I think I have confused myself, so de-confusion - both maths and logic - would be most welcome.

Last Friday's PHE surveillance report contained this table.

It shows for each watch list LA the number of individuals tested per day per 100,000 population, seven day moving average.

It also shows this for England overall.

Alongside this it also shows the instance per 100,000 weekly - I'm assuming "instances" is confirmed cases.

Given we know the rates per 100,000 and the associated populations the number of individuals with a processed test(s) - as opposed to tests processed - can be calculated both for England overall and for each of the named authorities.

As such England - 56.287 million people (rounded); 94.9 test rate. This equates to c.53,416 per day or over 7 days approx 373,914 individuals.

Over the same period England has a confirmed case rate (weekly) of 7.8 per 100,000. This equates to 4390 individuals with a confirmed case.

4390 / 373914 is 1.17%

Now published processed tests (P1+P2) for the same 7 day period are 734,336 or with a 3 day lag applied 749,310.

4390 / either of those numbers gives 0.60% and 0.59%.

However I note that dashboard tests processed has the following caveat This is a count of test results and may include multiple tests for an individual person .

Which takes me back to my is this really apples with apples? and should not infection rate be shown in terms of percentage of individuals testing positive rather than processed tests?

Your collective patience, understanding and clarity is as ever hugely appreciated as I fear I may have got lost in the numbers on this one.

boys3, because only number of tests a day are published, not people tested, this is all that can be use to calculate a positive rate on a daily basis.

As part of the weekly test and trace report (next one published tomorrow I think?) numbers of people tested in the relevant week is published, and this can then be compared week by week to assess the trend in positivity rates. It will be interesting to see whether the rate has gone up when this is published tomorrow.

It has been around 1.2% I think for the last couple of weeks for which figures are available. I’ve noticed that as a rough estimate, it’s about double the rate calculated from tests processed.

Which takes me back to my is this really apples with apples? and should not infection rate be shown in terms of percentage of individuals testing positive rather than processed tests

Yes, it should, but that's not what the dashboard shows, or data that has ever been published on a daily basis - it's always been number of tests processed, which indeed will be more than the number of people.

It would appear that the Dashboard is showing Tests, but that PHE table is based on people, which is new I think. I think you have been able to back calculate it from the cumulative numbers of people tested they publish in the PHE weekly reports, but only on a weekly basis, and by the time the report comes out the data is already 4-5 days old so it's not much good at really up to date data.

Those ratios fit with what I had worked out a little earlier for Scotland- -positivity rate on people of 1.3% vs your 1.17% and for on tests of .54% for today for Scotland vs UK 0.6%.

The ECDC daily summary states the 14-day incidence for the UK = 14.1 cases / 100,000

but again, I'm not sure whether this is cases or people

The age-related graph is another useful reminder of how much the age of infected people has dropped
The curve for age 80+ is dramatic

https://www.ecdc.europa.eu/en/interpretation-covid-19-data

•	"Testing policies and the number of tests performed per 100 000 persons, vary markedly across the EU/EEA and presumably even more so among third countries. More extensive testing will inevitably lead to more cases being detected.

•	The 14-day notification rate of new COVID-19 cases should be used in combination with other factors including testing policies, number of tests performed, test positivity, excess mortality and rates of hospital and Intensive Care Unit (ICU) admissions, when analysing the epidemiological situation in a country. Most of these indicators are presented for EU/EEA Member States and the UK in the <a class="break-all" href="http://covid19-country-overviews.ecdc.europa.eu/" rel="nofollow" target="_blank">Country Overview</a><a class="break-all" href="http://covid19-country-overviews.ecdc.europa.eu/" rel="nofollow" target="_blank">(link is external)</a> report."

ECDC also have a useful test positivity graph I check often
The scale shows the trend, if not the detail

.

@AprilLady , @PatriciaHolm thanks for the responses

You are quite right last week's PHE surveillance report was the first time I have seen a table from which allows the number of people tested to be identified over a given period. My calculation is directly from the info in that table, table 1 in last week's report, as nothing more than area population is needed. The "same" table in the prior week just listed LAs, status and direction of travel. So it will be interesting to see what this week's report includes.

@BigChocFrenzy

.

@BigChocFrenzy is there access to the data behind that graphs. rather like many of the PHE surveillance graph an axis scale is used which makes reading the graph - when we are at very low numbers - very difficult.

That ECDC number, for the UK at least, is reported cases over a fortnight per 100,000 of the population.

Are we able to talk about other countries on this thread? Spain 1772 cases identified yesterday seem still very high, it's (intentionally?) difficult to work out because of them adding antibody tests making it 2953, but hard to compare old versus new figures.

Also, does anyone have insight into what the situation is with Pakistan? Are they hiding figures or has positivity really dropped?

That is the % ofPeople Testedwho were positive - not tests

Ah, of course. Thank you for explaining it so clearly, PatriciaHolm.

Upbeat but realistic (I think?) article from the Beeb:

www.bbc.co.uk/news/health-53656852

Useful graphs. Discusses how more testing means more infections found, t&t helping even if not brilliant, local lockdowns show system is working.

Do the Epidemiologists of Mumsnet agree??

@boys3 The data file is here

www.ecdc.europa.eu/en/publications-data/covid-19-testing

The graph is the number of announced cases over tests, per week. The last datapoint is the week ending July 26, so the last 4 data points are pretty much all 0.5%! But you can't see that as the scale is bonkers because they start so far back when we were testing much less so positivity is much much higher (40%!)

Place mark. Thank you

Placemarking for later

@MRex

Are we able to talk about other countries on this thread? Spain 1772 cases identified yesterday seem still very high, it's (intentionally?) difficult to work out because of them adding antibody tests making it 2953, but hard to compare old versus new figures.

Also, does anyone have insight into what the situation is with Pakistan? Are they hiding figures or has positivity really dropped?

Yes - it was 1,178 new cases yesterday (positive cases not inc antibodies). It's still very focused though - 73% in Madrid, Catalunya and Aragon, with Navarra (next to Aragon) having a small relative number but a high incidence per 100,000.

Pakistan - covid.gov.pk/stats/pakistan seems to show quite a significant drop in cases with a fairly stable level of tests. Pakistan Today claims -

“Our strategy was based on speedy testing, contact tracing and isolation, ensuring services to admitted patients and observing shortages and immediately addressing these shortages through use of domestic technology,” an official at the Health Ministry said.
The NCOC moved scientifically and identified top 30 cities as Covid-19 hotspots with max Infected using auto trace and NITB maps. At an average 3-8 million population with 10-30% of infected people remained under 300-500 smart lockdowns (SLDs) on rolling basis since adoption of this policy in entire country."
"

@JulyBreeze

Upbeat but realistic (I think?) article from the Beeb:

www.bbc.co.uk/news/health-53656852

Useful graphs. Discusses how more testing means more infections found, t&t helping even if not brilliant, local lockdowns show system is working.

Do the Epidemiologists of Mumsnet agree??

Personally - not as an epidemiologist but a bit of a numbers nerd, and to be fair one of life's eternal optimists - yes, I do think it's pretty realistic, and quite refreshing to see especially from the BBC.

"Is the UK in a better position than we think?" The answer to that question depends on who "we" is. If it's the ones running around saying it's all a hoax then no. If it's a young mother terrified for her firstborn then yes. Personally I dont trust anything coming out of Oxford now as they have been unduly optimistic in the past with daft claims about herd immunity.

If track and trace was working well then the percentage of positive tests would go up very rapidly and then fall. You'd see the Isle of Anglesey picture - a very large spike in cases followed by almost no cases afterwards. It's actually the only place with that pattern.

So the picture is not all doom and gloom but track and trace is not working that well either.

At the moment the infection rate is too high in certain parts of the country and it has stopped falling in quite a few other areas. However the vulnerable have mostly managed to avoid infection so hospitals are able to do other work. But they still have to take massive precautions to avoid patients becoming infected and that makes it impossible to deal with the backlog before winter.

We are still relying on people being sensible.

[quote PatriciaHolm]@boys3 The data file is here

www.ecdc.europa.eu/en/publications-data/covid-19-testing

The graph is the number of announced cases over tests, per week. The last datapoint is the week ending July 26, so the last 4 data points are pretty much all 0.5%! But you can't see that as the scale is bonkers because they start so far back when we were testing much less so positivity is much much higher (40%!)[/quote]
that's great, thanks patricia

@PatriciaHolm I was using the term as a very broad collective phrase!

[quote JulyBreeze]@PatriciaHolm I was using the term as a very broad collective phrase![/quote]
yes! I was being tongue in cheek too ;-) I'm not sure we have any real epidemiologists here - do we? Maybe they are in hiding ;-)