Meet the Other Phone. Protection built in.

Meet the Other Phone.
Protection built in.

Buy now

Please or to access all these features

Covid

Mumsnet doesn't verify the qualifications of users. If you have medical concerns, please consult a healthcare professional.

Daily numbers, graphs, analysis thread 16

999 replies

BigChocFrenzy · 28/08/2020 18:44

Welcome to thread 16 of the daily updates

Resource links:

Uk dashboard deaths, cases, hospitals, tests - 4 nations, English regions & LAs
MSAO Map of English cases
[[https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/909430/Contain_framework_lower_tier_local_authority__14_August_2020.pdf
Slides & data UK govt pressers
UK added daily by PHE & DHSC
R estimates UK & English regions
PHE Surveillance report infections & watchlists every Thursday
ONS England infection surveillance reports
ONS UK death stats released each Tuesday
ECDC rolling 14-day incidence EEA & UK
Daily ECDC country detail UK
WHO dashboard
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 📈 📉 📊 👍

OP posts:
Thread gallery
90
Littlebelina · 31/08/2020 15:06

@Firefliess

Thanks *@Littlebelina*. So that's not a problem that would affect our overall count of cases (as we only count new cases) It would be a problem for keyworkers who needed a negative test to return to work.
No it shouldn't overestimate cases. I see it potentially causing issues in a couple of areas, one (as you say) is if a negative test was a requirement to return to work (I don't know if any employer requires this? Does anyone?). Another potential area is if there is wide spread testing of people without symptoms (due to hotspots for example). You could potentially end up with people (and their families) isolating, taking time off work/school etc when they had the infection several weeks prior. It could result in school bubbles shutting down. Ultimately a positive test shows there has been an infection so is useful from a case counting point but if this is correct it might result in authorities chasing down problems that actually happened a few weeks prior.

I have seen criticism of the test sensitivity somewhere before as well but it might have been from a "it's just bad flu" type source so I was waiting to see more evidence.

Littlebelina · 31/08/2020 15:12

mobile.twitter.com/apoorva_nyc/status/1299705092178956288

The author has summarised the article in tweets for those (like me) have exceeded their nytimes article count

AnyFucker · 31/08/2020 15:44

.

BigChocFrenzy · 31/08/2020 15:56

We need to distinguish between surveillance studies of populations, when almost all will NOT be infected
vs
tests of people with symptoms, who still probably don't have COVID, but do have a significantly increased chance.

I tried out 2 different scenarios to get a very rough idea of the consequences of false negatives and positives

Using a flow chart from Spiegelhalter where he calculated assuming 6% infected, what false negatives & positives would come from tests,

I copied his basis and took 6% infected for the situation of people tested because of symptoms
For ease of calculating both scenarios, I multiplied by 10 to get a 10,000 sample (and crossed out his figures for 1,000)

==> For the 6% infected (with symptoms) this gives 510 positives in total instead of 600, i.e. 90 too few positives

and then in red I added my figures assuming a general population with only 0.1 % infected (so no symptoms)
==> which gives 106 positives instead of 10, i.e. 96 too many positives from a sample of 10,000 people

This illustrates why population surveys need to use serology tests, not just swabs - hence look at the weekly surveillance reports for absolute numbers, with the daily numbers more for trends
and also why
the ONS COVID-19 Infection Survey will be expanded from regularly testing 28,000 people per fortnight in England to 150,000 by October
and eventually to 400,000 people, also adding the other 3 UK nations.
This will greatly improve confidence levels

However, an individual person with symptoms is probably much more likely to have a false negative than a false positive

  • saliva or blood tests will hopefully improve accuracy
Daily numbers, graphs, analysis thread 16
OP posts:
MRex · 31/08/2020 16:10

Far beyond my knowledge to know what is or isn't appropriate and the impacts. It's interesting in that I wonder if it impacts on different countries' figures, undoubtedly they will all have a different approach as with everything else.

I just looked up and NHS protocol looks to be 45 cycles.
I can't link but via Google the document is called
"Guidance and standard operating procedure
COVID-19 virus testing in NHS laboratories", it's in Appendix 5 that it says 45 cycles.

A few challenges I can think of to her theory even if it's correct (I'm not qualified to judge): 1) for tracing purposes we want to know who's unwell even if they are less infectious, 2) Early stage infectiousness is also useful to know, reducing false negatives.

BigChocFrenzy · 31/08/2020 16:10

I have read both 0.1% and 1% for false positives, depending on which test is used
Obviously makes a big difference when doing population surveys, not so much for an individual being tested

I found this PHE benchmark study for antibody tests, but not yet an assessment of UK swab tests (US CDC tests in the early stages were not adequate, so I don't know if they are currently at UK levels)

PHE: Evaluation of sensitivity and specificity of four commercially available SARS-CoV-2 antibody immunoassays

Apparently lab tests are assessed by both their ability to detect a positive case (sensitivity)
and their ability to determine a negative case (specificity).

https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachmentdata/file/898437/Evaluationoffsensitivityanddspecificityoff4commerciallyyavailableSARS-CoV-22antibodyimmunoassays.pdf

Sensitivity was evaluated on 536 positive samples from unique adult individuals with laboratory-confirmed SARS-CoV-2 infection at ≥20 days post-symptom onset;

specificity was evaluated on 994 pre-pandemic (2015-2018) specimens < hence definitely negative > from unique, healthy adult individuals.

All 4 tests met specificity criteria, but only Siemens met sensitivity as well, see table:

Daily numbers, graphs, analysis thread 16
OP posts:
MRex · 31/08/2020 16:23

1406 cases, 2 Deaths. Retrospective on specimen date looks like cases have flattened off, but could still go either way.

BigChocFrenzy · 31/08/2020 16:28

@Littlebelina

mobile.twitter.com/apoorva_nyc/status/1299705092178956288

The author has summarised the article in tweets for those (like me) have exceeded their nytimes article count

... Ah, thank you Brew

This refers to USA "state labs" so we definitely need to know what the UK and the rest of Europe does

One issue:
It looks like lowering the CT would reduce the sensitivity of tests, i.e. increase false negatives - which are already about 30% according to the NHS staff testing info I read

Also, if we look at when we had high deaths, there was still only a small minority infected (going by about 7% of the population atm with antibodies)
==> reducing the rl number of positive cases would increase the calculated fatality rate, which would require a lot more evidence to do if the change is more than tiny

OP posts:
CoffeeandCroissant · 31/08/2020 16:50

Might be of interest to any programmers on here? In any case, should see more interesting use of the data from those who know how to do so...

"The UK was among the worst in class when it came to early COVID-19 data access—but it has now released a full API coronavirus.data.gov.uk/developers-guide to access its cases, deaths, and testing data. "
mobile.twitter.com/redouad/status/1300364677642358786

ShootsFruitsAndLeaves · 31/08/2020 17:05

Much caution is needed with global covid numbers.

India testing more than 1 million samples per day

indianexpress.com/article/explained/india-covid-19-death-count-6577136/

Positivity rate is around 8%

Less scientific countries such as Indonesia have pathetic testing rates of 20,000 per day (relatively 10 times lower) and 16% positivity

www.thejakartapost.com/news/2020/08/29/indonesia-sees-record-high-in-covid-19-cases-for-third-day-running.html

I'm not sure the exact relationship between positivity and testing, but some areas won't be getting any tests at all (Indonesia tests many people using antibody tests, which don't count in its statistics at all), however clearly the daily case rate is many, many times higher.

Note that the pillar 1 positivity rate in the UK (for health workers) is only 0.2%.

I'm not sure if it will ever be possible to calculate the true global rate of covid-19 death whether it is undercounted currently by say a million, or more.

BigChocFrenzy · 31/08/2020 17:09

At best we can compare groups of developed countries, preferably with similar population density and consider the % positive tests for each

It makes no sense to compare developed / energing / developing countries, with vastly different demographics, testing regimes, health srvices etc

OP posts:
BigChocFrenzy · 31/08/2020 17:16

Good perspective from Spiegelhalter on statistics, statisticians and abuse of either

Relates a horrendous misuse of stats by the FDA chief recently
< the same chief who is considering allowing an overly rushed vaccine approval before the November elections, to help Trump >

https://www.theguardian.com/commentisfree/2020/aug/31/politicians-covid-19-statistics-statisticians

It is vital that we don’t just count cases, but look at their ages,
and the latest report from Public Health England reveals the highest rate of positive tests is in the 15-44 age-group.
< in Germany, the RKI state it is the 10-30 age group, but I don't know if this is a genuine difference or just different choice of age bands >

If 5,000 30-year-olds get infected, we might expect just one or twoo^ to die, and so this helps explain the lack of severe consequences so far.

But it’s a different picture for the frail and elderly, who need proper protection.

OP posts:
ChristmasinJune · 31/08/2020 17:17

Apologies if this has been discussed upthread but does anybody know why numbers in hospital and on a ventilator in England have dropped significantly over the last few days?
I know they've been declining steadily but this looks sharper, have there been more adjustments?

Patients in hospital 28/8 430 to 31/8 305 (initially dropped to 280 then jumped a bit)

Patients on ventilators 28/8 52. 31/8 33

IceCreamSummer20 · 31/08/2020 18:28

Spiegelhalter’s article is excellent. If there is ever a man to rely on for numbers and common sense, it is him!

MarshaBradyo · 31/08/2020 18:29

I always like hearing his name on R4

Quite fond of a few other regulars too

ShootsFruitsAndLeaves · 31/08/2020 18:32

It makes no sense to compare developed / energing / developing countries, with vastly different demographics, testing regimes, health srvices etc

Not sure how that works with the UK's quarantine rules .

For example, Jamaica was just added because cases exceed 20 per 100k in 7 days.

However Indonesia's case rate is only 8.

Indonesia should NOT be quarantine-exempt because it's a chaotic mess.

But it does make you wonder how much testing they were doing in Jamaica, compared to say Switzerland.

Also not sure how meaningful stats can be for countries such as Antigua (population 80,000)

PatriciaHolm · 31/08/2020 18:39

@ChristmasinJune

Apologies if this has been discussed upthread but does anybody know why numbers in hospital and on a ventilator in England have dropped significantly over the last few days? I know they've been declining steadily but this looks sharper, have there been more adjustments?

Patients in hospital 28/8 430 to 31/8 305 (initially dropped to 280 then jumped a bit)

Patients on ventilators 28/8 52. 31/8 33

In reality, probably lack of complete data reporting at the weekend - not all trusts report at weekends, and probably not on bank holidays too. The numbers have bobbled up and down a couple of times like that before and I think it's because some trusts aren't included. It will sort itself out over the next couple of days.
MRex · 31/08/2020 18:42

@ShootsFruitsAndLeaves - Indonesia isn't on the exemption list though? For the reasons you state, many countries aren't. Jamaica has a fairly pitiful health service, but cases up to August were negligible, while there is a fairly large Jamaican population in London who may have needed to travel to see family; blocking that in the context of averaging 5 cases/ day would have appeared very unfair.

boys3 · 31/08/2020 19:03

@MRex

1406 cases, 2 Deaths. Retrospective on specimen date looks like cases have flattened off, but could still go either way.
Just over 80% of today's cases in England, so less than the population proportion of the UK. Scotland and NI additions today, would be both over 1600 cases added if population same as England.

North West still a problem, and some of those eased out of local restrictions showing signs of struggle.

North West extract sorted by cases per 100,000 with specimen date for w/e today - which of course is an incomplete week at the moment; but if LAs already matching or higher than the previous full week figure then that tells it own story

Daily numbers, graphs, analysis thread 16
boys3 · 31/08/2020 19:14

and an updated seven day moving average - week ending yesterday shaping up to be pretty similar case number in England to the two previous weeks; w/e 16th and 23rd August; 6877, and 6743 cases respectively

Daily numbers, graphs, analysis thread 16
AlecTrevelyan006 · 31/08/2020 19:36

Tab3 here:
www.england.nhs.uk/statistics/wp-content/uploads/sites/2/2020/08/COVID-19-total-announced-deaths-27-August-2020-weekly-file.xlsx

Just 1,390 healthy people (i.e had no pre-existing condition) have died from Covid

whatsnext2 · 31/08/2020 19:39

Interesting paper predicting Covid risk by country by adjusting for age and sex:

www.sciencedirect.com/science/article/pii/S0305750X20302977

BigChocFrenzy · 31/08/2020 19:39

NI
Cases and hospitalisations are increasing, although from a low base,
but the R number is around 1.3, which is concerning if it stays there for long

Population 1.9 million

. The current estimate of R is 1.0 - 1.6
• Average number of new positive tests per day last 7 days - 56.3
• 7 day incidence based on new positive tests - 20.7 / 100k population
• 14 day incidence based on new positive tests - 36.9 / 100k population
• 7 day average of total tests (pillar 1 and 2) which are positive - 1.34%
• 7 day average number COVID occupied hospital beds - 17.0

www.health-ni.gov.uk/news/current-r-number-estimate-27-august-2020

OP posts:
Littlebelina · 31/08/2020 19:45

@MRex

Far beyond my knowledge to know what is or isn't appropriate and the impacts. It's interesting in that I wonder if it impacts on different countries' figures, undoubtedly they will all have a different approach as with everything else.

I just looked up and NHS protocol looks to be 45 cycles.
I can't link but via Google the document is called
"Guidance and standard operating procedure
COVID-19 virus testing in NHS laboratories", it's in Appendix 5 that it says 45 cycles.

A few challenges I can think of to her theory even if it's correct (I'm not qualified to judge): 1) for tracing purposes we want to know who's unwell even if they are less infectious, 2) Early stage infectiousness is also useful to know, reducing false negatives.

To your first challenge, if people are testing postive well into their illness (and past the point where they are very infectious) it could be that the wrong contacts are traced. ie with these people it's not the people they've been in touch with a few days before we need to worry about but those from a few weeks back. Obviously by this point it's probably too late for them to isolate but it might show where other cases might emerge? Highly complicated though and a logistical nightmare.

At the moment the pcr test is still the best option for tackling active cases and if ( as big choc says) reducing cycles increases false negatives I don't see it being an option. I think it's right to question its limitations and seek out better alternatives but there needs to be caution not to encourage the "covid is a hoax" lot who might see this a flag to wave to claim the test is faulty. It's not, it's just, like most things, has it's limitations. Not seen this from the author of the nytimes piece, think she is just calling for smarter testing in the US rather than reducing testing which seems to be the government's approach.

herecomesthsun · 31/08/2020 19:48

@boys3

and an updated seven day moving average - week ending yesterday shaping up to be pretty similar case number in England to the two previous weeks; w/e 16th and 23rd August; 6877, and 6743 cases respectively
@MRex and @boys3

more of an uptick