Covid 19 outbreak - arguing about Stats (again)

Use this board for general non-cycling-related chat, or to introduce yourself to the forum.
reohn2
Posts: 40169
Joined: 26 Jun 2009, 8:21pm

Re: Covid 19 outbreak - arguing about Stats (again)

Postby reohn2 » 22 Nov 2020, 7:42pm

thirdcrank wrote:
The utility cyclist wrote: ... When you say carriers do you mean people who can infect others? ...


I am saying:-

There is a virus

It's passed from people who are infected with it to those who are not infected.

Some people who are infected with the virus do not display symptoms of infection.

But if you isolate those with the virus,whether asymptomatic or not,you isolate the virus,of course there are those with other ideas :roll:
-----------------------------------------------------------

User avatar
horizon
Posts: 10246
Joined: 9 Jan 2007, 11:24am
Location: Cornwall

Re: Covid 19 outbreak - arguing about Stats (again)

Postby horizon » 22 Nov 2020, 7:56pm

Jdsk wrote:What does "asymptomatic people are not infectious" mean? Does that refer to asymptomatic infected people or asymptomatic noninfected people? Because it's obviously true for the second case.

Jonathan


thirdcrank wrote:
If the suggestion is that somebody testing positive may not transmit the virus then IMO until that's resolved for certain then it's best to err on the side of assuming they are infectious.


AIUI, there are people who test positive (and lots of people who aren't yet tested) who don't have symptoms. The people who tested positive (AIUI) are currently assumed to be infectious - that is to say, they can pass the virus on to someone else. The other people might have the virus without knowing and are therefore possibly infectious. IIRC, this was mentioned on Radio 4 as a downside of the Moonshot: lots of asymptomatic people were being tested and found to be either negative or positive when in fact it made little difference (if any?) either way to the spread of the virus. The lady professor in question said that the cost of the programme (£20 bn) was therefore wasted and that testing should take place when symptoms are present or at the very least in situations where they were likely.

I haven't re-checked my sources so apologies if I have got this wrong.
When the pestilence strikes from the East, go far and breathe the cold air deeply. Ignore the sage, stay not indoors. Ho Ri Zon 12th Century Chinese philosopher

User avatar
horizon
Posts: 10246
Joined: 9 Jan 2007, 11:24am
Location: Cornwall

Re: Covid 19 outbreak - arguing about Stats (again)

Postby horizon » 22 Nov 2020, 8:01pm

A bit more Googling, but I haven't read it in full (and never could!):

https://www.nature.com/articles/s41467-020-19802-w

There were no positive tests amongst 1,174 close contacts of asymptomatic cases.
When the pestilence strikes from the East, go far and breathe the cold air deeply. Ignore the sage, stay not indoors. Ho Ri Zon 12th Century Chinese philosopher

User avatar
horizon
Posts: 10246
Joined: 9 Jan 2007, 11:24am
Location: Cornwall

Re: Covid 19 outbreak - arguing about Stats (again)

Postby horizon » 22 Nov 2020, 8:20pm

horizon wrote:IIRC, this was mentioned on Radio 4 as a downside of the Moonshot: lots of asymptomatic people were being tested and found to be either negative or positive when in fact it made little difference (if any?) either way to the spread of the virus. The lady professor in question said that the cost of the programme (£20 bn) was therefore wasted and that testing should take place when symptoms are present or at the very least in situations where they were likely.

I haven't re-checked my sources so apologies if I have got this wrong.


It was here on Inside Science (first half of the programme). It isn't quite how I remembered it but still interesting and BTW, the cost of Moonshot is £100 bn :shock: :

https://www.bbc.co.uk/sounds/play/m000ph4n
When the pestilence strikes from the East, go far and breathe the cold air deeply. Ignore the sage, stay not indoors. Ho Ri Zon 12th Century Chinese philosopher

Jdsk
Posts: 3921
Joined: 5 Mar 2019, 5:42pm

Re: Covid 19 outbreak - arguing about Stats (again)

Postby Jdsk » 22 Nov 2020, 8:23pm

horizon wrote:
Jdsk wrote:What does "asymptomatic people are not infectious" mean? Does that refer to asymptomatic infected people or asymptomatic noninfected people? Because it's obviously true for the second case.

thirdcrank wrote:
If the suggestion is that somebody testing positive may not transmit the virus then IMO until that's resolved for certain then it's best to err on the side of assuming they are infectious.

AIUI, there are people who test positive (and lots of people who aren't yet tested) who don't have symptoms.

Yes.

horizon wrote:The people who tested positive (AIUI) are currently assumed to be infectious - that is to say, they can pass the virus on to someone else.

It depends where you mean, but for current UK contact isolation: Yes.

horizon wrote: The other people might have the virus without knowing and are therefore possibly infectious.

Which "other people"?

Jonathan

thirdcrank
Posts: 30108
Joined: 9 Jan 2007, 2:44pm

Re: Covid 19 outbreak - arguing about Stats (again)

Postby thirdcrank » 22 Nov 2020, 8:26pm

We risk using the "innocent until proven guilty" ethic inappropriately.

IMO, with something like this, that's wrong.

Re wasted tests, mass testing has only become an apparent necessity here because of inaction at the outset. ie If other policies such as controlled entry to the country had been adopted from the start, then mass testing would not have been necessary. If MD is right, South Korea isn't vaccinating for a year because the measures promptly taken there mean they feel they have this under control without vaccination.

Jdsk
Posts: 3921
Joined: 5 Mar 2019, 5:42pm

Re: Covid 19 outbreak - arguing about Stats (again)

Postby Jdsk » 22 Nov 2020, 8:27pm

horizon wrote:IIRC, this was mentioned on Radio 4 as a downside of the Moonshot: lots of asymptomatic people were being tested and found to be either negative or positive when in fact it made little difference (if any?) either way to the spread of the virus. The lady professor in question said that the cost of the programme (£20 bn) was therefore wasted and that testing should take place when symptoms are present or at the very least in situations where they were likely.

Muir Gray on what's happening in Liverpool:
https://www.bmj.com/content/371/bmj.m4436

Jonathan

Statement of interest: He's a long-term colleague and friend.

User avatar
horizon
Posts: 10246
Joined: 9 Jan 2007, 11:24am
Location: Cornwall

Re: Covid 19 outbreak - arguing about Stats (again)

Postby horizon » 22 Nov 2020, 8:37pm

Jdsk wrote:
horizon wrote: The other people might have the virus without knowing and are therefore possibly infectious.


Which "other people"?

Jonathan


Everyone else wandering around who doesn't have symptoms but who hasn't been tested i.e. most people. But they might be positive (without knowing it) - hence masks and SD. The question then is, even if they are positive but provided they are symptomless (i.e. asymptomatic), does it matter? Yes, according to current policy, no, according to some people.

BTW I had to think again to clarlfy to myself that asymptomatic means positive but with no symptoms. It doesn't mean someone who is virus free - that is simply, well, virus free - time to check this I think!
When the pestilence strikes from the East, go far and breathe the cold air deeply. Ignore the sage, stay not indoors. Ho Ri Zon 12th Century Chinese philosopher

Jdsk
Posts: 3921
Joined: 5 Mar 2019, 5:42pm

Re: Covid 19 outbreak - arguing about Stats (again)

Postby Jdsk » 22 Nov 2020, 8:40pm

horizon wrote:BTW I had to think again to clarlfy to myself that asymptomatic means positive but with no symptoms. It doesn't mean someone who is virus free - that is simply, well, virus free - time to check this I think!

That isn't right. Asymptomatic if unqualified simply means having no symptoms. Asymptomatic carrier or asymptomatic infected or asymptomatic positive mean what they say. Asymptomatic cases requires definition, but will often mean with symptoms or a positive test.

Jonathan

User avatar
horizon
Posts: 10246
Joined: 9 Jan 2007, 11:24am
Location: Cornwall

Re: Covid 19 outbreak - arguing about Stats (again)

Postby horizon » 22 Nov 2020, 8:57pm

asymptomatic
/eɪˌsɪmptəˈmatɪk/
Learn to pronounce
adjectiveMedicine
adjective: asymptomatic

(of a condition or a person) producing or showing no symptoms.
"infection is usually asymptomatic"

Translate asymptomatic to
Use over time for: asymptomatic
Definitions from Oxford Languages


I took it from the above but I agree with you that asymptomatic should mean just that unless qualified - I'll be listening out for careless use now!
When the pestilence strikes from the East, go far and breathe the cold air deeply. Ignore the sage, stay not indoors. Ho Ri Zon 12th Century Chinese philosopher

Psamathe
Posts: 11581
Joined: 10 Jan 2014, 8:56pm

Re: Covid 19 outbreak - arguing about Stats (again)

Postby Psamathe » 22 Nov 2020, 11:03pm

Jdsk wrote:
horizon wrote:BTW I had to think again to clarlfy to myself that asymptomatic means positive but with no symptoms. It doesn't mean someone who is virus free - that is simply, well, virus free - time to check this I think!

That isn't right. Asymptomatic if unqualified simply means having no symptoms. Asymptomatic carrier or asymptomatic infected or asymptomatic positive mean what they say. Asymptomatic cases requires definition, but will often mean with symptoms or a positive test.

Jonathan

I (maybe incorrectly) have assumed that "asymptomatic" is a question of degree and personal opinion. An infected person coughs a couple of times and some will declare "they've got C-19" and others will dismiss it and ignore it. I lost my sense of smell months ago (apparently) and never noticed until another condition had me doing a formal SIT test and I got lowest possible score (nothing to do with C-19). So I think of a continuum on top of which different people self-diagnose differently and recognise symptoms differently.

And on-top of that, US recognised "symptoms" different from the Zoe lis of symptoms and the UK list seems very short. And the UK official symptom list has changed sing the early days so people who were "asymptomatic" back in March might now be retrospectively symptomatic...

Ian

Jdsk
Posts: 3921
Joined: 5 Mar 2019, 5:42pm

Re: Covid 19 outbreak - arguing about Stats (again)

Postby Jdsk » 22 Nov 2020, 11:14pm

Yes, you have to be careful with definitions whether the purpose is clinical diagnosis, population surveillance or trials of interventions.

For example here's the WHO's for public health surveillance:
https://www.who.int/publications/i/item/WHO-2019-nCoV-Surveillance_Case_Definition-2020.1

And here's the current UK governmental advice to clinicians, which includes a definition:
https://www.gov.uk/government/publications/wuhan-novel-coronavirus-initial-investigation-of-possible-cases/investigation-and-initial-clinical-management-of-possible-cases-of-wuhan-novel-coronavirus-wn-cov-infection

Jonathan

mjal
Posts: 27
Joined: 4 Jul 2011, 2:22pm

Re: Covid 19 outbreak - arguing about Stats (again)

Postby mjal » 22 Nov 2020, 11:24pm

horizon wrote:A bit more Googling, but I haven't read it in full (and never could!):

https://www.nature.com/articles/s41467-020-19802-w

There were no positive tests amongst 1,174 close contacts of asymptomatic cases.


Whenever I see a result of 0% (or 100%) with reference to a biological system I am immediately suspicious.

Here is a quote from the comments below the Nature paper :

""There were no positive tests amongst 1,174 close contacts of asymptomatic cases."

How is this possible? Even if all those contacts were indeed uninfected, where are the false positive tests? Even assuming just a 1% false positive rate, the chance of getting zero positives out of 1,174 is less than 1 in 100,000. In the "source data" section, there's an even more improbable claim: zero positive results out of 85,884 tests conducted in East Lake Scenic Area of Wuhan.

Isn't this clear evidence that the data being used is unreliable?"

The false positive rate for PCR tests in SARS-COV-2 is a subject of some contention ; some virologists maintain that there is a nil rate in the strict sense but do admit that e.g. contamination, clerical error etc can introduce false positives into the data base. I have seen a figure of 0.5% offered.

Now, the Nature article seems to make no mention of "false positives". Are the researchers using different criteria for positive/negative? The obvious one is the number of "cycles" for which the test runs ; <37 was positive, >40 was negative, in between attracted a retest. The cycle figures seem to be the accepted standard. Another factor may be the "target genes", in this case 2 in number. Many tests use just 2 but some use 3.

Even if the percentage of false positives was barely above zero, there should still be a substantial number in almost 10 million tests. Is the testing procedure so good that false positives have essentially been eliminated or is there some other explanation?

My usual ending : Covid-19 : we just don't know...yet

Jdsk
Posts: 3921
Joined: 5 Mar 2019, 5:42pm

Re: Covid 19 outbreak - arguing about Stats (again)

Postby Jdsk » 22 Nov 2020, 11:59pm

mjal wrote:How is this possible? Even if all those contacts were indeed uninfected, where are the false positive tests? Even assuming just a 1% false positive rate, the chance of getting zero positives out of 1,174 is less than 1 in 100,000.

The FPR in the UK during the summer was < 0.05% because the total positive rate was about that.

Scroll to 8' 21":


Jonathan

mjal
Posts: 27
Joined: 4 Jul 2011, 2:22pm

Re: Covid 19 outbreak - arguing about Stats (again)

Postby mjal » 23 Nov 2020, 12:32am

Quick reply (at 0030) :

"there's an even more improbable claim: zero positive results out of 85,884 tests conducted in East Lake Scenic Area of Wuhan."

This is a "quote from the quote" attached to the Nature paper.

Any thoughts?