Jump to content
DOSBODS
  • Welcome to DOSBODS

     

    DOSBODS is free of any advertising.

    Ads are annoying, and - increasingly - advertising companies limit free speech online. DOSBODS Forums are completely free to use. Please create a free account to be able to access all the features of the DOSBODS community. It only takes 20 seconds!

     

False Positives and what it actually means


Recommended Posts

So - false positives ....

Some say (JHB for example when interviewing Matt Hancock) that most(?) of the current positive cases are down to false positives.  Doing the maths on that kind of rings true with c1%.

But surely the gov would have explained this away and not let this become a 'fact' if it was bollocks???

Independent article about it doesn't go into much detail apart from saying that JHB is wrong - but doesn't explain why and what is therefore actually happening.

https://www.independent.co.uk/news/uk/home-news/coronavirus-false-positives-testing-covid-19-test-b550133.html

Yours

Confused of Tunbridge Wells.

 

Link to post
Share on other sites
2 minutes ago, NewryH said:

But surely the gov would have explained this away and not let this become a 'fact' if it was bollocks???

 

 

It works both ways. If there is a 100% reliable trustworthy test, it would most likely show that only a tiny number of people have the 'Rona. But with a slightly unreliable test, the number of those appearing to have the Rona is multiplied, and since the test can't be 100% sure, the government can use that uncertainty as a justification for treating all 'possibles' as 'positives' and force them to obey the rules on the precautionary principle.

It's a bit like speed cameras; not all of them work but as you don't know which ones work and which ones don't, you treat them all as if they do.

Link to post
Share on other sites

False positives occur in every medical diagnostic test. It is one of the reasons why all good diagnostic algorithms require at least one confirmatory test using a completely different technology/principle. Because if the confirmatory test works on a different principle the chances of someone being FP in both is vanishingly small. Unless of course the test(s) have a very high rate of FPs.... its all maths. Very good diagnostic algorithms require 2 confirmatory tests each working on different principles, to make the maths even better in favour of eliminating FPs.

So an antibody test would require confirmation with say a PCR and/or Western blot.

Of course, if all you have is one shitty test method, this is all a bit moot....

At the simplest level, false positives occur because most molecules look a bit like other molecules - a tabby cat looks a bit like a Scottish wild cat and how easily you can tell them apart depends on the accuracy of you 'kitty diagnostic test'.

 

PCRs and rapid tests, ie what everyone seems to be putting their faith in for Rona, are notorious for both high FP and FN rates....

Edited by Melchett
Link to post
Share on other sites
3 hours ago, NewryH said:

So - false positives ....

Some say (JHB for example when interviewing Matt Hancock) that most(?) of the current positive cases are down to false positives.  Doing the maths on that kind of rings true with c1%.

But surely the gov would have explained this away and not let this become a 'fact' if it was bollocks???

Independent article about it doesn't go into much detail apart from saying that JHB is wrong - but doesn't explain why and what is therefore actually happening.

https://www.independent.co.uk/news/uk/home-news/coronavirus-false-positives-testing-covid-19-test-b550133.html

Yours

Confused of Tunbridge Wells.

 

The Huff Post tries to debunk the 90% claim here https://m.huffingtonpost.co.uk/entry/false-positives-coronavirus_uk_5f686da4c5b6de79b677e909?guccounter=1&guce_referrer=aHR0cHM6Ly93d3cuZ29vZ2xlLmNvbS8&guce_referrer_sig=AQAAADarmMjIPtHqle8SOVk2iebrHC9pgVoAncDGxmEN2kS4qk0Bwsr6JTZPXQKfADrbjSJWMWnp0JZEOG2xF-asLrSrwn_Q4WSZqMGOE4hVkgtY6Sd4TKdVf4A7fMBH3E2eY8dizUGbNozN9vwYGA7IU_N299oa7DnmQ3xjyTjch1Gt

“Imagine 1,000 people turn up to the testing centre, and only one person has Covid. That one person has a positive test. 

“Of the remaining 999 people, if the FPR is 1%, then you’ll get another nine positive tests from these 999

“So now you’ve got 10 positive tests, but only one of them has covid, so 90% of the positive tests don’t actually have covid.”

Both Yeadon and Heneghan, and in turn Hartley-Brewer, Toby Young and John Redwood, make one huge assumption – that the prevalence of coronavirus in the population tested is 0.11% like the ONS has said.

But this is not representative of the population that is actually being tested and whose results make up the material presented by the government and scientists of evidence of a second wave.

The ONS figure is based on a weekly survey of households representative of the UK as a whole, while the evidence of a second wave is based on tests on people who have sought one out.

Watson told HuffPost UK: “If you took the UK population as a whole and randomly picked one person out of it, the probability of them having Covid is actually very low at it has a reasonably low prevalence. 

“But if you turn up to a testing centre you’re already thinking: ‘I might have Covid’ and if you turn up with a cough and a fever then it’s probably quite a high probability that you have Covid.”

Norman Fenton counters this argument in an article I will try and find.

 

Link to post
Share on other sites
1 hour ago, Melchett said:

False positives occur in every medical diagnostic test. It is one of the reasons why all good diagnostic algorithms require at least one confirmatory test using a completely different technology/principle. Because if the confirmatory test works on a different principle the chances of someone being FP in both is vanishingly small. Unless of course the test(s) have a very high rate of FPs.... its all maths. Very good diagnostic algorithms require 2 confirmatory tests each working on different principles, to make the maths even better in favour of eliminating FPs.

So an antibody test would require confirmation with say a PCR and/or Western blot.

Of course, if all you have is one shitty test method, this is all a bit moot....

At the simplest level, false positives occur because most molecules look a bit like other molecules - a tabby cat looks a bit like a Scottish wild cat and how easily you can tell them apart depends on the accuracy of you 'kitty diagnostic test'.

I've looked (superficially) but not been able to find the primer sequences. I assume they are proprietary, and so I assume we don't know either the primer sequences (to do some fairly basic alignments to check the assurances in the kits' descriptions about primer specificity, or which bit of the SarsCov2 sequence is amplified. That's a bit unfortunate as it would enable some scrutiny.

The false positive rate. I haven't been following the testing results or where the testing is happening. Assuming the testing is countrywide and that the false positive rate is standard, then any difference in numbers between for example, regions, should reflect a real change in cases, with the big assumptions that there either isn't a huge variation in operator error across test centres, or that different test kits with different error rates are being used in different places.

Edited by Hopeful
Link to post
Share on other sites
1 hour ago, Gloommonger said:

The Huff Post tries to debunk the 90% claim here https://m.huffingtonpost.co.uk/entry/false-positives-coronavirus_uk_5f686da4c5b6de79b677e909?guccounter=1&guce_referrer=aHR0cHM6Ly93d3cuZ29vZ2xlLmNvbS8&guce_referrer_sig=AQAAADarmMjIPtHqle8SOVk2iebrHC9pgVoAncDGxmEN2kS4qk0Bwsr6JTZPXQKfADrbjSJWMWnp0JZEOG2xF-asLrSrwn_Q4WSZqMGOE4hVkgtY6Sd4TKdVf4A7fMBH3E2eY8dizUGbNozN9vwYGA7IU_N299oa7DnmQ3xjyTjch1Gt

“Imagine 1,000 people turn up to the testing centre, and only one person has Covid. That one person has a positive test. 

“Of the remaining 999 people, if the FPR is 1%, then you’ll get another nine positive tests from these 999

“So now you’ve got 10 positive tests, but only one of them has covid, so 90% of the positive tests don’t actually have covid.”

Both Yeadon and Heneghan, and in turn Hartley-Brewer, Toby Young and John Redwood, make one huge assumption – that the prevalence of coronavirus in the population tested is 0.11% like the ONS has said.

But this is not representative of the population that is actually being tested and whose results make up the material presented by the government and scientists of evidence of a second wave.

The ONS figure is based on a weekly survey of households representative of the UK as a whole, while the evidence of a second wave is based on tests on people who have sought one out.

Watson told HuffPost UK: “If you took the UK population as a whole and randomly picked one person out of it, the probability of them having Covid is actually very low at it has a reasonably low prevalence. 

“But if you turn up to a testing centre you’re already thinking: ‘I might have Covid’ and if you turn up with a cough and a fever then it’s probably quite a high probability that you have Covid.”

Norman Fenton counters this argument in an article I will try and find.

 

To me that sounds like an argument that the real infection rate is lower than we are being led to believe, ie backing up Yeadon and Hartley- Brewer. The people turning up for testing are self selecting as they think they might be ill, so the percentage of positives will be higher than in the general population.

Link to post
Share on other sites
55 minutes ago, Mirror Mirror said:

To me that sounds like an argument that the real infection rate is lower than we are being led to believe, ie backing up Yeadon and Hartley- Brewer. The people turning up for testing are self selecting as they think they might be ill, so the percentage of positives will be higher than in the general population.

touché

Link to post
Share on other sites

Why is anyone interested in people who aren’t ill?

The testing isn’t reliable, so why waste resources doing it on people who don’t need it?

Shield the vulnerable, using the testing to help with that, ie testing care staff and the like, and test people who exhibit symptoms and their contacts, no-one else. 

This could have been mostly over by now. 

Link to post
Share on other sites
3 hours ago, Mirror Mirror said:

To me that sounds like an argument that the real infection rate is lower than we are being led to believe, ie backing up Yeadon and Hartley- Brewer. The people turning up for testing are self selecting as they think they might be ill, so the percentage of positives will be higher than in the general population.

Interestingly in Victoria, as the Comrade-general in charge has said that lockdown only ends when numbers are below 5 a day, the volume of people volunteering for testing has been dropping.  Anecdotally I know of people who have had a cough but have not gone to get tested.

Why?  Because they think that if they do have it, it's probably not that bad and they can just self isolate for 2 weeks without anyone knowing.  

They'd rather not contribute to the numbers going up again and the lockdown extending.

 

I could have predicted this would happen if the government started talking about targets re lockdown.  Humans are complex and self interested.

Link to post
Share on other sites

You all need to google "false positive paradox". It is common to all medical testing. You cant just calculate test efficiency using total tested and multiplying by false positive rate, and then claim the testing is swamped by false positives numbers and therefore its rubbish, like the video posted on here. These people just dont understand the statistical maths behind the testing interpretation, they are making misleading claims which is dangerous and misleading people to beleive the testing results are rubbish.

 

Link to post
Share on other sites
19 hours ago, NewryH said:

So - false positives ....

Some say (JHB for example when interviewing Matt Hancock) that most(?) of the current positive cases are down to false positives.  Doing the maths on that kind of rings true with c1%.

But surely the gov would have explained this away and not let this become a 'fact' if it was bollocks???

Independent article about it doesn't go into much detail apart from saying that JHB is wrong - but doesn't explain why and what is therefore actually happening.

https://www.independent.co.uk/news/uk/home-news/coronavirus-false-positives-testing-covid-19-test-b550133.html

Yours

Confused of Tunbridge Wells.

 

I would direct you to google "false positive paradox" to see why JHB is embarrasingly wrong she doesn't understand the maffs. 

 

Link to post
Share on other sites

There's a few reasons for false positives.

1. The mathematical false positives from no test being 100% accurate.  https://www.mathsisfun.com/data/probability-false-negatives-positives.html has a simple explanation (I find the tree diagram the easiest way to visualise it).

2. The detection of positives where an individual is no longer contagious (i.e. the virus is no longer 'alive' within their body).  These come about as PCR tests can detect dead virus RNA strands in the blood for up to 90 days after infection.

3. Running too many cycles.  This is closely related to 2 above.  Every time you run a PCR cycle for a single test specimen, the amount of RNA present doubles.  20 cycles may result in showing up an active virus while 40 cycles may result in a single fragment of RNA becoming a massive amount.  As such it's important to calibrate your testing by establishing a reasonable number of cycles for each run.  NHS don't disclose how many cycles they are using. 

Basic explanation of PCR tests here ... https://www.sciencelearn.org.nz/resources/2347-what-is-pcr 

All this shows why PCR tests on their own are not a good indicator of how wide the virus is circulating in the community.  In fact the inventor of the PCR test says it should not be used to make a diagnosis.

Link to post
Share on other sites
46 minutes ago, mosstrooper said:

You all need to google "false positive paradox". It is common to all medical testing. You cant just calculate test efficiency using total tested and multiplying by false positive rate, and then claim the testing is swamped by false positives numbers and therefore its rubbish, like the video posted on here. These people just dont understand the statistical maths behind the testing interpretation, they are making misleading claims which is dangerous and misleading people to beleive the testing results are rubbish.

 

The paradox is that when the virus is at low prevalence in the community (as Covid is now), the test gives more false positives than actual positives.

Personally, I think the increasing 'cases' (as shown in the press) are a combination of increased testing, false positives and an actual increase in the number of infections.  Looking at the number of infections per 100,000 tests gives a much clearer view of what is going on.  See graph below which I haven't updated for a couple of days.

  image.png.4613b5b98f6d5d7184efd848b6d17558.png

Edited by PaulParanoia
Link to post
Share on other sites
42 minutes ago, mosstrooper said:

You all need to google "false positive paradox". It is common to all medical testing. You cant just calculate test efficiency using total tested and multiplying by false positive rate, and then claim the testing is swamped by false positives numbers and therefore its rubbish, like the video posted on here. These people just dont understand the statistical maths behind the testing interpretation, they are making misleading claims which is dangerous and misleading people to beleive the testing results are rubbish.

 

Matt Hancock, is that you?

Ive just done as you suggest, and what it says concurs exactly with what @PaulParanoia, I and others are saying here, and what JHB, Norman Fenton and others are saying on the vids which have been linked here.

Quote

...how surprised people are by the false positive paradox, situations where there are more false positive test results than true positives. For example, it might be that of 1,000 people tested for an infection, 50 of them test positive for having it, but that is due to 10 truly having it and 40 mistaken test results, because only 10 people of those tested actually have the infection but the test sometimes gives false results. The probability of a positive test result is determined not only by the accuracy of the test but by the characteristics of the sampled population.[2] When the prevalence, the proportion of those who have a given condition, is lower than the test's false positiverate, even tests that have a very low chance of giving a false positive in an individual case will give more false than true positives overall.[3]The paradox surprises most people.[4]

 

 

Edited by Mirror Mirror
Link to post
Share on other sites

There are two problems here, the false positive problem, and the sampled population problem.

1.  False positive problem.

To illustrate, consider 100,000 tests, 0.1% incidence rate in the population and 1% false positive rate.

Then you'd have 100 people with the disease, but 1,000 people with a positive result who don't have the disease.

This is a real result -- the test really is suggesting that there's 11x the infection rate than reality.

I believe this is the JHB suggestion.

But

2. The sampled population problem

The tests aren't taken from people in the general population -- they test people 'likely to have the disease' -- either through symptoms or reasonable contact with someone known to be infected.  Taking this into account, the actual numbers might be:

100,000 tests, 2.5% incidence rate in the sampled population, 1% false positive rate.

Now, for the tested population, you have 2,500 with the disease and only 1,000 false positives.  Thus the false positives are now fewer in number than the ones who really have the disease.

I believe this is the response to JHB.

Link to post
Share on other sites
33 minutes ago, dgul said:

2. The sampled population problem

The tests aren't taken from people in the general population -- they test people 'likely to have the disease' -- either through symptoms or reasonable contact with someone known to be infected.  Taking this into account, the actual numbers might be:

100,000 tests, 2.5% incidence rate in the sampled population, 1% false positive rate.

Now, for the tested population, you have 2,500 with the disease and only 1,000 false positives.  Thus the false positives are now fewer in number than the ones who really have the disease.

I believe this is the response to JHB.

As @Mirror Mirror pointed out earlier in the thread.  If the test subjects are self selecting, the sampled population no longer reflects the general population.  Therefore, results can not be used to imply the level of CV within said general population.

Either way you look at it, Pillar 2 PCR community testing results need to be treated with extreme caution when using them to gauge the level of CV in the general population and should not be used as the basis of government scare tactics or policy. 

Edited by PaulParanoia
Link to post
Share on other sites

The only numbers of any interest or value IMO are the hospitalisations and the number of ICU beds taken up with Covid patients. The fact that these are rising a bit would suggest that the case numbers are not all false.

How old they are would also be of interest to me, but in this hysterical world no-one can be allowed to die, even if they're 95 with dementia and can't recognise anyone.

Link to post
Share on other sites
2 minutes ago, swiss_democracy_for_all said:

The only numbers of any interest or value IMO are the hospitalisations and the number of ICU beds taken up with Covid patients. The fact that these are rising a bit would suggest that the case numbers are not all false.

How old they are would also be of interest to me, but in this hysterical world no-one can be allowed to die, even if they're 95 with dementia and can't recognise anyone.

Old people not allowed to die.

The rest of us not allowed to live.

Link to post
Share on other sites
23 minutes ago, swiss_democracy_for_all said:

The only numbers of any interest or value IMO are the hospitalisations and the number of ICU beds taken up with Covid patients. The fact that these are rising a bit would suggest that the case numbers are not all false.

How old they are would also be of interest to me, but in this hysterical world no-one can be allowed to die, even if they're 95 with dementia and can't recognise anyone.

Info here:

https://www.dailymail.co.uk/news/article-8790645/Boris-Johnson-warns-costly-NATIONAL-lockdown-ruled-out.html

Link to post
Share on other sites
52 minutes ago, JoeDavola said:

Old people not allowed to die.

The rest of us not allowed to live.

No, you really have it wrong Joe

Old people are allowed to die, in fact they are helped to do so.

Young people are helped to die too, from otherwise preventable illnesses.

It's simply that nobody is allowed to go to hospital.

 

 

Edited by Hopeful
Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

  • Similar Content

  • Recently Browsing   0 members

    No registered users viewing this page.

×
×
  • Create New...