False positive tests increase when other coronavirus are present in the sample

[Yew]

It seems that the German medical accreditation body has found that PCR tests have more false positives when other coronavirus are present, so today's measures could be literally down to people getting colds.

The false positive rate rose from 1.4% to 2.2%-7.6% when a Hcov was present (alpha and betacoronaviruses) This suggests that PCR tests are detecting Hcov infections in the same way serology tests are. Conducted by the accreditation body of Germany

https://twitter.com/ImDarrenDunn2/status/1308169854575349762

The original study:

Bericht zur Coronalage 28.05.2020

Die neuesten Zahlen bestätigen ziemlich eindeutig die These, dass wir es auch dieses Jahr mit einer Winterepidemie zu tun hatten, die im April ausläuft. In einigen Ländern hat sie mehr Opfer gefordert als die Influenza 2018, in anderen weniger, so wie es auch 2018 Länder gab mit seinerzeit...

www.achgut.com

 

[MikeWM]

And more people will be getting colds right now, because of the return of schools etc.

We're chasing a phantom, and tearing down the whole of our society in our fruitless attempts to catch it.

 

[Yew]

Admittedly, the study is in german, but I've read nothing about scientists and analysts even trying to consider this factor.

 

[birchesgreen]

To be honest i generally have a continuous series of colds from now until the spring and beyond. If this sets off the covid test i could end up having to isolate for ever.

 

[trebor79]

Wow. False positive 1.4%? I thought it was about 0.5%. So testing 250,000 a day, even if nobody has a cold that's still 3,500 false positives...

 

[Yew]

Wow. False positive 1.4%? I thought it was about 0.5%. So testing 250,000 a day, even if nobody has a cold that's still 3,500 false positives...

And yet we trust our future and our liberties to these tests.

 

[adc82140]

Wow. False positive 1.4%? I thought it was about 0.5%. So testing 250,000 a day, even if nobody has a cold that's still 3,500 false positives...

Not quite. The rate is a total based on positives returned, not tests done. So at 1.4%, based on today's figures the false positives will be about 60 cases.

 

[trebor79]

Not quite. The rate is a total based on positives returned, not tests done. So at 1.4%, based on today's figures the false positives will be about 60 cases.

No that's not right.
False positives don't change dependant upon how many real positives there are.

False positives is a measure of the proportion of tests that you carry out on uninfected people will return a positive.

That's why a few weeks ago when there were around 1,200 daily positives, people were pointing out a third of them would be false positives.

 

[big_rig]

No that's not right.
False positives don't change dependant upon how many real positives there are.

False positives is a measure of the proportion of tests that you carry out on uninfected people will return a positive.

That's why a few weeks ago when there were around 1,200 daily positives, people were pointing out a third of them would be false positives.

Yes, quite. There is a SAGE paper on this from June which considered this (my bolding of key points). Their minds do not seem to have changed since. There are ongoing references to this problem in their meeting minutes as well. If you add this German study to it then the idea of using these tests on groups of people 98-99.9% of whom do not have the virus (e.g people who present for tests, or the population as a whole) seems mad - file it with all the other things which are being done or have been suggested which sound like 'common sense' which seem nearly useless or in fact counterproductive once you think about it..

They also reference it in their paper on mass testing as well where my reading of the paper is basically 'if you must do this, then every positive test needs to be immediately followed up with a better one, and we also need to figure out how to sell that to the public.' They are too polite to say it might be a waste of time and money/a solution looking for a problem but that was also the direction I took from it.

What is the UK operational false positive rate?

The UK operational false positive rate is unknown. There are no published studies on the operational false positive rate of any national COVID-19 testing programme. An attempt has been made to estimate the likely false-positive rate of national COVID-19 testing programmes by examining data from published external quality assessments (EQAs) for RT-PCR assays for other RNA viruses carried out between 2004-2019 [7]. Results of 43 EQAs were examined, giving a median false positive rate of 2.3% (interquartile range 0.8-4.0%).

Why are false positives a problem?

DHSC figures [3] show that 100,664 tests were carried out on 31 May 2020 (Pillar 1 and 2 RT-PCR tests). 1,570 of those tests were positive for SARS-CoV-2 (1.6%). The majority of people tested on that day did not have SARS-CoV-2 (98.4% of tests are negative). When only a small proportion of people being tested have the virus, the operational false positive rate becomes very important. Clearly the false positive rate cannot exceed 1.6% on that day, and is likely to be much lower. If the operational false positive rate was 0.4%, 400 of the 1,570 positive tests would be false positives. That would represent 400 people being isolated when they are well, and much wasted effort in contact tracing. It is possible that a proportion of infections that we currently view as asymptomatic may in fact be due to these false positives. Unless we understand the operational false positive rate of the UK’s RT-PCR testing system we risk overestimating the COVID-19 incidence, the demand on track and trace, and the extent of asymptomatic infection

https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/895843/S0519_Impact_of_false_positives_and_negatives.pdf

Under mass testing, a larger proportion of positive results will be false positive than in symptomatic testing, even when using the same test, as infection prevalence is much lower in asymptomatic populations. The response to positive tests will therefore require careful consideration, including (i) whether rapid follow-up confirmatory testing is used to avoid prolonged isolation of large numbers of false positives, (ii) whether individual isolation requires household quarantine, (iii) how to communicate to the public the nature of mass testing and lower-confidence test results to avoid potential undermining of public perception and confidence in testing (iv) the possible impacts on individuals (loss of earnings) and groups (closure ofschools), and (v) how to define outbreaks when including asymptomatic tests.

https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/914931/s0712-tfms-consensus-statement-sage.pdf

 

[Richard Scott]

And more people will be getting colds right now, because of the return of schools etc.

We're chasing a phantom, and tearing down the whole of our society in our fruitless attempts to catch it.

It's like an episode of Scooby-Doo, just it isn't funny!

 

[Stephen42]

It seems that the German medical accreditation body has found that PCR tests have more false positives when other coronavirus are present, so today's measures could be literally down to people getting colds.

https://twitter.com/ImDarrenDunn2/status/1308169854575349762

The original study:

Bericht zur Coronalage 28.05.2020

Die neuesten Zahlen bestätigen ziemlich eindeutig die These, dass wir es auch dieses Jahr mit einer Winterepidemie zu tun hatten, die im April ausläuft. In einigen Ländern hat sie mehr Opfer gefordert als die Influenza 2018, in anderen weniger, so wie es auch 2018 Länder gab mit seinerzeit...

www.achgut.com

The tweet significantly misrepresents the the underlying report which is about the accreditation body's "Extra External Quality Assessment Scheme". It's a test of individual labs on cultivated samples rather than attempting to calculate operational false positive rates. The actual operational false positive rate is unknown, but the UK rate is unlikely to be anywhere near that high.

Firstly the 7.6% number is due to sample mixups from a small number of labs between a covid positive sample and a negative one. The updated report has figures excluding these labs for those samples is 98.1% correct so down to 1.9% not reported as negative - already a much smaller variation. Also the incorrect results aren't all positives, other failures include unreported/void tests and amounts above the detectable threshold, but below the positive. Actual positives on the unmixed samples were HCoV OC43 0.8% (8 positives reported) and no coronavirus 0.7% (7 reported). That variation could easily be by chance with such small numbers.

Secondly even the revised numbers above are unlikely to be accurate, both that the small number of false positives mean the actual rate could vary and it ignores any changes the labs reporting the false positives have made to improve. The labs were across 36 countries and it's unclear how the UK labs perform relatively against them. The ONS infection survey which is mass testing on a sample of the population last reported a modelled rate of 0.11% for testing covid positive, that suggests a very low operational false positive rate.

This isn't to say false positives aren't important, particularly when the prevalance rate is low and doing widespread community testing. Similarly false negatives are an issue when the prevalance rate increases. It seems unlikely that the daily cases have a large proportion of false positives because the ONS tests are done through the national testing programme in a subset of the labs used for other tests, even if you assume a variation by a factor of 2 due to the presence of other coronaviruses.

 

[big_rig]

Something about false positives I would love to know the answer to is how New Zealand tested approx 300,000 people between May and August when they had ‘eliminated’ the virus, yet recorded seemingly no false positives amongst their population (outside of potentials in cases identified in their quarantine hotels, but barely 100 people in those have had it). If the SAGE papers are right they should have had a thousand or so false positives but searching their news sites/Ministry of Health only references one such case identified in August.

 

[Yew]

@Stephen42 a remarkably odd place for a first post, but welcome nonetheless!

 

[MikeWM]

Something about false positives I would love to know the answer to is how New Zealand tested approx 300,000 people between May and August when they had ‘eliminated’ the virus, yet recorded seemingly no false positives amongst their population (outside of potentials in cases identified in their quarantine hotels, but barely 100 people in those have had it). If the SAGE papers are right they should have had a thousand or so false positives but searching their news sites/Ministry of Health only references one such case identified in August.

Do we know how many cycles they are using for the test? One of the issues here is that we appear to be using too many - and the more you do, the more false positives you pick up.

 

[talldave]

Do we know how many cycles they are using for the test? One of the issues here is that we appear to be using too many - and the more you do, the more false positives you pick up.

I've seen mention of 40 and 45 in articles I've read, but no idea whether those were facts or speculation.

 

[big_rig]

I've seen mention of 40 and 45 in articles I've read, but no idea whether those were facts or speculation.

This article today says 25, so perhaps that explains it. It also describes it in a bit of a different way, as if you can stop the process after a certain number of cycles and have a look at the results then (I don’t know if this is how it actually works or not) - perhaps they do that at 25 and make a call whereas we just set the machine to 45 and check the result then. Probably not how it works at all mind!

I still don’t know if this cracks it though because their bloke is talking about people who had the virus a long time ago being false current positives, instead of people who do not and have never had it being false positives.

Talking on Radio New Zealand on Wednesday, Bloomfield said that early in an infection the Ct value was lower, under about 25.

In other words, 25 or fewer amplification cycles would be needed for virus genetic material to be detected in those samples.

Around 30 “you start to wonder”, Bloomfield went on to say. “If it’s over 35 then you’re pretty confident this is an old infection.”

Bloomfield said earlier that international research had shown people with historical infections, weak positive test results and a high Ct value were not infectious.

When is a Covid-19 test result a weak positive, why more of them are being found, and how the cases are dealt with

Covid-19 weak positive results and why they have caught the attention of the Ministry of Health.

i.stuff.co.nz

 

[Trackman]

I don't know about cycles or weak positives, but my sister was contacted by track and trace and took a test - that came back a 'don't know' maybe it was a 50/50.
She took another one the day after the result and that came back negative.
What's all that about?

 

[Crossover]

I don't know about cycles or weak positives, but my sister was contacted by track and trace and took a test - that came back a 'don't know' maybe it was a 50/50.
She took another one the day after the result and that came back negative.
What's all that about?

A colleague had two inconclusive tests and gave up so stayed at home for two weeks. A family member of theirs who had similar symptoms tested negative, however

 

[Bantamzen]

I don't know about cycles or weak positives, but my sister was contacted by track and trace and took a test - that came back a 'don't know' maybe it was a 50/50.
She took another one the day after the result and that came back negative.
What's all that about?

Hold on? There's potentially a 'dunno mate' result? I've heard it all now.....