PCR False-Positive-Rate is likely between 2 - 7%.

[nopos]

Results of External Quality Assessment (EQA) studies evaluating PCR false-positive-rates (FPR) under ordinary lab conditions generally show FPR of 2-7%:
https://i.imgur.com/vInuYsb.png
https://doi.org/10.1101/2020.04.26.20080911

PCR tests are approved with a challenge test under optimal lab settings where most sources of false positives are eliminated, giving easily 0% false positives (100% specificity). Probably not using any same bottle, tool or personnel between subsequent analysis etc. This is not applicable to ordinary lab conditions because sampling environment, sample handling, transport etc are the main sources of cross-contamination and errors.

EQA prepared vs field-samples
Even EQA studies using prepared samples instead of field-samples compared to a gold standard test will circumvent potentially many or even most sources of false positives.

What creates false positives ?

”the likeliest source of these false positives is sample contamination or human error. Samples can be contaminated by a positive sample analyzed at the same time (cross-contamination), or more likely by target genes amplified from prior positive samples or positive controls (carryover contamination). False positives can also be produced by sample mix-ups or data entry errors.”

[Diagnosing COVID-19 infection: the danger of over-reliance on positive test results] https://doi.org/10.1101/2020.04.26.20080911


”Technical problems including contamination during sampling (eg, a swab accidentally touches a contaminated glove or surface), contamination by PCR amplicons, contamination of reagents, sample cross-contamination, and cross-reactions with other viruses or genetic material could also be responsible for false-positive results. 2 These problems are not only theoretical; the US Center for Disease Control and Prevention had to withdraw testing kits in March, 2020, when they were shown to have a high rate of false-positives due to reagent contamination.”

[False-positive COVID-19 results; hidden problems and costs]
https://doi.org/10.1016/S2213-2600(20)30453-7

[MikeAustin]

Thanks.

Although familiar with some of these reports, you have reminded me that the figure that I usually quote is the percentage of false positives among all positives, around 90%-95%. This is not actually the false positive rate, which is defined the percentage of false positives among all negatives - the 2% to 7% that you quote

However, I would suggest that my percentage is a more direct way of presenting the inadequacy of the testing programme as it deals specifically with the results that are being used by the government to such disastrous effect.

[MikeAustin]

nopos
I saw a post from you in which you said that the name of the percentage that I quoted was "False Discovery Rate." Strangely, that post has disappeared.
Anyway, I have updated the terminology on my calculator to suit:
https://drive.google.com/file/d/1f7X5sJ ... sp=sharing
This duplicates the BMJ calculator here:
https://sandpit.bmj.com/graphics/2020/c19test/
but allows for decimal places and incorporates a 'helper' to get true prevalence from measured prevalence (cases/tests).

[nopos]

I removed it. Just wanted to check if my understanding is how it's normally used. I think it is. Posting it again:

percentage of false positives among all positives

That would be False Discovery Rate (FDR).

FDR will vary by two factors: True Positivity and FPR.

FPR only relies on the accuracy of sampling and analysis.

https://i.imgur.com/H8iqfg3.png
https://www.medrxiv.org/content/10.1101 ... 1.full.pdf

[jmc]

I just read the paper and its obvious that none of the authors have a solid understanding of the mathematics involved. They use very heterogeneous data sources often with very low n value samples and try to draw conclusions from them. Plus comparing national testing results that often use very different testing protocols and with cumulative data set sample ranges of two or three orders of magnitude or more. etc. etc.

They treat false positives in isolation rather than as a part of a mass testing regime.

When you test many 10K's of people per day and the test target prevalence is about one order of magnitude less than the positive result error rate it does not matter too much what the actual cause of the false positive is, contamination etc. Be it 2% or 5%. Until the prevalence rate is much greater than the positive result error rate all positives are tentatively provisional. Which retesting only reduces, never eliminates.

Then when we add in the false negative rate due to great range of test sensitivity depending on which phase an actual active infection might be in, the number that actually matters is what percentage of those with active infections are successfully detected. And given that a 50% false negative rate for each testing round seem a reasonable estimate that would leave use with the mathematics telling use that mass testing using RT/PCT does not successfully identity the majority of people with active infections and the overwhelming majority of people with positive test results, even after retesting, have no active infection.

I was looking at the published literature on immunochromatography antibody tests and their level of accuracy. Both sensitivity and specificity. Antibody tests would at least reduce the sample prevalence problem. As more and more of the population had infections. But given the much lower sensitivity, looks like 80% is about as good as it would get with mass testing, and the much lower specificity, there are a lot of other human corona viruses out there, so 90% is very optimistic, despite what some people clam, it looks like the false positive problem does nt really go away until at least 15%/20% plus of the population have the antibodies. Which given the antibody fading issue would be around the time SARs CoV 2 has reached it general population equilibrium point. Plus due to the higher error rates with immunochromatography tests multiple test reduce the false positive rate at a much slower rate than with RT/PCR.

So mass testing for SARs CoV 2 is a complete bust no matter how you look at it.

Also the term "False Discovery Rate" is the technically more accurate term but when trying to communicate the problem to non-mathematicians the term "False Positive" is easier for others to understand.

[Splatt]

People are getting slightly confused with terms here.

EQU are using the correct terminology *but* the issue of rtPCR itself not being able to distinguish between infected and merely exposed is separate.

rtPCR genuinely returning RNA present for someone is not a false positive - the test is looking for specific genetic sequences that are present. The test is working as designed, not a false positive.

So we have 2 issues being merged and confused, the inability of rtPCR to see if someone is infected AND the false positive rate as estimated by the EQA (contamination and so on).

[Leggy12]

The only true way to know a false positive rate is to check the results against the 'Gold Standard Test'.

I have still never heard of the COVID-19 gold standard test, has anyone else?

[AfterAll]

I have still never heard of the COVID-19 gold standard test, has anyone else?

It's arbitrary, it's usually another PCR test, for example if you're testing a new PCR process then you will compare with one or more tests that have already been approved; here's an example for the Cepheid COVID-19 cartridges https://jcm.asm.org/content/58/8/e00926-20

The manual PCR lab tests can have higher false positive rates than the cartridge-based machines (more opportunities for errors to creep in, I suppose) for example in the Addenbrookes study: https://www.clinicalmicrobiologyandinfe ... lltext#%20

[nopos]

https://www.zerohedge.com/medical/who-f ... -positives

The World Health Organization released a guidance memo on December 14th, warning that high cycle thresholds on PCR tests will result in false positives.

https://www.who.int/news/item/14-12-202 ... -ivd-users

While this information is accurate, it has also been available for months, so we must ask: why are they reporting it now?
...

The answer to that is potentially shockingly cynical: We have a vaccine now. We don’t need false positives anymore.

Notionally, the system has produced its miracle cure.

So, after everyone has been vaccinated, all the PCR tests being done will be done “under the new WHO guidelines”, and running only 25-30 cycles instead of 35+.

Lo and behold, the number of “positive cases” will plummet, and we’ll have confirmation that our miracle vaccine works.

[MikeAustin]

https://www.zerohedge.com/medical/who-f ... -positives

The World Health Organization released a guidance memo on December 14th, warning that high cycle thresholds on PCR tests will result in false positives.

https://www.who.int/news/item/14-12-202 ... -ivd-users

While this information is accurate, it has also been available for months, so we must ask: why are they reporting it now?
...

The answer to that is potentially shockingly cynical: We have a vaccine now. We don’t need false positives anymore.

Notionally, the system has produced its miracle cure.

So, after everyone has been vaccinated, all the PCR tests being done will be done “under the new WHO guidelines”, and running only 25-30 cycles instead of 35+.

Lo and behold, the number of “positive cases” will plummet, and we’ll have confirmation that our miracle vaccine works.

Another, additional, answer is the legal case being brought against Drosten et al that has now started.