Peer-Reviewed Study Confirms Fatal Flaw in PCR Testing: 42% False Discovery Rate for SARS-CoV-2 nonQ-RT-PCR Test. This means COVID-19 Vaccine Outcomes Rate Data are Unreliable and Invalid
Peer-Reviewed Study Confirms Fatal Flaw in PCR Testing: 42% False Discovery Rate for SARS-CoV-2 nonQ-RT-PCR Test. This means COVID-19 Vaccine Outcomes Rate Data are Unreliable and Invalid
All COVID-19 Vaccine Studies Used nonQ-RT-PCR to determine case status. All of the estimates of outcome are unreliable. This is the most important study we will ever likely publish in our journal.
Thanks, good one. Kudos to all those who had the strength to say "No, no way" when their doctors broached the idea of using a Covid test for a presentation of what could be a viral illness (like allergies, lol.). Stay strong and Happy Halloween.
I avoided the damn tests like the plague. No need to b stigmatized based on a load of bollocks. Not to mention what for all I know the swabs themselves could be dosed with. Not to ass-u-me anything, but at this point nothing would surprise me.
According to Dr. Lee, all PCR positive test results should be verified by sequencing as the CDC advised for SARS-CoV-1 and the FDA advised for enterovirus
Nucleic Acid Amplification Assay for the Detection of Enterovirus RNA
It is *important* that studies like this one get into the arsenals of lawyers who are prosecuting against the mandates and government overreach. Dr. JLW, perhaps you’d like to reach out to pampopper@msn.com. She is affiliated with Make Americans Free Again, which has been filing lawsuits - and winning some! - against the entire Covid insanity. Other organizations, such as AFLDS, are also pursuing legal action. Your scientific studies are the fuel that can win this war!
I'm unvaccinated and because of that, I was required to get a PCR test prior to having a biopsy. If vaccinated people are getting covid anyway, then this is a clear discriminatory practice by a major hospital in New Jersey, and I'm sure elsewhere.
And simply stated, there was NO pandemic. No grounds for lockdowns. No grounds for schools to be shut down. And there were no grounds to approve an experimental vaccine using mRNA under the EUA. The entire thing was founded and based in lies. The damage is done, so now what will happen?
Thank you for this super-clear explanation. This non-scientist is able to grasp what you are saying, and now I can go and share this information with other non-experts.
One of the things I have most loved about Substack over this past year is that it is a place where expert specialists (credentialed, usually) and non-expert generalists can come together to share ideas and expertise, to cross-pollinate and inform. It has removed some of the former segregation of gatekeepers and paywalls that previously prevented "the rest of us" from having this deeper understanding of valuable information which should be available to all of humanity.
The problem compounds: when the prevalence the disease is low, most positive PCR tests will be negative. With a 42% false positive rate applied to 1000 people, 5% of whom are infected, only 50 positive tests can be true positives, but an expected 400 will be false positives. Under CDC’s “infection = disease” paradigm, 400 people without SARS-CoV-2 infections have to be quarantined for every 50 true infections.
There is no way that testing 1000 people led to 400 of them being declared positive for covid. There were many forced to test repeatedly, such as those on college campuses and in hospitals. To think that 400 of 1000 tested would be declared positive is simply unbelievable. Campuses would have shut down in such a scenario and almost half the hospital staff would have had to quarantine EVERY time they tested.
A 42% false positive rate would be the scenarios where if you test 100k people and 1000 were found to be positive, then approximately 420 of them would have been falsely labeled as positive. That means the remaining 580 or so of the ramaining 1000 positives would have actually been labeled positive correctly rather than the 50 this blurb mentions. This is still terrible and I expect it can be higher/lower than this value depending on the prevalence of the virus (there would be a higher FPR during summer for instance when the virus is not circulating). In other words, the false positve rate is not static. There was one university that actually tested positive students twice - I think it was Cambridge University - and found that of the 38 or so first found to be positive only 6 were positive after a second round. Also I know that the NBA (and likely other sports franchises) employed a test which was a lot more likely to have found a true positive with a test doing something akin to two rounds. I do agree with the conclusion that Sanger testing or at least another PCR test should have been applied for a confirmatory finding.
JustAPoster is quoting James, who created a scenario to illustrate what's being discussed in the fourth paragraph under the heading "Why This Matters 1: The Clinical Mess."
False positives are not a percentage of the positives that are false, but a percentage of the "true negatives" that are falsely tested as positives.
James' math in the example you're working off goes like this. If only 5% have the virus, then 950 of the people are true negatives. This means that when you test those 950 with a test with a 42% false positive rate, you will get 950 * 0.42 = 399, and James rounded that up to 400. That will then be a total of 450 people testing positive in a community of 1000 where 5% are infected.
Also, edited to add this:
Yes, the incredulity and "This is simply unbelievable" response is precisely the point to what James is getting at, even once you get the math right. Campuses, hospitals, state-managed nursing homes, &c did, in fact, get shut down; lots of people had to cycle out of work for days and weeks.
False discovery rate is the percentage of positives that are false. In this paper, 42% of the positive tests were negative, which would enable you to calculate the false discovery rate.
Sorry but your figures are unbelievable because they are.
I know many places that tested, including school campuses and work places. Nowhere was there even close to 400 people out of 1000 getting positive results. Usually it was about 2-3. If it hit 10, they would likely have closed down. Some campuses did because the positives were high enough to affect their ability to quarantine but they certainly did not quarantine half the students.
Here's an example.
Ontario actually reported the approximate number of tests conducted. From the article below:
Ontario reported 821 new cases of COVID-19 on Tuesday, the second-most on a single day since a resurgence of the illness began in the province in mid-August.
Notably, just over 24,000 tests were completed yesterday — the lowest number of tests Ontario has processed on a single day since Sept. 9. The province previously said it aimed to be processing 50,000 tests per day by mid-October, and as many as 68,000 daily by mid-November.
No this was back in October 2020 before vaccines were even around. Ontario actually used 30 as their max cycle threshold which is likely why they didn't have the number of covid cases America produced at that point in some states. Later, when variants of concern started to emerge they did go to 35 as a cycle threshold but only for those samples with a variant of concern (which eventually were all samples).
The point is that if the number of false positives were that high that you test 1000 people and 400 would come back positive, no hospital could actually function or any school. They would have all been shuttered.
Okay, these are not "my figures" but the hypothetical given by James and the conclusion from Lee's comparison of Sanger sequencing to samples testing positive using RT-qPCR. He found that 21 of 50 positive samples did not contain the actual genetic material he would have found if SARS-CoV-2 were present. See the section in the paper titled "3.2. SARS-CoV-2 was detected by RT-PCR and Sanger sequencing in only 29 of 50 RTqPCR positive reference specimens"
As for many people getting tested and showing up negative when many people should have? That's interesting: James points out the Ct values do matter, and that is also the upshot of that section in Lee's paper. By rolling up or down the number of cycles, it's possible to rig the outcomes in select industries, geographies, occupations.
Again, look at the paper in that section, and judge for yourself.
I agree with you but such a high FPR seems off because if the number of false positives were that high, no campus or workplace that tested everyone weekly or twice weekly would have been able to function and yet there were places I am very familiar with that didn't shut down. They never even had more than 2-3 positives at the height of omicron last winter.
Wow! We knew all along that the entire narrative rested on the tests.
I informed the county health director in Genesee County that even if the false positive rate were only 1%, at any point in time during several months in Michigan, there would be over 10,000 residents unlawfully detained by the quarantine orders for close contacts.
All the harm that was done based on the unreliable tests and the lies about asymptomatic infection/transmission!
I'm a bit dense today. Can someone point me to the math for this conclusion: 42% false positive rate -> for every 5 true positives, 400 false positives will be reported? Doesn't this depend on the percent of true positive in the sample?
If we start with hypothetical 1000 people, and we say 100 are true positive and the test is 42% false positive. Then among the 900 true negative, we get 900*0.42 = 378 false positive. So for 5 true positive there are 18.9 false positive.
If we start with hypothetical 1000 people, and we say 10 are true positive and the test is 42% false positive, then among the 990 true negative, we get 990*0.42 = 415.8 false positive. So for 5 true positive, there are 207.9 false positive.
If we start with hypothetical 1000 people, and we say 5 are true positive and the test is 42% false positive, then among the 995 true negative, we get 995*0.42 = 417.9 false positive. So for 5 true positive, there are 417.9 false positive.
So the claim "for 5 true positive we get 400 false positive" is contingent on the percent of people tested being truly positive equaling exactly 0.5222%, right?
Is that what the real-world data really show? Shouldn't you mention the ratio is contingent on that? How are you confident the true positive ratio is exactly 0.5222%?
What I'm gathering is if there are 100 positive tests out of 1,000 people, only 58 were actual cases. That's how 42% would be false discoveries. Therefore, it follows that 58 out of every 100 deaths with COVID "confirmed by PCR testing" actually HAD COVID.
Whether or not COVID would be a properly attributed cause of death even for those who actually had it is another matter (hence, Italy's 97% reduction of the toll attributed to COVID).
On top of that, imagine how greatly the actual death rate would be reduced if effective early treatment measures such as the Brownstein protocol (upon which I based my own home treatment that kicked it to the curb easily - my having beaten COVID confirmed by T-cell response testing a few months later, and to my knowledge had not a single long COVID case among his patients) were universally implemented? I could see the probability of actually dying FROM COVID being less than 1/1000 of the attributed toll, i.e., <1,000 per million reported.
"Dr. David Brownstein, who has a clinic just outside of Detroit, Michigan, has successfully treated over 200 patients with what has become my favorite intervention for COVID-19 and other upper respiratory infections, namely nebulized hydrogen peroxide.
A peer-reviewed consecutive case series of 107 COVID-19 patients treated with nebulized peroxide and other remedies, including oral vitamins A, C and D, iodine, intravenous hydrogen peroxide and iodine as well as intravenous (IV) vitamin C, along with intramuscular ozone, was published in the July 2020 issue of Science, Public Health Policy, and the Law.1 All patients survived."
hmm some of that is pretty safe and understandable, not sure about injecting h2po4 or o3 or iodine. Hydrogen peroxide and ozone are powerful oxidizers which will have eeehhh lots of effects. You know how people tout 'anti-oxidants' well. Yeesh.
Can't say for sure it's inadvisable but yeesh.
It's a bit like the 'lysol and bleach sterilize surfaces, so let's inject it'. That kinda yeesh. Are powerful oxidants selective to viruses? Well, no.
Yeah all 200 survived a disease where on average one might have died. Hmm.
All having survived is nothing. Not a single long COVID case that got into the cytokine storm phase, even with some of the patients having been previously hospitalized, is most remarkable.
Vitamin C and some of the other supplements in the protocol are strong antioxidants.
Brownstein has explained the dance between oxidation and reduction in videos.
What I care about most if that the home care version with oral supplements and huffing the H2O2 in saline mist using a simple spray bottle worked for me.
If, according to #4, the number of "cases" via positive PCR has been overstated by a factor of 80:1, does that mean the number of covid deaths have also been overstated by that factor? Or is it even worse because sometimes a positive test wasn't even involved in the decision to label a death as a covid death?
If the number of 'cases' are overreported by 80:1 then the case-fatality rate is under-reported by 1:80. Meaning the true death rate of people with real covid infections is much higher than reported.
The case-fatality rate is not the number used to support the lockdowns, business closures, masks, vaccines, etc.
It was always CASES! DEATHS! in raw numbers, highly inflated, to justify the idea of a deadly pathogen from which no one is safe, and requiring the government to step in.
And how many of the deaths in people who were truly infected - actual cases - were caused by medical malfeasance like lack of early treatment, withholding certain treatments, forcing dangerous treatments, withholding hydration and nutrition, prolonged isolation, etc.?
So the case fatality rate is also corrupted beyond any sort of reliability.
Hmmm... I wonder if a de-Googled android can run the software? But I do very much like Google maps. And don't do anything controversial on my phone, anyway.
The problem compounds: when the prevalence the disease is low, most positive PCR tests will be negative. With a 42% false positive rate applied to 1000 people, 5% of whom are infected, only 50 positive tests can be true positives, but an expected 400 will be false positives. Under CDC’s “infection = disease” paradigm, 400 people without SARS-CoV-2 infections have to be quarantined for every 50 true infections.
This analogy is way off. There is no way that 400 people out of 1000 tests were positive. Lots of companies and schools tested and they certainly did not see this high rate. Of the 1000 tested, it was likely maybe 2 to 3 would be positive. A false positive rate of 42% would mean that of the 2, 1 would likely be a false positive though given the small numbers, possibly both or neither. That is believable but saying that if you're testing 1000 people, you'd get 400 false positives is just not believable. If that happened on any campus, the campus would not be able to keep up with their quarantines and would have shuttered in 2021/2022.
Well, with a sample of 1000 and 5% true positive rate, you have 50 true positive. Thus a 42% false positive test yields:
Among the 950 true negative, 950*0.42 = 399 false positive. For every 5 true positive, you get 39.9 false positive.
So in your case the ratio is 8:1, not the 80:1 as James states here. See my comment 42 minutes ago. But is your hypothetical 5% true positive reaslistic? I think 0.5% is closer to real-world in a mass-test.
So with 0.522% true positive, and 42% false positive, James' 80:1 ratio computes. But does 42% false-positive match what we saw claimed by the mass-testing regime? You're quite right that we didn't see 40% tests claimed to be positive. Something doesn't quite add-up here.
I think your definitions in the article are wrong:
TPR = True positive rate = TP/(TP+FP)
FPR = False positive rate = FP/(TP+FP)
You are calculating the true discovery rate and the false discovery rate. I made that mistake as well when I gave the example of 1000 people tested in which 2-3 would test positive, 1 likely being a false positive (though with such low numbers it could be 0 or all being false). This is not the False Positive Rate but the False Discovery Rate.
TPR is TP / (TP+FN)
FPR is FP/(FP + TN)
You need to find out the number of true negatives to calculate the false positive rate.
This raises questions about the review process this paper underwent. I would expect a serious publication to withdraw a questionable paper for further review.
If the author, and the reviewers, didn't know the difference between trivial concepts such as FDR and FPR, how certain can we be it the rest of the results?
The FPR is not 42% chance of a false positive if tested.
It's the % of samples that DO test positive that are false.
Right. Of all the positives, the ones that are falsely positive.
In other words, you are calculating FP/(TP + FP) but this is not the false positive rate.
The false positive rate is FP/(TN + FP)
The analogy you have for the 1000 is wrong because you are claiming 50 true positives in 1000. Your analogy would be more accurate if you said of 1000 people that tested positive, 420 of them would be false positives.
I will have to look at the estimate but there is simply no way there is a 42% false positive rate. Think of all the people you know who have tested. This claim would mean close to half would have tested positive. This is simply not the case. Lots of work sites and campuses tested everyone (both vaccinated and unvaccinated) and they had maybe a 10%-15% rate at most though usually more akin to 2-3%.
The false positive rate is calculated as FP/FP+TN, where FP is the number of false positives and TN is the number of true negatives (FP+TN being the total number of negatives). It’s the probability that a false alarm will be raised: that a positive result will be given when the true value is negative.
I think you're calculating the False Discovery Rate:
This is so timely. "Bring down the house of Fraud"! I was just rudimenting over this and the Askewed death certificates, suffering untreated Patients early on and the true cost of life. Thank you and for also adding that powerful speech which introduced me to you.
Haha, a subtle psychological trick to motivate people to learn something new, eh?
Thanks, good one. Kudos to all those who had the strength to say "No, no way" when their doctors broached the idea of using a Covid test for a presentation of what could be a viral illness (like allergies, lol.). Stay strong and Happy Halloween.
I avoided the damn tests like the plague. No need to b stigmatized based on a load of bollocks. Not to mention what for all I know the swabs themselves could be dosed with. Not to ass-u-me anything, but at this point nothing would surprise me.
According to Dr. Lee, all PCR positive test results should be verified by sequencing as the CDC advised for SARS-CoV-1 and the FDA advised for enterovirus
Nucleic Acid Amplification Assay for the Detection of Enterovirus RNA
https://bit.ly/3DnRpQn
Or you could use cell-culturing if you care about accuracy and not clinical treatment--for research purposes.
Yes, and that's been done. See https://www.cebm.net/study/duration-of-infectiousness-and-correlation-with-rt-pcr-cycle-threshold-values-in-cases-of-covid-19-in-england/
I'll see your study.
https://link.springer.com/article/10.1007/s10096-020-03913-9?fbclid=IwAR3qSrFSmnA_b9QSzQlAdBZdazYsZW4BMZgeYnsGqvwjpclC5f6kcpmTH30
I've followed Raoult's work on Covid since Zelenko mentioned him. Raoult is very competent, but also disliked by academia.
If someone could figure out a way to speed up cell culturing from days to hours, they'd make out like bandits.
Why would you sequence for SARS-CoV-1 when you are looking for SARS-CoV-2 ?
No one is. In 2003, CDC and FDA used Sanger sequencing on SARS-CoV-1.
It is *important* that studies like this one get into the arsenals of lawyers who are prosecuting against the mandates and government overreach. Dr. JLW, perhaps you’d like to reach out to pampopper@msn.com. She is affiliated with Make Americans Free Again, which has been filing lawsuits - and winning some! - against the entire Covid insanity. Other organizations, such as AFLDS, are also pursuing legal action. Your scientific studies are the fuel that can win this war!
Thank you, eternally!
Dr. Popper is teaching for IPAK-EDU next semester!
I'm sure it will hit her radar :)
I'm unvaccinated and because of that, I was required to get a PCR test prior to having a biopsy. If vaccinated people are getting covid anyway, then this is a clear discriminatory practice by a major hospital in New Jersey, and I'm sure elsewhere.
Just WOW!
Thank you so much!
And simply stated, there was NO pandemic. No grounds for lockdowns. No grounds for schools to be shut down. And there were no grounds to approve an experimental vaccine using mRNA under the EUA. The entire thing was founded and based in lies. The damage is done, so now what will happen?
Thank you for this super-clear explanation. This non-scientist is able to grasp what you are saying, and now I can go and share this information with other non-experts.
One of the things I have most loved about Substack over this past year is that it is a place where expert specialists (credentialed, usually) and non-expert generalists can come together to share ideas and expertise, to cross-pollinate and inform. It has removed some of the former segregation of gatekeepers and paywalls that previously prevented "the rest of us" from having this deeper understanding of valuable information which should be available to all of humanity.
Thank you for taking the time.
Except it's wrong. This claim in particular:
The problem compounds: when the prevalence the disease is low, most positive PCR tests will be negative. With a 42% false positive rate applied to 1000 people, 5% of whom are infected, only 50 positive tests can be true positives, but an expected 400 will be false positives. Under CDC’s “infection = disease” paradigm, 400 people without SARS-CoV-2 infections have to be quarantined for every 50 true infections.
There is no way that testing 1000 people led to 400 of them being declared positive for covid. There were many forced to test repeatedly, such as those on college campuses and in hospitals. To think that 400 of 1000 tested would be declared positive is simply unbelievable. Campuses would have shut down in such a scenario and almost half the hospital staff would have had to quarantine EVERY time they tested.
A 42% false positive rate would be the scenarios where if you test 100k people and 1000 were found to be positive, then approximately 420 of them would have been falsely labeled as positive. That means the remaining 580 or so of the ramaining 1000 positives would have actually been labeled positive correctly rather than the 50 this blurb mentions. This is still terrible and I expect it can be higher/lower than this value depending on the prevalence of the virus (there would be a higher FPR during summer for instance when the virus is not circulating). In other words, the false positve rate is not static. There was one university that actually tested positive students twice - I think it was Cambridge University - and found that of the 38 or so first found to be positive only 6 were positive after a second round. Also I know that the NBA (and likely other sports franchises) employed a test which was a lot more likely to have found a true positive with a test doing something akin to two rounds. I do agree with the conclusion that Sanger testing or at least another PCR test should have been applied for a confirmatory finding.
JustAPoster is quoting James, who created a scenario to illustrate what's being discussed in the fourth paragraph under the heading "Why This Matters 1: The Clinical Mess."
False positives are not a percentage of the positives that are false, but a percentage of the "true negatives" that are falsely tested as positives.
James' math in the example you're working off goes like this. If only 5% have the virus, then 950 of the people are true negatives. This means that when you test those 950 with a test with a 42% false positive rate, you will get 950 * 0.42 = 399, and James rounded that up to 400. That will then be a total of 450 people testing positive in a community of 1000 where 5% are infected.
Also, edited to add this:
Yes, the incredulity and "This is simply unbelievable" response is precisely the point to what James is getting at, even once you get the math right. Campuses, hospitals, state-managed nursing homes, &c did, in fact, get shut down; lots of people had to cycle out of work for days and weeks.
Again, that's the point of what he's getting at.
False discovery rate is the percentage of positives that are false. In this paper, 42% of the positive tests were negative, which would enable you to calculate the false discovery rate.
Thank you!
No problem. Amazing what you can learn in from Wikipedia in 15 minutes...
Actually evaluation science is very cool - and powerful if paired w/machine learning.
Sorry but your figures are unbelievable because they are.
I know many places that tested, including school campuses and work places. Nowhere was there even close to 400 people out of 1000 getting positive results. Usually it was about 2-3. If it hit 10, they would likely have closed down. Some campuses did because the positives were high enough to affect their ability to quarantine but they certainly did not quarantine half the students.
Here's an example.
Ontario actually reported the approximate number of tests conducted. From the article below:
Ontario reported 821 new cases of COVID-19 on Tuesday, the second-most on a single day since a resurgence of the illness began in the province in mid-August.
Notably, just over 24,000 tests were completed yesterday — the lowest number of tests Ontario has processed on a single day since Sept. 9. The province previously said it aimed to be processing 50,000 tests per day by mid-October, and as many as 68,000 daily by mid-November.
https://www.cbc.ca/news/canada/toronto/covid-19-coronavirus-ontario-october-20-update-1.5769138
Thus, you can see most tests returned a negative (false or otherwise).
If they used CDC's "Ct <27 or less" for the vaccinated, this would be the expectation.
No this was back in October 2020 before vaccines were even around. Ontario actually used 30 as their max cycle threshold which is likely why they didn't have the number of covid cases America produced at that point in some states. Later, when variants of concern started to emerge they did go to 35 as a cycle threshold but only for those samples with a variant of concern (which eventually were all samples).
The point is that if the number of false positives were that high that you test 1000 people and 400 would come back positive, no hospital could actually function or any school. They would have all been shuttered.
Okay, these are not "my figures" but the hypothetical given by James and the conclusion from Lee's comparison of Sanger sequencing to samples testing positive using RT-qPCR. He found that 21 of 50 positive samples did not contain the actual genetic material he would have found if SARS-CoV-2 were present. See the section in the paper titled "3.2. SARS-CoV-2 was detected by RT-PCR and Sanger sequencing in only 29 of 50 RTqPCR positive reference specimens"
As for many people getting tested and showing up negative when many people should have? That's interesting: James points out the Ct values do matter, and that is also the upshot of that section in Lee's paper. By rolling up or down the number of cycles, it's possible to rig the outcomes in select industries, geographies, occupations.
Again, look at the paper in that section, and judge for yourself.
I agree with you but such a high FPR seems off because if the number of false positives were that high, no campus or workplace that tested everyone weekly or twice weekly would have been able to function and yet there were places I am very familiar with that didn't shut down. They never even had more than 2-3 positives at the height of omicron last winter.
https://popularrationalism.substack.com/p/prevalence-fp-tp-and-biased-fptp
https://popularrationalism.substack.com/p/prevalence-fp-tp-and-biased-fptp
https://popularrationalism.substack.com/p/prevalence-fp-tp-and-biased-fptp
I wish these threads were better. I would rather see a conversation in proper sequence.
Regarding Ontario.... no. It was on their website.....They cycled at 38.
We know because we all compared it when Florida changed their testing. We were complaining why Ontario didn’t do the same.
We have it from HHS in an email that anything above 35 is "basically meaningless".
And, being expert in PCR, the problem is NO universal "cycled at" is appropriate for RT-qPCR.
Yes. Everyone knew it was too high but no one listened. Here is a link from 2021 and the Ontario government considered positives at 38 and 40!!!!
This testing ruined so many lives here. Shutdowns. Schools shutdowns. Kids in masks. Mandatory vaxes. No proper funerals. I could cry!!!
I’m so mad and sad.
https://www.publichealthontario.ca/en/About/news/2021/Explained-COVID19-PCR-Testing-and-Cycle-Thresholds
Sounds like a hearing is in order.
How can you claim to be an expert, if you don't know the difference between FDR and FPR?
I know the difference. It was some minor slip up.
https://pubmed.ncbi.nlm.nih.gov/17697328/
Wow! We knew all along that the entire narrative rested on the tests.
I informed the county health director in Genesee County that even if the false positive rate were only 1%, at any point in time during several months in Michigan, there would be over 10,000 residents unlawfully detained by the quarantine orders for close contacts.
All the harm that was done based on the unreliable tests and the lies about asymptomatic infection/transmission!
Dani,
Thank you, that is most gratifying!
I'm a bit dense today. Can someone point me to the math for this conclusion: 42% false positive rate -> for every 5 true positives, 400 false positives will be reported? Doesn't this depend on the percent of true positive in the sample?
If we start with hypothetical 1000 people, and we say 100 are true positive and the test is 42% false positive. Then among the 900 true negative, we get 900*0.42 = 378 false positive. So for 5 true positive there are 18.9 false positive.
If we start with hypothetical 1000 people, and we say 10 are true positive and the test is 42% false positive, then among the 990 true negative, we get 990*0.42 = 415.8 false positive. So for 5 true positive, there are 207.9 false positive.
If we start with hypothetical 1000 people, and we say 5 are true positive and the test is 42% false positive, then among the 995 true negative, we get 995*0.42 = 417.9 false positive. So for 5 true positive, there are 417.9 false positive.
So the claim "for 5 true positive we get 400 false positive" is contingent on the percent of people tested being truly positive equaling exactly 0.5222%, right?
Is that what the real-world data really show? Shouldn't you mention the ratio is contingent on that? How are you confident the true positive ratio is exactly 0.5222%?
In the hypothetical example given, percent of true positive in the sample is 50 out 1,000.
When the prevalence is low, the issue exists.
Irritatingly, this issue was known for a long, long time in diagnostics. And even relatively early in COVID-19:
https://pubmed.ncbi.nlm.nih.gov/33087255/
How did you calculate the false positive rate?
In the set of PCR+s = FPR is the # of TP/(TP+FP)
No - that would get you the True detection rate.
The False detection rate is FP/(TP+FP)
And the FPR = FP/(TN+FP)
I couldn't find the true negative rate anywhere in this paper
And you won't find the true positive rate in the paper about vaccine efficacy that was published in NEJM, either.
oops, that's what i meant, of course.
I typed TP/(TP+FP) meaning FP/(TP+FP).
See https://popularrationalism.substack.com/p/prevalence-fp-tp-and-biased-fptp
Using very simple math, only 58 actual positives out of every 100 positives if the FPR is 42%, correct?
No, if the FDR is 42%...
The original article mis-used FPR, it should have been FDR and is now corrected.
Your math is correct.. I think one part of the issue is that those "For every 5 true positives" statements are likely typos.
Pretty much. Even typos have consequences! Thanks.
Thank you BongoBen for your comment - this will clarify nicely
https://popularrationalism.substack.com/p/prevalence-fp-tp-and-biased-fptp
What I'm gathering is if there are 100 positive tests out of 1,000 people, only 58 were actual cases. That's how 42% would be false discoveries. Therefore, it follows that 58 out of every 100 deaths with COVID "confirmed by PCR testing" actually HAD COVID.
Whether or not COVID would be a properly attributed cause of death even for those who actually had it is another matter (hence, Italy's 97% reduction of the toll attributed to COVID).
On top of that, imagine how greatly the actual death rate would be reduced if effective early treatment measures such as the Brownstein protocol (upon which I based my own home treatment that kicked it to the curb easily - my having beaten COVID confirmed by T-cell response testing a few months later, and to my knowledge had not a single long COVID case among his patients) were universally implemented? I could see the probability of actually dying FROM COVID being less than 1/1000 of the attributed toll, i.e., <1,000 per million reported.
What a mad, mad planet!
"Dr. David Brownstein, who has a clinic just outside of Detroit, Michigan, has successfully treated over 200 patients with what has become my favorite intervention for COVID-19 and other upper respiratory infections, namely nebulized hydrogen peroxide.
A peer-reviewed consecutive case series of 107 COVID-19 patients treated with nebulized peroxide and other remedies, including oral vitamins A, C and D, iodine, intravenous hydrogen peroxide and iodine as well as intravenous (IV) vitamin C, along with intramuscular ozone, was published in the July 2020 issue of Science, Public Health Policy, and the Law.1 All patients survived."
hmm some of that is pretty safe and understandable, not sure about injecting h2po4 or o3 or iodine. Hydrogen peroxide and ozone are powerful oxidizers which will have eeehhh lots of effects. You know how people tout 'anti-oxidants' well. Yeesh.
Can't say for sure it's inadvisable but yeesh.
It's a bit like the 'lysol and bleach sterilize surfaces, so let's inject it'. That kinda yeesh. Are powerful oxidants selective to viruses? Well, no.
Yeah all 200 survived a disease where on average one might have died. Hmm.
All having survived is nothing. Not a single long COVID case that got into the cytokine storm phase, even with some of the patients having been previously hospitalized, is most remarkable.
Vitamin C and some of the other supplements in the protocol are strong antioxidants.
Brownstein has explained the dance between oxidation and reduction in videos.
What I care about most if that the home care version with oral supplements and huffing the H2O2 in saline mist using a simple spray bottle worked for me.
If, according to #4, the number of "cases" via positive PCR has been overstated by a factor of 80:1, does that mean the number of covid deaths have also been overstated by that factor? Or is it even worse because sometimes a positive test wasn't even involved in the decision to label a death as a covid death?
If the number of 'cases' are overreported by 80:1 then the case-fatality rate is under-reported by 1:80. Meaning the true death rate of people with real covid infections is much higher than reported.
The case-fatality rate is not the number used to support the lockdowns, business closures, masks, vaccines, etc.
It was always CASES! DEATHS! in raw numbers, highly inflated, to justify the idea of a deadly pathogen from which no one is safe, and requiring the government to step in.
And how many of the deaths in people who were truly infected - actual cases - were caused by medical malfeasance like lack of early treatment, withholding certain treatments, forcing dangerous treatments, withholding hydration and nutrition, prolonged isolation, etc.?
So the case fatality rate is also corrupted beyond any sort of reliability.
Any setting that uses "PCR+ = COVID19" will overestimate the number of COVID19 cases, so, yes, you are correct.
But not by all that much...
Adjusted in the toy example to 8:1, but at any given time given the prevalence, yes, that it what it means - see https://popularrationalism.substack.com/p/prevalence-fp-tp-and-biased-fptp
I can’t send this from my iPad . I text it to friends and it won’t go through.
Please go ahead recommend what I should be using.
Thank you! I will watch.
Hmmm... I wonder if a de-Googled android can run the software? But I do very much like Google maps. And don't do anything controversial on my phone, anyway.
? One thing that has also occurred to me is that spyware can also be misled or confused. I get lots of spam for the most ridiculous things.
You're a star, Dr. Jack. Proud to be your student.
Shared, shared and shared again.
The problem compounds: when the prevalence the disease is low, most positive PCR tests will be negative. With a 42% false positive rate applied to 1000 people, 5% of whom are infected, only 50 positive tests can be true positives, but an expected 400 will be false positives. Under CDC’s “infection = disease” paradigm, 400 people without SARS-CoV-2 infections have to be quarantined for every 50 true infections.
This analogy is way off. There is no way that 400 people out of 1000 tests were positive. Lots of companies and schools tested and they certainly did not see this high rate. Of the 1000 tested, it was likely maybe 2 to 3 would be positive. A false positive rate of 42% would mean that of the 2, 1 would likely be a false positive though given the small numbers, possibly both or neither. That is believable but saying that if you're testing 1000 people, you'd get 400 false positives is just not believable. If that happened on any campus, the campus would not be able to keep up with their quarantines and would have shuttered in 2021/2022.
Well, with a sample of 1000 and 5% true positive rate, you have 50 true positive. Thus a 42% false positive test yields:
Among the 950 true negative, 950*0.42 = 399 false positive. For every 5 true positive, you get 39.9 false positive.
So in your case the ratio is 8:1, not the 80:1 as James states here. See my comment 42 minutes ago. But is your hypothetical 5% true positive reaslistic? I think 0.5% is closer to real-world in a mass-test.
So with 0.522% true positive, and 42% false positive, James' 80:1 ratio computes. But does 42% false-positive match what we saw claimed by the mass-testing regime? You're quite right that we didn't see 40% tests claimed to be positive. Something doesn't quite add-up here.
And I notice you're the only other commenter so far who has *thought* about these numbers. Kudos.
Thanks for the explanation. You're correct.
FPR = FP/(FP + TN)
vs:
FDR = FP/(FP + TP)
We don't have the true negative value. I think this is where the problem lies. James may be using FDR numbers as FPR numbers.
Not that complicated. A simple typo on my part for the toy example led me off.
Here's a clear explanation: https://popularrationalism.substack.com/p/prevalence-fp-tp-and-biased-fptp
I think your definitions in the article are wrong:
TPR = True positive rate = TP/(TP+FP)
FPR = False positive rate = FP/(TP+FP)
You are calculating the true discovery rate and the false discovery rate. I made that mistake as well when I gave the example of 1000 people tested in which 2-3 would test positive, 1 likely being a false positive (though with such low numbers it could be 0 or all being false). This is not the False Positive Rate but the False Discovery Rate.
TPR is TP / (TP+FN)
FPR is FP/(FP + TN)
You need to find out the number of true negatives to calculate the false positive rate.
It's not a simple typo
He was referring to the 80:1 false positives vs true positives actually being in reality 8:1 which was a typo.
Thanks for your explanation on everything!
Because this paper is trying to calculate the false discovery rate, not the false positive rate....
"For every 5 positive"
should be "For every 50 positive"
8:1 is about right. I need to update this part. Thank you.
Use any low prevalence you like. My math was "for example"
No, it's not... Based on this paper, for every 50 positives, you would get 36 false negatives... Seems like 1:0.7, not 1:8...
The FPR is not 42% chance of a false positive if tested.
It's the % of samples that DO test positive that are false.
The Marine Study result provides ample evidence that as proscribed, we're in that neighborhood.
They could not successfully sequence 37% of the PCR+ samples.
"we obtained more than 95% complete viral genomes from 36 specimens obtained from 32 of 51 participants (62.7%)"
https://www.navy.mil/Press-Office/News-Stories/Article/2413465/navymarine-corps-covid-19-study-findings-published-in-new-england-journal-of-me/
https://www.nejm.org/doi/full/10.1056/NEJMoa2029717?query=featured_home
So why did you say that you found the FPR, not the FDR...
See https://popularrationalism.substack.com/p/prevalence-fp-tp-and-biased-fptp
But you still haven't corrected you post
Must have missed one. Thanks.
This raises questions about the review process this paper underwent. I would expect a serious publication to withdraw a questionable paper for further review.
If the author, and the reviewers, didn't know the difference between trivial concepts such as FDR and FPR, how certain can we be it the rest of the results?
The FPR is not 42% chance of a false positive if tested.
It's the % of samples that DO test positive that are false.
Right. Of all the positives, the ones that are falsely positive.
In other words, you are calculating FP/(TP + FP) but this is not the false positive rate.
The false positive rate is FP/(TN + FP)
The analogy you have for the 1000 is wrong because you are claiming 50 true positives in 1000. Your analogy would be more accurate if you said of 1000 people that tested positive, 420 of them would be false positives.
The 42% estimate comes from Dr. Lee's study, not my toy example.
I will have to look at the estimate but there is simply no way there is a 42% false positive rate. Think of all the people you know who have tested. This claim would mean close to half would have tested positive. This is simply not the case. Lots of work sites and campuses tested everyone (both vaccinated and unvaccinated) and they had maybe a 10%-15% rate at most though usually more akin to 2-3%.
(TP+FP) = all who tested positive
TP/(TP+FP) = True Positive Rate (Detection Rate)
FP/(TP+FP) = False Positive Rate
The False Positive Rate:
FPR = FP/(TN + FP)
https://www.split.io/glossary/false-positive-rate/
The false positive rate is calculated as FP/FP+TN, where FP is the number of false positives and TN is the number of true negatives (FP+TN being the total number of negatives). It’s the probability that a false alarm will be raised: that a positive result will be given when the true value is negative.
I think you're calculating the False Discovery Rate:
FDR = FP / (FP + TP)
https://en.wikipedia.org/wiki/False_discovery_rate
The total number of rejections of the null include both the number of false positives (FP) and true positives (TP). Simply put, FDR = FP / (FP + TP).
> PCR Testing: 42% False Positive Rate for SARS-CoV-2
Sorry, incorrect.
There is no COVID PCR "test" which has ever been calibrated to anything in the Real World, i.e., viral isolates obtained from a sick patient.
ALL PCR positives "for COVID" are false!
💯%
This is so timely. "Bring down the house of Fraud"! I was just rudimenting over this and the Askewed death certificates, suffering untreated Patients early on and the true cost of life. Thank you and for also adding that powerful speech which introduced me to you.