numberwang said:
Germany has increased testing to ~100K PCR tests a day (!) to create 99% negatives with a typical 0.2-5% margin of error for false positives. Nothing but statistical noise. https://bgr.com/2020/03/26/coronavirus-testing-explains-germanys-low-covid-19-fatality-rate/ |
Where does that article say anything about false positives? PCR tests are extremely accurate and almost never give a false positive to my knowledge so I'd be interested to know where you get the information that up to 5% of results can be false positives.
From MITmedical if anyone is curious:
"How accurate is the diagnostic test for COVID-19?
The PCR test for COVID-19 works by detecting genetic material from SARS-CoV-2, the virus that causes COVID-19. The genetic material from SARS-CoV-2 cannot be confused with the genetic material from other viruses, so the COVID-19 diagnostic test is highly specific. This means it almost never gives a false positive. If you are tested for COVID-19, and the test comes back positive, you can be very sure that you are infected with this virus. The new antigen test for COVID-19 is also very specific and rarely gives a false positive.
Unfortunately, neither test is equally sensitive. If the specimen collection is not done perfectly, or if you are in an early stage of infection or already partially recovered, your nasal-swab sample might not contain enough viral material to come back positive. There are many stories about patients who tested negative soon after their symptoms began, only to test positive on a test done later. It is clear that the PCR test is more accurate at detecting early-stage infections, and there are early indications that the antigen test may be better at identifying patients who are already recovering.
However, because false negative results on diagnostic tests happen relatively often, a negative result should not give you a sense of false security. If you have any symptoms of COVID-19, it is safest to assume you are infected and act accordingly, even if your diagnostic test comes back negative."
Basically, PCR is a technique that is hard to get to work sometimes so false negatives can be quite common, but false positives almost never occur. The standard PCR primer is usually 15-20 bp long at least and so the chance of a different virus or other strand of DNA matching enough to bind the primer is exceedingly low. Then you have two primers, which are distanced a certain length apart so you would have to have both primers bind (unlikely) and then make a similar sized fragment that would get mistaken for the correct length on a gel (unlikely). This is why I can use this technique to get a gene out of an entire genome and be confident that I get the gene I was hoping for if both primers worked and the PCR product is close to the correct size.
...
