"We studied 72 persons, all of whom had a previous positive RT-PCR test but were symptom-free for >3 weeks before blood was collected for testing (Table)." Rea

"We studied 72 persons, all of whom had a previous positive RT-PCR test but were symptom-free for >3 weeks before blood was collected for testing (Table)." Reading this page I'm not sure how they are handling the possibility of false positives. That is, some percentage of their 72 persons with a positive PCR test were actually false positives and never had the disease and thus don't have antibodies. Just googling for "RT-PCR covid false positive" turns up this result[1] which says "They combined the results of multiple studies (some looked at PCR testing for SARS-CoV-2 specifically, and some looked at PCR testing for other RNA viruses). They found false positive rates of 0-16.7%, with 50% of the studies at 0.8-4.0%." If you assume 1% false positive rate, and 80 million tests[2], then that means 80,000 are false positives. 6,873,739 total positive tests, so false positives should make up ~12% of positive cases. Given the range of possible values for the false positive rate it seems reasonable to me to suspect that this might be entirely or mostly explained by false positive rate. Unless I've missed something where they account for this. edit: From their appendix: "While we cannot formally exclude false positive RT-PCR results for some participants, PCR contamination is highly unlikely as an explanation for our findings for several reasons. First, serologic non-responders were identified by three different diagnostic laboratories (Appendix Table 1), all of which employed stringent quality control measures to guard against false-positive results. Second, we were able to independently amplify SARS-CoV-2 sequences from a subset of the original nasal swab material. Analyzing 4 samples from seropositive and 8 samples from seronegative persons (Appendix Table 1), we amplified full-length spike genes with intact open reading frames from four specimens, including two from seronegative persons (Appendix Figure 5). Finally, RT-PCR positive seronegative persons have also been identified by several other groups" Not really sure how to integrate their explanation with my intuition described above. If samples were contaminated that might explain why repeat testing of the samples continued to show infection. I feel like this definitely merits a follow up investigation with a larger sample size than 71. 1 - https://theconversation.com/why-are-some-covid-test-results-false-positives-and-how-common-are-they-162163 https://theconversation.com/why-are-some-covid-test-results-... 2 - https://www.cdc.gov/coronavirus/2019-ncov/cases-updates/previous-testing-in-us.html https://www.cdc.gov/coronavirus/2019-ncov/cases-updates/prev...