A serology test is used to show whether a person has antibodies against a germ. If a person has antibodies against the coronavirus, it means they were exposed to the COVID-19 virus at some point.
Results of COVID-19 antibody tests may not always be accurate, especially if the test was done too soon after infection or the test quality is questionable. They can show whether someone’s immune system has encountered the virus. But it is unknown what level of antibodies, if any, confers protection against the coronavirus, so the test results can’t indicate whether a person is immune to a future infection. And it is unknown how long such immunity might last. So, without learning more about COVID-19, serology tests alone cannot be used to determine when to go back to work.
Furthermore, the accuracy of a serology test depends on both its quality and how common COVID-19 is in the population. Take, for example, a test with a sensitivity of 94% and a specificity of 96% that is used in a town with 10,000 people in it, where 1% of the people had COVID-19. If that 1% - 100 people - who had COVID-19 also had antibodies, 94 (94% of 100) of them would test positive. Of the 9,900 uninfected, 9,504 (96% of 9,900) would correctly test negative. However, the remaining 396 (4% of 9,900) would test positive, even though they don’t have antibodies. So, 396 out of the 490 (94+396) positive results (4 out of every 5) would be incorrect. These are called false positives. Even if 2% of the town had had COVID-19, 2 out of 3 people who receive positive results wouldn’t actually have antibodies. When disease prevalence is low, mass population screening runs into this false positive problem. For a more thorough discussion of this issue, please visit this resource.
Drafted 20 October 2020