Bayes’ rule applied on Covid 19 Immunity testing
Many persons experienced unconfirmed Covid-19 symptoms in the early months of 2020, not knowning whether it was truly Covid-19 or a classic flu.
With the interest of knowing if meanwhile immunity was gathered, many are now performing a serological test for SARS-CoV-2 antibodies.
In general, a positive antibody test is presumed to mean a person has been infected with SARS-CoV-2, the virus that causes COVID-19, at some point in the past. However, in many cases, the test turned out to be negative. The negative result of this test made us think about the probability for truly not having had Covid-19 or not being immune to the disease. The exercise described in this paper, follows a Bayesian Approach, using conditional probability to evaluate the chance of (not) being immune, and to learn if anything can be concluded on whether or not Covid-19 was contracted months ago.
Read the full paper: