Interesting idea: What to do when the only antibody test available has a 5% false result rate, meaning a 5% chance the test says you have antibodies but you don't, or it says you don't have them but you do?

Take the test twice, either a few days apart or test both people of a couple. If you get a consistent result, the false result value should drop to 1 in 400. Which is one quarter of the P=0.01 result that usually is quoted in scientific papers. Or if you have a family that was in close contact, test everybody. It's very safe to assume everybody caught it if anybody had it. If everybody tests positive or everybody tests negative for antibodies, the false result level drops exponentially depending on the number of tests.

It gets a little more complicated when repeated tests have different results. This might show an increase or decrease of antibodies over time. Or it might show a very low level of antibodies that is near the threshold of the test. In any event, repeating the test filters out the random element in test failure.