Question: The other day when you brought up the issues related to “immunity certification”, you mentioned a couple of “science” questions related to antibody testing — “What is the sensitivity and specificity of the tests? How comfortable are we with people being told they are immune and they actually aren’t?” Would you please elaborate for us non-public health people?

Answer: Oh, thanks for asking! Sensitivity and specificity are epidemiology terms that sound more complicated than they are. The concepts are about measuring the precision of a diagnostic test. Sensitivity is the ability of a diagnostic test to capture everyone who has the condition. Specificity is the ability of a test to identify everyone who does NOT have the condition. As sensitivity of a test increases, specificity decreases and vice versa — it’s a trade-off between the two. Copied below is a little table they use to teach us these concepts in epidemiology class (Table 1).

Now, just because a test has high specificity does not mean it’s a good test. You could have a totally bogus test tell *everyone* that they are positive for a disease and the test will be 100% sensitive — it captures everyone who actually has the disease, but it’s worthless because it tells *everyone* they have the disease. So, an acceptable test must be both reasonably sensitive and specific. But there’s yet another sticky wicket here. The lower the prevalence of a given condition, the higher the number of false positives we’ll find (see Table 2). In the hypothetical examples I’ve made up in Table 2, if 10% of a population of 1,000 actually has COVID antibodies, the antibody test will incorrectly identify 90 people as having antibodies (false positive). But if 50% of a population of 1,000 actually has COVID antibodies, only 50 people will be incorrectly identified as having antibodies. Since scientists still believe that a preponderance of our population has NOT been exposed to COVID and therefore does not have COVID antibodies, using antibody tests to issue “immunity certifications” could be exceptionally problematic as a number of people will mistakenly believe that they have markers of immunity that they actually do not have. Imagine all these people taking fewer health precautions, visiting nursing homes, and spreading disease. So many problems!

All that said, antibody testing can be extremely beneficial on the population-level rather than the individual-level. For example, we want to know how far the coronavirus has spread, so we need to know the proportion of people who have been infected. This is a very different use than the individual-level use that folks supporting “immunity certificates” are promoting. Finally, if you want to learn more, NPR had a really good write up on this published just today! Also, if this kind of stuff excites you and you want to learn more, check out Johns Hopkins free open course, Fundamentals of Epidemiology.

Table 1. Sensitivity and Specificity

Table 2. Put into Practice