1/7 Lots of folks trying to understand sensitivity and specificity. Maybe the most important concept to understand right now is that of POSITIVE PREDICTIVE VALUE.

Given a test result is positive, what are the chances you were actually infected?

Biostats 101 primer on the topic.

2/7 There are two types of positives for antibody tests:

- True positives (TP), which occur when someone has been infected, and the test gets it right.

- False positives (FP), which occur when someone has not been infected, and the test gets it wrong.

3/7 Positive predictive value (PPV) is simply the proportion of positives that are true positives:

PPV = TP/(TP+FP)

This depends on how good the test is at detecting infections (sensitivity), and how good the test is at accurately classifying non-infections (specificity).

4/7 In practice, PPV is the value that matters most to patients and clinicians. If a test says you have antibodies, is that result reliable?

Importantly, PPV depends not just on sensitivity and specificity, but also on the prevalence of infection.

5/7 An easy analogy is to cancer screening with prostate-specific antigen (PSA). This biomarker is pretty unreliable, but sometimes it's used in high risk groups (men with a family history of cancer). That is because we expect more true positives than in a low risk population.

6/7 In a low risk population, false positives quickly outnumber true positives that the results are very difficult to interpret. Even with a positive test, you're more likely than not to be negative. This all has implications for the use of antibody tests for "immune passports."

7/7 Many antibody tests have flooded the market. Highly specific tests (that yield few false positives) are important here. Coupled with other findings (known symptom history, CT scan), we might trust results more, but we still have a ways to go. END

Addendum: My old thread on sensitivity/specificity as applied to detecting acute infection. https://twitter.com/nataliexdean/status/1238518203736903680?s=20