It has been non-trivial (at least for me…) to find much information on test characteristics for Covid-19. In particular, your kid (or you, or…) is exposed-do you get a test ASAP? Or do you wait for the viral load to increase to avoid a false negative? Is that even the right question? And then, what do you do with the result? These are all FAQ's, but it's hard to find the answers. Here's an article that, to me, is very helpful: https://www.acpjournals.org/doi/10.7326/M20-1495
Variation in False-Negative Rate of Reverse Transcriptase Polymerase Chain Reaction–Based SARS-CoV-2 Tests by Time Since Exposure
Annals of Internal Medicine, Aug. 18 2020
It's open-access. Spoiler alert: if I'm reading this correctly, the false negative rates are atrocious. A negative test doesn't mean you're not infected. Over the course of days, the lowest false-positive rate was 20%. Russian roulette with a 5-chamber revolver.
This is a meta-analysis with all the inherent pitfalls. Also, we all know that the positive and negative predictive values of a test depend on both test characteristics (sensitivity and specificity) but also on disease prevalence. It is hard to define prevalence since it is always subgroup-specific, and what you really want is a pre-test probability. Population disease prevalence is the right pre-test probability only for a random sample from an undifferentiated population.
From a public health & transmission standpoint, testing with these marginal test characteristics would be useful. For individual patients and personal decisions, not so much.
The authors also touch on a critical aspect of all test characteristics-the degree to which re-testing improves test quality depends crucially on the degree of correlation between measurements (i.e., why was the test incorrect in the first place?). The question also arises of the relationship between false negatives and infectivity. This paper is a good read.
I would be very interested in hearing of other studies and other perspectives!
------------------------------
Richard Spencer MD, PhD
National Institute on Aging
Baltimore, MD
spencer@helix.nih.gov------------------------------