Dang - coronavirus testing numbers, at least in the U.S., are basically a useless crock.
State wide numbers in at least several major states, including Virginia, Pennsylvania, Georgia, and Texas, as well as the CDC overall, are simple summations of
- viral tests (Do you have an active infection now?), with
- antibody tests (Have you ever had it?).
Furthermore, the antibody test is notoriously unreliable, so only good for estimating herd immunity, not for guiding individual treatment.
But, by blending these two numbers together, indiscriminately, one gets numbers that, like in Texas now, will never come down. If say about ten percent of the population has already contracted COVID-19, and if most of the tests are the antibody test, then at least about ten percent of the test results reported will be positive, “forever”.
(It is a bit understandable. As someone with the math and computer background and skills who could competently analyze such data, I for damn sure would not expect anyone with such skills to want a job at a state, federal or CDC organization. There are way too many job openings in businesses that pay better and provide much better opportunity to enjoy one’s work. Such statistical jobs are one of the top jobs in demand.)
If you’re feeling too chipper, you can get a quick depressant, reading the details at https://www.theatlantic.com/health/archive/2020/05/cdc-and-states-are-misreporting-covid-19-test-data-pennsylvania-georgia-texas/611935/?utm_source=pocket-newtab