Now consider this same phenomenon-a higher chance of false positives than of real ones-applying to a large group, or even a whole country. The way I was using the test, any positive result was nearly certain to be wrong. What this meant is that my chance of a correct positive when I took the test was also essentially zero, while my false positive chance remained 2% like everyone else’s. The second source of trouble I didn’t anticipate is what is known as “pretest probability.” As I said, I don’t socialize, so my probability of actually having covid in first place was very low, maybe even zero. By the time my review of the home tests was complete, I’d tested five times in two days, accumulating 1 in 10 odds of being told I had covid when I didn’t (a 2% chance of a false positive each time, multiplied by five tests). The first way is through repeat testing, the kind I did.

What I didn’t realize-and what your everyday CVS shopper won’t either-is that there are two ways that less-than-perfect specificity can get amplified into a bigger problem.

For the home tests I tried, that figure is about 98%, with a corresponding 2% rate of false positives. The tricky part of unrestricted testing, I learned, comes instead from the concept of “specificity,” or the rate at which a test correctly identifies negatives. But if the alternative is no test at all, then none of those infections would be caught. That is, they catch about nine of every 10 infections, a metric called the test’s “sensitivity.” Some people have said that any missed cases are a worry, since a person with a false negative could go out and infect someone else. The issue with home tests is accuracy, which is between 85% and 95% for detecting covid.

Maker of covid test dr roger hodkinson