I don't get "positivity rate" as a metric. It seems like there's so much variance in the denominator -- how many tests there are, access to tests, whether people feel the need to get tested, etc. I don't really know what it's telling us. What am I missing?
So you are right, you have to know a bunch of other things to put it in perspective, but its mainly a guide first to how sensitive your surveillance system is for cases and then if all else stays stable (testing rates) it is a guide to disease activity. I am working on COVID-19 in Australia and monitoring stats like this. You can look at total cases and total deaths to understand the sensitivity of your surveillance system for cases and then look at positivity. If you have a high case to death ratio (setting aside clinical performance/or adjusting for clinical performance), you have an indication of sensitivity, but you also know that if you are testing adequately, i.e. testing all the mild cases then the % positivity should be low. In Australia we would be very concerned with a 3% positivity, suggests not enough testing is happening. Over the last 2 weeks our % positivity in NSW has been 0.002% and 0.009%. I know WHO recommends that % positivity should be less then 5% but that is a very low bar for low income countries. High income countries should be able to keep it well below 1%.
My understanding is that it mainly tells us whether or not we are administering enough tests, it's not a direct measure of spread due to the factors you mentioned.
In California's case we've had high sustained levels of testing so we can also infer that were reducing spread as the positive test rates drop but that doesn't seem as strong of a signal as other metrics.
