This is kind of old news, but I hadn't heard of it till now, and I'm quite sure that most others haven't, either. It seems that the CDC has been, or was, mixing PCR virus tests with antibody tests in their reported numbers for some states. This is a huge f-ck-up, because they test very different things, and if they're conflated, neither one becomes reliable. The PCR test tells us who is infected right now, and is supposed to be the source for all listed confirmed cases. The antibody test tells us who may have been infected in the past.
One problem with conflating them is that the confirmed tests are inflated, since far more people have been infected than get tested by PCR, and even worse, most of these are past infections, which means that current infections are in effect being double-counted to some extent. At least some of the people who have tested positive for antibodies will also have tested positive for the virus, so both tests enter the records, though they really refer to just one individual.
Another problem is that since most people who are tested for antibodies are not considered at risk, adding these data will reduce the positivity rate, or % of tests that are positive. Back when this problem was first reported, 15-20% of virus tests in the U.S. were reported to be positive. But if a large fraction of those tests were for antibodies, which at that time would typically be 5% positive, the result would be greatly underestimating the proportion of people who were actually testing positive right then. The real number would be much higher than 15-20%.
Since this is from an article published more than a month ago, i assume the problem has been addressed. But it has implications for our current situation. If some states had been underestimating the proportion of people testing positive in the past, then these states were even less ready to be re-opened than was thought.
https://www.theatlantic.com/health/...-test-data-pennsylvania-georgia-texas/611935/
I first learned about his from this current article:
https://justthenews.com/politics-po...ronavirus-data-punishes-testing#digital-diary
But the author doesn't really understand the situation. First, he seems to assume that the conflation of virus and antibody tests are continuing, which I'm pretty sure is not the case. Second, he argues that all these antibody positives and double counting are responsible for the surge in cases now being reported in many states. But even if that were the case--that the results of the two tests are still being mixed--that wouldn't be good news, because it would mean the positivity rate, which is very high in some states, is being greatly underestimated. In fact, this would completely undercut his first claim, that we're testing at a much higher rate now. If many of these tests were for antibodies, it would mean we're actually not testing at a higher rate.
There are other problems with this article. His main point is that Europe in unfairly blocking travel from the U.S. Like some other authors, he argues that the death rate of several major European countries is higher than that in the U.S. But in the first place, that doesn't really matter. Europe is not restricting travel from the U.S. because it's afraid that visitors will die over there. It's doing it because visitors will bring infections over there and spread them, at which point the death rate in the U.S. will be totally irrelevant.
The second problem is that all the European countries with a higher death rate than the U.S., except Sweden, have a much higher population density (3-10 times) than the U.S. As I've pointed out here before, population density is highly correlated with case rates and death rates. Sweden, the exception, is one of the few countries in the world that never had a lockdown, and it's clear that this failure in policy has contributed greatly to its deaths.