This chart was done by Youyang Gu, who does great analytical work on the epidemic and can be found on Twitter, contains his estimates of the current seroprevalence or actual total infected people in each state. He has an algorithm that he uses to estimate this. Note that for Minnesota he is estimating around 14%. Wisconsin has a much higher estimate of 21%. And look at the Dakotas up there at the top. I try to avoid predicting because I really have no idea about the actual impact of population immunity, contact patterns and pre-existing immune protection on the epidemic. But I believe transmission has to be slowing substantially in those largely rural states and I wouldn’t be surprised that they have peaked and will begin to fall. Minnesota is likely only a week or so behind. You look at the curves and it is clear that interventions don’t make much difference. For whatever reason, when CV-19 gets going, it just ramps up incredibly fast, runs out of steam in a few weeks and plummets down to a long tail til the next surge. This is a highly, highly transmissible pathogen and I really believe the infectious dose, especially for anyone who is vulnerable, is very low. And why it slows, I don’t know either. Fewer targets obviously; attacking the most vulnerable with visible disease first; behavior changes, all likely play some role.
In any event if that 14% or so number is right for Minnesota, we have had about 800,000 cases, or over three times the current reported total cases of 257,000. That would be down quite a bit from the one in ten cases the state and the CDC were telling us were detected in the spring. I suspect it is close, I can’t imagine that we are missing 90% of cases with the testing levels we currently have. Helps me think about how to “normalize” cases for testing over the course of the epidemic as well. Could do a sliding scale over time going from ten times reported cases in the spring to three now.