Read on Twitter
Here's my take on the new MN COVID model: 1) the data is very old. Missing about 25-30% of the pandemic. Data ends April 25. 3 month pandemic, missing latest 3 weeks of data? REALLY?! @StefanG_HEP @MDHCommMalcolm @GovTimWalz @thauserkstp
2) By ignoring "hotspots" and smoothing the results over the population, you ignore that 1% of population accounts for 80% of deaths. No amount of smoothing will give you good model of entire population. 99.25% of deaths come from a group that represents a tiny fraction of MN
3) Because of #2, @StefanG_HEP has implicitly assumed complete failure of 5 step plan for LTC facilities. If it works, 80% of deaths go away. 80%! Suddenly the MN population is wildly different from model. Totally ridiculous. Any 12 yo can see that not taking into account...
Not taking into account hotspots ignores completely, entirely, 100% that mortality in COVID is DRIVEN BY HOTSPOTS. People don't randomly die from COVID, so assuming MN population characteristic as if we were mostly nursing home residents is NUTS. @MPRnews
