Fun With Numbers in Minnesota, Part 2

By October 13, 2020Commentary

Testing has reached phenomenal levels in Minnesota.  To what end I am not sure.  Supposedly all this testing is to identify asymptomatic cases, contact trace them and quarantine the person and their contacts before they infect others, although the research is quite mixed on how much asymptomatic transmission exists.  Just like the mask mandate, the testing and contact tracing program isn’t making one bit of difference, other than likely identifying a few false positives and a lot of low positives.  We have seen no reduction in community transmission of unknown origin.

I actually got a real table together for this one, to show you both how much testing has increased and what cases might have looked like if we had the same level of testing earlier.  I am going to take a couple of weeks near the peak in May, a couple of the relatively complete recent weeks and a random week in other months.  You can see the steady rise in testing.  I calculated the positivity rate, the percent of cases in relation to tests.  The last week I used in September probably is pretty complete on testing and cases but a few more could dribble in. (I have a separate analysis I will share on the lag or completeness factor.)

I then took the rough average of weekly tests for the September weeks, which was 148,500.  I applied the positivity rates for the weeks other than the September ones, to get a number of cases we would have seen if we had been doing the same amount of testing as we did in September.  If you have qualms about whether we would have had the same positivity rate with a larger number of tests, remember that the CDC and the state were estimating ten times the actual number of cases to what was detected, so my method is actually significantly below that number.  What the analysis does show is that cases would have been far higher in May than they are now if we were doing the same amount of testing then.  It also means that the hospitalization rate and case fatality rate were much lower than it appeared.

More to come.

Week StartingTestsCasesCases as % TestsCases at Sept. Testing
April 2728245392613.90%20641
May 431779415713.10%19453
May 114653044689.60%14256
June 228068331414.70%6979
July 2710176547664.70%6979
August 2410495753075.10%7573
September 1413107963074.80%
September 2114783072684.90%
September 2816676472194.30%

Leave a Reply