dreamshark: (Default)
[personal profile] dreamshark
I don't entirely understand why, and with the combination of protests and "re-openings" I suppose we may turn back the other way, but the Minnesota stats for the past week are just what you want to see at this point. Most importantly (IMHO) the number currently hospitalized for COVID peaked on May 28 at 606 and has done nothing but decline since. The number currently in ICU has declined proportionately (198 today), so this does not suggest that the criteria for hospital discharge has changed substantially. Rate of arrival of new cases is clearly declining also, although it varies so much from day to day that you have to zoom way out to see the trend. Daily new cases peaked May 18-21 at about 965. Now we're seeing roughly half that.

Over the same period, our rate of testing kept going up but the percent of tests returned positive went steadily down (which is important for some reason, but I'm not sure I grasp why). That percentage had been 10-11%, is now down to 8%. One of the criteria for re-opening is for that number to be below 10%

See for yourselves, if you're interested.

Date: 2020-06-08 09:05 pm (UTC)
From: [personal profile] quadong
I am also a little fuzzy on why the positivity rate is taken to be a good metric. I think it's something like: if it's high, it means you're only managing to test the desperately ill, so if it's low, you're probably flooding the system with enough tests to catch most of them, which means that the total number of cases is probably similar to the number of known cases. Which is fine as part of the picture, but not as the most important metric. Maybe I'm missing something.

Date: 2020-06-08 09:34 pm (UTC)
sraun: portrait (Default)
From: [personal profile] sraun
The initial percentages being high meant we were low on tests, and only using them as confirmation. As long as the percentage is going down, that means we're testing more people, and getting a better handle on how many cases there really are. When the percentage stabilizes, that's a pretty sign of how many cases there really are.

The test numbers can also be used to approximate R0, the 'infect others' rate, but you need to know you've reached the point above first.

Profile

dreamshark: (Default)
dreamshark
April 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 2026

Page Summary

Style Credit