Games with Models, 5/27/20 Data

COVID19-hospitalizationsandprojections-052720-featured

Here’s another milestone with today’s COVID-19 data release from the RI Dept. of Health:  Hospitalizations (lagging case and death data by two days) fell below 10 newly admitted patients for the first time since the end of March.  Remember that this includes people admitted for other reasons but testing positive for COVID-19, as well.  That brings people in Rhode Island hospitals down to its lowest number since April 9th.

Turning to fatalities, the daily toll yesterday continued the streak below 10.  And for total cases recorded, the 14-day infection rate fell below 20%, meaning, basically, that every five people who had the illness two weeks ago infected only one person between them.

COVID19-hospitalizationsandprojections-052720

(See here for my original methodology and here for a subsequent modification I made.)

Projections versus actuals (date of report).

  • Cases:
    • Projection for 5/27: 14,314
    • Actual for 5/27: 14,353
    • Projection for 5/28: 14,449
  • Hospitalizations:
    • Projection for 5/27: 219
    • Actual for 5/27: 218
    • Projection for 5/28: 212
  • Deaths:
    • Projection for 5/27: 639
    • Actual for 5/27: 655
    • Projection for 5/28: 659

Note that the Dept. of Health revises numbers going back days, so for example, the actual number of deaths may be 16 higher than projected, but that doesn’t mean that many more people died in one day.  Rather, it can mean that the baseline from which the projection was made turned out to be higher.



  • Mario

    201 new cases, 7 deaths, and hospitalizations stay at 218 (+19, -15, -4). I just can’t agree that we can say anything about the amount of infections based on the number of known cases. I have tried to figure out the number of total cases separately, and new cases as a percentage total cases seems to have been falling behind since about May 11th, and hasn’t recovered. We are getting maybe 2.5% per day, where we were getting 3.5% in mid-April. I don’t think we’re testing enough people, or maybe just the right people, to actually bring the infection down rather than letting it slowly burn itself out. Now I have the deaths ending on July 14th at 862, but that date keeps moving forward and the number keeps going higher, so there’s no reason to think we’ll get away with fewer than a thousand. Maybe it’s just me, but it seems crazy high.

    If we keep going like this it will eventually get better, but think the recovery could be sped up quite a bit with more aggressive testing, particularly dragooning random people. I think the main reason why the positive rate went down is because we exhausted our supply of the easily tested rather than being evidence of a falling infection.