Games with Models, 5/20/20 Data


A notable change in today’s COVID-19 report from the RI Dept. of Health is that hospitalizations were up again, which would be very concerning except… The number of people newly admitted to the hospital was the second lowest it’s been since the end of March.  And the number of deaths in the hospital has remained about the same.

The reason hospitalizations are up is entirely that very few people diagnosed with COVID-19 were discharged — only 6 on Monday (the latest day reported), when the average for the past two weeks is 25 per day. If these numbers aren’t just in error, why the difference?

One explanation would be that, somehow, people in the hospital with COVID-19 now have more-serious cases than before, but then one would expect in-hospital deaths to be going up, too.  Not only that, but the number of COVID-19 patients in intensive care is going down, which definitely doesn’t jibe with the notion that they’re more serious.

Maybe hospitals are lengthening the stays of COVID-19 patients for some other reason, but I don’t think that’s what’s going on, here.  Remember that the hospitalization number now counts anybody in the hospital who tests positive for the virus, no matter their reason for being there.  With the governor now encouraging people to return to hospitals for elective surgeries and other things, it could be that their stays are getting longer because they’re not in the hospital because of the virus.


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

Projections versus actuals (date of report).

  • Cases:
    • Projection for 5/20: 13,076
    • Actual for 5/20: 13,356 (includes a large number of revisions to prior days)
    • Projection for 5/21: 13,492
  • Hospitalizations:
    • Projection for 5/20: 236
    • Actual for 5/20: 257 (includes a large number of revisions to prior days)
    • Projection for 5/21: 241
  • Deaths:
    • Projection for 5/20: 537
    • Actual for 5/20: 538
    • Projection for 5/21: 543