An Epidemic of Certitude and Games with Models, 6/30/20

COVID19-hospitalizationsandprojections-063020-featured

I dipped my (non-COVID) toe in a local Facebook group’s discussion of COVID-19 trends this morning, and one takeaway is that many people are absolutely certain they know what’s going on.

There is a lot of talk about how facts and science are not matters of opinion, but with a large gap between what the facts show and the claims being made on that basis.  For example, point out that in some places where cases of COVID-19 are on the increase, deaths are actually decreasing, and the response is apt to be:  “Oh, that’ll change! There’s a lag between cases and deaths!”  Well, maybe.  It depends who is being tested, who is catching the disease, what capacity and treatments the local hospitals have, and so on.

The reality is that we won’t be able to come to anything approaching conclusive judgment about what states did well and which did poorly until after the pandemic and the resulting economic downturn have run their courses, and even then there will be areas of debate.  Even if we don’t want to talk about economic trade-offs, we might find that some places that look like danger zones now will have lighter second waves.  We’ll see.

One comparison that came up on Facebook is worth a closer look.  Somebody asserted that Sweden’s approach didn’t work out.  Again, maybe.  Take a look at the data and compare it with, say, Massachusetts.  With smaller population, the latter has had significantly more deaths.

But let’s just focus on trends.  Massachusetts was seeing about 1,500-2,000 new cases per day back in April and early May, which has now dropped to the 200s.  Sweden’s daily new cases hovered in the 500-800 range until June, and now it’s up more in the 1,300-1,500 range.

Now look at daily deaths.  Sweden has seen a reasonably straight downward trend from 90-100 per day in early April down to around a dozen.  Massachusetts, over roughly the same time period, went from around 170-180 and is now fluctuating around the 30-deaths line.  I’m sure there are all sorts of factors that would help explain these differences, but one possibility is that the massive shutdown in Massachusetts had the effect of stretching out deaths.  In that case the lock-down didn’t protect the vulnerable, and the disease is still finding them.  Maybe the herd immunity idea was preferable.  Again, we’ll see.

We’ll also see whether Rhode Island Governor Gina Raimondo’s gut decision to renege on her promise to allow social gatherings in the 100s during Independence Day weekend proves justified, but at the moment the data suggests not.

New cases remain limited.  Daily deaths remain in the low single digits.  Intensive care units continue to open up.  And hospitalizations only held steady because the number of discharges was limited, which probably has to do with the fact that the state is now counting people who test positive for COVID-19 no matter why their in the hospital, so they haven’t been discharged because they’re in the hospital for some other reason.

COVID19-hospitalizationsandprojections-063020

(See here for my original methodology and here for a subsequent modification I made. A thorough explanation of the chart is included in this post.)

Projections versus actuals (date of report).

  • Cases:
    • Projection for 6/30: 16,803
    • Actual for 6/30: 16,813
    • Projection for 7/1: 16,851
  • Hospitalizations:
    • Projection for 6/30: 69
    • Actual for 6/30: 74
    • Projection for 7/1: 72
  • Deaths:
    • Projection for 6/30: 948
    • Actual for 6/30: 950
    • Projection for 7/1: 952