Post #1076: Virginia’s COVID-19 hospitalization rate: Home alone?

Posted on March 26, 2021

This is about the rising hospitalization rate of diagnosed COVID-19 cases in Virginia.  It’s about new hospitalizations per newly diagnosed case.

This posting is long enough to require an executive summary.  Briefly.

  • Hospitalizations for COVID-19 are concentrated among the elderly.
  • The age of Virginia COVID-19 cases has been dropping.
  • That trend toward “a young person’s pandemic” sped up rapidly with the onset of COVID-19 vaccination, focused first on vaccinating the elderly.
  • And yet, the rate of COVID-19 hospitalization per new case has risen sharply.
  • How sharply?  Among the oldest old, half of those newly diagnosed with COVID-19 are now being hospitalized for it.

Why has the hospitalization rate per new COVID-19 case skyrocketed among the oldest old?  My guess is that the un-vaccinated geriatric population has a high fraction of individuals who had difficulty getting the vaccine.  And I’ll bet that’s because a lot of them are living alone, on the margins of living independently.  That population — elderly, living alone, no formal caregiver — is the population that physicians will choose to hospitalize, for safety’s sake, rather than let them try to ride out a moderate illness at home.


Even as the elderly have been getting vaccinated in Virginia, those in that age group who do get infected with COVID-19 now have an extremely high likelihood of being hospitalized.  What do I mean by extreme?  For the oldest old (age 80+), half of newly-diagnosed COVID-19 cases in Virginia end up in the hospital.

I rarely get a result that makes me stop and say, that can’t possibly be right.  Or, at least, I rarely get one that doesn’t turn out to be an outright mistake.  But I’ve checked, and that’s right. It’s obvious if you just scan down the raw data.  If you know to look for it.

If I had to guess, I’d guess this is a behavioral effect.  I don’t think the virus has become more virulent.  Or that the practice of medicine has changed.  I think it has more do with the people who did and didn’t get vaccinated than with any change in the virus or in medical practice.

Among the elderly — those who have had excellent access to vaccine — new COVID-19 cases are now being concentrated into the small portion of the population that didn’t get vaccinated.  Either didn’t bother to, or for some reason, was unable to.  And I have to suspect that that residual, un-vaccinated population is quite different from remainder of the population.  And that’s what’s driving this.

And if that’s right, I think there’s a broader lesson here.  I’ve noticed that COVID-19 infection rates are falling among the elderly, but they aren’t falling in proportion to the fraction of the population vaccinated.  With (say) two-thirds now vaccinated, we haven’t seen the COVID-19 infection rate drop to one-third of its former value in that population.  Maybe that’s just due to time lags, or the fact that not everyone is fully vaccinated.  But I think that, in part, that may be due to the residual, un-vaccinated population being systematically different from those (with access to vaccine) who have gotten vaccinated.

All of that is pure guesswork.  Let me now report the facts.  And then formalize those guesses a bit.

The issue:  Why aren’t COVID-19 hospitalization rates falling?

The issue here is that coronavirus has become a pandemic of the young.  The age of the average case began shifting downward, slightly, in Virginia, as far back as June 2020.  And now, here and elsewhere, that shift is accelerating through a combination of high rate of spread among young adults, and high rate of vaccination among the elderly.

It’s well established that hospitalizations for COVID-19 are concentrated among the elderly.  As a consequence of the shift in the age of the average case, all other things equal, you’d expect to see a decline in hospitalizations per new case.

In addition, but less importantly, once you are vaccinated, if you get sick, you are supposed to have a vastly lower chance of getting severely ill.  I say that’s secondary because I expect to see few symptomatic COVID-19 cases among the vaccinated.  In effect, the vaccinated population largely drops out of the analysis, due to what ought to be a very low rate of diagnosed COVID-19 cases among that population.

And yet, here’s a graph of new COVID-19 hospitalizations per new COVID-19 case, seven-day moving averages for both, for Virginia.  The upshot is, the hospitalization rate of newly diagnosed cases is rising rapidly.

The first graph starts 1/1/2021, the second graph starts 5/1/2020.  The second graph shows how steep the runup is, against what had been a long-term secular decline in the hospitalization rate.

By age.

Well, the argument was that the younger age of the typical case should be driving down the hospitalization rate.  The obvious thing to do is split this by age.

Same deal as above.  First graph is short run (since 1/1/2021), second graph is longer-run perspective (since 5/1/2020).  The dip in the line for the oldest old (circa 3/10/2021) is clearly a data reporting glitch.  (It first showed up in my plots of new cases by age in just-prior posts).  Note the scale.

Just from the timing of it, and the strong impact on the oldest old, you’d have to guess that this is connected with COVID-19 vaccination.  (And remember, this is all per-case, not per 100,000 residents).

And my only guess is that the people who are left over, after the vaccinations are done, are systematically different from the rest.

Let me go on an do one further bit of speculation.  There are two obvious competing hypotheses about why the un-vaccinated elderly remain un-vaccinated.  Roughly speaking, butt-headedness versus frailty.

By butt-headedness, I mean what is now politely termed “vaccine hesitancy”.  Which, of course, is a nearly-complete misnomer.  All those Republican men who refuse to get vaccinated, there’s no “hesitancy” about that decision.  All the nutsos who believe in fill-in-the-blank craziness, ditto.

So, one hypothesis is that you’ve just got a bunch of old people who shake their heads and say, nope, no Bill Gates microchip for me, thanks.

But if that were true, there wouldn’t be any particular reason for an increased rate of hospitalization per case.  There might be more cases than you’d expect, if they were equally in denial about (e.g.) mask use.  But no reason to suspect a greater need for hospitalization.

The other hypothesis is frailty or lack of mobility as a driver of the non-vaccinated elderly population.  And here, given what I know from having studied hospitalization, what we’re probably really talking about is old people who are living alone.  But who are just at the edge of being able to cope with that.

I say that because that would, indeed, drive the hospitalization rate up.  In Medicare, the single largest determinant of whether an elderly person will require post-acute care, following hospitalization, is whether or not they live alone, versus having somebody in the house who can take care of them.  (Well, that, and whether or not the hospital owns an acute-care facility.)

And so if Mr. Jones, age 80, comes down with a moderate case of COVID, the likelihood of hospitalization will depend strongly on whether or not he’s living alone and living independently.  If Mrs. Jones is still around, still mentally competent, and still able to drive, it’s safer to keep Mr. Jones out of the hospital if possible.  Presumably Mrs. Jones will act if Mr. Jones goes south.  But if Mr. Jones is alone, maybe can’t drive, and maybe isn’t as sharp as he used to be, there will be a greater propensity to hospitalize him, for safety’s sake.

And that sort of person — living alone, with some minimal assistance perhaps, but not really fully capable, possibly no longer able to drive — that’s just the sort of person who might find it difficult to get vaccinated.

Anyway, my bottom line is that maybe the un-vaccinated elderly population has a significant fraction of people who live alone.  And that’s the driver here.

Pure guesswork.

In any case, as vaccination has progressed rapidly among the elderly, the hospitalization rate for those who do get infected with COVID has shot up.  I’m not quite sure what that means.  I know it was totally unexpected.  But that’s what explains the apparent paradox of increasing hospitalization rates overall (new hospitalizations/new cases), despite a falling average age of the typical new COVID-19 case.