Post #788: Virginia universities and COVID-19: A more systematic look

Source:  Analysis of data found in Wikipedia.

In Post #786, I looked at the low risk of infection and hospitalization faced by a typical community-resident college-age person in Virginia. Of the things that might result in hospitalization of a person that age, in Virginia, coronavirus hardly makes the list.

I then went on to look at infection rates at a handful of Virginia universities, to judge the amount of additional risk (if any) and generally to see how they were dealing with coronavirus.

Of the ones I looked at, only Radford stood out as appearing to have an outbreak. They’re aware they have an issue and are taking some steps, starting with limiting all gathering to 10 or fewer people.  Edit:  And they’ve shut down a fraternity, suspending eight members, for some spectacularly stupid behavior.

That last post wasn’t anything like a systematic analysis of Virginia schools.   For one thing, the list of universities in the prior post was a “sample of convenience”, meaning, it was whatever I had some interest in and could easily find the information for.

In this post, I get a little more systematic.  I start with this Wikipedia page, a list of universities and colleges in Virginia, and work my way down it in order of enrollment, stopping when enrollment dropped below 5000 students.

It’s still not quite as orderly as I would like, because there’s a lot of variation in what colleges and universities are doing.  But here is what I gleaned from reading through the re-opening plans and COVID-19 dashboard of those schools.  I’d say that the main finding is that, so far, this appears to be working out OK in most colleges and universities.

Detail follows

Outbreaks?  No other general outbreaks to speak of, yet.  Of institutions with 5000 or more enrollment, the only other thing I saw that looked like “an outbreak” was that Virginia Commonwealth University had an outbreak (a cluster of cases) on their athletic teams.  A total of 44 athletes tested positive for COVID-19. Next closest was a large university that tracked 11 cases to a single large party.  Otherwise, so far, for those universities that are reporting infection rates, things look fairly calm.  Near as I can tell, the only university to have “pulled back” from their original fall plan so far is Radford University.

Take that for what it’s worth, given that it’s not yet September 1, but I found that kind of heartening.  It does seem to be possible, in almost cases, to restart college classes and college life without an immediate crisis-level COVID-19 outbreak.  How robust that is, I guess we’ll find out.  But so far, so good.

But not all residential institutions are reporting infection rates.   Notably, the two big Christian universities (Liberty and Regent) report nothing about COVID-19 infections in their student bodies.  If there were outbreaks within the various non-reporters, we’d have no way of knowing it.

Let me quantify that.  Of academic-degree institutions (i.e., not technical schools), that are not community colleges (commuter schools), and have 5000 or more enrollment, and have not gone to all-on-line classes, I find that about two-thirds of them are actively reporting their COVID-19 infection rate.

None of the commuter-only schools (i.e., community colleges) is testing for COVID-19.  None of those schools is reporting infection rates.  At least, none of the half-dozen large ones that I checked.  Some of them have gone fully to on-line classes, but otherwise, I’m not sure what to make of that.  Presumably that follows state guidelines.  And so, presumably, the state must think that risk of transmission within classroom settings must be fairly low, or at least, low compare to the risk of transmission within dorms and so on.

Virginia created voluntary guidelines for college reopening They are available on this web page In particular, the guideline for re-opening college campuses can be found here (.pdf), and the guideline for testing the student body can be found here (.pdf).

These guidelines don’t require much in the way of testing or reporting. Virginia does NOT recommend testing all students, or even testing a sample of students.  They don’t even recommend testing all incoming students before they arrive.  The reason given is the potential for false positive and false negative test results.

Not all institutions are testing students before they return to campus.  Some (e.g., William and Mary) tested everybody.  I’d say that the bare majority of large residential institutions did NOT test students prior to return to campus.  One (Old Dominion University) tested 15% of returning students.  One (Longwood University) was very clear about why they weren’t testing prior to return, and that most centered around the difficulties of administering that testing on a timely basis.

The most common policy is to test symptomatic individuals only, which is, in fact what the Commonwealth of Virginia guidelines suggest.  Only a few colleges appear to have a goal of testing all students.

Some but not all colleges are testing a random sample of students on an ongoing basis.  Mostly, it’s a “voluntary” random sample, whatever that means.  Presumably, this is to keep an eye on the true prevalence of active COVID-19 cases on campus.  (And, honestly, the most common sampling rate mentioned — 5% — didn’t seem anywhere near large enough to allow prevalence to be tracked, given how few cases there are at present.  I believe that the Commonwealth does not recommend routinely testing anyone, including routinely testing a random sample of students.

Many or most institutions are routinely testing (screening) all athletes, regardless of their testing policy for other students.   Apparently the term of art is “high contact groups”, meaning, I guess, individuals who have no choice but to be in close contact with others.  The upshot is that the typical university testing policy explicitly acknowledges the risk of disease transmission in college sports.

Looks like the median policy is “some on-line classes”, but there is a lot of variation.  George Washington University is all on-line, University of Virginia delayed opening and is still considering going all on-line.  Several (e.g., Norfolk State University, University of Mary Washington) have the first two or three weeks of classes on-line only.  William and Mary has upper classes returning late and taking the first two weeks of classes on line.  Many of the community colleges seem to have gone to all or nearly all on-line.  Some colleges (e.g., Christopher Newport University) the mix of on-line and in-person classes was explicitly dictated by space constraints.

Several schools (U.VA., ODU) delayed the start of classes.  (This is as opposed to other universities that started classes on time, but on-line).  I’m not sure what these schools expected to gain, other than to see whether other campuses of other schools immediately blew up with high infection rates.

Pooled testing is non-existent.  As an aside, nobody seems to be doing what I consider to be the obvious thing, which is scanning the student population on a continuous basis using pooled testing.  (See Post #605, Post #731).  I’m not quite sure why, because dorms/lectures, where every student has a uniform ID card, would seem to be an ideal setup for that.  Possibly the burden of doing that (you do have to track who is in each pooled sample) is excessive, possibly pooled testing has not been approved in the Commonwealth even on an emergency basis (Post #761).  Whatever the rationale, I think this is a missed opportunity.

A few random notes:

The diversity of testing protocols makes it hard to interpret any COVID-19 case counts.  Screening an entire cohort — testing every individual — is going to reveal more cases than testing symptomatic individuals only.  (If for no other reason than the fact that a significant fraction of cases appear never show symptoms, particularly cases among younger people.)  By contrast, screening only those who are symptomatic should yield lower rates of known infection.  And that largely appeared to be true.  And so, I attributed the higher know infection rate at some colleges to the decision to screen everyone, and not to some instantaneous outbreak at the start of classes.  To be clear, that’s an untested assumption on my part, but it seems to make sense.

Obviously, the lack of reporting from some schools makes it impossible to draw blanket conclusions about the overall infection rates.

Schools that have a mix of live-in and commuter students muddy up the statistics.  Near as I can tell, the typical approach is to test all students, but test the commuters at a lower rate.

The reason I couldn’t find a COVID-19 dashboard for U.VA. in my last post is that they hadn’t put it up yet.  Per this newspaper article, they finally did that yesterday.  The dashboard is at this link.  They have delayed the start of the semester, so I don’t think there’s much useful information there yet.  U.VA. may yet go all on-line, apparently that hasn’t been decided.