Post #897: Preparing for a hard winter, #9: Recent COVID-19 trends in Virginia, cold spots as hot spots, and extrapolating to mid-winter

Posted on November 21, 2020

It has been more than a month since I tabulated the trends in new COVID-19 cases in Virginia.  It’s no secret that the trend has been up.  Here’s Virginia, in the national context, as of yesterday’s data reporting.


Record new cases, but not record hospitalizations or deaths.

In Virginia, we are setting records for new cases per day.  But, as you can see, this is only by a modest margin.  It’s nothing like the scale of the outbreak going on currently in the Midwest.  So the headlines of “record new cases” really mean different things here, as opposed to there.

Here’s the same trend diagram I’ve been showing for months, updated to yesterday.  As you can see, the trend in Fairfax County is similar to the trend for Virginia as a whole.

Source for data:  Virginia Department of Health COVID-19 website.

That one-month trend line doesn’t really put this in context, so let me show the trend since April 1.  In both of the graphs below, what I’m showing is the seven-day moving average, to smooth out the data.

The trend for total diagnosed cases is sharply upwards, for the past three weeks or so.  We are making new records almost every day.  But, as importantly, we are not (yet) making new records in terms of hospitalizations or deaths.

Source for data:  Virginia Department of Health COVID-19 website.

Deaths take a while to occur, and so tend to lag the uptick in cases.  But hospitalizations are contemporaneous.  And deaths tend to track hospitalizations, on average, in Virginia (Post #871, Age-adjusted trend in COVID-19 mortality rate in Virginia).  So it certainly appears that the average severity of COVID-19 cases newly diagnosed in Virginia has been trending downward.

You can’t attribute that change to a shift in the age distribution of COVID-19 patients.  I calculated that directly in Post #871.

And you can’t attribute this to some uptick in testing that is disproportionate to new cases, because a) that hasn’t occurred, b) the fraction of tests that are positive has remained stable for the past couple of months, and c) in Virginia, testing volume is mostly driven by the number of people who have COVID-19 symptoms.  As more people get sick, testing volume rises to match that.  So, while the total number of test has risen (bottom graph below, yellow line), the fraction of tests positive hadn’t changed since the start of July (top graph below, yellow line).

Source:  Virginia Department of Health.

Either the Virginia health care system has gotten much better at treating individuals outside of the hospital, or the average case presenting in Virginia is less severely ill than it was just a few months ago.

I will also point out that this same phenomenon is being observed all over the country now.  Case counts are up greatly, but hospitalizations and deaths have not risen in proportion.   That gave the crisis states in the Midwest a little bit of breathing room before they completely overwhelmed their hospital capacity, as is starting to happen now.

And that same phenomenon gives us a little more space here in Virginia.  Space between the growth in new case counts, and the strain on our health care and mortuary systems.


Our cold spots are now our hotspots.

The most recent NY Times map of new weekly cases shows that the hotspots in Virginia (darker colors below) are clustered in the mountain-and-ridge portion of the Commonwealth.  As a crude rule, the further you get from sea level, the higher the current new case load.

Source:  NY Times coronavirus map for Virginia.

As with the US as a whole (shown in Post #894), our hotspots are occurring  where low temperatures meet low mask use. 

Here’s the NY Times mid-summer snapshot of mask use, for Virginia.  Here, lighter areas represent low propensity to wear a mask.  As you can see, mask use is lowest in the mountainous southwest corner of the state, and is generally low (outside of major cities) for the entire Appalachia plateau.  If you flip back and forth between the two maps, you can see that they aren’t quite the exact inverse of one another, but they are close to it.  Light areas on the mask-use map now correspond to dark areas on the hotspots map.

Source:  NY Times map of US mask use, edited down to Virginia.

You can see the more systematic analysis of this point in the key graphs of Post #882, Does mask use really reduce COVID-19 spread.

But a key point is that mask use has been low in those areas since mid-summer.  Areas that are now showing high new case counts, as fall turns to winter, are areas that have been low-mask-use areas all along.

And, as it turns out, those are also the colder parts of the state.  Indoor relative humidity will have been dropping there, earlier than in the rest of the state.   (Because, as a rule, the lower the outside temperatures, the lower the relative humidity in heated indoor air.)  Separately, the Tidewater region has the further, modest humidifying effect of being located near so much open water.

While I could not find a map showing typical temperature differentials across the state, I did find tabulated data (from this source).  When I boil that down, I get the table below.

My point here is a two-part prediction.  First, the temperatures we should expect to see in all of Virginia, next month, are colder than the temperatures seen in the Ridge-and-Valley areas of Virginia, this month.  If current weather patterns hold, and if the historical average differentials match what happens this year.  And that means that indoor relative humidity in the populous areas of Virginia will soon be as low as it is in the Mountain areas now.

Second, if cold weather is driving the current uptick in cases (via low indoor humidy, Post #894), the only thing that stands between our current new case rate for Virginia as a whole, and the high new case rates shown in the Appalachian areas of Virginia, is our generally higher mask use. 

When I sift through all of the evidence, my best guess is that the uptick we’ve seen in the past three weeks is going to continue for quite some time. 

And new case counts are likely to rise quite a ways.  If the populous areas of the state end up with rates anywhere near those currently prevailing in the mountain areas, we should expect daily new cases to be somewhere around triple the current rate.  Or maybe 6000 new cases per day?

My point is not doom-and-gloomYou can get unquantified versions of that any day, just reading the news.  My point is not to freak out as the case counts rise.  It’s not random, it’s not a moral failure.  It’s just the change in seasons.  And we can always go into an actual lockdown (as opposed to the relatively weak restrictions seen so far in Virginia) if things get really bad.

And it’s almost certainly not going to exceed the capacity of the Virginia hospital system.  At least, not based on what we’ve seen so far.  At present, no hospitals appear truly stressed, and Virginia had the foresight to put in a lot of surge capacity at the start of the pandemic.

It’s just better to be prepared than not.  If it’s going to be a hard winter, it’s better to understand that than to be surprised by it.  All I’m saying is, don’t be surprised to see record-setting case counts in Virginia this winter.


ADDENDUM:  FWIW, based on recent trends, I’m getting more cautious in what I will and will not do.  I’ve gotten pretty slack about (e.g.) just popping down to the store for an item or two.  About doing the full 20-second hand-washing routine after every shopping trip. Now I’m back to re-thinking every trip into a public space, and minimizing those trips.

And, as I shop, I note that a lot of other people are slacking off as well.  Last trip to the grocery store, there were lots of people having loud close-quarters conversations.  That’s exactly what you don’t want to do.  When I went to pick up citrus fruit at the recent Lions sale, with my windows tightly rolled up, order information on a piece of paper sitting in the window, I was the only one who did that.  Everybody else rolled down the windows and talked to the people filling the orders.  Which pretty much guts the entire rationale of doing that from your car.

We seem to have lost a lot of the fear factor that drove cautious behavior early in the pandemic.  I think maybe we need to recover some of that attitude now, to get through this winter.  What was relatively low-risk behavior, a month ago, may not be low-risk a month from now.