Source: Calculated from NY Times Github data repository, data reported through 1/17/2021.
This is a post from my narrow perspective as a resident of Northern Virginia. If you’re not from around here, this probably won’t interest you.
Right now, Virginia has the steepest upward trend in new COVID 19 cases per day in the country. And this is the steepest one-week change in trend that Virginia has yet seen.
What particularly stands out is that a) this runs sharply contrary to the downward trend in almost all other states, b) this puts Virginia grossly out-of-sync with Maryland and DC, which is a first for the pandemic, and c) by eye, this is the sharpest increase in new cases that Virginia has ever seen.
All suggest that something has changed fairly radically. Northern Virginia is so cosmopolitan, it makes me wonder about that new UK strain of COVID, and whether that’s here in numbers but Virginia simply hasn’t looked for it yet.
But, having dealt with health care data all my life, this also suggests that this might just be a data reporting error or change, in whole or in part. When things seem too unusual to be true, they are often are not true.
So this is an odd situation. And for me, it’s close to home. Let me see if there’s anything identifiable within the recent sharp increase.
Drill down: Counties.
Chesapeake, VA
(I apologize in advance for flipping back and forth between raw counts and counts / 100,000 / day. This is more-or-less a narrative of what I did, and I didn’t bother to clean it up.)
The first result is that this increase is highly concentrated in just a few counties. One-quarter of the total increase is due to Chesapeake, VA. Chesapeake is a city of about a quarter-million population, just south of Norfolk/Portsmouth. That increase in reported COVID cases roughly eight-fold disproportionate to the population.
In the past week, the seven-day moving average of cases more than tripled, from a nice, steady 40 (new cases/100,000/day) to 140. Chesapeake is large enough to rule out this being some small-numbers fluke. And by inspecting the raw data, this isn’t due to some single-day reporting correction. For some reason, new cases jumped from about 100/day a week ago, to around 400/day now, plus or minus.
Lack of confirmation via news stories. For previous localized upticks in Virginia, I’ve been able to find discussion of (e.g. ) a meat packing plant or correctional facility. Here, there’s nothing in the news about this. No mention of a crisis in that area. No discussion of a single large employer. No discussion of an influx into the local hospitals. No internet chatter that I can see.
No uptick in hospitalizations. Comparing the hospitalization data, COVID hospitalizations actually declined there, from last week to this week (and the week prior). That doesn’t make sense.
Uniform across ZIP codes within the reporting area. By ZIP code, the numbers are not plausible because they are too uniform. There are six ZIP codes in Chesapeake. All six showed increased in new cases /100,000 per day from around 40 or so, to around 200 or so, over the past week.
No spillover to nearby reporting districts. Finally, there appears to be absolutely no spillover into the adjacent Portsmouth/Norfolk area.
So, we have an apparent city-wide outbreak, uniformly across all ZIPs, quadrupling new cases per day, uniformly, everywhere in that one city? But nothing in the adjacent areas. And no uptick in hospitalizations.
The only thing I can think of is that maybe they began to report cases among military personnel? There’s a large naval presence in that area (E.g., Norfolk Naval Shipyard is, I believe, the largest U.S. Navy shipyard.) Even so, the increase seems far too large for that.
So this has the look of some sort of data reporting issue. Either an outright error, or a major change in what’s being reported. Solely within this one district (Chesapeake).
Prince William, Manassas, Manassas Park
Prince William County (PWC, excluding the independent cities of Manassas and Manassas Park) accounted for just over 20% of the total increase in Virginia’s COVID 19 new case rate. PWC accounts for about 5% of the population of Virginia, making this four-fold disproportionate to its population.
Here, this is due mostly to a one-day (1/17/2021) jump in the count of new cases. They saw about 1400 new cases reported that day, against an underlying trend of about 275 cases/day. That one-day jump appeared across almost all the ZIP codes, all on the same day. It also appears in the data for Manassas City, on that day But not in the data for Manassas Park.
So this, too, might plausibly be just some sort of data reporting issue. It looks like they had some sort of one-day catch-up of cases that had not been reported previously? In any case, most of the uptick in Prince William, and its ZIP codes, and in Manassas City, can be attributed to that one-day 6-fold jump in new cases.
Loudoun County.
I’m only including this because a) it’s in Northern Virginia, and b) it accounts for about 10% of the increase in Virginia. But it also accounts for about 5 percent of the population of Virginia. So its impact is only two-fold disproportionate to its size.
There, the low new case rate (around 30/ 100,000/ day) did not quite double over the last week, to 54/ 100,000/ day. (Both of those are a seven-day moving average). As far as I can see, that’s due to a smooth increase in the case counts over this past week.
Without going any further into it, that has the look of a true increase in cases.
Summary
This past week, Virginia had a sudden upward spike in new daily COVID-19 cases. This puts it at odds with, well, the entire rest of the U.S., pretty much. As well as with neighboring states with which it has been more-or-less in sync since the start of the pandemic.
Upon inspection, I don’t think that’s real. Or, at least, large parts of it don’t appear to be real. They appear to be some artifacts of data reporting in a handful of counties. I’ve let the Commonwealth know what my opinion is, and we’ll see if these stand, or are subsequently corrected.
