Post #912: COVID-19 trends: No clear Thanksgiving effect yet; ND data anomaly

Posted on December 10, 2020

Yesterday, North Dakota decided to add persons with a positive antigen test to their counts of new positive COVID-19 cases.  Previously, they did what every other state does, and counted PCR tests only.   (See diagnostic test on this FDA webpage for the difference between antigen tests and the standard PCR tests.)

The backlog of those individuals added thousands of new cases to their totals, yesterday, and generated an apparent spike in cases.  For anyone looking for the effect of Thanksgiving travel in the data, that certainly caught the eye.

But if you look net those cases out of the data, to make the current data consistent with the historical data, ND continues on its prior downward trend.  That’s the red line segment added to the ND data line above.

That’s all by way of getting to my main point.  So far, while the trends are in general not good (outside of the states that had high peak rates last month), the trends are no different today than they were a week ago.  Today is exactly 14 days after Thanksgiving.  And, to may eye at least, so far, there’s no evidence of a widespread Thanksgiving-driven increase in the number of new cases.

Yet. 

Instead, right now, outside of the Midwest/Mountain states that had those high peak rates, most of the rest of the country looks like some variation on the above.  Rates started to rise as colder weather — or, really, low-indoor-humidity weather — started to arrive.  (Where the big recent uptick in southern California tests that rule about temperature versus humidity as the cause for the upticks.)

But, so far, any massive increase in cases from Thanksgiving has yet to show up in the data.  At least, by eye.  To me.

And it’s getting to be just about the time when that ought to be happening.   My rule, based on what I think I’ve observed in Virginia and elsewhere, is that there’s a median of five days from exposure to symptom onset (that’s based on hard data), and then roughly another seven days for seeking care, getting tested, having the test reports returned, and having those get processed and show up in the officially tabulated data.

So, I’d say, a median of 12 days’ lag, from the day of infection to the day that’s reported as a new case.  I’ve seen other discussions that suggest a two-to-three week lag between infection and reporting, without reference to any specific data to back that up.  I’m not sure who’s correct.

But either way, that wave of Thanksgiving infections should be showing up soon.  Depending on state-to-state variation in the second part of the lag (between the time a person seeks a test and that test is reported), by my estimate, the bulk of brand-new infections caused directly by contact during Thanksgiving travel and celebration should have started showing up in the data earlier this week.  And, by others’ estimates, they might show up in the coming week.  But nobody suggests that the lag would be longer than that. 

Every new day when we don’t see that Thanksgiving surge adds to the feeling that maybe that won’t happen.  Or won’t happen to a large enough extent that it will matter against the generally upward background trend.

And it’s important to track that.  Day after day, the newspapers throw out warning after warning regarding risky COVID-19 behavior.  And some of those warnings matter.  And some don’t.  And a lot of the time, if nothing bad happens, that never gets reported out.

This summer, I thought it was important to document it when that happened.  When a lot of tsk-tsking went on, but in fact, there was no bad outcome.  (E.g., Post #709, Perhaps cheese protects against coronavirus).

If we see a Thanksgiving surge, that’s important information.  And if we don’t see one, that’s equally important information.

This in no way denigrates all the individuals who warned against traveling and meeting for Thanksgiving.  Given what we knew at the time, that seemed to be incredibly sound advice.

But we still need to know, with 20-20 hindsight, whether that was in fact sound advice.  Particularly with another set of holidays right around the corner.  Was that advice sound, looking after-the-fact, with the benefit of the results in hand?  Did all that travel and meeting result in a big bump-up in cases?

As far as I can tell, it doesn’t look like it has happened yet.  And, maybe we’ll never be able to separate that from the ongoing upward trends.   All I can say is, I’ll keep tracking this.