Post #1037: Trend through end of February 2021: flat.

Posted on March 1, 2021

Starting 1/21/2021, the trend went from strongly downward to essentially flat.  Some part of that was due to the loss-and-recovery of data reporting in Texas and similar areas.  The rest is unexplained.

If you study the thin lines on the chart just above, you’ll see that many regions have returned to a fairly strong downward trend.  (I.e., most of the thin lines are once again sloping downward in this log-scale graph).

The exceptions are the Northeast (circled) and the South Central region.

The South Central region is the center of the data reporting anomaly following the mid-February cold wave.  That may still be in the process of getting sorted out.  Today it’s Arkansas’ turn to perturb the trend, per the Arkansas Department of Public Health website:

This shows up as the start of yet another “speed bump” in the data, this time for Arkansas.  That will remain in the data over the next week as the seven-day moving average moves past this reporting anomaly for Arkansas.

 

 

 

 

The Northeast is the other region that is not trending down.  And there appears to be something more systematic happening there.  No state is trending down, New Jersey is trending up, and the rest of the states are flat.

 

 

 

The upshot is that, outside of the Northeast, the national picture looks good.  In most areas of the country, new case counts continue to trend down.

The New York Variant (B.1.1.526)

The particular concern for the Northeast is that they have their own home-grown variant strain taking over in that region.  Depending on whom you believe, this new variant (B.1.1.526) accounted for over 25% of cases sequences in the New York City area in February.

Reporting in Forbes shows the key data.  This is actually for two related variants.  I believe it is for New York City only.

Source:  Forbes.

People are not yet saying whether this is more infectious that prior variants, or how much more infectious.  The table above shows that it’s clearly out-competing existing strains.  Given the size and heterogeneity of the New York City population, the simplest explanation for that would be that it’s a lot more transmissible.  (As opposed to “sheer luck”,which can happen if a strain just happens to get picked up by superspreader event or similar.)

I note that new COVID-19 cases fell just 20%, in NYC, from the peak of the U.S. COVID third wave.  That’s roughly half as much as for New York State as a whole, or one-third as much as for the Northeast as a whole. 

Source:  Calculated from NY Times Github COVID-19 data repository count-level file, data reported through 2-28-2021.

In fact, if you look at the graph above in detail, you’ll see that new case counts haven’t really fallen in New York City since February 1.  And that is distinctly different from New York State as a whole, or the Northeast as a whole.

And so, empirically, New York City has shown much smaller reduction in case counts than nearby areas.  And the reduction in case counts ended three weeks earlier than in nearby areas.

Without doing the math again, I’m just going to to a quick comparison to show that the growth of the New York variant is similar to the growth projected for the U.K. variant.  This uses my simple model of likely spread of the U.K. variant (e.g., Post #1021).

The U.K. variant, at 40% more transmissible, would have been expected to increase from 6% of cases to 34% of cases in a single month.  The New York variant was, in fact, observed to increase from about 6% of cases to 28% of cases in a month.  Given sample sizes and variability, I can’t rule out that this New York variant is just about as infectious as the U.K. variant.

Source:  Forbes, red markings added.

So that’s interesting, in an ominous way.  Let me just recap.

  • There’s a new variant spreading in the New York City area, B.1.1.526.
  • It has all the biological earmarks of being more infectious.
  • It’s taking over at the rate you would expect if it’s about as infectious as the U.K. variant (i.e., 40% more infectious (transmissible) than the existing strains).
  • New York City shows far smaller reductions in case counts from the peak of the U.S. COVID third wave than either NY State or the Northeast as a whole.
  • New York City cases new case counts stopped falling a month ago.
  • The extent of spread outside of NYC is unknown.

This will be well worth keeping an eye on.  Note from my “simple model” table above, that the turnaround in case growth, from a more transmissible strain, is slow and subtle.  That’s the right-most column.  You don’t see some sharp, well-defined turnaround.  You see a broad bottoming as the new strain takes over.

Well, if NYC is about to be overtaken by a more-infectious strain, we ought to know soon enough.  If that U.K. strain model is any guide, by the end of March, the NY strain should account for the preponderance of cases in NYC and case rates should be climbing.


Think of it as evolution in action.

That’s a quote from a science-fiction dystopian novel by Larry Niven and Jerry Pournelle. As I recall it, it was graffiti in a high-mortality-rate poverty area.

At this point, we’ve seen enough new variants to understand that these successful mutations aren’t one-off unique events.  And to expect the successful ones to be more contagious.

In other words, this is just evolution in action.  Genetic mutations are random.  Out of that huge pool of random mutations, some will prove to be more infectious.  Those will then, almost by definition, take over.

We’ve even reached the point where we’re seeing a form of “convergent evolution”.  That’s when Mother Nature re-invents the same traits, independently, in different species or locations, because those traits confer a survival benefit.

And so, the same successful genetic mutations are popping up in many different places, independently.  For example, the New York mutation shares several DNA-level changes in common with the South African mutation.  Not because there was any cross-contamination, but because those changes, randomly arrived-at, allow the virus to succeed.