Post #1045: Update on the U.K. variant B.1.1.7 in the US

Posted on March 5, 2021

To understand the methods here, look for prior posts using the Helix corporation data to track the U.K. coronavirus variant.

This information is updated weekly by the Helix Corporation.  Data sourced from the Helix® COVID-19 Surveillance Dashboard. Accessed at on 3/5/2021.

Briefly:  By their data, B.1.1.7 now accounts for 42% of Florida COVID-19 cases (42.49% as of 3/3/2021) and 27% of U.S. cases (27.03% as of 3/4/2021), per their Daily Percent SGTF of Positive Samples table.

My simple model is running a bit too fast, relative to the observed U.S. number.  Compare the 34% circled below with the actual figure of 27% highlighted above.  My model is about four days too fast, based on the estimated U.S. doubling time for the U.K. variant as a fraction of all cases.

That’s consistent with the way I set this up, using an estimated doubling time of around 10 days.  I now looks like that’s maybe doubling every 13 days in the U.S. on average.  So my model runs a bit faster than it should.

Here’s something that should really freak you out if you think about it:  How much of a further increase is already baked in, due to the lags in reporting?  I’m not entirely sure, but I think it takes Helix about a week to sequence the samples.  The samples themselves are from infections that occurred roughly a week prior to that, owing to the lag between infection and symptom onset, and then the lag between symptom on set and testing.  So these figures could be a snapshot of the infections that took place about two weeks ago.  Since the doubling time is about 13 days, the U.K. variant as a percent of infections occurring right now should be about twice what’s shown above.  Roughly.  And, because nobody is doing anything to slow it down, it’s pretty much a certainty that we’re heading for an acid test of whether or not the greater infectiousness of the U.K. variant will drive up new-case rates.  If things stay on track, the U.K. variant should account for around 80% of the infections occurring today, in Florida, the data for which will be reported a couple of weeks from now.

If the fraction-of-cases estimates are correct, Florida should begin seeing increasingly positive rates of new-case growth soon.  All we can do is keep an eye out for that.

Near as I can tell, looking at the Helix data and the CDC raw case counts referenced in earlier posts on this topic, we should have already seen a fairly big “wedge” appear between the Florida growth rate and the rest of the South Atlantic states.  (Because, near as I can tell, Florida has a much higher fraction of the presumably-more-infectious B.1.1.7 strain.)  No such thing has happened yet (see below).

I did not, for an instant, think that researchers in both the U.K. and the U.S. who estimated that this U.K. strain is vastly more infectious than existing strains could possibly all have been wrong.  So I still have full faith that the aggregate case growth data will eventually reflect the presumed substantially higher infectiousness of the U.K. strain.

That said, the data are what they are.  This is now much like the wait for the various post-holiday surges.  Every day that we don’t see it, the less likely it is that we’ll ever see it.