Post #1040: U.K. variant: The fuse remains lit. Keep your eyes on Florida.

Posted on March 3, 2021


Putting aside for a moment the New York COVID-19 variant (which people claim is more contagious, a seemingly plausible claim based on the new-case trend in New York state),

and ignoring the California COVID-19 variant (which ditto, but if so, it’s surely not affecting the new case trend there, as California has the largest rate of decline of new COVID-19 cases in the nation),

and just forgetting about the South African variant as not being common enough in the U.S. yet,

(and fill in other variants at will here, because you’ve got plenty to choose from),

The granddaddy of COVID-19 variant problems for the U.S. is the U.K. variant, B.1.1.7.  There’s reasonably solid research from the U.K. to suggest that this is, in fact, more contagious than currently-prevalent strains.  Independent U.S. research also suggests that it’s substantially more infectious.  It was the principal reason Great Britain went into lockdown in early 2021, where they will remain until April-ish.  (Although, in hindsight, part of what they were seeing in terms of increased new case counts was simply the seasonality of COVID-19).  And it’s already highly prevalent in the U.S. and spreading rapidly.

So that one gets the ticking-time-bomb/lit fuse award.  If you have to pick one to worry about, for the U.S. as a whole, that would be it.

Key questions are:  Just how fast is that fuse burning?  How rapidly is it spreading in the U.S. now?  And how will we know if this really will matter in the U.S.?

Why even question whether it will matter?  Look at that California variant again.  Everything said about the U.K. variant was said about the California variant:  Spreading rapidly, widely prevalent now, lab data suggests it should be more infectious, and so on.  But if that’s having an impact on the new-case rates in California, it sure hasn’t shown up in the data yet.  So it’s not unreasonable to be a bit skeptical.

How will we know?:  Florida.  After looking over the scant data available, that’s my answer.  If it really is as infectious as claimed, and really is a prevalent as claimed, then we should see Florida new-case counts begin to take on a new, upward trajectory sometime this week.  Maybe as late as next week.

Source:  Calculated from NY Times Github COVID-19 data repository, data reported through 3/2/2021.

Now let me explain how I got to that conclusion.

Helix Corporation data is all we’ve got.

Data in this section is sourced from the Helix® COVID-19 Surveillance Dashboard. Accessed at on 3/3/2021.

I believe that the only real-time information on this comes from the Helix corporation.  They test a large number of COVID-19 samples every day, testing quickly for a specific mutation that turns out to be an excellent marker for the U.K. variant.  They also directly confirm the presence of the U.K. variant in those samples by more extensive DNA sequencing, although those more extensive DNA tests appear to take about a week longer to complete.  About once a week they update the website where they present the latest counts.  If you want to read the most recent formal writeup of their research in this area, you can find that here.

The upshot is that you can find data, typically less than a week old, showing the fraction of positive COVID-19 tests that had this marker for the U.K. variant. And, as the U.K. variant has spread rapidly, this marker is approaching 100% correct in flagging only the U.K. variant.  The result is that you get a snapshot reading of the U.K. variant as a fraction of all U.S. cases.  With less than a one-week lag, on average, in the data.

The single largest drawback to this data source is that it’s the fraction of all the tests that Helix processed.  It’s a sample of convenience, and by no means a random sample of the U.S.  There’s no guarantee that their sample of cases is a good proxy for the U.S. as a whole.  And they make no attempt to (e.g.) re-weight their state-level data to reflect the proportion of total new U.S. COVID-19 cases by state.

That said, they appear to be the only game in town for this task.  And thank goodness they are willing to share.

By contrast, you can see the CDC’s variant web page here.  The CDC tracks the U.K. variant, but a) provides simple counts of cases, nothing that can be used to estimate the fraction of all cases, and b) only provides current data, not historical (trend) data.  Apparently, the CDC is in the processing of getting hundreds of millions of dollars of federal funding for “a down payment” for a “gold standard” system, which, I suppose, they might have in place in time for the next once-per-century pandemic (Post #1021a). 

There is one more oddity about the use of the Helix data.  The numbers are climbing so rapidly that you can’t use a “moving average” technique to smooth the numbers.  I mean, sure, you can, but you actually induce more error when you do that than if you simply took each day’s raw numbers.  The reason for this is that their moving-average data are now consistently 20% below the raw data, due to the strong upward trend.  The raw numbers may have a little day-to-day random transient component in them.  But their five-day moving average has a built-in 20% downward bias.  Of these, the (apparently) small random daily component appears to be, by far, the lesser of two evils.

Now let me present two factoids from the current (3/3/2021) version of the data as presented by Helix.  This is again from from the Helix® COVID-19 Surveillance Dashboard.

As of 2/27/2021, roughly 25% of U.S. cases appear to be the U.K. variant.  That’s from “Daily Percent SGTF of Positive Samples”, 2/27/2021, 24.54%.  And an assumption that, by now, more-or-less 100% of those are the U.K. variant, based on the trends and levels shown in the table just below that “B.1.1.7 as % of sequenced SGTF”.

This means that the original predictions that this strain would become the dominant U.S. strain appear to be dead on target.  Here’s my “simple model” of the U.K. strain, from Post # 1007, February 14 2021.  (The case growth numbers aren’t on track, given the abrupt change in trend that occurred 2/21/2021.  But the fraction-of-cases numbers are right where they were projected to be.)

Probably more importantly, this new strain appears to be concentrated in just a few states.  And of those, Florida stands out as having 36% of its 2/27/2021 tests indicate the likely presence of the U.K. strain.  The growth rate there is just slightly less than the doubling-every-ten-days reported for the U.S. in earlier research published by Helix staff.  Given that, half the cases in Florida should be the U.K strain sometime around March 4, 2021.

And if all of that is reasonably accurate, then we should be seeing an inflection point in the Florida new cases data right about now.  Per the simple model above.  And so, if all of the estimates above are reasonably accurate, and the U.K. strain is as infectious as has been estimated, then Florida will soon begin to diverge from the pack.

Caveats, I’ve had a few.  But then again, too few to mention.


Let me boil it down:  Based on the Helix data, Florida should be well ahead of the rest of the U.S. in showing the effects of the U.K. variant. 

How far ahead?  Well, we should start seeing those effects right about now.  If the Helix numbers are somewhere near on-target.

Are there caveats?  Well, yes and no.  There are some caveats regarding the extent to which the Helix data are representative of Florida COVID-19 cases as a whole.  But … that’s about it.

Normally, you’d also have to hedge.  Well, sure, the U.K. variant is one of many things going on.  If all other things are equal, and blah-blah-blah, then you’ll see the effect.  So normally, you have every reason to prevaricate.

Oh, yeah, and if the Governor of Florida takes this threat seriously and works pro-actively to contain it, then that could matter.  But Governor “Right-to-Party” DeSantis is such a (see illustration above) that we can pretty much rule that out (Post #825; Post #864).

Otherwise, I don’t think the normal caveats apply.  If the story that has been told so far is true, and the U.K. variant really is that much more infectious, and it really is already that prevalent in Florida, then it will simply dominate any other plausible effect.  It should begin to swamp Florida with new cases, until such time as so many have been infected/vaccinated that herd immunity kicks in and shuts the whole pandemic down.

So this is, I think, a fairly straight-foward strong prediction.  No weasel-wording allowed.  If we’re going to have the next wave of COVID-19 courtesy of the U.K. variant, then that wave ought to start in Florida.  And it ought to start in the next week or so.