Post #1356: Three-and-a-half weeks.

Posted on December 19, 2021


This morning’s news seems to be a full-court press on the COVID-19 doom and gloom front. Accordingly, I though I might bring the graph below to your attention.

As you know, Omicron was discovered in South Africa and adjacent regions about a month ago.  Today, Google “COVID South Africa” and you’ll be given the following graph, using Johns Hopkins as the underlying data source.

That appears to be a peak.  I wouldn’t normally point that out so early, but this has been predicted for almost a week now.  That prediction was based largely on the trend in test positivity rate that, a factor that has been associated with peaks throughout this pandemic.  Finally, starting yesterday, multiple news outlets have called that as the peak of the Omicron wave in South Africa.

If you look for it, you can track down a string of reporting first predicting a peak, and then more-or-less confirming it.

Self-limiting waves and a weakness of standard epidemiological models.

Finally, December 18 New York Magazine, here’s a view of this that absolutely mirrors my own.  This is “Gauteng’s Omicron Wave Is Already Peaking. Why?,

That article is a long read, but I found it worth the time.  Particularly if you need some cheering up about the situation, from people who clearly are well within the mainstream of scientific thought.

Most importantly, the author points out the apparently self-limiting nature of these COVID-19 surges, something I have noted repeatedly in prior postings.  His point being that all COVID-19 waves prior to this one stopped well short of exhausting the vulnerable populations.  That is, well before traditional epidemiological models said they should stop, based on the estimated level of immunity in the population and the ability of the virus to spread.

To put it another way, those waves stopped well short of the point where the population had herd immunity.

The way I’ve put it is that we always get right to the point where the hospitals are full, and the COVID-19 wave crests.  Despite the fact that plenty of targets for new infection remain in the population.  And that has always puzzled me.

No one is quite sure why, but the best guess that I read in this article is that, as a population, we don’t “mix” thoroughly enough to allow full spread.  To some degree, we each seem to have our own social or business strata.  Chains of infection travel within those strata, but find it much harder to cross those strata.  And as a result, chains of infection do not spread as far or as thoroughly as standard epidemiological models expect them to.

Badgering the witness

Separately, I particularly liked this exchange regarding the virulence of Omicrion, where the author actually got a trained epidemiologist to say, reluctantly, that yeah, if the numbers out of South Africa are real, then it probably has to be due to a less virulent virus:

"Does that seem plausible to you? Some of the numbers that we’ve seen today — the hospitalization rate may have fallen by 91 percent, the death rate of people in hospital may have fallen by two-thirds — do you think you could get that large in effect just through additional acquired immunity? This is just compared to the last wave, so it’s not all that long ago. Could you get that scale of an effect just through acquired immunity? It would seem to be that reduced virulences has to be at least a part of the story, no?
Yeah, if those numbers are real, then I think you would have to have lower intrinsic severity."

My two cents:  If that virus is stuck on fast-forward, what does that imply for duration of this wave?

In any case, here’s my cheerful observation:  The South African Omicron wave took three-and-a-half weeks, from first case to peak rate.  That seems to be how it’s playing out in that location.

Not to be too crude about it, but it appears to me that if the virus spreads about twice as fast, then the whole wave unfolds on fast-forward.  I don’t think that’s exactly rocket science.

But consider what that implies if, as was the case with prior waves, there is some limit to the number of persons that can be infected in any one wave.  If that limit is anywhere near where it has been for prior variants, you’ll run through this wave’s vulnerable population in half the time.

Let me emphasize that:  It’s not like the numbers are going to go straight up for the next two months, because the last wave lasted two months.  If they go straight up, best guess, they’re going to go straight up for a much shorter period of time.  In other words, my best guess — now that we’re seeing it in South Africa — is that there’s going to be a tradeoff between rapidity of spread and total duration of the wave.  It’s going to spread like wildfire, but only for a few weeks, not for two months.

Prior COVID-19 waves averaged about two months from start to peak.  Prior COVID-19 variants had a doubling time of about 4.5 days.  This COVID-19 variant has a doubling time that’s about twice as fast.  And so?  It hits its limit — it goes as far as it can go in this wave — in a far shorter amount of time.  Something like half the time.  Or less.

But won’t there have to be a higher peak?

But if there is a limit to the number of people infected in this wave, won’t it be higher than prior waves.  Yes.  To be clear, traditional epidemiological models tell us that, all other things equal, the greater transmissivity of Omicron should result in more people being infected, in total, than with Delta.  In effect, the pandemic should run further with a more transmissible variant.

But I’m guess that, in fact, it’s not going to be hugely more.  Wishful thinking?  No.  Aside from what we’re seeing in South Africa now, that’s based on our own history in the U.S.

Recall that Alpha was more transmissible than the native strain, and Delta was more transmissible than Alpha.  And the peaks for both of those waves, in the U.S., were well below that of the winter 2020-21 peak, which was the native strain.

Here’s the pandemic, through 12/17/2021, with the peaks labeled.

Data source for this and other graphs of new case counts:  Calculated from The New York Times. (2021). Coronavirus (Covid-19) Data in the United States. Retrieved 12/18/2021, from”  The NY Times U.S. tracking page may be found at

One thing that is not held equal is the level of immunity in the population.  Alpha was shut down by vaccination, and Delta faced a population with high existing immunity both from vaccination and from prior infections.  As a result, we have not yet seen anything approaching the peak of the 2020-21 winter wave, despite spread of new variants that are far more transmissible than the native (Wuhan) strain.

A second factor not held constant across waves is the level of COVID-19 hygiene.  And my observation is that that has pretty much gone to hell.  At least around here.  Even in the upscale DC suburbs, I’m seeing more people in the stores unmasked, and more workers with perfunctory masking (e.g., mask hanging below the nose).

In the past, COVID-19 hygiene has lagged the waves of COVID.  People only masked up once things got really bad.  So I don’t have much hope that we’ll get people masked up for Omicron pro-actively.  Currently, about 60% of Americans say that they routinely wear masks in public spaces, down from more than 90% just past the peak of the 2020-21 winter wave.

Source:  Carnegie-Mellon COVIDcast.

On the other hand, people know the drill now.  At least, responsible people.  For example, even though there are no government mandates, we’re seeing a lot of big group events being shut down in this area.  Voluntarily.  E.g, my wife and I like to sing Handel’s Messiah at this time of year, and the one sing-along that had been offered in Northern Virginia has now been cancelled due to Omicron.  (Which, I think, was the right thing to do, given that singing is about as bad as coughing in terms of spreading airborne droplets (Post #678, 4/30/2020).  I see similar reporting of London bars and restaurants voluntarily closing in the face of this latest wave.

So, we’ll see.  Maybe there will be a much higher peak.  Maybe not.  All other things equal, sure, there would be.  But all other things aren’t equal.

In any case, if we’re going to blow through last winter’s peak in the U.S., we ought to see that soon enough.  The one thing that everybody agrees on is that this Omicron spreads fast.

Nobody ever got fired for not making a mistake.

Just a bit of realpolitik here.  Nobody in any official position in a public health organization could possibly state that Omicron cases appear to be much milder, or that that Omicron wave is likely to be shorter than prior waves.  It’s simply not going to happen. So don’t hold your breath waiting to hear that.

If things turn out better than public health experts predict, then, well, from their viewpoint, there’s no harm done.  In fact, there will be lives and hospitalizations saved if they can scare people into more responsible behavior than would otherwise occur.  From that viewpoint — minimizing harm from this deadly disease — the only ethical thing for them to do is run with worst-case scenarios.

Now that they have the population’s attention, regardless of the facts, this is an opportunity for some impactful public health messaging.  And it would be unethical of them to waste it.

And consider the alternatives.  If they underplay this wave, and the situation is much worse than predicted, there would be repercussions.

I surely know which way I’d lean if I were in their shoes.

But for me, how many times, in how many different countries, have we seen what appear to be remarkably lower hospitalization and death rates from Omicron?  And in every case, officials brushed them aside with the same sets of potential alternative explanations.  And so, somehow, we’re supposed to believe that between Delta (last month) and Omicron (this month), every population, in every country, has managed to develop much more resistance to COVID-19?  All in just about exactly the same proportion.

I’m an Occam’s Razor kind of a guy.  Either there’s some unique set of circumstances in each country that, separately, manages to explain the data.  And that, somehow, all just happen to have just about the same magnitude.  Resulting in remarkably similar outcomes across-the-board.

Or Omicron is a lot less virulent than Delta.

I’m going for the latter option.

To which I’ll now add, and it sure looks like the wave is going to be a lot shorter, too.  Based on the idea that the virus moves faster, and if there is some natural limit to each wave (as has been true for all past waves), we’ll hit that sooner.

As a result, my opinion now falls far, far into the fringe.  I’m betting that the U.S. Omicron wave will be short and sharp, that it might lead to more peak cases than the just-prior Delta waves (but not vastly so), and that despite that, we’ll see far fewer hospitalizations and deaths than we’re seeing under Delta.

Just to keep you up to date, for the past week, in the U.S., here’s what we’re seeing this week under Delta:

Source:  Carnegie-Mellon COVIDcast,  Notations in red are mine.

So, if you manage to catch any data on case hospitalization or case mortality in Europe, for Omicron, compare it to this most recent week’s U.S. rates for Delta.  No matter how I slice it, absent some absolutely incredibly large number of new cases, we’re going to see hospitalizations and deaths fall as Omicron displaces Delta.