Post #1348: U.S. winter wave stalls, low South African death rate validated in news reporting.

Posted on December 14, 2021

I apologize, but I have nothing to scare you with today.

The U.S. winter wave has stalled out pretty much everywhere but the Northeast.  In particular, the regions that led last year’s wave (Mountain and Midwest) now look like they’ve peaked.

And we now have a South African academic doing the math showing that Omicron cases are vastly less severe than Delta cases, on average.

I’m sure you’re going to see a lot of reporting to the contrary.  I address one piece of that below.  But as far as I can tell, all the information now points to Omicron having vastly lower average severity per case.

US winter wave stalls

Not much more to say than that.  Case counts are high and rising in the Northeast.  Maybe down to the mid-Atlantic region.  But pretty much nowhere else.

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 12/14/2021, from”  The NY Times U.S. tracking page may be found at

In particular, the two leading regions from last year — Midwest and Mountain — look like they might have peaked for the year.  And that’s not hugely inconsistent with the timing from last year, as shown below.  It’s getting late enough in the yer that it’s time for them to peak.

Really, as I showed in an earlier post, the critical difference between this year and last year is California.  That was a major contributor to the winter wave last year.  This year, nothing much is happening there.

Omicron?  Not yet. 

Take cold small comfort in this stall-out of the winter wave in the face of Omicron.  Fact is, Omicron isn’t really in play yet, in the U.S.  But it will be next week.

The CDC’s most current estimate is as follows (and I have to apologize to the CDC, because I saw no indication that they could or would do these estimates in such a timely fashion):

“The CDC estimated omicron was 2.9 percent of sequenced U.S. cases for the week ending Dec. 11, up from 0.4 percent one week earlier.”

Those numbers (from 0.4 to 2.9 in a week) are quite consistent with the roughly 2.5-day doubling time that was reported in both South Africa and Great Britain.  It’ll take just over three more doublings to top 50%.  So Omicron should be the dominant strain here about a week from now. 

In other words, if there is going to be a “tidal wave” of cases here, as was asserted for Great Britain, we ought to know it by Tuesday next week.

That said, with Great Britain, I’m still hard pressed to identify the Omicron impact on total cases, by eye.  Let alone the tidal wave.  Presumably, that’s coming, but it hasn’t shown up yet.  Based on reporting from four days back, they should , with an estimated one-third of new cases in Great Britain being Omicron.

In any case, Omicron currently accounts for just under 3% of U.S. cases.  So whatever is causing this stall-out in the U.S., it’s unrelated to Omicron.  But we should know within a week whether Omicron is going to have a major impact on total U.S. cases.


South Africa

A 12/13/2021 story in Fortune seems to validate everything you’ve read here about the extremely low average severity of illness of Omicron cases.  In that story, a South African analyst did the math on total harm from Omicron versus Delta.  (That is, not just cases, but cases x hospitalization rate per case).  I think it’s worth putting in as a direct quote from that story (emphasis mine).

From: “Omicron cases may already be peaking in South Africa, less than a month after the COVID-19 variant first surfaced“, by Vieienne Walt, December 12, 2021, Fortune Magazine.

But South Africa’s data shows Omicron has indeed been much less deadly there than previous COVID-19 strains—at least so far—even though only about 23% of the population is fully vaccinated. 

“Even if peak daily case levels exceed Delta by a factor of three (34,000 per day), the number of patients ending up on ventilators will only peak at 140,” said Streicher, the University of Johannesburg analyst, adding his calculation is based on a 10-day lag between people being infected and then placed on ventilators.” The number, he says, “is extremely low.

In other words, your odds of ending up on a respirator, after being diagnosed with Omicron, are about 11% of your odds of doing the same, under Delta.

It is worth noting that almost all COVID-19 decedents go through a period of respiratory failure.  That is, they need ventilator support.  Accordingly, if the count on respirators is low, that virtually guarantees that the ultimate mortality it rate will be that low as well. 

And so, I’ll go ahead and say it.  The “missing deaths” on this chart from a few days back still appear to be real.  The version below is a few days fresher than what I posted earlier.  Those deaths still haven’t appeared.  There is no peak in South African COVID deaths deaths matching the recent peak in Omicron cases.

Source:  WHO, accessed 12/14/2021.

A final aspect of that story is that South Africa is showing all the signs of nearing a peak.  The increase in new cases is slowing down, the fraction of tests that come back positive is falling, and hospitalizations are down.  All that, with just about 23% of the population vaccinated.

Contrary reporting on severity

By contrast, reporting from today’s Washington Post makes it appear that there is only a small reduction in case severity.  They report a study from the largest South African insurer, of cases between showing that “risk of hospital admissions among adults who contracted covid-19 was 29 percent lower than in the initial pandemic wave that emerged in March 2020.”

One of those two views of this must be materially wrong.  Either the risk of serious illness is vastly lower, or it’s slightly lower.  It can’t be both.

Other reporting indicates that the time period was “November 15 to December 7“, and provides critical details.  In particular, they state that “Hospital admissions for adults diagnosed with COVID-19 are 29% lower compared to the wave that South Africa experienced in mid-2020, after adjusting for vaccination status, according to the analysis.”

The New York Times adds further critical detail:  “A snapshot of the first three weeks of each of the four waves of infection shows that hospital admissions are significantly lower during the Omicron-driven fourth wave — 38 admissions per 1,000 compared with 101 during the Delta-driven wave, and 131 per 1,000 when the Beta variant was dominant, the study showed.”

So the crude hospitalization rate for Omicron is actually 62% lower than under Delta.  That’s getting closer to my estimate of 89% lower case ventilator rate.

Maddeningly, nobody provides a link to the actual study.  But I’ll give you my take on it.  Based on the full reporting, here’s what I think they did.

  1. The took all persons, with that insurance, who had a positive COVID-19 test during the period November 15 to December 7.
  2. They have no way to separate out individuals with Omicron and those without.
  3. They calculated the fraction who were hospitalized, separately for those vaccinated and not.
  4. They inferred that any drop in hospitalization must be due to the fraction of cases that were Omicron, and so back-solved for the drop you’d see if 100% of the cases were Omicron.

As I see it, this study share the same problem as prior reporting on total “COVID-19” hospitalizations in South Africa.  Most of the people who were hospitalized weren’t sick from COVID-19. Over three-quarters were persons hospitalized with COVID, not persons hospitalized for COVID.  The (presumably asymptomatic) COVID-infections were uncovered only during routine testing after admission to the hospital for some other, unrelated condition.To see that, take a look at the most recent statistics on South African COVID-19 hospitalizations, reported here.  They report, on 12/14/2021:

  • 6,895 patients with any diagnosis of COVID-19
  •  469 in ICU
  • 187 on ventilators

Just to give an example, Virginia also tracks all persons in the hospital with any positive COVID-19 diagnosis.  But in Virginia, the corresponding numbers are:

  • 1260 patients.
  • 301 in ICU
  • 178 on ventilator

Source:  Virginia COVID-19 dashboard, accessed 12/14/2021.

However you slice it, South Africa has about five times as many “COVID-19” cases in their hospitals, but roughly the same number in ICU or on ventilator.  That tells me that far larger share of those South African cases are likely being treated for something other than COVID-19.

My upshot is that I think the contrary reporting is correct on its facts, but incorrect on the implied level of case severity.  The commonly-reported number is “29%”.  But that’s a comparison to the original strain, and removes some factor to account for vaccination.  More detailed reporting reveals that the crude hospitalization rate for Omicron is 62% lower than for Delta.  And I am virtually certain that understates the true impact by commingling those admitted with COVID-19 and those admitted for COVID-19.


In short, I’m sticking with my original estimate.  As I see it, Omicron infections aren’t slightly less severe, as that 29%-lower-case-hospitalization-rate figure would imply.  That aren’t even much less sever, as the 62% difference in raw hospitalization rate implies.  As I read the data, they are far, far less severe, on average, than Delta.  Best guess, in South Africa, they are seeing one-tenth the hospitalization rate for COVID-19.  And so far, the impact on the mortality rate is almost too small to be seen.

And so, I conclude that Omicron is by far the lesser of two evils.  Even if Omicron boosts total case counts by a factor of three, relative to what Delta would have done, we’re coming way ahead by having Omicron drive Delta out of the population.   We’ll end up with less crowded hospitals, and fewer deaths.