Post #1370: CDC estimate of Omicron share revised sharply downward

Posted on December 28, 2021


As you may recall, about a week ago, the CDC announced that Omicron accounted for an estimated 73% of new U.S. COVID-19 cases, for the week ending 12/18/2021. 

That came as a true surprise.  That was easily four-fold higher than I had expected, based on the prior week’s data.  I kind of doubted it at the time (Post #1360).  I pointed out that it was based on a highly uncertain projection, based on a tiny number of actual cases  (Post #1362).  And I’ve repeatedly said that it doesn’t square with the still-high hospitalization rate of current new COVID-19 cases.  I went as far as to say that the current rate suggested something closer to a 50/50 mix of Omicron and Delta (Post #1367).

In short, that number has weighed heavily on my mind.

Today the CDC updated its estimates.  Turns out, that 73% is now estimated actually to have been 23%.  That is, the CDC’s new estimate, for the week ending 12/18/2021, is that Omicron accounted for just 23% of cases.  Not 73%.

For the most current week (week ending 12/25/2021), the CDC now estimates that Omicron accounts for just 59% of cases. 

That’s a huge difference.  One more week of growth should have taken that 73% estimate into the 90-percents.  Instead, it’s now estimated to be 59%.

I have learned the hard way that just because two measurements disagree, that doesn’t mean that the more recent one is correct.  But the newer estimate makes vastly more sense to me.  The prior estimate meant that the U.S. was, in effect, already done with Omicron prior to Christmas.  The new one says that Omicrion was only just getting started at that time.  The old one suggested that Omicron still had a very high case hospitalization rate.  The new one squares the observed overall case hospitalization rate data that we are currently seeing in the US with estimates of the Omicron case hospitalization rate from South Africa, England, and Scotland.

In theory, this is bad news.  Though, oddly enough, I’ll bet that the popular press reporting gets it backwards.  This tells us that we’re nowhere near done with Omicron.  But anybody who has looked at the new case data already knew that.

But most, to me, this new estimate makes sense of the data.  The latest CDC 59% estimate would have been achieved on or about 12/23/2021.  So we’ve had another four days of Omicron growth since then, to get to today’s data.  Best guess, that would put it in the mid-80-percents today.  Which would again square with an acceleration of the new case trend through today.  (Plus-or-minus the differences in data reporting lags between the CDC test swab date and the new case report date).

In other words, it looks like we’re just now getting into peak new case growth.  And that now squares with the CDC estimate, because we are just now getting into peak Omicron share of cases.  And it squares with a case hospitalization rate that has not yet fallen enough to bring total hospitalizations down, despite the rise in cases.