early estimates on omicron prevalence in the US revised WAY down

week of 12/18 drops from 73% to 23%

well, that’s quite a revision:

Jason Gallagher @JGPharmD

CDC estimates of circulating variants including week of 12/25. Notably week of 12/18 estimate of Omicron revised from ~73% to 22.5% (just a tad different!). Now saying 58.6% of variants are Omicron nationwide. 1/

it turns out omicron was not quite as everywhere as was claimed. it seems likely we’ll see similar revisions to current front week data as well.

some seem to find this worrying:

Jason Gallagher @JGPharmD

Highest proportion of Omicron is in NY/NJ/PR. Honestly, I preferred last week’s shocking 73% estimate because this implies to me that there’s more room for Omicron to grow. 2/

personally, i find it in many ways reassuring as it proves an answer to the question “if omicron is so much milder, why are hospitalizations so high in much of US?”

it’s because they are not really being driven by omicron yet.

(well, and our determined over-testing and nosocomial case mining to maximize gravy train)

it’s also just media running wild with models and error bars. the CDC nowcast has a 33 percentage point 95% confidence interval for omi share of the US viral market.

this is starting to get a little “mr magoo” to be a basis for too much severity analysis just yet.

but the places in which it is higher can give us some sense.

i happened to have the denmark data sitting around, so i used that.

i took the weekly hospital admissions number and divided it by the trailing 7 day avg for cases normalized to 12/22/21 testing levels to mitigate the sample rate variance artifacts.

it’s not clear to me that this % maps to case hospitalization rates in a comparable fashion, but it does provide useful comparison to itself and thereby give us a fairly reasonable sense of the rate at which cases are driving hospitalization.

the news looks quite good. (i’m not sure what that weird spike in april was. seems probably it’s a data artifact driven by a sudden drop in reported case counts over a ~5 day period that then reverted)

it seems to find concordance to what many are reporting in south africa.

Vijayanand - Covid Data Analytics @vijay27anand

South Africa new cases, deaths and hospitalization trend. Hospitalization is down trending, flattening now as the daily new cases is declining, small increase in death rate but did not see a spike like earlier Delta wave. #omicron

recall that SA is a very low vaccination country, so this points to lower severity rates being a function of the variant, not the vaxxes.

this would seem to imply that more rapid spread of omi in the US is actually desirable to the extent that it crowds out and outcompetes more virulent variants. it seems to surge rapidly, but do little harm in terms of severe outcomes.

it looks like 100 middling colds and flus we’ve weathered before.

they never used to keep us home…