Other parts of the article were fine, but the part I quoted explicitly gets the math wrong.
Looks like it peaked. I don’t think there’s any reason the positivity rate should change much due to additional data coming in.
Today:
Yesterday:
Well, now here’s an optimistic projection in today’s news (SD Union and probably elsewhere). I hope their model is right. Others have said basically the same thing in the past few days. The reason that’s cited used to be called “herd immunity”, an expression that seems to have gone out of fashion.
The University of Washington’s own highly influential model projects that the number of daily reported cases in the U.S. will crest at 1.2 million by Jan. 19 and will then fall sharply “simply because everybody who could be infected will be infected”
In other words, when omicron gets starved for new people to infect, it’s transmission rate will fall. The projection is for the US overall, however, so there will continue to be local “hot spots”, presumably for some time after the national peak. The overall US rate of decline won’t be extremely sharp, I would think, for that reason.
That’s what I’ve been saying since Omicron showed up.
Exponential spikes fall as sharply as they rise. Nationally, US case counts have doubled every 7-10 days since mid-December. The daily average is approaching 1 million (and that’s an undercount) and there are only 330 million people in the country. I think the places that started spiking late are mostly not very populous so their later peaks and falls won’t have much effect on the national total.
That’d be very good.
1 Like
How is that defined? The unvaccinated? Covid deniers? Those in riskier occupational or living situations?
Anyone have the breakdown of previously-infected vs. never-infected at this point? I would think that by January 19, there well still be a majority who still have yet to be infected.
Omicron will infect a substantial majority of people who spend time indoors with strangers, unmasked or for an extended period of time with a relatively ineffective mask, including everyone who’s eaten or drunk indoors at a restaurant or bar since Omicron came to town, and everyone they live with.
Having had Covid from a previous variant provides little or no protection against Omicron.
I think that the way this was put in the article I read (and the quote I posted) is overstated. Not everybody that could be infected (which is actually everybody, vaccinated or not, boosted or not) will have been infected when the drop-off starts. I think the sense of the writer was that the number of remaining uninfected people would be too small for the virus to spread without limitation, i.e., that Ro would be less than one, and virus transmission would drop off. I guess I picked a bad quote, for which I apologize, but the way that other similar comments are written means, basically, that conditions for approaching herd immunity will have been attained about a week (or so) from now.
Notwithstanding the unfortunate wording (which I glossed over because I knew what they intended), the article was interesting in the degree of specificity of this particular model prediction, both in terms of number of new cases at the peak and the timing of the peak, and also that their predicted peak was so imminent.
Not necessarily. Also, while the rise of cases in the plot that you posted is very steep, it isn’t a true, simple exponential. Initially, maybe, but not overall. Even if it were, the strength of a subsequent exponential decay wouldn’t have to be the same as that of the rise.
The trajectory of the previous two waves both had inflection points about a third to halfway to their peaks. It’s harder to see on the long time scale of this plot, but looking closely I think we’ve passed the current inflection point.
Here’s an expanded plot of the current wave in South Africa:
It takes some visual smoothing of the wrinkles in the dark red curve to see it, but if you do that you can see an inflection point around the end of the first week in December. The smoothed peak would be right about mid-month, and then the curve drops off. However, the rate of drop-off, while fast, hasn’t been as rapid as the rise.
The angles of the up and sown slopes are about the same. The rapid down ended and something else has been happening with a much higher level of cases than before Omicron showed up.
Exponential decline is inevitable after exponential growth, but it doesn’t necessarily take you back to the pre-exponential baseline.
Yes, but…
Of course, since people are often sick for months. That doesn’t explain why deaths doubled in two weeks. Could be delayed reporting over the holidays.
Good chance of that.
There was also a very steep decrease in reported vaccinations over the holidays, at least here in SD County. The good news is it’s now going back up again – steeply. I thought that some unvaccinated or un-boosted people may have had a “come to Jesus” moment (as they say) but been avoiding the aftereffects of the vaccines during the holidays so to not to completely disrupt all of their plans. Some people get pretty wiped out for a couple of days after getting one of the mRNA shots. The timing of the omicron outbreak couldn’t have been worse.
What I’ve noticed on some of these and similar plots (the case plots in South Africa, for example), is the extremely regular pulses on the daily bar graphs – “skyscrapers and sidewalks” – probably the latter at least being a weekend effect on data collection.
I wish the average curves shown on these graphs were taken over more than 7-or even 14-days, to smooth out the wrinkles more. You have to eyeball through the wrinkles (or print out a screen capture and pencil in a smooth curve as I do) to see the real trends.
Seven-day averages are usually plenty to see the trends, especially with data that tends to clump on certain days of the week. If you can’t see a trend it’s probably because there isn’t one.