Take a look at this chart of COVID cases and deaths in Great Britain. Apologies for the hated dual-axis chart, but I didn't have much choice:
In January, cases peak and then, a couple of weeks later, deaths peak. In the summer, cases begin to rise steadily and then, a couple of weeks later, deaths also start to rise steadily.
But in December, cases begin to spike enormously and then, two weeks later. . .
Nothing. The death rate is going down.
The same thing is true in Canada and the United States. But in France, both cases and deaths are up. And in Germany, cases are down but deaths are up.
I don't know what this means. But Omicron sure is different from Alpha and Delta.
There has not been enough time for a significant number of omicron patients to die in the UK. Deaths lag cases by almost a month. Be patient, Kevin, you did this follow up throughout the pandemic and ought to know that.
What this data means beyond omicron is: Local conditions matter and using studies from foreign countries to make prognoses for the US is inherently risky. In a few weeks we will have US data (including deaths) and will be able to properly model.
Actually, if the disease is generally milder with omicron than earlier variants you'r expect the time interval between cases and deaths to be even longer.
I agree it's too early to see the spike in deaths. I think the best fit for the lag is three weeks or a little more, in the United Kingdom as well as the United States.
Just to make a wild guess: If the existing level of transmission is high the relative "harmlessness" of omicron wins out and death numbers decline. If on the other hand the existing level of transmission is low omicron's ability to break through immunity wins out and deaths rise (albeit from a very low level) ins spite of the "harmlessness".
The December '21 (Omicron) wave is affecting a much different population as the number of people with immunity ( via vaccine, prior infection or a combo) is at an all time high. Hard to compare with what 2020 might have looked like if 80-90% of the population had immunity.
This hypothesis contradicts at least somewhat the fact that omicron has much higher break through rates than earlier variants. And only a small fraction of people has been boosted yet and is "fully" protected (i.e. as fully as achievable at this point).
The December '21 (Omicron) wave is affecting a much different population as the number of people with immunity ( via vaccine, prior infection or a combo) is at an all time high.
We probably saw this with the Delta wave, too. Deaths in September (the peak of that wave, which began in early July) were lower than the peak of the previous wave (the week Trump left office).
So, here's my wild guess: the death peak of the current wave, which I reckon we'll see in January (deaths have been rising for about a month in the US at this point) won't be as high as late September's peak.
The gradually strengthening effects of mass immunity is what is normal in most pandemics. To not see this effect kick in eventually would, I believe, be unusual. (And also extremely unfortunate).
Also, I was doing a bit of research yesterday, and I saw that the rate of rise in deaths (I use the 7-day moving daily average) in the US has been slowing throughout these waves, too (by my count four: the peaks were April 2020, January 2021, September 2021 and the current one).
So, yes, as time goes by, more people have immune systems capable of greater defense of this coronavirus.
Since it has immune escape, a lot of those cases (seemingly the vast bulk, if the data are right - particularly from western Europe) are among those who are vaccinated.
According to this interview at NY Mag, the AZ vaccine provides basically 0% protection against infection from Omicron while 2-dose Pfizer provides about 30%. Who knows what the booster protection is. https://nymag.com/intelligencer/2021/12/is-omicron-a-new-wave-or-a-parallel-pandemic.html
Meanwhile the protection from severe/life-threatening COVID is still very high, seemingly comparable to other strains. So while you're much more likely to get it now, regardless of being vaccinated, the vaccines still provide enormous protection from death and severe symptoms.
But all of these people who are vaccinated and have been hanging out with mostly only vaccinated people, and therefore not really taking much in the way of precautions in those settings because the risk of transmission was low... That's the population within which we're seeing the Omicron tsunami.
And so of course the Omicron death rate is going to appear lower, when in actuality it is still as deadly as regular old COVID and Delta among the unvaccinated population.
Probably true that mortality to unvaccinated folks is still high as before, but it's much lower among vaccinated or previously infected people who are now testing positive for new infections. We'll know a lot more in another month.
Mortality among vaccinated folks was already very low.
Yes. It's (still) much lower among vaccinated people than among unvaccinated people, to be clear.
And so of course the Omicron death rate is going to appear lower, when in actuality it is still as deadly as regular old COVID and Delta among the unvaccinated population.
That may be the case. But it may not be. We know from epidemiological history that other coronaviruses have become A) endemic and B) less dangerous. This was very likely the case with the Russian "flu" virus (1889) — a pathogen that once killed millions, but is now relatively harmless.
I'm obviously speculating at this point (aren't we all?) but I reckon there's a fair chance that the declining lethality of omicron vis-a-vis previous variants (if this is indeed, as one hopes, an actual trend) will have resulted from a blend of stronger immunity in people and the variant's intrinsic attributes.
But in any event we won't know for some time.
It means that glass houses notwithstanding, the UK's death rate is still above the level you used to consider "ok."
It also seems to suggest that what I assume is 7 day trailing average for deaths per million per day is back in favor.
Re different countries,
Uk;
First hard to come to any relative conclusions from uk numbers. Their patterns are just different than any other nation. Cases have been way way higher for months than other nations ( some go higher for a while but uk always near the top) but deaths have stayed low for months too. Their cfr has just been much lower than everyone else.
And seems clear that their testing policies are just totally different. They are testing maniacs testing more than anyone. And so their confirmed cases pick up a huge percentage of actual infections and might even contain a decent number of false positives.
Re comparing us confirmed case counts with other European nations, that might be different due to testing differences but in the range of comparability. The uk is just too different. That does mean the uk cfr is getting closer to true ifr . Also makes me estimate our % of missing cases a bit higher. And ifr maybe really was just .25% even with delta.
Germany and France and also the more northern states in usa . Still were at delta peak or just coming off of it when omicron hit . And, especially as delta is more deadly and they lag cases , the deaths being reported today are still almost all from delta wave . And also somewhat for uk although they just have had a consistent high number and not a wave for delta.
Not true that totally too early to see some pattern for omicron deaths . Especially for uk where it hit earlier . Some people die quicker so lag is variable. How many of those few initial omicron deaths we see would indicate something even early.
But problem is that they are still being swamped by the tail of delta deaths from prior delta cases. So any small indication we could see from number of omicron deaths early is just " noise " in total deaths with delta.
Probably really need to wait three weeks or so after omicron cases overtook delta before total deaths are going to be a meaningful signal . When the variability in expected delta deaths is small enough that the excess attributed to omicron will have a small enough min max range to show something. Today, the range of possible omicron deaths subtracting expected delta is like zero to large , because the range of delta deaths is near or at total anyway.
Smart to look first to fla and deep south and Hawaii for first indication. Since they had reduced their delta to such low levels recently, new deaths can be attributed more to omicron sooner . Still a little early but fla could tell us something in a week or so. But won't because uncertain holiday distortions will confuse the issue ( are reported deaths low on jan 4th because omicron is so mild or due to holiday delay?).
But looking at rapid rise in cases now , seems almost impossible that deaths will not increase temporarily. Maybe omicron really is only 10% as deadly - some scientists are saying just that much . But if have 20× the cases due to rapid rise, still get twice the deaths. 20× would leave us around 2,000,000 confirmed cases a day vs. Before omicron . Yes , o think we might hit that. And maybe real infections of 5,000,000 a day ! Or more! And maybe really ifr will only be .05%, 1 in 2000. But that is still 2500 deaths a day .
But , at that level the wave will peak and plummet fast and rhis will be almost all over by February.
Note I can accept and live with deaths of 2500 a day or maybe up to 5000 , even if hospitalizations increase proportionally by more compared to year ago. That level is severely straining medical system but probably not breaking it . As long as this is just a very short period, ok.
Above that level or so , asking for serious trouble like what we saw in NYC and Italy in spring 2020. Real chaos.
And also one worth nobody is discussing. If omicron produces ten times more cases but one tenth of the hospitalizations, because so many mild cases , that is not the same stress on hospitals. More because now 10× more hospital staff out with mild sickness.
Even if omicron only caused mild sickness and zero hospitalizations, having 10% of the USA mildly sick at same time is pretty serious causing major disruptions.
With 7-day trailing average per-day, per-million seemingly back in favor, those countries with that statistic above 1.0 through the 22nd of December, 2021:
Rank Population (Millions) Country Deaths/Day/Million 7-dav Avg
1 1.39 Trinidad and Tobago 18.02
2 5.46 Slovakia 15.11
3 9.68 Hungary 13.79
4 4.00 Georgia 12.22
5 4.13 Croatia 12.00
6 37.89 Poland 11.75
7 7.00 Bulgaria 10.45
8 2.76 Lithuania 8.33
9 10.47 Greece 7.82
10 145.87 Russia 7.04
11 43.99 Ukraine 6.96
12 3.30 Bosnia and Herzegovina 6.92
13 10.69 Czechia 6.35
14 2.08 North Macedonia 5.83
15 1.91 Latvia 5.17
16 2.08 Slovenia 5.09
17 83.52 Germany 4.35
18 4.04 Moldova 4.35
19 444.97 EU 4.30
20 11.54 Belgium 4.18
21 329.06 US 3.96
22 512.50 EU w/o Brexit 3.96
23 8.96 Austria 3.73
24 10.10 Jordan 3.55
25 8.77 Serbia 3.47
26 2.96 Armenia 3.14
27 19.36 Romania 2.84
28 1.33 Estonia 2.80
29 17.10 Netherlands 2.77
30 65.13 France 2.49
31 8.59 Switzerland 2.46
32 96.46 Vietnam 2.42
33 83.43 Turkey 2.16
34 60.55 Italy 2.12
35 32.51 Peru 2.08
36 6.86 Lebanon 1.96
37 10.23 Portugal 1.75
38 67.53 United Kingdom 1.67
39 9.45 Belarus 1.66
40 1.15 Eswatini 1.62
41 4.88 Ireland 1.61
42 11.51 Bolivia 1.55
43 5.77 Denmark 1.53
44 5.38 Norway 1.46
45 7.04 Paraguay 1.44
46 10.05 Azerbaijan 1.44
47 5.53 Finland 1.42
48 2.88 Albania 1.39
49 51.23 Korea, South 1.39
50 18.95 Chile 1.30
51 1.18 Cyprus 1.21
52 127.58 Mexico 1.09
53 6.78 Libya 1.05
54 31.95 Malaysia 1.04
Just want to say thanks to Rick for posting this and all the prior times he has. It must be a good amount of work and I know i do not have the computer skills to pull this type of thing together.
Actually Rick's posts last year when he was giving updates so often is what first got me reading Kevin's blog every day.
My pleasure.
I dunno, to me it means Delta and Omicron are existing side by side. Delta remain a variant of the unvaxxed and Omicron being a variant of the vaxxed.
This would be very VERY bad for us.
Not really , vog .
Your worry about that seems mostly to be relieved by recent data , at least.
The % of sequenced cases that are delta and imputed delta cases do seem to dramatically go down when omicron hits.
You cannot expect that to happen overnight. In fact it is happening faster than it " should be" modeling based on assumption delta will be dominated by omicron and will be driven out . Kevin posted on that hard to explain too rapid decline the other day .
If omicron and delta really cannot exist long term side by side , you still would have some period where they both are existing side by side during the process of omicron becoming dominant. That is today. Delta still spreading but at R under 1.0 so cases going down fast.
But also you cannot use fact that delta deaths are happening at same time as omicron deaths as evidence infections are happening same time . Even if , magically, delta disappeared in one day when omicron hit, you would have overlapping deaths because deaths lag infections at a greatly variable amount . We often use 3 weeks as an average ( I think a bit longer for delta ) but really range is less than a week to 6 months.
On say Jan 31, we will have some delta deaths from infection in September and some omicron deaths from infection mid January.
Would add that omicron being at least similar enough to delta is also indicated by fact that vaccination ( developed for original) still is significantly effective, although reduced, for omicron ,especially if boosted . If vaccine antibodies looking for original spike can still recognize omicron at all , with all its spike mutations, there is still enough the same that it is not like a totally different virus . And should mean natural immunity from delta infection should be reduced even less by omicron , as that not focused on heavily mutated spike and more of unmutated parts to recognize.
But in one respect and one way bad news. Original antigenic sin. Means that an omicron infection is not going to reset immunity to be against omicron instead of delta if delta infected prior ( or vs. Original if vaccinated). The immune system stays most ready to fight the first strain it saw . So we will never had as good omicron immunity as someone never vaccinated and never infected before omicron. Not that I think there are many of those.
But maybe still good news . If immunity still remains focused on delta and an omicron infection improves delta immunity as much as omicron immunity, then delta ain't ever coming back . If an omicron immunity switched focus to omicron, delta could come back once immunity was better against delta..and they could take turns having waves, just not at same time.
I will happily accept worse immunity against less virulent omicron for better against delta, even if delta extinct . If that is why delta extinct.
I am expecting the deaths to be lower this time around, but the statistic we should really be reviewing at the moment is hospitalizations. It's kinda strange it's not very simple to get this health surveillance data, as well as the proportion of vaccinated v. unvaccinated (and boosted, and previous symptomatic COVID). We need to do a massive upgrade to our disease surveillance system, from hospitals all the way to wastewater, worldwide.
An emerging concern this go-around is that even if the death rate is much lower, with the suspected contagiousness and number of infections, combined with general fatigue by health care workers, we could see emergency health care systems getting very stressed, and thus the death rate rise as a result. At least we've got much more experience with treatment so less people are dying for that reason as well.
I realize it's nearly impossible to keep up to date on data, but, there's a reason why Omicron is resulting in lower hospitalizations: Its viral load is concentrated in the upper respiratory system not the lungs.
It will present as the flu or the common cold, not initially as pneumonia. Most certainly, lots of people with transition to a bacterial pneumonia infection, but most won't.
At any rate, compare hospitalizations to infections -- https://covid.cdc.gov/covid-data-tracker/#trends_dailycases_currenthospitaladmissions
You can clearly see how different this wave is.
Thanks for this information.
Thanks for the link. Had not gone there before. Useful.
But think it does not really show much yet , especially if trying to compare cases and deaths. The death lag is just mostly too long to really see much yet as omicron just started a little while ago . And those omicron deaths would just be starting to barely come in now and still any data there is swamped by delta deaths.
But , for first indications, I suggest look at fla as their delta cases before dropped so low which limits the delta interference with omicron trends . And look at hospitalizations, not deaths. Hospitalizations happen before deaths - should be much less than a week lag there. So should see plenty of omicron hospitalization by now starting . And with lower lag, the old delta hospitalization is much less of an issue especially in fla .
If you look at fla comparing cases and hospitalization graphs and eyeball the short lag , and look at the super obvious scary omicron wave, you do not see much of a hospitalization bump up even given the lag . Still early and possible omicron hospitalization just has longer lag and will just has not shown up later. Give it another week re fla cases va hospitalizations and same data and I will be much more confident.
Fla data , with such low prior delta and a definite rapid omicron wave is the canary in the coal mine for usa . And so far , at least for virulence, good signs that the canary will live.
But look at the slope of thar recent case line in fla . Yikes. Without as much delta going down offsetting it , the rapidity of omicron growth just stands out. That I'd what scares me. Possible number of cases at peak even if sickness per case is quite low . Original or delta never threatened the massive scale of a peak that omicron might. Hopefully we will see at least an infection point soon ..
Well, as soon as NICD data from SA came out, I made the controversial point that Omicron might be our Goldilocks way out of the pandemic.
If a virus is too infectious, it shortens the time that it can accumulate mutations to create a new strain to evade the immune system of the folks who'd just been infected.
Granted, nothing ever goes as planned -- it's a hell of a notion -- but even Pharaohs turn to sand like a drop in the ocean. So, I guess we're at the mercy of virologists and governments who know better, right?
https://www.biospace.com/article/omicron-vastly-outpaces-previous-strains-in-replication-/
https://www.google.com/search?q=omicron+upper+respiratory+%22cytokine+storm%22
https://www.telegraph.co.uk/global-health/science-and-disease/omicron-may-not-damaging-lungs-covid-variants-says-study/
Hi Kevin.
It could be just a case of who's getting it.
Most of the cases seem to be in the 20-39 and 40-59 groups, predominantly the former. The younger group is less likely to get vaccinated, but also less likely to develop severe symptoms and hospitalisation. The older groups are more likely to get vaccinated so will either be less likely to contract the disease or develop severe symptoms if they do. But it's all a big f*#$ing gamble!
I looked that this:
https://www.theguardian.com/world/2021/dec/21/covid-uk-coronavirus-cases-deaths-and-vaccinations-today
And scrolled down to "England: coronavirus case rates by age group".
Thanks very much for the blog, by the way. It's been very interesting.