Omicron vs DeltaTagged:
Will the Omicron variant outcompete Delta? Starting to look like it. Will that be a bad thing? Dunno, could go either way depending on reinfection rate and severity.
Comparing Omicron’s Breakout Speed with Delta’s
Via Your Local Epidemiologist come pointers to people looking at the primary data in the African outbreak of Omicron.
First comes Ridhwaan Suliman, a senior researcher and applied mathematician at Cambridge University. He shows the 7-day rolling average of COVID-19 cases in Gauteng, vs time in days since outbreak. You can see the enormous Delta wave in blue; but in black is the Omicron wave rising faster than Delta did:
Different versions of this graph showing 7-day rolling avg #COVID19 cases in Gauteng floating around. For those interested, here's the latest update tonight:— Ridhwaan Suliman (@rid1tweets) December 5, 2021
Will share an analysis, including a look at hospitals and deaths, as soon as I get a chance.. #Omicron #Gauteng #4thWave pic.twitter.com/w2VWcsmLkl
That’s… enormously fast. Faster than Delta, which was enormously fast in its day.
Next, consider Tom Wenseleers, a professor of biology and biostatistics at KU Leuven. Just picking out the high points of a thread:
- I think he’s quoting an estimated Omicron $R_0 \sim 6.2$, with 95% confidence limits $[4.0 - 9.7]$!
- The things being traded off are: infectiousness, immune/vaccine escape, and severity of disease. If all 3 get optimized at once, it would be disastrous. If just one of them, then we can cope.
- Less conservative estimates give crazy numbers like $R_0 \sim 40$, when the record-holding measles virus is only 12 - 18.
- Other sources estimate 1.3x Delta, so $R_0 \sim 6.5 \times 1.3 = 8.45$. That means, to explain the rapid spread, Omicron must have more immune evasion.
He also shows some interesting graphs, showing how the variant waves take over from each other, driving their predecessors to near extinction:
We can also use GISAID sequence data to make another independent estimate of the growth rate advantage of Omicron over Delta. Instead of a logistic regression, we then use a multinomial (2 df) spline fit, as in https://t.co/3hdjojBnwW. pic.twitter.com/itQyAJUG7t— Tom Wenseleers (@TWenseleers) December 2, 2021
Delta is the purple wave. You can see on the right it is being totally eclipsed by the red wave of Omicron. And the Omicron wave is rising faster than Delta did.
This gives a slightly more conservative estimate of Omicron having a 0.30 [0.24-0.37] 95% CLs growth rate advantage per day over Delta, corresponding to it infecting 4.1x [3.0-9.1x] more people over the course of a single generation time of 4.7 days. pic.twitter.com/SJEdVIUEqP— Tom Wenseleers (@TWenseleers) December 2, 2021
I don’t know exactly how to interpret his “infecting NNN more people over the course of a single generation time of 4.7 days”; it’s clearly not $R_0$. But still, you can see Omicron is taking over in South Africa. So I’d put the probability of an Omicron takeover at north of 90% now.
Now, there have been anecdotal reports that Omicron cases have been milder illnesses. If that turns out to be true, and we do not know that yet, then this could be good: Omicron outcompetes Delta but is itself more survivable. Since the evidence here is so scanty, I won’t try to put a probability on it.
The other fact to gather from here, is that being freshly boosted is your best bet at immunity. Delta & Omicron specific boosters will probably happen next spring. Here, my probability estimate is maybe 80% or so?
Notes & References