A computational biologist weighs in on Omicron, the future of vaccines, and the CDC’s variant forecast

What do the data so far tell us about Omicron and whether it causes milder disease than previous Covid-19 variants? What can we expect to see as Omicron infections crash up against the country’s health care system? Why do Omicron waves seem to decline so quickly after scaling such heights?

We don’t know. So we asked Trevor Bedford, a computational biologist at the Fred Hutchinson Cancer Research Center in Seattle, for his thoughts. 

Bedford believes that while there may be something intrinsically different about the way Omicron viruses attack human bodies, much of what is being reported in terms of mildness of cases can be explained by the fact that many people being infected have some immunity to the SARS-CoV-2 virus, either because they were previously infected or they’ve been vaccinated.

That’s the good news. The bad news is that because Omicron’s mutations allow it to slip past antibodies people have acquired over the past two years, he thinks potentially half the globe may contract Omicron over the coming weeks and months.

How concerned is Bedford about Omicron? He said he uses a “freaking out” scale of 1 to 10 to rate pandemic developments. The Delta variant was a 6. He’s still not sure how bad Omicron is, saying at this point it falls somewhere between 3 and 8.

This transcript of our conversation has been lightly edited for length and clarity. 

Everybody’s looking at what’s happening with Omicron in South Africa and the U.K. and the Scandinavian countries. But what’s happening in Israel? It would seem to me that that would tell us an awful lot about, for instance, how good boosters are at turning this into a less threatening event.

That’s something that I’ve been interested in as well. If you look at the genomic data … we can see the early takeoff in Israel, Denmark, the U.K., the U.S., Switzerland — a bunch of places. It doesn’t look any slower in Israel in that two-to-three-day doubling [of new cases]. That is very much the same. 

What will matter more is as it starts to climb, do we see a difference in hospitalizations, do we see a difference in the ceiling of the wave? But the initial takeoff seems remarkably similar across all of these countries, despite different vaccine histories, different levels of first doses, different levels of infection, different levels of boosters.

And remarkably similar, also in terms of severity of disease?

If you calculate the case-hospitalization ratio as something like one-third or one-half of Delta, I still think this is entirely consistent with many more cases, if not most cases, being reinfections or breakthrough infections and immunity being the thing that is driving the decrease in severity.

The average case in South Africa and the average case in the U.S. and the average case in Denmark will all be less severe than the average case of Delta, just because of [immunity in the population].

But if that is the case, then there are some big pockets of this country that are at serious risk, are there not?

I actually don’t think so. If you do some really simple math, we’re about 35% of the country having been infected, very roughly. We’re about 70% of the country having had at least one dose [of vaccine]. Factored together, that’s 80-odd% [of people with some immunity]. And then, if you look at the seroprevalence work from the Centers for Disease Control and Prevention — this is only in people 15 and older — you get 90% of the country having antibodies. That range of 80% or 90% is about where we are so, yeah, I think we’re most of the way there towards everyone having some immunity.

Based on your read of the way this variant is playing out, would you then expect a lot of cases to be substantially milder?

I’m still expecting lots of infections, lots of cases, but that the average case will be less severe because it’s in a person with immunity.

Part of the huge confusion here has been when people talk about severity, they’re not being clear enough whether it’s severity of the average case or intrinsic severity [of the virus]. And so we have Imperial College’s [statement] “It’s no different.” But there they controlled for vaccination status and breakthrough status in their estimates. When you control for breakthrough infections and reinfections you see that Omicron appears to be no less severe than Delta, than previous variants — which makes sense. But in some ways that’s not fair. If we’re going to think about how impactful the wave is, most infections will be in people that have some immunity.

One of the things I find puzzling is just this massive peak of cases and then this really sharp drop-off. How do we make sense of that and do we expect it to be a one-time thing or would you expect Omicron to take off again?

That is actually probably more puzzling than the severity question. … Because … we expect something with a higher initial Rt [effective reproductive rate] to have a wave that’s bigger than something with smaller initial Rt. And when you look in South Africa, we see that Delta comes in with an Rt of 1.5 and Omicron comes in with an Rt of 3. We should expect a much larger wave with that Rt of 3. And that’s the thing that needs to be explained. 

I do think that, with more mild disease — through whatever mechanism — that we’re going to be having a smaller fraction of infections detected as cases. If you were actually able to measure infections, rather than cases, then you’d see perhaps twice as many infections at the peak of the Omicron wave in [South Africa’s] Gauteng province compared to cases. And then the other thing that I think is entirely possible is that if the way that spread is faster is not due to one person infecting more people but due to kind of faster turnaround between cases, so you can have more doublings in the same 10-day interval, then you can get initial rapid take-off that is super fast, but that wave will crest and fall and crash more quickly then a wave that’s driven by just increased transmission alone.

People are really flipping out about Omicron. Is there a possibility that we’re overreacting?

I don’t think we’re overreacting. When I’ve had this kind of similar conversation, I’ve tried to rank things on a Freaking Out scale of 1 to 10. Delta was a 6. Whereas here I’m still between 3 and 8.

In a recent Q&A you did with New York magazine, you talked about maybe this is what endemicity looks like. Is it possible that this is the first winter wave of endemic Covid?

It’s this 80% or 90% immunity stat. At endemicity, we’ll be at 100%. So, next year.

But 80% or 90% is very close to 100%. And the worry is that through things like Omicron and [antibody] waning and everything else that’s going on, that every winter we get significant attack rates. So I can imagine easily over the course of 2022 that 50% of people will catch Omicron or some other variant. So lots and lots of infections. Maybe we’ll stop counting cases as much because there’ll be a bunch of asymptomatic and mild infections that people don’t even bother to get tested for. A small proportion of severe disease and death. But I still could easily imagine 100,000 people [in the United States] dying every year from Covid in the endemic state just because you have a very big number of infections, even if the infection fatality rate is quite low.

You were saying on Twitter on Monday that we need to be working on updating the vaccine strain. How will we know when to do that? This variant is so different from other variants that if we put our eggs into the Omicron basket, what happens if what comes next isn’t a good match for Omicron?

We can expect winter waves of Covid. I think that is fair. And so then we want a vaccine [booster] campaign in September, October. And so, it’s just walking back from there. In order to have doses available in September you’ll need to have made your strain selection decision for an mRNA vaccine, say, in June.

We could be in June and Omicron will have swept the world and it’s only Omicron viruses. We could be in June and there could be Omicron and Delta and something else. But in June we would want to make that decision.

Anything I should have asked you but didn’t?

I am annoyed by the 73% news out of the CDC.

What happens is there’s an algorithm ticking along behind the scenes at the CDC website. It’s a fine algorithm. Because the genomic data takes two weeks to come in, it tries to do a “nowcast” to predict what the variant distribution is today. And then that algorithm does its update and comes up at 73% Omicron with some wide confidence intervals [CIs], and now the media cycle is reporting that the CDC data finds that 73% of current cases are Omicron. 

It’s a model. It’s not data. And they’re wide CIs. We don’t know what proportion of cases are Omicron in the U.S. right now, because genomic surveillance lags.