Yeah, I think he meant “at the population level” but even there it definitely seems like it would matter, and for individual outcomes it’s obviously a huge issue. Not sure why that bit is what got excerpted though, this is the part I thought was most interesting, regarding what’s going on with the SA data:
I’m curious to talk to you about the state of play generally, but I wanted to start with what we’re seeing in Gauteng, where the wave already appears to be cresting, though many fewer people appear to have been infected than are vulnerable — especially when you consider that, while people who’ve been infected or vaccinated already may be protected against severe disease, there is hardly any protection against infection per se. Which means, in terms of infection, this is almost a virgin population, so to speak. How do you understand that?
Yeah, it’s a really good question. I’ve been looking into this a bit. The basic idea is that we can measure Rt, and there’s a very simple equation that will convert Rt into your population attack rate: how many people will be infected in the entire epidemic wave. That projection is quite linear. With Delta and the Delta Rt of 1.5 — when it was coming in — I was able to convert that to an attack rate that ultimately matched what we saw. I did that by assuming it would really be mostly targeting the people that aren’t vaccinated or infected previously.
With Omicron, and its initial Rt being three-ish, that same equation should give you something like 90 percent of the population infected. But from what we’ve seen in South Africa, it seems like the wave is crashing well before that. So something is going on.
What do you think it is?
The options that I have been thinking about — there’s five of them. They’re non mutually exclusive. So to go through …
Please.
First, there’s the simple limit to testing capacity. As things increase, our testing capacity doesn’t increase as fast, and so we’re missing more and more cases. That can give you a distorted picture — it could look like a plateau in Gauteng, but you could imagine it’s really a much higher crest.
Like the top of a mountain has been chopped off by bad testing.
I also bet we can expect a lot more underreporting of Omicron, compared to previous wave, because it’s more mild, either through existing immunity or through actual reduction of intrinsic severity. And if, on average, you’ve reduced the severity of cases, there’d be a lot of people that don’t bother to come to the hospital or to get tested. And so as a rough guess, you might go from like one in ten cases reported in South Africa to one in 20 or even one in 30 cases — that wouldn’t seem unreasonable to me. And that makes it so that at the same caseload of Delta versus Omicron you could actually have three times as many infections with Omicron.
We could also have a change in generation interval. If we have Omicron kind of doubling at this very fast two- or three-day rate, you don’t actually have to have Rt be three. You could have actually just made the whole thing faster without having the number of secondary infections being much higher. And we don’t have no way of knowing that at this moment.
The last two are, it might not be that the entire population is susceptible to Omicron. Maybe half the population is susceptible. And then, finally, I think there’s a network effect — that as things kind of percolate through the community, you can imagine those transmission chains circling back on themselves and hitting someone that has already been exposed.
