
Originally Posted by
Trebuchet
Grant, as a doctor, do you believe that herd immunity can be achieved other than through vaccination; in other words by enough people getting infected? I keep seeing different stories; such as that that was what Sweden was planning on.
Anders Tegnell is on record saying that Sweden's aim was never to achieve herd immunity, in the sense of deliberately setting out to get everyone infected. I've no reason to doubt him. (Sweden, of course, is the subject of another set of mutual contradictory narratives.)
In general, I think people became hypnotized from very early on by the naive calculation of the herd immunity threshold, (R
0-1)/R
0, which comes out to 2/3 of your population if R
0=3. From that, they looked at the infection fatality rate, did a little arithmetic, and start yelling about hundreds of thousands of deaths. So from March to August this year, no-one has really been able to say the words "herd immunity" without becoming the focus of a toxic reflex twitter-storm.
But we
know that infection is not distributed at random throughout the population (which is a key assumption of the naive calculation); and we
know that R doesn't really resemble R
0 in any society in the world at present, because we've made huge social changes
specifically to achieve that end. So we
know that the 2/3 threshold is an overestimate.
People who carry out more realistic modelling exercises of how populations actually behave are coming up with "effective herd immunity" levels as low as 20%. There's a
recent preprint here, for instance. To that, we can add recent observations that there appears to be a degree of cross-immunity between common cold coronaviruses and SARS-CoV-2, which means that some individuals may be immune without having been infected with Covid-19.
Now, there are already places in the world with seroprevalence of Covid-19 antibodies of 20%. (Bit of wiggle room there, however, in that we still don't know what level and kind of antibodies are protective, and we do know that some people may have cell-mediated immunity in the absence of antibodies.) And there is suggestive epidemiology--the way the case rate in Sweden is declining without any changes in disease control measures, and while still far short of that 2/3 seroprevalence everyone got so exercised about; the way London seems to have (so far) avoided becoming a UK hotspot as lockdown relaxes, despite being earliest and worst hit in the first wave.
That said, there are places where seroprevalence is
higher than 20%. If 20% was some sort of magic herd immunity wall, then that wouldn't have happened. But we'd need to model the social structure
in those places to estimate the effective herd immunity threshold
for those places, before we could truly understand the figures.
So, basically, I don't know. What I do know is that it's complicated and messy, and that "60-70% for herd immunity" is a figure that only makes sense if you're planning to vaccinate a population at random, not if you're dealing with a real disease transmitting through a real population. I think we're seeing a little scintilla of evidence that in some settings "useful herd immunity" might already be occurring, while in other settings it very much isn't. And I know that, by any reasonable expectation, we're going to see a lot more cases this winter, so we're going to get more evidence, one way or another, of how close some societies are to helpful levels of seroprevalence developing in their populations.