The Office for National Statistics’ Covid infection survey suggests that about 92% of the adult population would test positive for antibodies — that is, has either had the disease, been vaccinated, or both. That sounds good, but 92% of the adult population is only about 74% of the total population. About a fifth of Britons are children.

But a lot of children (including my eldest!) have had Covid. The Cambridge MRC Biostatistics Unit estimates that about five million, or 80%, of five- to 14-year-olds have had the disease, and on top of that, a bit less than half (you can get the data here) of 12- to 15-year-olds have been vaccinated. Maybe that’s an overestimate, but we can confidently say that more than half of children have some immunity. On this naive calculation, then, we have very likely reached the herd immunity threshold, probably with a bit to spare.

There’s a complication, which is that “immunity” isn’t binary. I know lots of people who’ve had the vaccine but still got Covid; they’ve generally been fine, at worst having a nasty flu-like condition for a few days. But they still got it. The vaccine reduces your risk of getting the virus, and your risk of getting very ill if you do, but it doesn’t reduce it to zero. We can’t just say “90% have been vaccinated or had the disease, ergo we’re at herd immunity”. That said, Ferguson’s guess that we’re near it seems plausible.

It makes sense because we opened up so much earlier than other countries, meaning that we’ve gained a lot more immunity through infection (at the expense of a lot of lives lost). And it certainly would explain the wobbly plateau of case numbers that we’ve seen for the last few months. 

But here’s the note of caution. Herd immunity isn’t a fact about the virus: it’s an interaction between the virus and our behaviour. As an epidemiologist told me the last time I wrote about Covid case numbers, it’s much harder to predict the course of a pathogen when R is near 1. You can get unpredictable outbreaks in local patches, sub-populations with lower vaccination levels. Minor behaviour changes (more people socialising indoors as it gets colder? Christmas?) can push R up and alter things significantly. 

In response to that piece of mine, the forecasting site Metaculus put together some predictions for what was going to happen in the coming months — will we have further restrictions like Plan B or something more extreme; how high will the number of people in hospital get over winter, that sort of thing. 

Metaculus’s forecasts have a pretty good record in the pandemic so far. They think that Plan B-style restrictions are more likely than not before February 2022, and that there’s a good chance (≈25%) of more stringent ones, including a ban on household mixing at Christmas (≈20% chance). They also think that hospitalisations are likely to go up fairly significantly from their current level. (You can see their other relevant predictions here and here.) 

So while we may be at or near herd immunity, we’re also in a knife-edge situation where changes in behaviour could change things quite a lot, and change the threshold for what counts as herd immunity. Numbers could still go up, and we could still see ourselves facing restrictions. I think it’s vanishingly unlikely that we’ll see anything like last winter’s numbers. But this winter could still have some unpleasant surprises in store.

---

Tom Chivers is a science writer. His second book, How to Read Numbers, is out now.

TomChivers

---