Thread by @TheFerrariLab

Only a fool would wager on the herd immunity threshold for COVID after the year we’ve had … but I would wager this — 10 years from now we will still be debating the HIT for SARS-CoV-2 🧵1/

Five reasons why COVID herd immunity is probably impossible nature.com/articles/d4158…

The HIT is a magnificent, elegant theoretical result. It is very powerful for planning — it is the reason that we can make plans now to open schools and businesses in later even if we won’t have 100% of the population immune (vaccine or exposure) 2/

But you don’t need to know the exact magnitude of the HIT to plan … and frankly, you’d be a fool to plan around any exact magnitude. The magnitude of the HIT when you start planning is unlikely to be the magnitude of the HIT when you get there - Why? 3/

Because the HIT depends on the pathogen AND the population. The world in late 2021 or mid 2022 won’t be the same as the world in early 2020 when this discussion started. It will be a world with more mask wearing, changed social norms, more HVAC filtration 4/

The virus is unlikely to be the same either — evidence is that B.1.1.7 is more transmissible, so the HIT would (in theory) increase. Perhaps that will be replaced by an immure escape mutant that we don’t yet know about 5/

Placing in the HIT in a range is useful for planning (on the order of 50% is WAY different that on the order fo 90%). But the HIT is a moving target and we won’t know what it is when we get there until we get there … and even then we’ll still be debating 6/

We can’t know we’ve reached the HIT except retrospectively — and it’s not just as simple as the % immune when incidence goes to 0. Incidence can go to 0 before the HIT because of stochasticity. It can go to 0 after the HIT because of momentum/overshoot 7/

Some great work on critical slowing down, led by @drake_lab suggests that we might be able to see dynamical signatures of the approach to the HIT … but, and I think this might be an overly technical term, it is “hard”, so the operational value is still TBD 8/

(Brief aside) - @drake_lab HAS shown some cool potential for critical slowing down to be useful in detecting emergence; that is, approaching the HIT from the other side in re-emerging pathogens 9/ (journals.plos.org/ploscompbiol/a…)

As I’ve said before — the HIT is one of the coolest phenomena I know — it is elegant and makes for wonderful intellectual debate and we’ll be talking about it late into the evenings for decades 10/

The HIT has utility in public health planning as a concept, not as a target. It is an aspiration to motivate programs and evaluate relative investements. But the location of the HIT is better diagnosed by a historian after the fact than by an epidemiologist now. /end