What the Modellers Still Don't Understand About Herd Immunity

Bristol’s Professor Philip Thomas has a new piece in the Spectator this week. Readers may recall that I criticised his previous pieces for what seemed in my view to be wildly over-the-top predictions of the likely scale of the Delta surge.

In June, he predicted “an enormous final wave“, in which the virus “would quickly seek out the one-in-three Britons who are still susceptible: mainly the not-yet-vaccinated” and peak in the middle of July (the bit he got right) “at anywhere between two million and four million active infections“. According to the ONS, around 951,700 people in the U.K. were PCR positive in the week ending July 24th, and that appears to be the peak, which is less than half of Professor Thomas’s lower estimate.

He now admits: “The situation is better than I bargained for at the beginning of June and also better than my estimates a month later.” In fact, it’s so much better, that he thinks “the decline in active infections can only mean that England is about to reach the herd immunity threshold for the Delta variant”. By which he means that “around 86% of England’s adults and children must now be immune”. On this basis he argues that it is “extremely unlikely” that there will be a new Covid surge in the coming winter.

The problem with this analysis is it is still based on the SAGE assumption that herd immunity is a once-for-all-time thing, that was made harder to reach by the more transmissible Delta variant, but which we have now just achieved, mostly through vaccination, and it will now keep us safe.

This is a mistake. It fails to recognise that herd immunity is variant-specific, is much closer than you think due to prior immunity, and has already been achieved several times over. Herd immunity is the reason that each wave or surge has peaked prior to lockdown each time in England. The original wave peaked and declined ahead of the first lockdown due to herd immunity (possibly assisted by the warmer weather), and thus there was no exit wave in summer 2020 as restrictions were lifted. We did have new surges as new variants arrived (and the seasons changed) – first the ‘Spanish’ variant in the autumn and then Alpha in the winter. Both these surges peaked and declined ahead of their respective lockdowns as herd immunity was reestablished. Thus there was again no exit wave in spring 2021, but only a new surge later on when Delta arrived (just as U.S. states had no exit wave as they reopened in the spring but are having a Delta surge now). That too peaked and declined despite the easing of restrictions, again reflecting a restored herd immunity.

This means there is no reason there shouldn’t be a new Covid surge of some size (presumably smaller) in the winter, assuming a new variant shows up by that point. If it doesn’t, we may well have a wave of a different seasonal respiratory virus, perhaps our absent friend influenza.

The point is that herd immunity isn’t some single final endpoint that we have been gradually moving towards since early 2020, and which only mass vaccination has allowed us to reach. This is clear if only from the fact that the fully vaccinated are still being infected and transmitting the virus in large numbers – Imperial College this week estimated that 44% of infections between June 24th and July 12th were in the fully vaccinated.

The modellers, whether at Bristol or on SAGE, keep making the same mistakes, assuming universal susceptibility until infected or vaccinated, that lockdowns have been responsible for controlling infections, and that there will be a huge exit wave when restrictions are lifted. They never seem to notice that outbreaks consistently peak before or without lockdowns, and that exit waves never occur (except when there’s a new variant about at the same time).

So this is my view: Professor Thomas will be wrong about there being no Covid surge this winter (unless by some good fortune no new variant shows up by then) for the same reason he was wrong to predict an “enormous” exit wave this summer. Herd immunity is not a once-for-all-time thing, but variant (and season) specific. This means it is both closer than you think, and never final. But it doesn’t need to be, because such a small proportion of those infected are made seriously ill and die from this virus, whatever the variant.