What’s the cost of acquiring herd immunity to Covid-19?

Many are proposing a way ahead with Covid-19 based on letting the infection spread naturally through populations to allow them to develop herd immunity. How feasible might that approach be in the population of England?

There’s a well-known target required for this to work in any population: the Herd Immunity Threshold (HIT), the proportion of the population who must have immunity before spread slows and stops. This depends on the R0 for the infection: HIT = 1 – (1/R0).

For flu, R0 is typically around 2, making HIT 50%. Half the population must have immunity to that type of flu before herd immunity cuts in and that epidemic stops as a result. Note this R0 isn’t the same as the effective R which varies in time and place.

Estimates for R0 for SARS-CoV-2 vary widely, up to 4, but it’s unlikely to be much below 2. So for natural herd immunity to take effect, at least half the population need to have had Covid-19 and currently be immune to reinfection.

Next we need to know what proportion of the population are currently immune to Covid-19. There are several studies covering England, including REACT which gives 6% in early July, and latest figures from blood donors give 5.7% to mid September.

So to reach HIT, at least 44% more of the population need to have Covid-19 and develop immunity to it. For England, that’s just under 25 million from a total of 56 million. We’ll give them 3 months to get infected and immune, generously 100 days.

That works out at a quarter of a million (246,400) fresh cases of Covid-19 per day on average, for a period of 100 days in England alone, to reach the HIT of 50% to provide herd immunity by natural infection.

We now need to work out the consequences of that sustained high rate of fresh cases. The percentage who require care in hospital varies widely with population, anything from 1-5% or more. That’s anything from 2,460 to over 12,000 cases each day.

Overall death rate (case fatality ratio) also varies. Worldwide it has been over 3%, and 6% in the UK first wave, but could be as low as 1% now. That means there’d be an average of 2,460 deaths per day for 100 days in England alone. Total dead 246,000.

Even halving these figures for the cost of attaining natural herd immunity in England to bank on a miracle, it would require daily hospital admissions of well over 1,000 for 100 days, and the deaths of 123,000 people on top of the 46,717 already dead.

More likely those figures would be worse still, with daily averages of 6,000 hospitalisations and 3,000 deaths, for a total of 350,000 dead. That’s in England alone, which doesn’t have that hospital capacity.

That’s why society has invested so heavily in immunisation, which can produce herd immunity with almost no need for hospital care or death. But some seem hell-bent on putting the clock back to before 1770 and killing people when there’s a better way.

Next time someone suggests just letting Covid-19 spread so we acquire natural herd immunity, get them to show you figures for its impact. If they don’t, then they’re just selling you worse suffering and death. There’s a special place in hell for them.

I leave it as an exercise for those in other countries to step through the same calculations to apply to their own situation. Looking from here it’d be no better.

A final note: models also show that vaccinating when the effective R is below 1 makes it quicker and easier to confer artificial herd immunity. So out of control infections will make it harder to vaccinate effectively too.


https://www.the-scientist.com/features/why-r0-is-problematic-for-predicting-covid-19-spread-67690 on R0; REACT study https://www.medrxiv.org/content/10.1101/2020.08.12.20173690v2; Halloran ME et al 2009 https://www.springer.com/gp/book/9780387403137

This is a compilation of my Twitter thread.