Herd immunity may only need 10-20 per cent of people to be infected

Since mid-March there has been an assumption that herd immunity against Covid-19 would not be achieved until around 60 per cent of the population has been infected. It is a figure which gave rise to the now-famous paper by Professor Neil Ferguson of Imperial College, which claimed that a herd immunity policy (which the government denies ever following) would result in the deaths of 250,000 people in Britain. That figure has been challenged by scientists who have questioned some of the assumptions behind it – for example, it assumed a mortality rate of 0.9 per cent which Imperial College itself has since revised downwards to 0.66 per cent, and some believe is lower still.

Most popular

Katja Hoyer

How the Magdeburg Christmas market attack will change Germany

But today comes another challenge. A team led by Gabriela Gomes of the Liverpool School of Tropical Medicine argues that it is wrong to assume that herd immunity will only be achieved when 60 per cent of people have been infected. It is more likely, they argue, that the true figure lies between 10 and 20 per cent. The 60 per cent figure, they say, is based on the idea that we are all equally likely to contract the virus. In reality, there is a wide variation in an individual’s susceptibility to becoming infected. People who are frail or who have greater exposure to the virus – perhaps because they are working in an intensive care unit – are in practice far more likely to contract the disease. As the epidemic progresses the pool of easily-infected individuals dries up and the virus has to search out new victims who are less-easily infected.

Modelling by Gomes’ team aims to calculate the ‘coefficient of variation’, which quantifies the variability in individuals’ susceptibility to the virus – with zero denoting no variability at all (ie we are equally likely to be infected). If this coefficient really were zero, say the scientists, then herd immunity would only be achieved when over 60 percent of the population has been infected. If the coefficient were four, on the other hand, it would be achieved when 10 per cent of us were infected. The team then looked at real-life data to try to deduce what the coefficient of variation really is and concluded that it is in the range of just under two to just over three. That would mean herd immunity could be achieved when between 10 and 20 per cent of us have been infected.

The usual health warnings apply. Gomes’ work is theoretical modelling and, in common with a lot of material on Covid-19 that is being pre-published at the moment (including Ferguson’s paper of 16 March), it has not been peer-reviewed. But it is interesting that it gives an estimate for herd immunity of between 10 and 20 per cent, because that echoes real-life experience. The closest we have to a controlled experiment on the spread of Covid-19 was the cruise ship Diamond Princess, where the disease was able to spread uncontrolled in January, and almost all were later tested for the disease. Out of the 3,711 passengers and crew, 712 – or 19 per cent – were infected.

If herd immunity really is achieved at between 10 to 20 per cent it could mean that many parts of the world are approaching it – or are there already. A study of 1,000 residents in the North West German town of Gangelt in early April suggested that 14 per cent had already been infected (many without even knowing it). A study of 1,300 New Yorkers in late April suggested that 21 per cent have been infected.