Controlled infection with COVID-19 as a safer means of establishing herd immunity

Update: it seems (although clarity is lacking) that the UK has moved away from the herd-immunity strategy. I very much welcome this. It renders the argument below irrelevant, except as an illustration of the weaknesses of a strategy that seeks to vaccinate a population using an epidemic.

Disclaimer: I am neither a virologist, nor an immunologist, nor an epidemiologist. This post is written with confidence, because too many caveats hinder clarity, but should be understood to be the thoughts of a layperson.

This week the policy of the UK government on COVID-19, under advice from scientists, has been articulated clearly for the first time. The government intends for around 60% of the population to be infected with COVID-19 in the anticipation that after the majority who survive recover, the population as a whole will have ‘herd immunity’ to the virus, since if only 40% of the population can pass on the virus it will be unable to spread.

One can argue about the wisdom of the policy, which is in disagreement with the position of the WHO and many other countries. This post is not about that. What I am interested in is: given this policy (to allow 60% of the population to become infected), what is the least bad way to achieve this? For instance, should you let the virus spread at random through the population or should you actively infect people with the virus under very controlled conditions to minimise the risks?

The current approach

The government’s approach is to allow the virus to spread with relatively few restrictions initially. At a point before it overwhelms the healthcare system, it will introduce further social-distancing measures to attempt to “flatten the curve” and spread the 40 million infections over a longer time. There will also be some measures to try to ensure that the most vulnerable – older and immunocompromised people – are isolated, but it is unclear how effectively this can be achieved.

Risks of the current approach

**Infecting the vulnerable. ~**20% of the UK population is over 65. The fatality rate appears at least 10x higher in this demographic group, hence the government’s attempts to isolate them. At a point where a substantial portion of the population has the virus it is likely to be very difficult to achieve this. If the 60% of the 3.2 million people over 80 became infected, and the fatality rate of 14% for this age-group from China applied then that would represent 270,000 deaths in this age group alone.

Overloading the healthcare system. Exponential growth is a powerful adversary and small miscalculations could result in overloading the healthcare system with the result that many people die simply because there are no ventilators available to keep them alive. There is concern among some experts that it will not be possible to flatten the curve enough to ensure that  everyone receives adequate care. Exponential growth is difficult to avoid in natural infections because people who get the virus will not initially know that they have it, and so will unknowingly spread it to others.

Viral load. There are reasons to expect that the more particles  of virus a person is exposed to initially, the more likely they are  to experience severe disease. In the government’s approach there is no control over this and so it is possible that young and healthy people may develop severe symptoms due to exposure to a large viral load.

An alternative approach

We have seen that the government actively intends for 60% of people to get the virus. It is almost as if it were to simply go and infect people, as at a chicken-pox party.  So why not actually do this? Could there be advantages to actively exposing 60% of people in a controlled way to the virus to generate the same herd immunity?

Advantages the controlled exposure approach

Not infecting the vulnerable  – only fit and healthy people would be selected to receive the virus. Almost no children under 10 have developed severe symptoms from COVID-19. It would therefore seem very safe to put all immunocompetent children into the 60% who get the virus. Other young people are also comparatively resilient to the virus, and so a utilitarian argument would suggest they would  be the safest people to inoculate with the virus. We will see below that this approach may be safer for the people  who get infected, than the alternative policy of letting the virus pass through the population.

Not overloading the healthcare system  – there are two ways in which this controlled approach would avoid overloading the healthcare system. Firstly, by avoiding exponential growth: those infected would know they had been infected and could be isolated afterwards until they were no longer shedding virus. Once resistant, they could go out into the population, and safely work in crucial roles where they were likely to be exposed to the virus. Secondly by reducing the hospitalisation rate: because these people would be drawn from a much more resilient population the chances of them developing severe symptoms would be much lower. The end result of this would be fewer people in hospitals, meaning that the small(er) number of people who did develop severe symptoms would be able to get adequate care. The end result might be that a 100% chance of being infected, and receiving adequate care if required, might be safer than a 60% chance of being infected with the possibility of developing severe symptoms in a degraded healthcare system, which is dealing with exponential growth and vulnerable patients.

Controlled viral load, and possibility of an attenuated strain – there are several lines of evidence to suggest that the amount of virus one is initially exposed to may be a determinant of whether someone develops severe symptoms. By controlling the infection one could minimise the amount of virus someone is exposed to (while ensuring they develop an immune response). This  might again substantially reduce the risk compared to a 60% chance of receiving a random viral load (determined by how near you are to an infected person and how much virus they are shedding). Finally, although proper “no-risk” vaccines are unlikely to be available for some time, my non-expert brain wonders whether it is possible to rapidly develop a version of the virus that would be expected to be mildly attenuated by, for example, leaving the proteins the virus is encoding exactly the same, but altering the “codons” of DNA such that these proteins are likely to be produced less rapidly. The effectiveness of this could not be guaranteed, but on the balance of probabilities it seems very unlikely to be more dangerous than the wild-type virus.


The government are conducting an experiment in which  they will allow 60% of people to be infected. To my mind this is in substance the same as the government randomly choosing 60% of people and actively infecting them with a random load of wild-type virus, but not telling them they are infected with the result that they may spread it to vulnerable people. An alternative approach in which selected resilient people were infected in a controlled manner with a low-dose of an attenuated virus might be safer for the people infected, as well as society at large, than the current strategy.


I am neither a virologist, nor an immunologist, nor an epidemiologist. I would welcome explanations of where this argument falls down from such people, or anyone else. Furthermore, I am not arguing for this approach as superior to a containment strategy, merely as superior to a strategy of natural infection. Finally, this is a suggestion for an organised governmental approach, not a DIY ‘chicken-pox party approach’ which would not control viral load and would therefore have great risk.

Theo Sanderson
Theo Sanderson
Sir Henry Wellcome Fellow

Biologist developing tools to scale malaria reverse genetics.