The ethics of age-selective restrictions for COVID-19 control

By Bridget Williams, James Cameron, James Trauer, Ben Marais, Romain Ragonnet and Julian Savulescu.

One of the major controversies of the COVID-19 pandemic has been disagreement about whether age-selective measures should be introduced, with greater focus on preventing infection in older people but tolerance of some transmission amongst younger people. Some have advocated a path of focusing efforts on protecting those most vulnerable and tolerating transmission in younger people. Others have argued for minimising community transmission. This debate involves important empirical uncertainties; including the feasibility of effectively isolating older people and the consequences of allowing infection in a large number of younger people, as well as the feasibility and consequences of alternative measures such as strict border control and quarantine. It also raises ethical considerations, including whether introducing age-selective restrictions is unjust, and whether it is acceptable for a policy to tolerate foreseeable harms.

Here we address these ethical questions and suggest that, although the appropriateness of age-selective approaches requires further consideration of the empirical evidence, ethical concerns should not prevent its consideration as a policy option.

An age-selective liberty restriction approach in the context of vaccination

With global dissemination of SARS-CoV-2, it seems inevitable that herd immunity – which we define as the point at which population immunity is sufficient to maintain the effective reproduction number below one in the absence of movement restrictions – will be reached in the coming months to years. While SARS-CoV-2 may continue to circulate and cause seasonal outbreaks in the future, the effects of these outbreaks will likely be lessened by having some population immunity. The current challenge is mitigating the impact of the virus in a fully susceptible population.  The world’s scientists have delivered several safe and effective vaccines against SARS-CoV-2, which make the initial pathway to herd immunity much safer. However, as vaccines are rolled out, it is inevitable that some people will continue to become infected with SARS-CoV-2. Although host immunity to the virus remains incompletely understood, evidence thus far suggests robust immune responses following infection. Therefore, in many countries, herd immunity will be reached through a combination of vaccination and infection. The outstanding question now is the relative contribution of vaccination and natural infection as pathways to immunity in different settings, and which sections of the population reach immunity through infection and which through vaccination.

A modelling analysis by Ragonnet et al. presents the outcomes of different approaches to managing the COVID-19 pandemic. It suggests that an age-selective mixing strategy that minimizes contacts of those over 50 years while allowing those under 50 to interact close to normal levels, may make it possible to achieve herd immunity with a much lower level of mortality than without age-selective policies. The analysis included the six European countries that were worst-affected in the March-April epidemic peak. The analysis suggested that for most countries, had an age-selective approach been taken from October 2020, herd immunity may have been achieved over a 6-month period with fewer deaths than those countries have seen already under their current approaches, through a concentration of infection in younger people.

This was a theoretical model, which did not include vaccination, and there remain important questions about the feasibility and full consequences of this approach. However, it makes clear that age-selective policies could have a profound impact on mortality in this pandemic. These principles remain relevant, both for ongoing management in COVID-19 and for planning in future pandemics. Vaccine roll-out will take many months, even in countries with the most privileged access to vaccines. While extreme pressure on health care systems (such as is currently being experienced in the UK and many European countries) requires maximal suppression, countries will look to ease restrictions once this has relieved. This may be done homogenously across the population, or there may be staged, focused restrictions for different age groups. An age-selective approach may minimize the impact of the pandemic as vaccines are delivered, and may be important for limiting the burden of future epidemic waves.

There are important reasons to be cautious of a selective isolation approach, including difficulty in effectively isolating those at high risk, infection causing lasting sequelae in some younger people, and the potential for virus mutations with increased viral transmission. Each of these considerations should be taken seriously, and are critical areas of uncertainty that will be essential to quantify in future research.  Nevertheless, we are concerned that the analysis required to properly evaluate these questions and the potential benefits of an age-selective approach, has not occurred due to vague concerns of the ethics of selective lockdowns, and an aversion to tolerating any COVID-19-related harms.

Ethics of age-selective restrictions

Savulescu and Cameron previously proposed that selective liberty restrictions are more likely to be justified when those who have their liberty restricted are those who benefit most from the measures. In the case of COVID-19, with the marked difference in susceptibility to severe disease across ages, we argue that age-selective restrictions do not constitute unjust discrimination and are likely to be preferable to liberty restrictions for all.

Recently Lawrence and Harris have disagreed with this analysis, suggesting that liberty restriction places a greater burden on the elderly, which is not justified if pandemic control can be achieved through other means. They are right to point out the additional difficulties that may be faced by the elderly, and indeed these difficulties should be addressed by governments as part of their COVID-19 response. However, Lawrence and Harris fail to acknowledge the key differentiator between younger and older people in the pandemic; that greater susceptibility to severe COVID-19, and therefore greatly increased resource needs, creates different harms.

Liberty restricting measures are often justified on the basis they will prevent harm to others. Mill argued that harm to self is not sufficient grounds to interfere with liberty, but that harm to others is the sole ground for interference in liberty. This recognizes that people should be free to make their own decisions, including to identify and weigh risks to their own health. One challenge of infectious diseases is that people are not just the victims, they are also the vectors, and so their infection poses a risk of harm to others. But in a pandemic, this challenge is amplified, as people pose a risk to others through the potential spread of the disease and by contributing to overwhelming the healthcare system if they become ill. During the COVID-19 pandemic, a number of liberty restricting measures have been justified on the basis that they will limit the spread of the disease and so prevent the health system from being overwhelmed. On these grounds, various coercive measures could be adopted, including quarantine, isolation, lockdown, and surveillance. Under this framing, the extent to which liberty restricting measures are justified depends on the level of risk and potential severity of the harm.

In the COVID-19 pandemic an elderly person with COVID-19 poses a risk of harm to others in two ways: through acting as a potential vector of transmission, and through being more likely to need health care services and thereby contributing to the risk that services will be overwhelmed. While younger people also act as vectors of transmission, they have significantly less risk of having severe illness and requiring health care services. So, while we don’t normally take “use of limited health resources” to be a decisive factor in restricting liberty, in a pandemic we assert that it can be. And it is on this ground that those who are more at risk present a kind of harm that others do not present (even if both present the same risk of spreading the virus). And it is on this ground their liberty can be restricted.

Accepting some harms to prevent worse harms

Another aspect of an approach that tolerates infection amongst younger people is that it involves accepting foreseeable harms due to COVID-19. The analysis conducted by Ragonnet et al. suggested that if such an approach had been taken since October 2020, it could have been expected to result in additional COVID-19 deaths: approximately 5000 in Belgium, 28,000 in France, 45,000 in Italy, 23,000 in Spain, 3,000 in Sweden, and 48,000 in the UK. Each of these deaths would mean the premature loss of a person, who has intrinsic value themselves and who would likely leave behind many bereaved friends and family. This is a significant cost, and many may think that a policy that projects these outcomes should not be pursued. However, on the 24th of January 2021 the COVID-19 death toll since October in each of the listed countries is: 10,763 in Belgium, 41,212 in France, 49,567 in Italy, 23,650 in Spain, 5,112 in Sweden, and 55,906 in the UK. In all countries the death toll is already higher (and continuing to rise) than the projections from the modelled age-selective approach. In Belgium it is already roughly double. As mentioned, this is a theoretical model and there are important questions concerning this approach, including that the new variants of SARS-CoV-2 may be more transmissible in children and associated with more severe disease, making several pathways more dangerous, and that effectively implementing such severe restrictions on specific interpersonal contacts would be unprecedented. However, it makes clear that careful manipulation of age-mixing had the potential to have a profound impact on mortality, and an approach of homogenous restrictions has resulted in significant mortality as well as profound socio-economic harms.

A reluctance to accept any deaths may have contributed to more deaths actually occurring. Similarly, a reluctance to accept any COVID-19 morbidity amongst young people may result in greater harms to them as a consequence of the socioeconomic impacts of a pandemic with greater mortality and widespread restrictions. Acknowledging the heterogeneity in risk across age groups and incorporating this into restrictions may be important for limiting future harms from COVID-19, both during initial vaccine roll-out and in future epidemic peaks.

Evaluating all options

Another option may be following the approach of China, Vietnam, New Zealand and Australia and aim for elimination with strong border restrictions and strict quarantine. However, the appropriateness of this approach for other countries needs to be critically assessed. Countries face different challenges in this pandemic, and a successful approach in one country may not be optimal in another. Enforcing border restrictions and quarantine is likely to be much more difficult for European countries than for countries of the Asia-Pacific. If such strict border restrictions are not implemented, even if a lockdown brought local transmission to a halt the population would remain vulnerable to virus reintroduction until population immunity is achieved. This is not to suggest that an elimination strategy with strict border control should not have been considered for the UK or other European countries. All options should be considered – and their benefits, costs, risks, feasibility and ethical implications assessed as comprehensively as possible. But if this approach is not taken, and instead countries plan to maintain population movement restrictions as vaccines are distributed, then they should consider age-selective restrictions as an alternative to restrictions across the population.


Much of the rhetoric around this issue has hinted that those who suggest consideration of an age-selective approach are ageist and unconcerned with the fate of older people; callously willing to sacrifice them for a return to normal life. However, an approach of age-selective liberty restrictions as vaccines are rolled out, and in future epidemics, may result in fewer deaths, specifically among the elderly, while also allowing the most important social and economic activities to continue. Further analysis would be required to determine whether such a policy would indeed be feasible and effective at minimizing the harms of the pandemic, and how best this should be combined with rapid vaccine roll-out. However vague concerns about ageism and a reluctance to accept any harm due to COVID-19 or refusing to acknowledge that harm will inevitably occur, should not prevent such analysis from occurring.

Authors: Bridget Williams1,2, James Cameron3, James Trauer2, Ben Marais4, Romain Ragonnet2, Julian Savulescu1,3

Authors’ affiliations:

  1. Uehiro Centre for Practical Ethics, University of Oxford, Oxford OX1 1PT, United Kingdom
  2. School of Public Health and Preventive Medicine, Monash University, Melbourne, Victoria 3004, Australia
  3. Murdoch Children’s Research Institute, Parkville, Victoria 3065, Australia
  4. Marie Bashir Institute for Infectious Diseases and Biosecurity, University of Sydney, Westmead, New South Wales 2145, Australia

Competing interests: None

Social media accounts of post authors: @bridgetw_au @juliansavulescu

(Visited 397 times, 2 visits today)