A trade‐off: Antimicrobial resistance and COVID‐19

Abstract As we combat the COVID‐19 pandemic, both the prescription of antimicrobials and the use of biocidal agents have increased in many countries. Although these measures can be expected to benefit existing people by, to some extent, mitigating the pandemic's effects, they may threaten long‐term well‐being of existing and future people, where they contribute to the problem of antimicrobial resistance (AMR). A trade‐off dilemma thus presents itself: combat COVID‐19 using these measures, or stop using them in order to protect against AMR. Currently, I argue, we are choosing to continue with these measures, and thus to prioritize combatting COVID‐19, without adequate ethical reflection on the AMR‐associated costs of these measures. I discuss the magnitude of the possible costs and benefits involved in making the trade‐off in favour of COVID‐19, and their distribution. I highlight two salient aspects of distribution that can help determine whether combatting COVID‐19 whilst exacerbating AMR produces justly distributed costs and benefits: distribution between current and future populations, and distribution between existing geographical populations. Adopting this account, I argue that based on the magnitude and distribution of costs and benefits of combatting COVID‐19, we have good reason to rethink this trade‐off, and instead consider prioritizing protecting current and future people against AMR, but jettisoning measures against COVID‐19 that also exacerbate AMR.


| RESPONSIBILITY FOR THE ANTIMICROBIAL COMMONS
The World Health Organization (WHO) predicts that AMR will cause some 10 million deaths per year by 2050. 4 Currently, WHO's antimicrobial stewardship framework recommends international programmes to promote the appropriate prescription and use of antimicrobials in medicine and agriculture. 5 The goal is to preserve the antimicrobial commons, protecting against AMR. The commons is the global shared good of effective antimicrobial therapies, for which we have collective responsibility and to which we contribute via responsible use of antimicrobials and biocidal agents. 6 Unlike pure public goods, the antimicrobial commons is not an entirely non-competitive shared resource, because the more it is used, the more it is depleted and therefore unable to be used by others. This makes it a rivalrous, common pool resource, and international coordination to maintain it is more difficult: it is in each country's interest to continue supporting its own population via the use of antimicrobials in agriculture and medi- cine. Yet the more we use antimicrobials (especially ineffectively), the more selective pressure there is on microorganisms to evolve new mechanisms to resist the drugs, the effects of which are global. With innovation to develop new antimicrobial agents flagging, this resource is becoming increasingly competitive and non-renewable. 7 Attitudes of responsibility for the antimicrobial commons vary globally. For example, although one important aspect of antimicrobial stewardship is the responsible prescription of antimicrobials in medicine, adherence to guidelines is inconsistent. In South Africa, there is some evidence that healthcare inequalities and cultural resistance to ideas of collective responsibility for the global antimicrobial commons correlate with disproportionately high levels of antimicrobial prescription compared to many other countries. 8 Prescription rates are particularly high in private healthcare, where national guidelines on the prescription of broad-spectrum antibiotics do not apply. For lower-income countries, however, prescribing behaviours resulting in higher antimicrobial use do not correlate with measures of the seriousness of AMR in that country according to the drug resistance index (DRI). Rather, a DRI indicating a large AMR problem in countries such as India is more affected by lack of access to effective antimicrobials to treat resistant infections in that country. 9 In higher-income countries, the correlation between the seriousness of AMR according to DRI measures and high antimicrobial prescription rates is stronger. In the UK for example, though the National Health Service (NHS) has pushed for comprehensive national guidelines regarding antimicrobial prescription, high levels of broad-spectrum antibiotic prescription still occur (perhaps due to a continued emphasis on individual patient health and the reduced visibility of the health effects of AMR). 10 This affects the DRI of the UK more than the spread of AMR infections due to lack of access to effective antimicrobials.
Some argue that international environment and trade law should hold countries accountable for their contributions to AMR. 11 The Access Watch Reserve (AWaRe) database developed by the WHO, for example, if linked to the enforceable mandates expressed in its constitution, could set standards to protect and limit uses of newlydeveloped antimicrobials for participating countries. 12 Currently, AWaRe classifies and recommends antimicrobial drugs as first or second choice prescriptions according to their harms or benefits and current pathogen resistance to them. Purely advisory programmes such as this, however, have been shown inadequate to address clinicians' prescription behaviours. 13 Changing clinicians' prescription behaviours to eliminate the non-beneficial prescription of antimicrobials is one aspect of protecting the antimicrobial commons.
The second aspect involves prescription behaviour when antimicrobials can be expected to provide clinical benefit to patients, but where this benefit-and patient expectations-must be weighed against costs to the antimicrobial commons. Clinicians' decisions are complicated by their duty of care to their own patients, which they may consider to outweigh their responsibilities for the antimicrobial commons. 14 In addition, clinicians' perceptions of patients' expectations to receive antimicrobials increase inappropriate prescription. 15 Yet, where antimicrobial use can be expected to provide less individual benefit compared to the cost imposed on future people by depleting the antimicrobial commons, prescription may be morally impermissible. To avoid these difficulties, it has been suggested that instead of relying on clinicians' decision-making alone, we should consider taxing antimicrobials. 16 Such a strategy aims to disincentivize beneficial but non-essential uses of antimicrobials (i.e., use in cases where an infection is mild or self-limiting) in highantimicrobial consumption, high-income settings. It also provides revenue for funding antimicrobial research and development and conservation programmes.
One problem that continues to loom for the antimicrobial commons is the immediacy of other threats that temporarily increase antimicrobial use, at a cost to the commons. The COVID-19 pandemic is one such threat.

| AMR DURING THE COVID-19 PANDEMIC
An international 'grand bargain' between countries to protect the antimicrobial commons does not yet exist, and is unlikely to in the near future, although work is being done to support such global agreements. 17 Hospitals around the world are using different standards to prescribe antimicrobials, and to maintain general hospital hygiene using biocidal agents-particularly during the COVID-19 pandemic. In the UK, the introduction of telehealth, prescription of antimicrobials to prevent or treat co-infections, and increased use of These costs to healthcare systems-not to mention increasing mortality rates-are predicted to increase as more drug-resistant strains emerge.
In hospital settings, increased prescribing rates have already been noted in the US during the COVID-19 pandemic by the CDC. 23 In fact, one recent metastudy showed that 72% of patients from hospital settings worldwide were prescribed antibiotics to treat suspected co-infections in early 2020, whilst only 8% of patients actually had bacterial/fungal co-infections. 24 This discrepancy means that the benefit of antimicrobial prescriptions in such cases is itself minimal where co-infection is not occurring. Such non-essential prescribing decisions, even where they provide a small benefit to the individual, may contravene the current WHO guidance on AMR stewardship during the pandemic, where the expected benefit is not sufficiently great to outweigh the cost to the AMR commons. 25 However, they also highlight potential feasibility issues with current guidance. The WHO's interim report recommends that patients with moderate COVID-19 symptoms be left to self-isolate or be admitted to hospitals on a case-by-case basis, and that antibiotics are not prescribed 'unless there is clinical suspicion of a bacterial infection'. 26 This guidance relies on the time, resources, and the ability to assess the severity of a patient's COVID-19 and other symptoms, in order to determine the significance of the benefit they may receive via antimicrobial prescription, without in-person consultation being possible, in some cases.
Whilst these prescription trends in UK and US telehealth and hospital settings may not be reflective of prescribing rates globally, or even in all HICs combatting the pandemic, insofar as these trends are illustrative of the issue, and insofar as these countries themselves have large enough populations to affect global AMR, they highlight an area of legitimate concern.
Biocides are cleaning agents that often use antimicrobials to indirectly through scarcity in the supply of hand sanitizer products, as recorded in the UK setting. 31 Household uses of some biocidal agents pose the discussed risks in terms of both expected contributions to AMR, and cross-resistance. 32 The benefit of using biocides is also reduced in household settings, assuming the households are representative of the broader population. The average individual is less vulnerable to being seriously affected by COVID-19 (with one study in the first wave estimating a fatality rate of 0.14% for those under 60 years of age) compared to older individuals often in hospitals and other institutional care settings (with estimated fatality rates between 6.4% for those over 60 and 13.4% for those over 80). 33  I turn now to discussing how the costs and benefits of the actions discussed here are distributed.

| Distribution across current versus future populations
Assessing the morally significant costs and benefits of our current actions can be difficult when these outcomes fall across time and generations. Often, short-termism prevails in such assessments, either by default or due to political pressure. Short-termism has been described as the maximization of short-term performance at a cost to future competitiveness or viability (usually of a firm, as the term arises in business ethics). 35 The concept can be usefully applied when we consider the current prioritization of combatting COVID-19, which benefits current people, at a possible cost to their future selves and future generations.
Analyses have been proposed to assess outcome distribution over time and generations-particularly when it comes to future harms. Hilary Graham notes that it is 'difficult for the public health community to direct attention to conditions for future health' because 'risk-factor epidemiology pinpoints risks in temporal proximity to the individual […] and economic evaluations weigh policies according to their value to the current population'. 36 I argue that we can and ought to adequately account for future harms without falling vulnerable to short-termism, which risks turning the current population into a temporally privileged minority that unfairly benefits from temporal health inequalities.
Intergenerational justice demands that, in order to respect future persons, we treat harms to their well-being as equally significant compared to harms to existing persons' well-being, at least once differences in certainty of those harms are accounted for by some discounting. 37 That is, future individuals must be treated as the moral equals of existing individuals and be given 'equal opportunities, and an equally high quality of life'. 38 Moral equality of individuals over time also holds that we ought to treat harms that may occur to the same person, but in their future, as equally significant compared to current harms occurring to them.
This contrasts with the 'discounting' that we commonly undertake, wherein 'individuals discount risks in their own lives-future risks being regarded as of lower consequence than current risks'. 39 Finally, environmental ethics uses the concept of sustainability to avoid discounting future harms, demanding the recognition of 're- economic or moral significance of effects on future people as they are increasingly more temporally distant, due to their reduced certainty-as long as the reduction is not hyperbolic, and as long as it does not discount future individuals' well-being as such. 41 The future costs of AMR can be to a small extent discounted in this way, but given how we already see the effects of AMR and will do so in the near future, the discount rate should be low. In the COVID-19 context, allowing continued over-prescription of antibiotics for the supposed prevention of co-infection, or in telehealth for beneficial but non-essential treatment that imposes too great costs to the antimicrobial commons, is highly likely to contribute to these harms, based on our knowledge of the effects of antimicrobial use. The corresponding benefits of preventing the effects of co-infection fall on current people, and not all those for whom antibiotics are prescribed as a preventative measure, at that (as some of these individuals would presumably not have been co-infected anyway). As for the use of biocidal agents, the cross-resistance effects detailed in the section above are highly likely to contribute to the predicted mortality rates from AMR in the future, posing a significant cost to future generations, whilst the benefits of reduced household spread of COVID-19 fall primarily on the current population. 42 Overall, we should consider all people-current and future, near and far away-as having equal moral status. In doing so, we assume basic moral equality as a grounding, above which an account of distributive justice may be applied. If this moral equality stands, then we have cause for concern when it comes to the COVID-19-AMR trade-off. The harms caused by prioritizing COVID-19 in the COVID-19-AMR trade-off are significant, probable, and fall disproportionately on future people. This may contravene the requirements of a theory of justice grounded in basic moral equality of persons. For instance, sufficientarian and prioritarian accounts of distributive justice require that we not impose the costs of an act or intervention on the worst-off, or at least not on those already below a minimum level of well-being. Whilst the status of current versus future people in terms of well-being levels can be difficult to determine, we may assume that, as a general, global trend, currently existing people who are worse off now are likely to remain so in the near future, whether this be a future with high levels of AMR or not. Similarly, those currently well off are likely to remain so. The pattern extends intergenerationally, wherein the descendants of these groups are likely to be similarly well off to their parents, following a trend of intergenerational transmission of poverty. 43 I acknowledge that this is a limiting assumption, in that my following argument only stands if it is indeed the case that levels of wellbeing can be expected to remain similar between generations, and thus that, due to the inapplicability of sufficientarian/prioritarian accounts of justice intergenerationally, we are to refer to the more basic idea of moral equality between individuals.
In contrast, the benefits of prioritizing COVID-19 are quite certain, and fall primarily on existing people-although it must be acknowledged that there is some benefit to future people of combatting the current pandemic more effectively, given an anticipatable trickle-down effect this has on them via the well-being of current people. For instance, if the effects of the current pandemic are mitigated, healthcare systems world-wide may be less depleted, leaving more healthcare resources for future generations. The same applies to economies, globally. The point remains that it is current people who benefit the most from our combatting COVID-19 using the measures I am discussing. Their magnitude may be less than that of the future harms of AMR, depending on the effectiveness of antimicrobial prescription and use of biocides on increasing well-being during the pandemic-a benefit that I have already discussed as limited in magnitude given the often non-essential use of these measures.
Subscribers to Parfit's non-identity problem 44 may object to my analysis, claiming that we cannot truly say that even a future with depleted antimicrobial commons would be against the interests of future people, or harm those particular people. Mirroring Parfit's depletion scenario, some might argue that the counterfactual circumstances-that is, prioritizing AMR over combatting COVID-19, and undertaking effective stewardship programmes that affect health behaviours-would cause different people (perhaps in number as well as identity) to exist in the future. If we assume that the lives of future people in a future with depleted antimicrobial commons would still be worth living, then it cannot be in those particular people's interests to not exist due to prioritizing effective AMR stewardship over COVID-19 action. Indeed, on some accounts of harm, they would be harmed by not existing if we were to effectively undertake future-changing AMR stewardship. However, as Parfit himself notes, there may be plausible claims regarding the impersonal costs of a (antimicrobial commons-) depleted future, or impersonal benefits of a (antimicrobial commons-) maintained future. That is, an antimicrobial commonsdepleted world may be worse off compared to a world (of different people) resulting from prioritizing AMR stewardship practices now. 45 In that case, we can consider simply the impersonal costs and benefits involved in the COVID-19-AMR trade-off. In aggregate, the impersonal harms associated with combatting COVID-19 using measures that exacerbate AMR may be greater than the aggregate harms of doing the opposite.
If this is so, then it appears that an ethical analysis of the dis- Then again, we may consider COVID-19 futures as well. The benefit of existence in a future world in which COVID-19 had been less effectively combatted due to the prioritization of protecting the antimicrobial commons would still hold for the particular individuals in that future. However, given the magnitude of difference made by AMR action compared to COVID-19 action, it is likely that the impersonal benefits of AMR action are greater than the impersonal benefits of COVID-19 action, for possible future worlds.

| Distribution across existing geographical populations
The distribution of costs and benefits of making the COVID-19-AMR trade-off in favour of combatting COVID-19 (as we currently are) must also be assessed across existing populations. Geographical distribution of costs is especially notable for COVID-19 and AMR as global phenomena whose effects defy borders. The global significance of both COVID-19 and AMR seems to demand a theoretical approach of moral cosmopolitanism, 46 requiring that the theory of justice subscribed to applies globally, rather than within national borders. Yet, we commonly undertake spatial social discounting as well as temporal discounting when it comes to the costs of current actions, resulting in an approach that employs a degree of moral nationalism. 47 The well-being costs of our actions for those who are far away are thus less salient than the effects for our neighbours. If we should avoid reliance on moral nationalism for issues with global morally relevant consequences, then the application of our account of justice should be applied globally. Two commonly accepted accounts in pandemic ethics that might be applied are prioritarianism, and sufficientarianism. Prioritarianism demands that the distribution of costs of an intervention not to fall on the worst-off. Sufficientarianism demands that they not fall on those already below a minimum threshold of well-being. 48  Based on emerging data, we can surmise that the prescription of antimicrobials during the COVID-19 pandemic and increased use of biocidal agents has primarily occurred in well-resourced HICs, where antimicrobials and biocidal agents are available. 50 The health and well-being benefits of these uses primarily accrue, then, to individuals who are already advantaged. This is because they have betterresourced healthcare systems to support them, and often do not suffer from other untreated diseases that may exacerbate the effects of COVID-19 infection. Additionally, given that elderly people are most vulnerable to serious health outcomes of COVID-19, it is older populations (therefore, primarily those in HICs) who we can expect to be most benefitted by these measures. The same trend in benefit falling on the already-advantaged may persist in access to antimicrobials and benefits of combatting COVID-19 within HICs as well, with low-SES background groups or individuals being less benefited than the better-off. On the whole, however, the trend holds better globally. The vast majority of those who are already below a plausible minimum threshold of well-being (say, the poverty line), are located in LMICs, and do not have as much access to the benefits of increased antimicrobial prescription and biocide use as those in HICs. 51 Thus, the well-being benefits of combatting COVID-19 may seem to fall primarily on the better-off, contra the requirements of prioritarianism and sufficientarianism.
Yet the contributions that current antimicrobial prescription behaviours and biocide uses make to AMR have more global costs. Although either COVID-19-associated harms or AMRassociated harms will be exacerbated by making the trade-off in one direction or the other, the magnitude of harm arising from AMR can be expected to be more significant than that of COVID-19. In 2020, COVID-19 killed around 1.6 million people. 54 However, the scope of the pandemic is limited, with effective vaccines already being distributed. Whilst more deaths are ex- The magnitude of expected harm is much greater for AMR than COVID-19, according to plausible estimates. In terms of likelihood, the costs of AMR in 30 years are less certain than the costs of COVID-19 in 2020-2021. However, we can extrapolate reliably from past and current effects of AMR, such that the predicted costs of AMR are still highly probable and of large magnitude, given current trajectories.
The harms of AMR fall primarily on future populations and our future selves. This runs counter to the standards set by moral equality and intergenerational justice. The harms fall around equally on the populations of LMICs compared to HICs, and thus advantaged and disadvantaged populations. In contrast, the benefits of combatting COVID-19 fall primarily on current populations and the populations of HICs (though these effects are somewhat mitigated by the global nature of the pandemic).
If the analysis I have proposed here is adopted, it seems clear that we have strong moral reason not to make the COVID-19-AMR trade-off in favour of combatting COVID-19. This case is only made more convincing by how small some of the benefits of current antimicrobial prescription behaviours and biocide uses are, particularly where over-prescription occurs and where biocides are used in an average household, as opposed to an institutional care setting.
One aspect of analysis I have not considered here is whether, where the costs of responsible antimicrobial stewardship are low, we (those combatting COVID-19 using antimicrobials) may have duties of 'easy rescue'-that is, taking on a small sacrifice in order to protect or promote significant broader interests or protect from significant harm to others. 56 It may be that such a duty should be enforced by governments in HICs, through some of the legislative measures discussed in the antimicrobial stewardship section of this paper.
Through combatting this pandemic, the damage to the antimicrobial commons may already be done, but my analysis applies also to future pandemics of similar scale. We must not fall prone to shorttermism or spatial discounting. Rather, we must be informed by ethical analysis, and act on that analysis effectively using legislative tools, in order to avoid acting impermissibly where our pandemic responses affect the antimicrobial commons.

ACKNOWLEDGEMENTS
I would like to acknowledge comments and discussion that contributed to this work from colleagues in the Uehiro Work in Progress Seminars and from supervisors Alberto Giubilini and Julian Savulescu.

CONFLICT OF INTEREST
The author declares no conflict of interest.