Rethinking the Ethics of Pandemic Rationing: Egalitarianism and Avoiding Wrongs

Abstract

This paper argues that we ought to rethink the harm-reduction prioritization strategy that has shaped early responses to acute resource scarcity (particularly of intensive care unit beds) during the COVID-19 pandemic. Although some authors have claimed that “[t]here are no egalitarians in a pandemic,” it is noted here that many observers and commentators have been deeply concerned about how prioritization policies that proceed on the basis of survival probability may unjustly distribute the burden of mortality and morbidity, even while reducing overall deaths. The paper further argues that there is a general case in favor of an egalitarian approach to medical rationing that has been missed in the ethical commentary so far; egalitarian approaches to resource rationing minimize wrongful harm. This claim is defended against some objections and the paper concludes by explaining why we should consider the possibility that avoiding wrongful harm is more important than avoiding harm simpliciter.

The Received View

Julian Savulescu, Ingmar Persson, and Dominic Wilkinson begin a recent paper with a striking claim.

There are no egalitarians in a pandemic. The scale of the challenge for health systems and public policy means that there is an ineluctable need to prioritise the needs of the many. ¹

Empirically speaking, they have a point. Across the world, wherever demand for acute care resources created by the COVID-19 pandemic has outstripped supply, rationing guidelines have been argued for and implemented that instruct health professionals to selectively deny certain individuals access to care, or withdraw it from them altogether, typically on the basis that they have been “objectively assessed” to have a lower survival probability, or a lesser capacity to benefit quickly from treatment, than their peers, or on the basis of very similar criteria. ^{2 , 3 , 4 , 5 , 6 , 7 , 8 , 9} These policies are straightforwardly inegalitarian—they selectively deprioritize for treatment those who meet certain criteria, and who would stand to benefit from treatment (i.e., these are not decisions regarding medical futility). John Harris, speaking out against what a casual observer might take to be a robust bioethical consensus, has correctly pointed out that prioritizing on such grounds, even in the name of “the aggregate health of the population,” involves not only overt discrimination, but overt discrimination against those intuitively “most in need of healthcare” ¹⁰ (i.e., the sickest patients).

Despite Harris’ conviction that “[n]o rational or halfway decent person will want or accept the terrible moral and human cost of abandoning those who need [clinician’s] assistance most…” ¹¹ the justification for these inegalitarian policies is supposed to be obvious. In cases where large numbers of people all require life-saving treatment, and there are not enough resources to go around, selectively prioritizing those with the greatest survival odds and/or those who are likely to benefit most quickly will, in all probability, save the greatest number of lives overall. This is because favoring those most likely to benefit quickly maximizes the number of potentially life-saving interventions that actually successfully save life (i.e., it minimizes the treatment “failure rate”), and maintains a steady “flow” of patients through intensive care unit (ICU). So, in fairness to those who Harris might consider neither “rational” nor “halfway decent,” not implementing a policy of selective prioritization on the basis of survival probability carries an important moral and human cost as well; significant, predictable excess death.

Although selective prioritization of this kind skews consequentialist in its reasoning, it is worth noting that a truly utilitarian outlook would also probably attempt to optimize the rationing system so as to save the greatest number of life years (i.e., selectively save those with the greatest post-treatment life expectancy) and adjust those life years for quality (i.e., deprioritise those who are expected to live for similar lengths of time as their peers, but at a lower level of assessed health or wellbeing). ¹² Typically, these latter two rationing criteria have been rejected for legal reasons—it is thought that implementing them would constitute, or be overwhelmingly likely to result in, unlawful (not to mention grossly unethical) discrimination on the grounds of age and disability. ^{13 , 14} Validated measures of in-hospital survival probability exist (e.g., SOFA or LAPS2 tests ¹⁵) and could be used to relatively objectively assess survival odds (although major medical authorities in the UK have suggested the use of far more dubious and subjective measures ^{16 , 17}). Medical assessments of life expectancy and future quality of life, however, seem to be far less accurate, and open to serious charges of bias and prejudice. ^{18 , 19 , 20} In keeping these kinds of decisionmaking out of rationing guidelines, healthcare systems have, it seems, tried to retain a modicum of egalitarianism as a check on what is otherwise a robustly economic logic—an imperative to make most efficient use of the limited resources we have.

The Problem with Unequal Treatment

Regardless of how egalitarian the existing policies are compared to those that true utilitarians might like to see, those of us who tend toward a strong egalitarian outlook in healthcare ethics have watched the implementation of these guidelines with no little horror. Maximizing lives saved is not the innocent goal that it may appear to be. For one thing, policies that aim at it may so grossly overburden certain sections of society with the resulting mortality and morbidity that the value of the initial goal is undermined.

To egalitarians like myself, how such burdens are distributed matters at least as much as the overall degree of burden. With a few notable exceptions, ^{21 , 22 , 23} the bioethical mainstream has largely sided against egalitarianism here, not seriously concerning itself with the fact that these policies risk (avoidably) distributing burdens of mortality and morbidity grossly unevenly across the population. Those with disabilities that alter their survival prospects and the elderly will take the brunt of such an uneven distribution (which, importantly, will stack on top of any uneven distribution that pre-existing vulnerability to infection and serious complications from COVID-19 may cause). Moreover, population and communities that have higher rates of prospect-altering chronic illness due to existing health inequalities will also predictably be exposed to increased burden (albeit indirectly), which, in the UK, will mean that such policies risk indirectly discriminating against minority ethnicities too, as well as those in the most socioeconomically deprived regions. ^{24 , 25} That such issues may expose a eugenicist logic underlying selective prioritization policies has been a point raised predominantly by disabled journalists and activists, ^{26 , 27} rather than bioethicists.

Although the use of the term “eugenics” to describe such situations is often considered inflammatory by the medical profession, it is not obviously unwarranted in this case. To the extent that eugenicist thinking is characterized by devaluing certain kinds of lives relative to others, the existing selective rationing policies are eugenicist, as they instruct clinicians to selectively disregard/discount the claims certain individuals have on healthcare, in ways likely to lead to their death, based on criteria over which they have no say or control, on the grounds that it will improve the aggregate health of the population. This is a serious devaluing of those claims, even if the disadvantaged group maps only indirectly and imperfectly (although in practice, significantly) onto protected categories such as age, disability, socioeconomic deprivation, and ethnicity.

Why Adopt a Different System?

With case numbers of COVID-19 once again rising sharply across Europe and the UK in particular, ^{28 , 29 , 30} it seems prudent to re-assess healthcare rationing policies, because there will again be widespread demand for critical care resources which cannot be met at current capacity. Quite apart from egalitarian worries regarding the distribution of the burden of mortality and morbidity that I have spelled out above, I think there is a more general consideration in favor of an egalitarian approach to the distribution of these resources, even where we know that fewer lives overall will be saved as a result. These considerations are classic ones in the philosophical literature on so-called “conflict cases,” ^{31 , 32} and involve minimizing wrongs done to individuals. One way of thinking about my goal for the rest of this paper is as attempting to clearly spell out exactly what, at root, the “terrible moral and human cost” ³³ of selective prioritization policies is, and to present a case in favor of avoiding this cost, even at the price of excess death, which goes beyond that provided by Harris. In short; selective prioritization policies inflict huge numbers of wrongful harms on individuals, and we have good reasons to avoid wrongful harms that do not apply to harms simpliciter.

A quick diversion. We intuitively distinguish situations where people are harmed, and those where they are wronged^{. 34}, ³⁵ If a slate falls off a roof and hits me on the head, then I have certainly been harmed (apart from anything else, because I experienced pain), but unless it was negligently installed or maintained I have not been wronged. Conversely, if I am locked in a room without my knowledge or consent, but I do not wish, and at no point attempt, to leave it (and so suffer no distress) then I have been wronged (by being detained) but not harmed. Many wrongs, of course, also cause harm (intentional, unprovoked physical assault is an obvious example). The harms that result from such wrongs may helpfully be termed wrongful. ³⁶

For our purposes here, we can avoid providing a perfect analytic distinction between these two concepts, and simply note two important dimensions along which the paradigm cases of each differ.

1. 1) A wrong, unlike a harm, requires an individual to be responsible for it. Harms can occur independently of human agency, but wrongs must be done to us. ³⁷

2. 2) A wrong can simply involve the violation of a right, without any individual being worse off (either than they would otherwise have been, or than they previously were) as a result of that violation. ³⁸

Typically, doctors want to avoid harm befalling their patients (whether directly caused by them or not), and also want to avoid wronging their patients. Rarely are these two aims in conflict. I argue, however, that in acute rationing situations they pull in opposite directions—harm reduction tells us to ration selectively, but this involves violating certain individuals’ claims to healthcare in a manner that clearly wrongs them. Let me explain.

I take it that most of us agree that patients who stand to benefit from a certain course of (ordinary ³⁹) treatment have a claim on that treatment—that is they can legitimately demand that healthcare providers fairly and equitably consider their welfare interests when deliberating over its use. ^{40 , 41} This is especially obvious in countries which fund universal health coverage in part or in whole by way of general taxation, but may well apply more widely than that.

Consider a simple rationing situation where an acutely ill patient (x) is denied a single available ICU bed on the basis of their having a lower survival probability than another acutely ill patient (y). In such a case, x’s claim to ICU treatment has been discounted to their detriment—regardless of the value of the goal that the health service is pursuing in so doing, x’s claim has been deliberately ignored in favor of y’s; they were given no chance at all to benefit from treatment, let alone an equal one to y. They have thereby been wronged, since their welfare interests in receiving the treatment were not treated as equally important as their competitor’s. Utility-minded individuals may, of course, think that this wrong is justifiable in pursuit of the overarching goal of saving lives. I will offer and motivate an alternative view in the final section. I submit, however, that (minimally) everyone should agree that x has been wronged, justifiably or otherwise. If x were thereby to die, their death would also be wrongful, in the sense that it was a clear consequence of an avoidable wrong being done to them (or, at least, the probability of their death was increased by their being wronged).

Some may object to this characterization of the situation. They may argue as follows. It is true that x’s claim to the resource means that their welfare interests must be fairly weighed when considering to whom the resource will be allocated. But since they have a lower survival probability than y, x thereby has a weaker interest in the resource than y does (since it is less likely to realize the benefit of saving their life). As a result, a fair weighing of x’s and y’s competing claims will find y’s claim to be stronger, without wrongly disregarding x’s claim. Thus, my opponents will claim, in the situation described above, x has not been wronged, as it is in fact fair to discount the strength of their claim.

I can see why some may find this line of reasoning tempting; it suggests that higher relative survival probability, in circumstances of acute resource scarcity, allows one’s claims to “trump” another’s. Thus, while harm may consequently befall the person with lower survival probability, this harm will not (so the argument goes) be wrongful. It will, instead, simply be the result of misfortune; they were unlucky that the weighing of their claim had to be done in a situation where means were scarce, and a stronger claim happened to be competing with theirs. This description of the situation makes it sound regrettable rather than dubious. But it also makes some strange assumptions about how interests and claims operate.

When it comes to claims on resources, we have interests in (potential) benefits—valuable states of being that the resources in question increase our odds of realising. We do not have an interest in the expected utility of an intervention that uses a resource on which we have a claim, because the expected utility is not a benefit that can be realized. It is, rather, an abstract representation of the value of that benefit discounted by the probability of its realization. Let me explain.

Rationing situations that prioritize on the basis of survival probability are not prioritizing on the basis of who will receive more benefit from the use of a resource. The potential benefit is the same in both cases, and equally valuable to each individual. The situation under discussion is one in which both individuals stand to (potentially) gain the same, enormous benefit. True, one individual has a lower chance of realising that benefit than another, if they are allocated the resource, and thus a lower expected utility from the use of it. But that is insufficient to show that they have a weaker interest in receiving the resource than their competitor. We have interests in the (potential) realization of benefits. Both individuals have an interest in realizing the same benefit here—namely, survival. We may further assume that both parties’ interest in survival is the same. Not only is it impossible to sufficiently quickly “rank” the value of survival in this way, it would be unreasonable, unethical, and potentially illegal, to ask doctors to habitually make decisions on the basis of whose survival is “worth more.” Moreover, as Harris rightly points out, any unwanted death is equally bad for any individual who experiences it, no matter their life expectancy or any consideration of quality of life; only this fact can explain why “murder is always wrong, and wrong to the same degree, regardless of the age or health state of the victim.” ⁴²

The expected utility of an intervention (unless zero) is immaterial to these individual’s interests, because neither stand to gain the expected utility as a benefit. What they stand to gain is survival. The expected utility of the intervention is merely an abstract representation of the value of survival, discounted by the probability that the value will be realized. It is not a benefit that either party might receive. Thus, it makes no difference to either party’s interests and cannot affect the strength of their claims. Wherever one’s chance of survival given that the resource is allocated to you is nonzero, one’s interest in being allocated the resource is the same.

So, a decisionmaking procedure that deprioritizes individuals on the basis of survival probability wrongs them, as well as harming them. In conditions of acute resource scarcity, some degree of harm resulting from prioritization decisions is unavoidable. But can we avoid wronging people?

Egalitarianism in Pandemic Rationing

Consider now a similar situation, where x is once again denied care in favor of y, but this time, on the basis of an egalitarian selection mechanism. Instead of discriminating against x or y, we give each of their competing claims an equal chance to be met (say, by flipping a coin or some other lottery-like system). y wins and so, once again, they receive care and x does not. We may assume that, once again, x thereby dies and y survives.

The important feature of this latter situation is that nobody has been wronged, ⁴³ although x has experienced the same harm in both cases and seemingly does no worse in either case. The harm in question is about as serious as harm comes, but a certain amount of it (for someone) was, by hypothesis, unavoidable. What was avoidable in the simplified rationing situation I put forward (and, I submit, was avoided in the latter response to it) was perpetrating any wrong.

It is impossible to meet both claims simultaneously in the situation as described. So instead, in order to treat these two individuals justly, we need to ensure that (1) each claim is given the same chance to be met, and furthermore that (2) this equal chance is the largest chance possible (in this simple instance, 50 percent). That is, we make a decision as to who receives treatment on the basis of the principle of Equal Greatest Chance (EGC). ^{44 , 45 , 46} In the case where EGC is met, x is not wronged, as their claim is given as much weight as is compatible with all the equal, competing claims being given equal weight, thus satisfying their claim on fair and equitable consideration. Similarly, any future claimants who would be saved by a less egalitarian policy (e.g., one where they would be able to “bump” y from an ICU bed if their survival probability is greater) are not wronged by such a decision either. Nobody has a claim on resources that cannot be provided without violating somebody else’s rights, so in the situation we are imagining, and may be confronted with, such individuals, regrettably, have no claim on resources at all, since none are available. This will change only if beds become available, or if an individual occupying one of those beds ceases to have any interest in continued treatment, say because they decline to the point that continued efforts are futile.

Some opponents may note that nothing in EGC tells us which benefits to base our prioritization decisions on. They may object at this juncture that we ought to look at a different, perhaps narrower, benefit/interest than mere survival when evaluating competing claims in an egalitarian way. Perhaps one ought to think instead in terms of each individual having an interest in surviving until, say, at least the age of 70, or without serious, ongoing chronic illness. So, for individuals over the age of 70, or who are guaranteed, if they survive, to live afterwards with a highly debilitating condition, it is either certain or impossible that they will gain the benefit in question. In the former case, a person over the age of 70 can be deprioritized in favor of somebody under that age since their chances of realizing the benefit is assured without intervention (since they have already done so!). In the latter case, a person who will survive only with a debilitating chronic condition can be deprioritized in favor of somebody who may survive without such a condition, because they have no chance of realizing the benefit in question. ⁴⁷

This way of thinking might be prima facie appealing to those who take the need to appropriately balance competing claims seriously, but are disturbed by the thought of allowing more harm than it may be possible to prevent. Yet such approaches are obviously ruled out on more straightforward egalitarian grounds—"selecting” the wrong benefits (and hence claims) to consider would mean overtly and directly discriminating against those with important protected characteristics (in this case, age and disability). This is, of course, part of the reason why hospitals and health services, in planning for such awful scenarios, have thus far focused their attention on survival probability as the key benefit (and thus interest) at stake; they are reluctant to prioritize on the basis of criteria that are obviously discriminatory (for legal reasons, and also because the overtly inegalitarian nature of such a process quite rightly horrifies many key actors). They have correctly identified that an interest in survival is shared by all parties in the most ethically challenging cases they may face, and so it is not (by itself) discriminatory against any group to focus on the potential benefit of survival when making decisions. That is to say, EGC does not dictate which benefits we should seek to distribute fairly, or under what description—it only tells us how to fairly allocate whatever scarce benefits we identify. Nevertheless, a broader egalitarian perspective obviously rules out focusing on benefits that will result in the systematic de-prioritization of particular groups. Any benefit capable of justifying current policy when distributed according to the EGC would likely do this.

The requirements of egalitarianism go beyond framing the benefits at stake in a neutral way, although this is necessary and achieved by focusing on survival. An egalitarian response also requires that we weigh those competing interests in a way that respects the equal strength of each individual’s claim on the resource (i.e., that respects EGC), regardless of survival probability. Not doing so threatens to impose an unfair burden of mortality and morbidity on groups (including the elderly and disabled) whose survival probability is likely to be lower on average from the start, essentially ensuring that this lower survival probability disadvantages them twice over—once as a matter of misfortune (and pre-existing structural barriers to health and healthcare access) and once as a matter of policy. ⁴⁸ It also means that deaths that occur as a result of being selectively deprioritized will be wrongful deaths, in the sense that the individuals died (in part) because their claims were unfairly treated as weaker than those of their competitors.

Why then should we care so much about the difference between death and wrongful death? After all, in both cases described above, x dies, and in the egalitarian case we may suppose for the sake of argument that others do as well. The fact that nobody has been wronged will, I concede, probably be of little comfort to those that die or their grieving families. Why care, then, about whether these deaths were wrongful or not?

We should care especially about all wrongful death, in my view, for the same fundamental reason that we prefer that people die natural deaths rather than be murdered, regardless of their age, disability-status, or life expectancy. Wrongful death is bad in a distinctive way that nonwrongful, although regrettable and harmful, death is not. The badness of murder is worse than the badness of death simpliciter, and is (moreover) bad in an additional, important way. Although the degree of the wrong in pandemic rationing may not approach the severity of paradigm cases of murder (although Harris makes a case that it does), ⁴⁹ death that is a result of unjust and inequitable treatment ought to leave us with a uniquely bad taste in our mouths. Just as with murder, death that comes about as a result of unfair discounting of a claim on lifesaving resources is worse than death simpliciter, and it is (moreover) bad in an additional, important way.

Moreover, policies that aim to minimize harm rather than avoiding wrongful harm make all those responsible for implementing and enforcing the policy agents of significant wrongs. ⁵⁰ This is not simply something that we ought to avoid doing as a matter of policy (although it is that). Being made to do wrong as a condition of one’s job, especially in the caring professions, has the potential to be emotionally scarring, or even result in significant moral injury. ⁵¹ It is one (very bad) thing to be exposed to an unusual number of seriously ill people, and to have to bear witness to their decline and eventual death without being able to help beyond palliative measures. It is quite another to be forced as a matter of policy to evaluate them and, on that basis, choose them for such a fate.

Moreover, the fact that it is impossible for one’s decisions to “do no harm” does not licence selectively ignoring individual’s claims, even in order to minimize overall harm. That it does so is implied by those views that hold that prioritization practises based on survival probability are justified by the state of emergency in which we find ourselves. The idea here is supposed to be that so much harm can occur in these situations that we are duty-bound to reduce it by means that we would not consider outside of times of emergency; in Harris’ terms, defenders of these policies are invoking a doctrine of necessity. ⁵²

We should not treat these purported moral consequences of emergency as obvious. For one thing, as previously noted, they necessitate making individual clinicians and clinical decisionmakers the agents of substantial volumes of wrongful harm. On the contrary, it is critical that when rights become hard to defend, we defend them all the more vigorously; that we do not, in Harris’ words “lose our grip on the values and the compassion that make individual and collective survival worth fighting for, or indeed worth having.” ⁵³ Otherwise, we risk abandoning our principles just when we need them the most, and inadvertently licensing the infliction of a huge number of wrongs in the name of avoiding harm. Where one cannot avoid doing harm, we should be open to the possibility that the next best action-guiding principle is to “do no wrong.”