On the Morality of Reallocating Life-Sustaining Interventions in Times of Scarcity

Abstract

I. INTRODUCTION

As coronavirus disease 2019 (COVID-19) made its way to the United States in the spring of 2019, health systems began to develop guidelines regarding the allocation of what was expected to be scarce medical resources. We began to see triage protocols based on patients’ prognoses and factors such as sequential organ failure assessment (SOFA) scores. Ethical conversations centered around how to make the most of our resources when we do not have enough for everyone who could potentially benefit from them. Considerable focus was given specifically to the allocation of life-sustaining resources such as ventilators. In accordance with triage protocols, the medical community braced itself for the possibility of having to make tragic decisions, unilaterally withholding life-sustaining interventions, that is, regardless of patient wishes, from those who, compared to other patients in need, were less likely to survive to discharge. Although such allocation decisions would be painful, bioethicists and policymakers believed that they would be morally justified when there is true scarcity and triage is absolutely needed to save lives (Emanuel et al., 2020). While the United States was hit hard by COVID-19, it was lucky enough not to have to face the kind of ventilator scarcity that we feared. Other countries, such as Italy and India, were not so fortunate. In the meantime, waves of COVID-19 continue to bring new variants. It may be difficult to predict public health crises and the strain they put on resources, but public health professionals expect more pandemics. While the medical communities in wealthy countries may be prepared to absorb the strain with minimal triaging of life-sustaining resources, those who work in poorer countries continue to make difficult decisions regarding how best to provide care to patients when resources are scarce.

Through all of this, a related question remained morally controversial. While providers may be morally justified in unilaterally withholding life-sustaining interventions from patients with relatively poor prognoses, would they similarly be justified in unilaterally withdrawing such support from one patient (whose life might be saved by its continuation but who would die without it) in order to reallocate that resource to some other patient who has a better prognosis (but who would die without it)? Although a number of bioethicists (e.g., Emanuel et al., 2020; White and Lo, 2020), including some who often write from theological perspectives (e.g., Daley, 2020; Eberl, 2020), support the reallocation of life-sustaining resources such as ventilators and extracorporeal membrane oxygenation (ECMO) machines in times of scarcity, others have rejected such proposals (Sulmasy and Sugarman, 1994; Chu et al., 2020; Sulmasy and Maldonado, 2021a). In general, they have feared that the seemingly consequentialist rationale behind such reallocation may be at odds with the values of medicine or in conflict with traditional Christian norms, such as human dignity or the sacredness of human life. However, claims of moral wrongness may be somewhat ambiguous if they are not clear about the specific deontic classification of the acts in question. It could be that ventilator reallocation is intrinsically wrong, that is, always wrong no matter what. Alternatively, it may not be intrinsically wrong, but still wrong (perhaps even gravely so) under nearly all circumstances. This might leave open the act’s permissibility under at least some conditions. No one would morally question the intention to save lives and steward resources if it were not at the expense of other important moral considerations. However, if saving lives and stewarding resources comes with clearly foreseeable harms, a moral justification is required. Given concerns about the possible intrinsic wrongness of reallocating life-sustaining interventions, the foreseeable harms, and their relation to the overall intention to save lives, the doctrine of double effect (DDE) has been applied to assess the morality of the act (Chu et al., 2020; Eberl and Donovan, 2020; Tham, Melahn, and Baggot, 2021). The DDE is rooted in the writings of St. Thomas Aquinas (1948, II–II, Q. 64, Art. 7) and is widely used in the Catholic tradition to determine whether actions that cause harm may be morally justified under certain conditions. The following is a typical rendering of the DDE (Mangan, 1949; Cavanaugh, 2006; Furton and Moraczewski, 2009; Austriaco, 2011; Westberg, 2015): It may be permissible to bring about harm or evil so long as. . .

• 1) The act, considered by itself, is not intrinsically evil.

• 2) Only the good effect is intended.

• 3) The bad effect is not the means to the good effect.

• 4) There is a proportionate reason to tolerate the bad effect.

In light of the DDE, it might be that reallocating life-sustaining interventions, in and of itself, is somehow intrinsically wrong (Condition 1). Alternatively, even if such reallocation is not intrinsically wrong, its impermissibility may stem from the end that the clinician is seeking (Condition 2). Or perhaps the concern is that the clinician is intending or causing some evil (e.g., the death of a patient) as a means of saving the life of another patient (Condition 3). Even if the reallocation is not sought for some evil purpose, it still may be wrong because there is no proportionate reason to do so under the circumstances (Condition 4). So, at a time when there is still disagreement among clinicians and ethicists regarding the moral status of reallocating ventilators, the DDE provides a needed focus to our dialog. Now even if the reader ultimately disagrees with my conclusions, we will have achieved a more precise understanding of the source of our disagreement.

In this article, I argue that the DDE may support the reallocation of life-sustaining resources from patients with a poorer likelihood of surviving to discharge to those with a better likelihood so long as it is tied to evidence-based, transparent triage criteria (e.g., benefit, need, SOFA scores) and comes with effective ongoing communication with patients, families, and communities (Antommaria et al., 2020). While concerns about threats to human life, dignity, public trust, and social justice are understandable, I propose that there is no obvious reason to classify the act as intrinsically evil or done for some evil means or end, even if we assume a broad third-person conception of intention. Ultimately, my application of the DDE suggests that the disagreement surrounding the reallocation of life-sustaining interventions is less about its absolute impermissibility or permissibility and more about the circumstances under which it might be justified and the manner in which it is done. In what follows, I speak primarily of the reallocation of ventilator support, but the analysis is intended to apply to any life-sustaining intervention that may be reallocated in times of scarcity. The term ventilator reallocation refers to the unilateral discontinuation (i.e., absent consent) of ventilatory support from one patient earlier than it would have been discontinued without a scarcity of that resource, so that it may be available to save the life of another patient who may have a more favorable prognosis. It is assumed that both the patient being extubated and the potential patient to whom the ventilator may be reallocated would almost certainly die without the resource.

II. REALLOCATION AND THE DDE

Now we turn to our question about the DDE and acts of reallocating ventilators. In this section, I provides a brief argument for why its conditions may support the reallocation of ventilators and addresses the likely objections in Sections III and IV. First, let us clear away an objection that DDE does not even apply to ventilator reallocation. Daniel Sulmasy and Fabien Maldonado (2020, 2170) argue that the DDE is only intended to evaluate single acts, but that the reallocation circumstances we are considering involve two separate acts—one of discontinuing and another of initiating ventilator support. This kind of objection is certainly understandable, but a red herring nonetheless. First, we have to keep in mind that the only act being morally assessed is that of withdrawing ventilator support from one patient for the purpose of making available a resource that may allow another patient to survive. The fact that the intended good effect (making available a scarce resource) requires further action seems irrelevant to the application of the DDE. Every action undertaken by an agent is specified by an object and done for the purpose of achieving some good. Goods may be intended as means or ends and are often willed along a means-end continuum in which an end also serves as a means to a further end. Seeking to make available a scarce resource is a chosen end, but is also a means to saving the life of another patient. Any physical action in the world almost unavoidably results in unintended side effects. In circumstances where an action may also produce harm or evil, we appeal to the DDE. In the case of reallocating a ventilator, the harmful side effect is the foreseeable death of the patient from whom ventilator support was discontinued. The DDE is designed to address precisely this kind of circumstance, so there is no clear reason why the DDE could not apply.

Having addressed the question of the applicability of the DDE, let us consider the first condition of the DDE. Recall that this condition requires that the act in question not be intrinsically wrong by itself, independently of the agent’s intended end or its consequences. Christianity has long held that certain kinds of behaviors are always wrong. In moral theology, a given act may be said to be intrinsically wrong by virtue of its object. In his 1993 encyclical, Veritatis Splendor, Pope John Paul II retrieved this Thomistic concept in his rejection of consequentialist and proportionalist moral reasoning.¹ As John Paul II explained, the object of an act is the “proximate end” of a “freely chosen kind of behavior.” (1993, 78). The object specifies that act as the kind of behavior that the agent is choosing to perform here and now (as a means to achieving some end). Some kinds of acts are intrinsically wrong because their very description entails an intention to undermine some basic dimension of human flourishing, such as life. For instance, an act is intrinsically wrong if its description, such as “murder,” necessarily entails the intention to kill a human being. In contrast, there are some kinds of acts that may culminate in the death of a human being, but are not intrinsically wrong. This is so because their description does not entail intentional killing. Consider Aquinas’s account of self-defense. He allowed that an assailant might be killed in the course of a proportionate defense, but the intention to kill is not a specifying feature of the act. In fact, he added that “it is not permitted that someone intentionally kill anyone in order to defend himself” (1948, II–II, Q. 64, Art. 7). To do so would be to convert it into an intrinsically evil act. Similarly, one might argue that killing in the context of just war is not intrinsically evil since its description does not entail intending the death of enemy forces, but rather only to render them harmless (Rhonheimer, 2002). The issue becomes a bit convoluted when addressing how the first condition of the DDE might permit such acts as capital punishment, in particular how its description does not entail the intention to kill. However, these are not instances of individual persons intending the death of innocent victims, but rather of states acting on guilty parties. In any case, applying all this to our question, is ventilator reallocation the kind of act that is intrinsically wrong independently of the agent’s intended end or its consequences? The answer seems to be no. There is no obvious reason why withdrawing ventilator support, by itself, is intrinsically wrong as a freely chosen kind of behavior. Its very description does not necessitate an attack on human life or some other feature of human flourishing. Extubating a patient is generally thought to be permissible when it is clear that it no longer benefits the patient or when the patient refuses it. None of this could be morally justified if the act of withdrawing ventilator support, by itself, were intrinsically wrong. The same could also be said of the act of “extubation without consent” or even “reallocation without consent” (Eberl and Donovan, 2020, 2). If these were intrinsically wrong clinicians would never be able to reallocate an intervention even if it no longer provided benefit to the patient (Eberl and Donovan, 2020). Certainly, the debate surrounding clinicians’ authority to withhold or withdraw interventions unilaterally against or absent the wishes of patients/surrogates has been contentious (Troug, Brett, and Frader, 1992; Schneiderman, Jecker, and Jonsen, 1996; Helft, Siegler, and Lantos, 2000). Providers are often caught between their desire to accommodate the demands of family and their obligation to provide only those treatments that are consistent with the norms of medicine. Still, hospital policies typically dictate that when a provider cannot resolve a conflict with a patient/surrogate, the provider offers a transfer of care to another provider (Panicola and Hamel, 2012). However, if none can be found, the intervention in question may be discontinued. Of course, every effort should be made to avoid such a decision. Often, if the intervention in question is not especially burdensome to the patient, a provider may allow more time for the family to process the reality of their loved one’s circumstance. Unfortunately, there are occasions when the intervention in question is so disproportionately burdensome or so clearly outside the standard of care that it must be withheld or discontinued. For instance, a clinician should not accommodate a request to perform CPR on an actively dying patient. So, despite the demands of families, patients, or surrogates, it is generally accepted that providers need not initiate or continue interventions that they believe to be medically inappropriate. Consider the discontinuation of chemotherapy when it becomes known that the patient has developed a serious infection. The patient may demand that the therapy continue, but the oncologist has a professional responsibility not to provide unsafe treatment. Also, imagine a family demanding the continuation of a protracted resuscitation effort when the patient is not responding. Responsible clinicians may discontinue that intervention. So, neither acts of “withdrawal” nor those of “withdrawal without consent” by themselves seem intrinsically wrong. Nor does it appear that the act of allocation to another patient by itself is intrinsically wrong. Therefore, if ventilator reallocation is wrong, it does not seem to be intrinsically so by virtue of its object.

The second condition of the DDE requires that the person acting only intends the good effect of the action (although she may clearly foresee a bad effect). An agent chooses to perform an action for the sake of achieving a good, and that is the intended end (Aquinas, 1948, I–II, Q. 12, Art. 1). Any action is wrong if the person performing it intends some evil end. We should bear in mind that the end we are considering here is the subjective purpose of the agent performing the act. It is not the proximate end that specifies the kind of behavior it is. Now what does all this imply with regard to our question about the permissibility of reallocating ventilators? It would appear that when a clinician, under circumstances of scarcity, extubates a patient, he does so in order to make that resource available so that it may ultimately save the life of some other potential patient (Bishop and Eberl, 2021). Thus understood, the intention, that is, the goal that the clinician seeks, is clearly good. The intention is not to bring about the death of the patient from whom the ventilator was withdrawn. Certainly, if that patient did not die, the clinician would feel a great sense of relief and he certainly would not feel as though he had been unsuccessful in achieving his goal. Therefore, the death of that patient could not have been his intended end. So, if ventilator reallocation is wrong, it is not because it is done for some evil purpose.

The third condition of the DDE requires that the foreseen bad effect(s) not be a means to the good effect. With regard to our reallocation question, this condition seems to be satisfied since there is no reason to think that the clinician intends the death of the one patient (the bad effect) as the means by which he saves the life of the other patient (the good effect). That is, the death of the patient is not sought as an instrumental good to preserving the life of the patient with the morally favorable prognosis. Again, if that were the case, the patient from whom the ventilator had been removed would have to die in order to save the other. Clearly, this is not part of the clinician’s plan.

According to the fourth condition of the DDE, the act in question must be supported by a proportionate reason. The good that one intends must outweigh (or at least be equal to) the foreseen bad effect. This fourth condition differs a bit from the first three. If any act fails the first three conditions, that means that the chosen act or the end for which it is chosen is always wrong (and no amount of good achieved or harm avoided can change that). However, some acts might very well satisfy the first three conditions, but may still be wrong if the intended good effect is not proportional to the foreseeable, but unintended, bad effect(s). The fourth condition takes into consideration the (additional) circumstantial features of the act, such as its consequences, the probability of those consequences, and all of the other contextual dimensions that may make the act more or less justified. Whether or not an act of ventilator reallocation violates this condition is difficult to say in the abstract, since it hinges in part on factual beliefs about possible future events that are only justified with some degree of probability. Some of those who question the moral legitimacy of ventilator reallocation worry about social consequences, such as undermining public trust and exacerbating racial tensions (Chu et al., 2020; Peterson, Largent, and Karlawish, 2020; Sulmasy and Maldonado, 2021b). These are important concerns that I address in Section IV. For now, if it is the case that reallocating ventilator support does not violate the first three conditions of the DDE, then we are at least entitled to conclude that acts of ventilator reallocation might be permissible depending on the circumstances and the proportionality of harm to benefit.

III. OBJECTIONS THAT VENTILATOR REALLOCATION DOES NOT SATISFY THE DDE

Let us now turn our focus to objections to my argument that DDE may support ventilator reallocation. One such objection is that ventilator reallocation constitutes an intention (either as an end or as a means) to bring about the death of the extubated patient. In responding to this, I mean to address not only the agent’s intention (in both the second and third conditions) but also the proximate intention that specifies the object of the act (in the first condition).

There have been a couple of recent applications of the DDE to ventilator reallocation that conclude that the death of the extubated patient is necessarily part of the clinician’s intention. Joseph Tham, Louis Melahn, and Michael Baggot (2021) present a hypothetical case in which four patients (A, B, C, and D) will die without ventilator support, but patient D has already been intubated. They argue that reallocating the ventilator to patient A, who has a more favorable prognosis, constitutes an intention to kill patient D. After pointing out that the ventilator is still of benefit to patient D, they conclude that

(S)aving patient A is now intimately tied to disconnecting patient D. Not only is the disconnecting of D now a necessary condition for connecting patient A, but—crucially—the medical staff must deliberately and voluntarily set out to disconnect patient D, while remaining free to refrain from doing so. Therefore, it seems difficult to argue that the disconnection (and hence the death) of D is not in the intention (in intentione) of the medical staff. It is something they have, on some level, willed . . . (Tham, Melahn, and Baggot, 2021, 209)

In general, nothing seems terribly amiss in Tham, Melahn, and Baggot’s account of the ventilator reallocation. They are certainly correct to say that the act involves a voluntary choice to disconnect a patient for the purpose of making available a scarce resource for some other patient. Clearly, the disconnection of the ventilator is part of the (proximate) intention. However, the critical question is whether the death of the patient is part of the intention to disconnect him from the ventilator. Tham et al. parenthetically assume this connection, but in doing so they pass over the primary issue at stake. Other opponents to ventilator reallocation who apply the DDE make a similar error (Chu et al., 2020). The choice or intention to extubate is not the morally controversial part of the issue; it is rather whether the death of that patient is part of that intention.

Anytime there is a direct causal connection between a given action and a bad effect, it understandably raises suspicion that the bad effect is intended and not merely foreseen. The foreseeable consequences that follow from an act, even if outside the intention, always have an impact on the overall morality of the act (Aquinas, 1948, I–II, Q. 20, Art. 5). As Aquinas noted, foreseeable consequences that follow always or frequently from an act are still willed, even if they are outside the agent’s intention (1963, Sententia Physicam, II. 8; Eberl, 2006). If the consequences are harmful, the agent still remains responsible (2002, De Malo, q. 1, a. 3; Eberl, 2006). However, there has been a long-running disagreement among commentators from different camps of natural law theory and moral theology about the extent to which these causal physical connections should be included in the agent’s intention. Part of this disagreement unavoidably stems from a dispute about the proper perspective from which to describe acts and intentions. Many scholars, not only new natural law theorists such as John Finnis, Germain Grisez, and Joseph Boyle (2001), but others including Patrick Lee (2017), Christopher Tollefsen (2000, 2012), and Martin Rhonheimer (2004) defend a first-person point of view, meaning that the nature of an agent’s chosen act and the end for which the act is chosen are defined in terms of the agent’s own perspective. The attractions of a first-person perspective are understandable. Perhaps the most compelling reason is that it seems consistent with the way in which we normally distinguish acts that are identical in terms of their physical causality, but distinct in their moral status. If I have caused someone to form a false belief by providing erroneous information, I cannot be said to have lied to that person if I did not intend to deceive her.² Consider also the difference between withdrawing life-prolonging treatment because it is disproportionately burdensome versus withdrawing that treatment for the purpose of bringing about the death of the patient. In either case, the outward physical causes and effects are the same, but the moral difference that distinguishes the former as an act of withholding overly burdensome treatment from the latter as an act of murder or euthanasia relates to the intention of the agent.

Nonetheless, even those who appreciate this point still worry that a strictly first-person account of actions allows an agent to define his/her intention so narrowly as to ignore the unavoidable physical causality of the act and thereby make an evil intention appear good (Austriaco, 2011; O’Brien and Koons, 2012; Gormally, 2013; Jensen, 2014). They stress that actions are necessarily physical and part of the causal connection of the material world. To allow the means and end of an act to be defined strictly from the mental states of the agent seems to ignore this fact. We have seen this kind of disagreement played out many times in connection with a wide range of morally disputed medical and surgical interventions, from craniotomies to the resolutions of ectopic pregnancies and more (Rhonheimer, 2009; Guevin, 2017). The defenders of the first-person perspective insist that none of the chosen acts or intended ends in such procedures entail the killing of anyone, but admit that the death of a human being is clearly foreseeable and all but certain. From this point of view, craniotomies may be described as reducing the size of the baby’s skull so that it may pass through the birth canal and salpingostomies may be described as the removal of an embryo from the fallopian tube (Moraczewski, 1996; Finnis, Grisez, and Boyle, 2001; Kaczor, 2009).³ However, scholars such as Steven Jensen (2014) and John Haas (2017) still insist that acts thus described should entail intending the immediate physical effects, including the death of a human.

Generally speaking, among those who adhere to a third-person-oriented account of intention, there has been a tendency to want to include the death of a human being as part of the intention when the act in question terminates in the body of the person killed (Long, 2013). However, even at the common-sense level, a direct physical effect on the body does not seem to equate to intending the death of a patient. Certainly, a surgeon, with the consent of the patient, may undertake a procedure in which the death of the patient is possible or even highly likely. If there are no better alternatives under the circumstances, the risk may be acceptable. If the patient dies in the course of the surgery, the surgeon cannot be said to have intended the death of that patient. More to the point, consider Aquinas’ own example of causing the death of a human being in the course of defending oneself (1948, II–II, Q. 64, Art. 7). Although the act of self-defense involves a direct impact on the body of the assailant, Aquinas does not consider the death of the assailant to be part of the intention. Granted, the case of withdrawing a ventilator is considerably different in that it involves a removal or discontinuation of something. Still, whatever disruption of bodily integrity is brought about upon extubating a patient, it does not entail that the death of the patient is part of the clinician’s intention.

Some proponents of the broader third-person account of intention offer specific criteria to help us determine whether the physical causality that follows from an act should be considered part of the intention. For instance, offering a plausible reflection of Aquinas, Matthew O’Brien, and Robert Koons suggest the following: “If realizing condition A in and of itself entails, by virtue of the laws of nature, the realization of condition B, and the agent knows this, then B is certainly included in his intention to bring about A” (2012, 668). Thus, according to O’Brien and Koons, if the laws of nature make it virtually impossible to achieve the good effect without bringing about the bad effect, then the bad effect should be considered part of the overall intention. They apply this guiding principle to craniotomies and gunshots to the head, that is, actions whereby the death of the subject, by virtue of the laws of human physiology, seems certain and therefore intended. So, if we apply O’Brien and Koons’ criterion to our ventilator reallocation question, the death of the extubated patient would count as intended if it was entailed by virtue of the laws of nature. Let us bear in mind that we have already conceded in our framing of the reallocation scenario that without ventilator support, both the current patients and those who may be arriving to the hospital in an expected surge would almost certainly die. Nevertheless, while it would seem to defy the laws of human physiology if crushing the skull of an infant or shooting someone in the head did not immediately bring about their death, the same does not seem to apply as obviously in the case of ventilator reallocation. In ventilator reallocation, the survival of an extubated patient is extraordinarily unlikely, but it would not defy physiological explanation in the same way as if an infant survived a craniotomy. While the latter entails the direct mutilation of a vital bodily structure, the former does not. So, even if we accept O’Brien and Koons’ broader account of intention, it is not clear that ventilator reallocation includes an intention to kill.

A broad version of intention centered around physical causality might also imply that when a clinician’s act leads to the inevitable death of the patient, then that death must be part of the intention. For instance, Thomas Cavanaugh (1997) has argued that contemporary versions of the DDE depart from the account rooted in Aquinas’ treatment of self-defense, in that they potentially allow for nearly all foreseeable negative side effects to remain outside the intention. In contrast, according to Cavanaugh, Aquinas did not consider foreseeable inevitable effects as outside the agent’s intention. Cavanaugh’s (1997) reading of Aquinas is that only in those cases where the act simply poses a risk of harm (where the negative effect follows sometimes, but not always), may the effect be considered outside the agent’s intention. Cavanaugh’s (1997) analysis focuses on Aquinas’s account of self-defense, and he himself confesses some uncertainty as to what it might imply with regard to other kinds of situations. That said, if Cavanaugh’s reading of Aquinas is correct, we might be forced to conclude that many or most cases of ventilator reallocation involve intending the death of extubated patients, in that their deaths may be inevitable. There are a few problems that arise if we apply this understanding of intention to ventilator reallocation. First, there may very well be cases in which the death of the patient is not inevitable, in which case the patient’s death need not be part of the intention. Second, and most importantly, if the foreseeable inevitability of death were the criteria by which we determine whether an effect is part of the intention, many kinds of interventions in healthcare that have been morally accepted as entailing foreseeable but unintended evils would now have to be considered impermissible. The death of an embryo in cases of salpingectomies and removal of cancerous uterus from pregnant women would be inevitable, and therefore part of the intention.

At this point, someone might argue that, although it might be possible to give an account of the physical causality surrounding ventilator reallocation without intending the death of the extubated patient, it is not possible to do so without intending injury to that patient. Consider Stephen Jensen’s (2014) critique of first-person descriptions of craniotomies. As mentioned above, technically, all that is needed in a craniotomy is a reduction in the size of the infant’s skull so that it may pass through the birth canal. Jensen, for the sake of argument, grants that the death of the infant may not be intended, but argues that crushing the skull of the infant surely constitutes intentional injury. So, perhaps the same point might be made of ventilator reallocation, that while death is not intended, there is still an intent to injure. Admittedly, depriving someone of supplementary oxygen and respiratory support seems injurious. However, there is a morally relevant difference between withdrawing ventilator support and crushing a skull. In a craniotomy, the injury is brought about directly by the physician. In the case of the COVID-19 patient (or any patient in need of respiratory support), it is the pathological condition itself that has brought about the injury, not the physician. So, neither the death of nor an injury to the patient is an obvious part of the intention when reallocating ventilators.

Leaving behind concerns that I have inappropriately excluded the physical causation connected to the death of (or injury to) the patient from whom the ventilator has been withdrawn, let us consider a related objection. It might be argued that the intrinsic wrongness of reallocation is not found in any intention to kill, but that other kinds of acts, while not as gravely immoral as killing, may still be intrinsically evil. Consider lying (or any intentional act of deceiving). These kinds of acts are thought to be intrinsically wrong (Aquinas, 1948, II–IIa, Q. 100; Tollefsen, 2012). So, maybe there is some kind of deception at the heart of reallocating ventilator support. One might argue that ventilator reallocation constitutes a kind of lie, in that patients may feel justified in assuming that they will have access to all resources up until they are no longer provided benefit. However, what patients are justified in expecting would depend in part on what has been communicated to them by providers, healthcare organizations, and public health authorities. A patient or her family may feel deceived if they are allowed to believe that there are no circumstances under which ventilator reallocation may occur, but they may not feel deceived if informed of this unfortunate possibility. It is somewhat analogous to the limits on patient confidentiality. Patient confidentiality is an important norm in clinical medicine, but not exceptionless. A patient who is led to believe that confidentiality is absolute may feel deceived if it is breached to prevent harm to outside parties. However, there would be no deception if the provider was transparent about the limits of confidentiality. There is no question of some directly intended evil means or end (i.e., no deliberate intention to lie or harm), but whether or not a breach is warranted under the circumstances. The same holds for ventilator reallocation. There is no reason to think that it entails a deliberate intention to lie, but it does require open communication with patients, families, and communities as well as careful consideration of the circumstances and consequences.

Although it does not appear that ventilator reallocation constitutes an instance of intentional killing or lying, perhaps it entails some other intrinsic evil such as theft. However, as I discuss in Section V, this does not seem correct, since patients in hospitals usually do not own their ventilators. They typically cannot claim a property right over the ventilator that they are using.⁴ It would make more sense to say that the hospital (or perhaps the community) owns the resource, and physicians are entrusted to be responsible stewards over the resource according to the norms of medicine, society, and social principles (United States Conference of Catholic Bishops, 2018). So when one patient’s ventilator is reallocated to another, the clinician cannot be said to have stolen it in any obvious sense.⁵

Without a doubt, these are complicated philosophical and theological issues that require careful attention, the debate about what should or should not be included as part of an intention remains alive in the scholarly literature. I have tried to suggest that even if we accept the broad third-person account of intention, it is not clear that ventilator reallocation entails an intention to kill (or lie to or steal from) the patient from whom the ventilator has been withdrawn. We have already noted that, while it is heartbreaking, the act of withdrawing ventilator support (or any other life-sustaining intervention) from a patient may be permissible when that intervention does not offer a reasonable hope of benefit or is excessively burdensome. In terms of its physical causality, the removal of the ventilator is directly related to the patient dying sooner rather than later, although it is understood that the patient’s death is not the intention of the clinician. Analogously, in the case of reallocating a ventilator, the death of the extubated patient is almost certain, but this does not require that the patient’s death is intended. The death of the extubated patient is not a good that the clinician seeks, neither as an end nor as a means. Granted, in the typical case of withdrawing nonbeneficial or medically inappropriate interventions, the good that the clinician seeks is to disburden the patient of a treatment that is no longer helpful and may be harmful. Obviously, in the case of ventilator reallocation, the good sought is not directed toward the extubated patient. Rather, the good sought is to make a life-saving resource available which later may be used to save the life of some other potential patient. Whether or not, taking everything into consideration, there is a proportionate reason to justify this is a legitimate question, but not one that speaks to an intention to kill. If the withdrawal of ventilator support in cases where it is nonbeneficial does not entail an intention to kill, then there is no good reason to assume that ventilator reallocation entails such an intention.

IV. THE CONSEQUENCES AND CIRCUMSTANCES OF VENTILATOR REALLOCATION

Up until now, I have been arguing that ventilator reallocation does not include an intention to kill or injure the patient from whom ventilator support has been withdrawn, or for that matter, to lie or steal. In the context of the DDE, this means that ventilator reallocation is neither intrinsically wrong nor is it done for some evil end. Furthermore, there has been no obvious reason to conclude that the death of (or injury to) the one patient is the means to saving the life of the other. If this is true, then the only remaining objections seem to be related to the fourth condition of the DDE, that, under the circumstances, the intended good effect(s) achieved by reallocation is/are not proportional to the foreseen bad effect(s). Hence, in this section, I consider some objections in connection with the social consequences, likelihood of harm, and overall circumstances of ventilator reallocation.

One objection to ventilator reallocation linked to the fourth condition is that reallocating ventilators (as a pandemic practice) would have a detrimental impact on the clinician-patient relationship and on the overall public trust in the medical profession (Hilliard, 2010; Chu et al., 2020; Eberl and Donovan, 2020). If ventilator reallocation is permitted as a practice, patients on ventilators might constantly worry that their use of the ventilator could be discontinued in the event of a surge of patients in respiratory distress. Moreover, one could easily imagine a slippery slope in which patients who use a ventilator at home fear that it could be reallocated if they go to the hospital (Reynolds, Guidry-Grimes, and Savin, 2021). We might then wonder what other life-sustaining technologies may be reallocated when there is a scarcity of that technology. These concerns are legitimate and suggest that we consider ventilator reallocation with great care. However, it does not mean that ventilator reallocation is somehow a directly intended attack on the public trust or the fiduciary bond of the clinician-patient relationship or any other core value. It is, rather, a worry about potential consequences to patients, families, the practice of medicine, and the public trust. The probability that these will come to pass or the exact extent of their detriment is unclear. Meanwhile, a triage team may at times have a high degree of confidence that it can save the life of a patient by reallocating a ventilator. None of this is to dismiss the concerns about the potential detrimental impact on the clinician-patient relationship or the public trust. However, medicine has always been practiced in the context of competing values, and so long as those values are not directly attacked, decisions about the appropriate priority of those values in light of the circumstances are inevitable. So, it does not seem unreasonable to hold that with transparency, public education, and public dialog that ventilator reallocation, done with evidence-based triage criteria and consistency, could occur with minimal impact on public trust and the clinician-patient relationship.

In a related objection, it has been argued that physicians’ primary obligations are to their individual patients and only secondarily to public health concerns (Hilliard, 2010). Allowing ventilator reallocation appears contrary to this traditional priority, a priority that deserves protection. However, medicine, like any profession, is a social institution, so its values and goals are not self-evident moral facts that exist independently of the needs and circumstances of society. The profession of medicine relies on input from society in the development and articulation of its norms and aims, and it is only because of this that it is allowed the power to withhold or withdraw treatments that run contrary to those values (Rubin, 2007; Tomlinson, 2007; Bhurga, 2014). So, while the priorities of medicine reflect a long tradition, they are not set in stone independently of society. Social circumstances, such as extreme scarcity, may require reprioritizing traditional goals of medicine, including the priority of the individual patient over the public good.⁶ This should not imply an abandonment of traditional values of medicine. Medical practitioners who work in areas of the world where a scarcity of resources is an everyday reality make heart-wrenching decisions, but still believe themselves to be practicing medicine in accordance with the norms of the profession. At the very least, it remains an open question how the priorities of medicine might shift under circumstances of scarcity.

Other consequence-oriented concerns related to ventilator reallocation point to the potential for or appearance of unfairness, discrimination, and potential negative impact on the poor and disabled. For instance, Sulmasy and Maldonado (2021b) ask the reader to imagine how it might be received, in the context of racial tensions, if a provider had to inform the family of an elderly African American woman that her ventilator is being reallocated to a young Caucasian man. Interestingly, some bioethicists, such as Eberl (2020), express just the opposite concern, that prohibiting ventilator reallocation may exacerbate injustice by privileging those who benefit from a “first come, first served system.” In any case, the potential to exacerbate racial tensions and to discriminate against marginalized populations are justified concerns, but not good reasons to close down the conversation about the permissibility of ventilator reallocation. Cases of discrimination on the basis of race, gender, disability, and socioeconomic status exist, but these are not unchangeable social facts. Societies should work to eliminate these systemic evils. These social ills make ventilator reallocation more ethically complicated than it otherwise would be, but these obstacles may be overcome with prudence and caution, adhering to transparent evidence-based triage criteria.

In all of this, we must bear in mind that when the pandemic began, the healthcare community had already accepted the unfortunate need for triage. This meant that patients who would have unquestionably received ventilator support when there was no scarcity, now may not receive it because of their relatively poor prognosis. Having set evidence-based triage criteria (e.g., SOFA scores) that guide allocation decisions about who is most likely to survive with the assistance of a ventilator and who is not, we have agreed that those patients who sadly do not satisfy the allocation criteria should not be provided ventilator support. So, when a patient who initially satisfied the triage criteria deteriorates to the point at which he no longer satisfies the criteria, it should not seem morally unjustifiable to discontinue ventilator support and to reallocate that ventilator to another patient who satisfies the criteria. National Catholic Bioethics Center ethicist John DiCammilo lends support for this. Speaking to the issue of reallocating ventilators, DiCammilo states that “informed consent is not always necessary if a patient’s triage score is in a range that does not qualify for the resource in question” (2020; emphasis added). While this may not be what opponents of ventilator reallocation have in mind, this too is a form of ventilator reallocation. If we accept that it would be permissible to withdraw ventilator support from a patient who no longer qualifies under the triage criteria, then we are accepting the permissibility of a ventilator being withdrawn earlier than it would have been in times of nonscarcity, thereby accepting ventilator reallocation.

I am not advocating the reallocation of ventilators from patients who still meet evidence-based triage criteria, but that is not because doing so would constitute an intrinsically evil means or end. It is rather that adhering to and articulating clear objective triage guidelines to patients, families, and communities is critical to minimizing any detrimental impact on the public trust and the provider-patient relationship (Peterson, Largent, and Karlawish, 2020). Absent those criteria, acts of reallocation would appear subjective and arbitrary. Opponents of ventilator reallocation, such as Sulmasy and Maldonado (2021a) and Chu et al. (2020), might wonder why my concerns about undermining public trust and traditional norms of medicine would not justify prohibiting ventilator reallocation altogether. My only response is that given that ventilator reallocation does not seem to entail any intrinsic evil, we are left to make judgments about the proportionality of good effects to bad effects. We try to safeguard as many core values as possible, while still honoring our commitment to promote and conserve human life. Framed in this way, lacking any absolute moral prohibition, we may cautiously permit ventilator reallocation under the conditions described above.

V. VENTILATOR REALLOCATION AND THE EQUIVALENCE THESIS

We have been discussing the permissibility of ventilator reallocation. However, given that we are examining a moral question that centers around the withdrawal of treatment, an issue that has attracted considerable attention over the past few decades, we must take care to consider related concerns. In particular, there had been a tendency to believe that withdrawing life-prolonging treatment was morally wrong or morally worse than withholding that same treatment under the same circumstances. However, that belief is now widely held to be mistaken and overcoming it has been hailed as an important achievement in healthcare ethics (Beauchamp and Childress, 2018; Lo, 2020). The rejection of this view is codified in the so-called equivalence thesis. The equivalence thesis roughly states that if it would have been morally permissible to have withheld a treatment under the current circumstances, then it would be equally permissible to withdraw that treatment under those same circumstances. If we apply this thesis to our ventilation reallocation issue, it seems to support the withdrawal of a ventilator from one patient in order to reallocate it to another. This is so because it would have been permissible to have withheld the intervention in the first place from the patient with the poorer prognosis under the triage circumstances. That is to say, given that one would have been justified in withholding the ventilator from the patient with the poorer prognosis, one would likewise be justified in withdrawing it.

Although largely accepted in the clinical healthcare ethics community, the equivalence thesis has been challenged over the years, most notably by Daniel Sulmasy and Jeremy Sugarman (1994). Sulmasy and Sugarman reject the claim that withholding and withdrawing treatment are morally equivalent. They ground their view in Robert Nozick’s (1974) principle of justice in acquisition, which implies that one is entitled to a holding if it is unowned and one happens to acquire it. Analogously, they argue that a patient has a claim to the continuation of treatment once it has been initiated. Based on this, they conclude that the equivalence thesis is false, since a patient on whom treatment has been initiated has a claim to it that another patient does not, even though both patients may have been equally entitled to it prior to its initiation.

If Sulmasy and Sugarman are correct, reallocating ventilator support might violate the extubated patient’s claim to continued use of the ventilator and thereby undercut my thesis for its potential permissibility. However, on its own, Sulmasy and Sugarman’s argument does not imply that such a claim may never be overridden by other considerations, only that withholding and withdrawing treatment are not morally equivalent. There may very well be other values in play, such as the stewardship of resources that override that claim. Additionally, the analogy with Nozick’s principle of justice in acquisition may be somewhat strained. As noted in Section IV, medical resources in a hospital setting are typically not private property that patients own. They are resources that doctors are entrusted to make use of according to the norms of medicine and society. Moreover, as Eberl (2020) points out, even Aquinas himself allowed that all things may become “common property” under circumstances of urgent need (1948, II–II, Q. 66, Art.7). In and of itself, this does not justify ventilator reallocation, but it does show that Aquinas held that property claims may be overridden under certain circumstances. In any case, even if we accept Sulmasy and Sugarman’s challenge to the equivalence thesis, it does not necessitate an outright prohibition of ventilator reallocation.

VI. CONCLUSION

I have argued that in times of scarcity the DDE may support the unilateral reallocation of life-sustaining interventions such as ventilators, provided that it is supported by transparent evidence-based triage criteria. In particular, there does not appear to be any compelling reason to classify the unilateral reallocation of a ventilator as intrinsically evil. Therefore, if it is morally wrong, it is so because it lacks a proportionate reason under the circumstances. I have suggested that our commitment to saving lives and stewarding resources may justify the act in times of scarcity. However, as always, it is far preferable to avoid moral challenges than it is to resolve them. So, taking steps to prevent the need for reallocating life-sustaining resources would be ideal. This would mean making sure that a surplus of ventilators and other key resources is available in case of emergency. Additionally, redoubling our efforts to promote advance healthcare planning helps by avoiding the use of scarce resources on patients for whom they would be nonbeneficial, even under ordinary circumstances. Nevertheless, in parts of the world where resources, such as ventilators and ECMO machines are scarce, reallocation may be unavoidable and should be considered permissible.

ACKNOWLEDGMENTS

I would like to thank Drs. Kenneth Richter, Nikolai Butki, and Juan Carlos Flores for their helpful comments on earlier drafts of this article.

REFERENCES

Footnotes