1. Introduction

Nowadays the bioethical debate on end-of-life issues seems to still be characterized by some problematic interpretations of moral responsibility. For example, within certain utilitarian approaches, the same moral responsibility is ascribed to a physician who practices euthanasia and to another who withholds or withdraws life sustaining treatments. In general, the equal ascription of responsibility is justified with the following argument: killing and letting die have the same consequence, that is, the death of others, and by performing these actions the physician contributes in the same way to the fatal outcome and can thus be considered equally responsible for the patient’s death. Within this framework the killing/letting die distinction is morally irrelevant and usually identified with the distinction between active and passive euthanasia. To specify, by maintaining that killing and letting die are morally equivalent actions, the act of withholding or withdrawing medical treatments is included within the range of euthanasia and a physician who has performed this practice (passive euthanasia) is always equally responsible to another one who has killed the patient, for example by lethal injection (active euthanasia). I will call this point of view the always equal moral responsibility argument (hereafter “the always equal argument”).

An opposite approach to the ascription of responsibility raises from the thesis that there is an absolute moral distinction between killing and letting die. Indeed, by affirming that killing is always morally wrong and letting die is always morally good, it could be said that it is never possible to ascribe the same responsibility to a physician who practices euthanasia and to another who withholds or withdraws medical treatments. I will call this thesis the never equal moral responsibility argument (hereafter “the never equal argument”).

Now, is moral responsibility correctly declined within the theses mentioned above? By killing the patient and by letting him die, does the moral agent always, or never, play the same role in the cause of patient’s death?

In order to answer these questions, I will firstly show that the always equal argument raises a problematic question of terminology and methodology. Indeed, by identifying the killing/letting die distinction with the active/passive euthanasia distinction, the argument erroneously describes the act of withholding or withdrawing medical treatments as euthanasia. Secondly, the implications that both a rejection and an unconditional defense of the killing/letting distinction could have in the ascription of responsibility will be addressed. In this respect, it will be argued that while the always equal argument calls for an over-responsibility of the physician, the never equal argument leads the agent to take less responsibility for his actions. In order to suggest an intermediate position which preserves a not absolute distinction between killing and letting die, I will then refer to other moral relevant distinctions that highlight the weaknesses of the always and the never equal arguments. To specify, by distinguishing the cause of an event from the conditions that contribute to the occurrence of an outcome, the causal differences between killing and letting die will be addressed. Through the distinction between negative and positive duties, killing and letting die will be described as breaches of different obligations. Given that these differences entail a different agent’s moral responsibility in performing these actions, the always equal argument will turn out to be invalid. Finally, through the distinction between morally culpable letting die and letting die for the patient’s good, it will be argued that in some cases letting die is morally equivalent to killing. In this way the invalidity of the never equal argument will be proved and the presence of ambiguous “grey zones” between killing and letting die will be shown. This aspect will justify the main conclusion of the paper: ascribing moral responsibility at the end of life means struggling with ambivalence, but maintaining all the distinctions mentioned above is necessary to acknowledge the complexity of the fundamental structure of morality.

2. A problematic question of terminology and methodology

The issue of whether killing a patient and letting him die should or should not be considered morally equivalent actions is a perennial question ^[1]. In particular, the killing/letting die distinction has become a main research topic since the birth of bioethics, especially for advocates of euthanasia. Indeed, in order to argue the moral legitimacy of euthanasia, many authors have denied the killing/letting die distinction through a disapproval of the distinction between active and passive euthanasia ^[2-^8].¹ From their point of view, since its origins medical ethics has expressed a different value judgment regarding the two practices of euthanasia by affirming that any act aimed to kill the patient is illicit, but in some cases it is permissible to withdraw treatments and to let the patient die.² In other words, the rationale underlying the illegitimacy of active euthanasia and the legitimacy, under certain conditions, of passive euthanasia would be the idea that there is a relevant moral difference between killing and letting die. By adopting a consequentialist approach and by arguing that killing and letting die are morally equivalent because both actions have the same consequences (the death of others), advocates of euthanasia have thus tried to show that the active/ passive euthanasia distinction is without moral significance.

Proposed by James Rachels, the Equivalence Thesis clearly exemplifies this position. Indeed, according to this thesis, the difference between killing and letting die is not in itself a morally important matter, and consequently, the fact that an action is characterized as a killing or as a letting die should not affect our judgment concerning its rightness or wrongness. Rachels illustrates the Equivalence Thesis by introducing parallel cases, identical in terms of the upshot and the agent’s intentions, yet different in that they are respectively a case of killing and a case of letting die (the Bare Difference Argument). In this respect, a well-known example is that of Smith and Jones, who both want their little cousin dead in order to get his inheritance; nevertheless, while the former actively drowns the child in the bathtub, the latter merely lets his cousin drown in the bath after he has slipped and hit his head. Given that we cannot say that Jones’s behavior is less reprehensible than Smith’s, Rachels concludes that the difference between killing and etting die in itself makes no moral difference to our moral assessments ^{[4, p. 79; 5, pp. 111-114]}.

Applied to end-of-life issues, the Equivalence Thesis maintains that active and passive euthanasia are morally equivalent per se and claims what I have called the always equal argument. Indeed, the rejection of the killing/letting die distinction leads to ascribe the same moral responsibility to a physician who kills the patient and to another one who lets him die by withholding or withdrawing medical treatments. Actually, Rachels affirms that the moral legitimacy of these actions is determined by other extrinsic features, such as respect for the patient’s wishes and/or the minimization of suffering. That is to say that, by these factors, in some cases it may be permissible to kill but not to let die, while in other cases it may be permissible to let die but not to kill. This means that the physician is worthy of praise and equally responsible for the patient’s death when killing and letting die have been performed to respect the patient’s will and/or to minimize suffering. Furthermore, a physician who lets the patient die although he had expressed the wish to be killed is worthy of blame and equally moral responsible to a healthcare professional who kills the patient although he had wanted to be let die. Within this framework, a different ascription of moral responsibility occurs only when the physician has acted with or without the patient’s consent and/or has minimized suffering or not.

The particular characterization of moral responsibility offered by the always equal argument will be discussed in the following paragraph because I will firstly address a problematic question of terminology and methodology raised by the link between the killing/letting die distinction and the active/passive euthanasia distinction.³ Indeed, by using the expression “passive euthanasia”, which is vague and confusing, the always equal argument includes withholding or withdrawing medical treatments within the range of euthanasia. Actually, euthanasia aims to intentionally end a life. In other words, when euthanasia is practiced, the physician acts with the specific intention of causing the patient’s death. On the contrary, letting a patient die could be aimed to respect his will and/or minimize his suffering, although the action might also be performed in order to hasten the dying process, and, as it will be argued in the fourth paragraph, this is a case in which letting die is illegitimate and morally equivalent to killing.

In general, advocates of euthanasia criticize this conclusion by affirming that euthanasia is aimed to respect the patient’s will and/ or minimize his suffering as well. Nevertheless, this view denies another relevant distinction, that is, the distinction between intention and motive: while the former is a sort of commitment for the agent to engage in a certain action and is fulfilled when the intended and wanted outcome is realized, the latter refers to the reasons that render the outcome so interesting and attractive. In other words, intention commits the agent to act in a certain way and motive is the moving power which impels him to do that action rather than another one. Now, before enacting euthanasia, the physician is committed to ending the patient’s life, so much so that the substances used for injection are lethal drugs. If the physician said that his intention, and not his motive, was to relieve suffering, his commitment would be engaged in the achievement of this goal, but not in the patient’s death. Of course, the distinction between intention and motive is a subtle one and there are problematic issues regarding the concept of intention because this notion refers to internal, and to some degree impenetrable, dimensions of the agent that can be verified only after the fulfillment of the action. However, intention is an important aspect in determining the rightness or wrongness of an action⁴ and is also relevant within the legal context in order to determine the offender’s degree of guilt.

Furthermore, by relating the killing/letting die distinction to the active/passive euthanasia distinction, the former could be wrongly identified with the action/omission distinction. In fact, while killing involves «doing», withholding or withdrawing treatments counts as «allowing». Nevertheless, when an agent allows something to happen he might be enabling it, for example by removing a plug and allowing water to flow, or refraining from preventing it, for example by not warning someone about a trap and allowing him to walk into it.⁵ Applying these considerations to end-of-life issues, we could say that euthanasia is always performed through acts and letting a patient die can sometimes be performed by actions (withdrawing treatments) and sometimes by omissions (withholding treatments or refraining from preventing the patient from dying) ^[17].

For all these reasons, the expression “passive euthanasia” is inappropriate, misleading and unhelpful, and should be avoided.⁶ Indeed, the adjective “passive” does not correctly describe the euthanasia practice, which is active by definition, and the term “passive euthanasia” creates ambiguities around the act of withholding or withdrawing treatment.

3. The killing/letting die distinction and the ascription of responsibility

A further weakness of the always equal argument concerns its characterization of moral responsibility. As already noted, the argument is focused on the consequences of the action and this aspect leads one to ascribe the same responsibility to a physician who practices euthanasia and to another who withholds or withdraws life-sustaining treatments. In this way, the ascription of responsibility neglects other relevant moral features related to the agent and to the act in itself.

For example, although killing and letting die have the same consequences, these actions cannot be considered morally equivalent because both the proximate cause of the patient’s death and the physician’s role are different. Indeed, by injecting a lethal drug, the physician creates a new, fatal pathology (for example hyperkalemia in case of the injection of potassium). In other words, when the agent kills someone, he produces a new and dangerous situation that leads to the victim’s death and then causes the fatal outcome. On the contrary, by withholding or withdrawing medical treatments, the physician creates a condition which enables the preexisting disease to yield its effect on the patient. The physician then contributes to the patient’s death, but does not cause it because he is faced with a preexisting fatal condition, that is, a death threat not ascribable to himself.

These considerations are based on the distinction between cause and condition that is morally relevant to ascribing responsibility. The cause brings about an effect and must be present for the effect to occur. The condition enables the cause to yield its effect but does not permit the occurrence of the effect in the absence of the cause. In answer to the question “What caused the fire?”, we are unlikely to reply “The presence of oxygen or flammable material within the building”. Indeed, these conditions do not allow a fire (effect) to occur in the absence, for example, of a short circuit (cause). Appling these considerations to end-of-life issues, we could say that killing a patient and letting him die are causally different actions because while the former is a sufficient condition for death, withholding or withdrawing life-sustaining treatments is not,⁷ although it could play a role in the cause of the patient’s death, obviously in addition to other and stronger conditions.

Furthermore, killing and letting die are breaches of different duties, respectively “Not to kill” and “To prevent others from dying”. This aspect is related to the distinction between negative duties (non-interference) and positive duties (helping others) whereas the former is stricter than the latter. Indeed, while we can always and completely fulfill our negative duties by refraining from killing or inflicting injury to everyone -always, even when we are sleeping- the duty to provide aid and to not let people die is impossible to fulfill all the time and toward everyone. Doing harm is thus harder to justify than not helping others or merely allowing harm. As a consequence, contrary to what is affirmed by the always equal argument, the physician cannot be considered equally responsible for the patient’s death on all the occasions because the fatal outcome could be the result of the breach of different duties, negative ones in cases of killing and positive ones in cases of letting die.

The distinctions between cause and condition and between negative and positive duties are both relevant to ascribing moral responsibility. On the one hand, by denying the former, the risk of a problematic explanation of any event is realistic because infinite causes of what has happened could be addressed. Furthermore, the rejection of the cause/condition distinction is counterintuitive because we could absurdly ascribe parents as being morally responsible for their son’s death as they gave birth to him ^{[20, pp. 85-86]}. It should be noted that the distinction between cause and condition cannot lead to an underestimation of the conditions because they are often necessary for explaining what happened, so much so that by their absence the cause would not produce all its effects. Coming back to the example previously proposed, with the presence or not of the flammable material in the building, the fire produces different effects. In other words, the identification of the cause is required but the explanation of an event cannot be reduced to it. This is true about the act of withholding or withdrawing life-sustaining treatments as well: in order to explain the patient’s death, the physician’s action must be mentioned but it should also be recognized that he has performed that act in the presence of a preexisting fatal condition not ascribable to himself.

On the other hand, denying the distinction between negative and positive duties and conferring them the same normative force means calling for an over-responsibility of the moral agent. Indeed, in order to not be considered a murderer, the agent should always fulfill positive duties toward everyone, which means that he should provide aid even when his help has been refused or injures a third party’s rights. Furthermore, such an approach requires the agent to underestimate his personal relationships and save all of humanity. This goal is obviously impossible to be achieved by an individual because it needs combined efforts as well as economic and political strategies. Similar conclusions can be drawn about the physician’s commitment in promoting health and saving lives. On the one hand, considering that a healthy status is susceptible to variations determined by subjective, cultural, social and environmental factors (the so-called determinants of health), the promotion of health cannot be considered a task exclusively requested to the physician. On the other hand, the respect for patients’ autonomy and the use of criteria to allocate the limited healthcare resources are required by those who perform the medical profession.

Now, by maintaining the two distinctions mentioned above, the following thesis might be advanced: there is an absolute moral distinction between killing and letting die because these actions are causally different, and by performing them the agent violates different duties. In other words, by adopting what I have called the never equal argument, it may be said that killing is always morally wrong and letting die is always morally good. Given that under no circumstances these actions can be considered morally equivalent, the conclusion would be that it is never possible to ascribe the same responsibility to a physician who practices euthanasia and to another one who with-holds or withdraws medical treatments. The never equal argument is problematic as well because it denies the legitimacy of killing committed in justified self-defense and affirms that the agent legitimately acts in any case of letting die, even when he refrains from rescuing a drowning victim although the action would not threaten his own life. As a consequence, by adopting this approach the agent is more likely to take less responsibility for his actions.

Actually, as suggested by Daniel Sulmasy, if it is true that except in cases of self-defense or rescue all killings demonstrate a pathognomonic sign of moral illegitimacy, letting die cannot be considered a morally legitimate act merely because it lacks this sign ^{[21, p. 58]}.⁸ By using an analogy, measles can be diagnosed to everyone who presents Koplik spots (small and white stains in the oral cavity) but the absence of this pathognomonic sign of measles cannot exclude the diagnosis. In other words, a person who does not show any spots in his mouth could have measles. In the same way, even if the pathognomonic sign of moral illegitimacy is absent, letting die could still be morally unjustified. As a consequence, our moral judgment regarding the rightness or wrongness of letting die cannot be univocal, and depending on the circumstances, this action or omission can be sometimes legitimate and sometimes illegitimate.

This is why the ascription of moral responsibility at the end of life requires a more thorough examination of the different types of letting die. The next paragraph will then deal with the following questions: from a moral point of view, under what conditions and circumstances can withholding or withdrawing treatments be justified? What moral features render letting a patient die legitimate or illegitimate?

4. The different meanings of letting die

Assessments about the moral permissibility of withholding or withdrawing medical treatments are generally based on the argument against imposing unwanted treatment on the competent patient or on his substitute decision-makers. Although respect for autonomy is widely acknowledged principle, the moral relevance of the withholding/withdrawing distinction might be argued in the following terms: refraining from intervening (omission) is a sufficient condition to respect the patient’s refusal of a medical treatment, but the request to withdraw the same treatment requires a course of action (act) performed by the physician. Withholding could thus be less problematic than withdrawing because, contrary to the former, the latter occurs when a treatment has already begun and, consequently, strong reasons are required to decline to provide it to the patient ^[22-^23]. Furthermore, it might be said that, when the patient’s will concerns life-support treatments, a different moral responsibility has to be ascribed to a physician who withholds and to another who withdraws that kind of treatment. For example, it could be argued that while by withholding artificial ventilation the physician refrains from intervening and saving the patient’s life, by withdrawing the life-support from a patient unable to breathe spontaneously the physician performs a causally decisive act for the patient’s death ^[24].

This framework is problematic for the following reasons: firstly, by considering the patient’s request to withdraw a treatment more problematic than a refusal of it, the risk of imposing medical treatments, although disproportionate, on the patient is realistic.⁹ Admittedly, a morally legitimate decision around withdrawing is made when the treatment turns out to be ineffective and burdensome for the patient. From this point of view, it could thus be argued that withdrawing is more justified than withholding. Secondly, the moral illegitimacy or permissibility of letting die cannot be argued by exclusively referring to the act/omission distinction. Indeed, when intervention could have prevented the death of others and the agent had the ability as well as the means to save the victim who had not refused help, omitting to act is equally wrong as doing something that lets someone die. Applying these considerations to end-of-life issues, we might then conclude that the moral relevance does not lie in the withholding/withdrawing distinction ^{[26, pp. 158-162] 10} because, as it will be shortly argued, it is the conditions and circumstances in which withholding and withdrawing take place that make a moral difference rather than action and omission as such.

For example, intention is a relevant aspect in determining the rightness or wrongness of letting die. On the one hand, a physician could let a patient die in order to respect his will as well as minimize his suffering. Under these circumstances, hastening the dying process is not intended, although it is foreseen,¹¹ and we can say that a letting die for the patient’s good has occurred. On the other hand, withholding or withdrawing treatment could be aimed to hasten the dying process (a morally culpable letting die) and be performed with different motives. For example, the physician might consider the patient’s life no longer worth living or believe that the healthcare resources currently devoted to the patient should be used by those who might receive greater benefit. In this case hastening the dying process is not only foreseen, but also intended.

When deciding to let a patient die, treatment decisions are at stake because an assessment of the proportionality or disproportionality of treatments is made.¹² Such an evaluation should be based on the criteria of the effectiveness of the treatment as well as of the benefits and burdens that it generates. To specify, a proportionate treatment is effective when it alters the patient’s current health-status in a positive way, beneficial if it is considered worthwhile by the patient himself, and burdensome when it imposes physical, psychological and social costs on the patient and his family in particular, and on the medical team or society in general ^[31-^32]. Given that benefits and burdens are not easily quantifiable, factual and value components as well as objective and subjective criteria are interconnected in proportionality judgments. As a consequence, the assessment of the proportionality or disproportionality of treatments should integrate the physician’s technical expertise and the patient’s (or surrogate’s) values, and cannot be a unilateral decision. In particular, the patient’s refusal of proportionate treatments should be always critically analyzed by the physician. Merely accepting and fulfilling this refusal means making abandonment happen. Abandonment leads to the erosion of the physician’s main duties (to care and to relieve suffering) and corresponds to an illegitimate letting die.

The decision to withhold or withdraw treatments should be made within the clinical encounter and be preceded both by the patient’s informed consent and by a critical analysis of the patient’s will. Indeed, the presence or the absence of these two elements is a relevant aspect in order to determine the rightness or wrongness of letting die. To specify, letting a patient die is legitimate when the patient’s refusal of medical treatments is fully informed and freely done, and when the reasons for this choice have been critically analyzed and discussed. On the contrary, a morally culpable letting die occurs when both the patient’s informed consent and a critical analysis of his will or one of these aspects are absent. Admittedly, in some particular circumstances, one or both of them might be “present” but the characteristics of this presence could lead to an illegitimate letting die. For example, following a voluntary, well-considered and informed request for euthanasia, the physician could critically analyze the patient’s wishes and accede to this request but at the same time reject killing the patient by lethal injection. In other words, by ascertaining physical pain and emotional stress caused by the disease,¹³ the physician could legitimate the patient’s choice and decide to withdraw treatments in order to hasten the dying process. In this way, an illegitimate letting die morally equivalent to killing occurs.

In order to distinguish a morally culpable letting die from a letting die for the patient’s good, a further analysis of the non-fulfillment of positive duty is required. As already addressed, all letting die actions are breaches of the duties “Help others” and “Prevent others from dying”, which are positive and less strict duties than the one that obligates us to not kill. Now, within a clinical context, the non-fulfillment of the aforementioned positive duties is unjustified when the physician possesses the means to save the patient’s life and these means have not been refused by the patient. This morally culpable letting die should be distinguished from the letting die for the patient’s good in which the non-fulfillment of duties is justified. For example, the physician could not provide help to follow the patient’s will, that is, in order to respect his informed refusal. Furthermore, refraining from aiding is legitimate when providing help requires an excessive commitment by the physician (for example, exposing him/herself to the risk of death), or injures a third party’s rights. In this particular case, the physician could act in order to maximize the overall good if he is, at the same time, committed to saving one life or more patients’ lives.¹⁴ In such circumstances, justice and healthcare allocation issues cannot be underestimated, even though, within a healthcare context, a purely economic logic should be avoided because it would impede an adequate safeguard of, and assistance to, vulnerable patients.

5. Conclusions

In order to ascribe responsibility at the end of life, the reference to some moral relevant distinctions seems to be unavoidable. The analysis of the killing/letting die distinction proposed in this paper confirms this aspect: to answer the question whether the same moral responsibility could be ascribed to a physician who practices euthanasia and to another who withholds or withdraws life-sustaining treatments, it has been necessary to distinguish because cause and conditions, negative and positive duties, morally culpable letting die and letting die for the patient’s good. Maintaining all these distinctions means recognizing the multiform elements of a moral human act (agent, fulfillment of duty, and consequences) as well as the different aspects that characterize the clinical context (promotion of patient autonomy and respect of his/her will, proportionality of treatments, and fulfillment of ethical and deontological duties required by those who perform the medical profession).

Taking all these aspects into consideration and finding the right balance among them is a difficult task, but a necessary one in order to avoid reductive approaches when dealing with end-of-life issues. For example, by recognizing the cause/condition distinction and by addressing the different roles of a physician in killing a patient and letting him die, moral responsibility is not defined in purely causal terms. Furthermore, the physician is less likely both to take over-responsibility for his actions and to become an underresponsible agent.

At the same time, dealing with the aforementioned distinctions means struggling with ambivalence because in some cases letting die is morally equivalent to killing. Moreover, in some particular clinical contexts a clean line between morally culpable letting die and letting die for the patient’s good might seem not to occur. For example, when the patient’s informed consent and a critical analysis of his will are “present”, the patient’s refusal of treatment and the physician’s non-fulfillment of the positive duty seem justified.

Attesting the presence of these ambiguous “grey zones” cannot bring us to the wrong conclusion that all cases of letting die are the same kind of action, or worse, that they are killing actions. On the contrary, this aspect should help in the acknowledgment of the complexity of the fundamental structure of morality and make us more aware regarding the conflicting decision-making process at the end of life.