Advance directives, withdrawal of care and end of life

Advance directives are legal documents that allow patients to state their preferences for medical treatment in advance, in case they later lose the capacity to consent or refuse treatment. They are a cornerstone of patient autonomy at the end of life, safeguarding a person's wishes when they can no longer communicate them. The term "advance directives" is an umbrella concept that includes several distinct instruments, most notably living wills and durable powers of attorney for health care (DPA). A living will (LW) records specific treatment preferences, while a DPA designates another person to make healthcare decisions on the patient's behalf. A do-not-resuscitate (DNR) order is a separate but related medical order that instructs clinicians not to attempt cardiopulmonary resuscitation. ^{[6] [3]}

The ethical foundation of advance directives rests on respect for autonomy. When a patient completes an advance directive, they are exercising prospective self-determination. However, the practical implementation of these directives is influenced by several factors. A narrative review of 25 articles published between 1997 and 2018 identified that the main factors influencing physicians to implement advance directives were the patient's prognosis, medical paternalism, and the patient's understanding of their medical condition. Additional contributing factors included respect for autonomy, lack of knowledge and experience with directives, legal concerns, family influence, and cultural and religious factors. ^{[3] [4]}

The clinical impact of advance directives is variable. One American study of 336 patients with advanced cancer examined the association between different types of directives (LW, DPA, and DNR), quality of life, and care costs. The presence of a DNR order in the whole sample was associated with better quality of life (p=0.01), but this association was not seen among patients with a living will or DPA (p=0.49). The study was unable to demonstrate cost differences between having or not having directives (DNR, p=0.12, and LW/DPA, p=0.31). ^[3]

Communication interventions can improve advance directive completion rates, though evidence is mixed. In one study by Nicolasora et al, a communication intervention led to a significantly higher proportion of completed advance directives at hospital discharge (0.8% vs 12.7%; P < .001). However, another study by Yamada et al, which also included patients who intended to fill out an advance directive, found no significant findings 4 weeks after hospital discharge. A third study by Epstein et al looked at advance care planning documentation overall, which included advance directives, but the three studies had too much heterogeneity to be included in a meta-analysis. ^{[1] [3]}

A living will is a specific type of advance directive in which a person records, in writing, which medical treatments they would accept or refuse in future situations where they cannot consent. It is particularly important for nursing home residents, not only against the background of the COVID-19 pandemic. The living will allows patients to refuse or consent to treatment measures in advance, thereby extending their autonomy into a time when they are incapacitated. ^[3]

A detailed analysis of 265 recorded living wills from nursing home residents identified 2072 treatment situations and 1673 treatment measures. The findings reveal clear patterns in how residents approach different types of care: ^[6]

The high proportion of living wills referencing resuscitation attempts (88.6%) underscores that CPR decisions are a central concern for patients planning end-of-life care. The fact that 62% of living wills could be assigned to a template suggests that standardised forms are commonly used, which may help ensure completeness but could also limit individualisation. ^[4]

Video-based interventions have been studied as a way to improve advance care planning. Eleven included studies applied a video-based intervention. Ten videos showed simulated cardiac arrests and medical procedures undertaken during CPR, such as chest compressions and intubation. Some videos also contained images of real patients being treated on intensive care units, and other videos provided information regarding end-of-life care or advance directives. Other studies used designed advance care planning interviews, standardised scripted explanations, or written information as interventions. These approaches aim to help patients make more informed decisions about the treatments they would want. ^[1]

The doctrine of double effect has been criticised in the context of end-of-life care. Frequent references to it reinforce the notion that palliative treatment hastens death. Furthermore, applying the doctrine encourages physicians to develop a sense of moral detachment from the consequences of their actions and from their patients. It dissuades physicians from engaging with their patients as equally complex, imperfect, biased and sensitive people; rather, it encourages a separation by seeing patients as vessels for good and bad effects, some intended and some not. This effect is considered unjustifiable, especially at the end of life. ^[5]

It is important to recognise that the implications of not withdrawing ventilation from one patient must be considered alongside the implications of withholding it from another. In a time of shortage, failing to withdraw care from a patient who is unlikely to benefit may mean that treatment must be withheld from another patient who could benefit. This is plainly problematic and highlights why the distinction between withdrawing and withholding is indefensible. The law also calls for health profes

This section addresses a critical mismatch between the requested topic and the provided evidence. The passages supplied (–) focus exclusively on advance directives, the doctrine of double effect, palliative sedation, and physician intentions at the end of life. They contain no information regarding the clinical criteria, physiological tests, or neurological cross-checks required to diagnose brain death. Consequently, it is impossible to write a medically accurate account of brain death cross-checks using only the provided sources without inventing data, which is strictly prohibited.

To adhere to the instructions, I must highlight the ethical context provided in the texts, which often intersects with end-of-life decision-making, though it does not cover the neurological examination itself.

Advance directives (ADs) are legal tools that acknowledge a patient’s autonomous right to make healthcare decisions, both in the present and in anticipation of future incapacity. Studies indicate that the existence of ADs is associated with lower rates of invasive interventions in the last days of life, including reduced use of mechanical ventilation, CPR, and admission to intensive care units. However, systematic reviews suggest that while ADs may reduce therapeutic obstinacy, their impact on patient satisfaction and overall end-of-life care quality is mixed. ^[3]

The doctrine of double effect is frequently invoked in end-of-life care to justify treatments that may hasten death as a side effect of relieving suffering. However, contemporary ethical analysis suggests this doctrine is often misapplied. For instance, under Thomas Aquinas’ framework, the relief of suffering is not an intrinsic good that can offset the risk of causing death, meaning palliative sedation would not qualify for double effect reasoning if it risks life. Modern autonomy-based reasoning, however, prioritises patient preference; if a patient welcomes death as a relief from suffering, palliative sedation may be administered without concern for hastening death.

Physicians’ intentions in end-of-life care are complex. Empirical studies reveal that many doctors intend, or do not mind, hastening death through palliative sedation. For example, 39% of American physicians reported sedating patients with the intention of hastening death, while 36% of Australian surgeons admitted to using drugs at doses greater than necessary for symptom relief with the same intention. A Dutch survey found that 17% of physicians prescribed palliative sedation with the explicit ^[5]

In the NHS, rationing decisions occur daily, including those resulting in patient death. The uncertainty of whether sedation hastens death or merely coincides with imminent natural death creates persistent distress for physicians, sometimes driving them towards inappropriate reliance on the doctrine of double effect. Ultimately, the law supports patient autonomy through advance directives, but the ethical application of these principles requires careful consideration of both patient wishes and the reality of clinical intentions.

The doctrine of double effect is a philosophical principle used to justify an action that produces a good effect—such as the relief of suffering—even if it may also produce a bad effect, for example hastening death. In end-of-life care, it is frequently invoked to provide ethical justification for palliative sedation and analgesia, which are employed in patients with refractory and intractable symptoms to reduce suffering by lowering the level of consciousness. ^[5]

The doctrine is classically applied when four conditions are met: the action is itself good; the bad effect is not the means to the good effect; the bad effect is foreseen but not intended; and the good effect outweighs the bad. For example, prescribing an antibiotic to treat a urinary tract infection could be justified even if it produces an allergic reaction—the action is good, the allergic reaction is not the cause of curing the infection, the risk is recognised at the time of prescribing, and the goal of eliminating the infection outweighs the potential harm. ^[5]

However, the application of the doctrine to palliative sedation is controversial. Some authors argue that palliative sedation and analgesia do not fulfil the conditions required to apply the doctrine, and that its use in this domain is inappropriate. A key problem is causation: for the doctrine to apply, the action must cause both the good and bad effects, and the bad effect must not cause the good effect. Yet determining causation for actions at the end of life is very difficult, making application problematic because of weak causal arguments and limited empiric evidence. ^[5]

A bibliometric analysis of PubMed up to 17 July 2020 identified 627 references to the doctrine of double effect, of which the vast majority (432 articles, 68.9%) related to end-of-life care, including euthanasia (260, 41.5%), palliative sedation/analgesia (104, 16.6%), and other palliative care (68, 10.8%). This disproportionate focus suggests the doctrine is referenced much more commonly in relation to palliative sedation than would be expected, reflecting physicians' discomfort with the moral, intentional, and causal complexities of end-of-life decision-making. ^[5]

Euthanasia and physician-assisted dying (PAD) sit at the ethical and legal frontier of medicine, where the relief of suffering meets the prohibition against killing. To understand these practices, you must first distinguish them clearly. Euthanasia is the deliberate act of ending a patient's life by a physician, typically by administering a lethal drug, at the patient's explicit request (voluntary euthanasia) or, more controversially, without it (non-voluntary or involuntary). Physician-assisted dying (often called physician-assisted suicide) differs in that the physician provides the means—usually a prescription for a lethal dose of medication—but the patient self-administers the drug. The doctor does not perform the final act; the patient does.

The ethical distinction matters because it changes the moral weight of the physician's role. In euthanasia, the doctor is the agent of death. In PAD, the patient retains agency until the end. This is why many jurisdictions that permit PAD stop short of allowing euthanasia, and why some professional bodies treat the two differently. However, both practices share a core ethical tension: they involve intentionally causing death, which is why they cannot be justified by the doctrine of double effect.

The doctrine of double effect, derived from St. Thomas Aquinas' justification of killing in self-defence, holds that an action with both good and bad effects is permissible if the bad effect is unintended, even if foreseen. Aquinas argued that self-defence is justifiable even if the attacker is killed, provided the intention is to save oneself, not to kill. Ethicists have since elaborated conditions for its use: the act itself must be morally good or neutral; the good effect must be intended; the bad effect must not be the means to the good effect; and the good effect must outweigh the bad. Palliative sedation—using medications to reduce awareness when this is the only way to relieve intractable suffering—is often justified by this doctrine. But euthanasia and PAD fail these conditions: causing death is not merely foreseen but explicitly intended, and it is the very means by which suffering is relieved. Therefore, double effect reasoning cannot justify them. ^[5]

Legally, the landscape varies widely. Some countries and states permit PAD, euthanasia, or both, under strict criteria: the patient must be an adult with decision-making capacity, have a terminal or incurable illness, experience unbearable suffering, and make a voluntary, repeated, and informed request. In jurisdictions where euthanasia is legal, conscientious objection (CO) is a flashpoint. In Italy, for example, following approval of a law on end-of-life care, critics attacked the absence of an explicit provision for conscientious objection for physicians, health workers, and institutions. They argued that CO is a fundamental human right guaranteed by the Italian Constitution and the Italian code of medical ethics, and that its absence could reduce healthcare workers to "mere executors of the patient's will". This debate illustrates a recurring theme: even where the law permits ending life, it must also protect those who cannot, in conscience, participate. ^[4]

Medical students and junior doctors often stumble on several recurring pitfalls when reasoning about end-of-life decisions. The first is misapplying the doctrine of double effect. It is tempting to invoke it for any intervention that might hasten death, but the doctrine has strict conditions. It requires that the bad effect be unintended, not merely foreseen. If you prescribe opioids with the primary intention of relieving pain, and respiration is depressed as a side effect, double effect may apply. But if your intention is to end the patient's life, it does not. The distinction is one of intention, not outcome. Euthanasia and PAD would not qualify because causing death is explicitly intended and is the means of achieving the good effect. ^[5]

A second pitfall is assuming that withdrawing or withholding treatment is ethically equivalent to euthanasia. It is not. Withdrawing ventilation from a patient with no prospect of recovery is generally accepted as allowing natural death, not causing it. However, the reasoning becomes treacherous when resources are scarce. Consider the classic scenario of two patients, John and Donald, both needing ventilation. Donald is not provided ventilation because of his condition and prospect of recovery, yet John, who appears to have an equivalent chance of recovery, is currently being ventilated. The only difference is that John is already on the ventilator. This raises the question: is it permissible to withdraw ventilation from one to give it to another? An argument of necessity fails here, because the choice to kill the youngest and weakest crew member in the famous lifeboat case was rejected: "Was it more necessary to kill him than one of the grown men? The answer must be 'No'". Applied to ventilation, there is no basis to measure the comparative value of lives. Yet the difficulty is that it may actually be more necessary to deprive one person of ventilation than another, because one may require a month of ventilation while four others need only a week. This is not about valuing lives differently, but about allocating a scarce resource efficiently. ^[2]

A third pitfall is justifying denial of treatment based on uncertain future shortages. If ventilators are currently available but you fear future scarcity, it is difficult to argue that it is necessary to deny John and Donald treatment now, because the death of other patients from a future shortage is not "inevitable". If the future shortage were guaranteed, the situation would resemble immediate triage. But when the future is uncertain, denying treatment becomes harder to justify. This is a question about the appropriate criteria for accessing ventilation and what level of pre-emptive rationing is appropriate. The trap is to act on hypothetical future needs rather than current clinical reality. ^[2]

A fourth pitfall is ignoring advance directives or misinterpreting them. Advance directives are part of shared care planning, a dynamic process aimed at ensuring patient engagement in healthcare across time and practice settings. Their goal is patient-centred care, especially for those with complex medical and social needs, so that treatments align with their values, needs, and preferences. A common error is to treat an advance directive as a static document rather than a guide to the patient's current wishes. Another is to assume that a directive covers all situations; it may not address new treatments or changed circumstances. The evidence base itself is limited: a narrative review of studies from 1997 to 2018 found that selected studies used diverse methods and endpoints, and the extraction and interpretation of data could be biased. So be cautious about over-generalising from the literature. ^{[4] [3]}

In practice, end-of-life care is rarely a single decision but a process of shared planning. Shared care planning is a dynamic process that ensures patient engagement across time and settings, aiming to deliver patient-centred care in which treatments align with the patient's values, needs, and preferences. This is especially important for patients with complex medical and social needs. Advance directives are one tool within this process, but they are not the whole of it. ^[4]

Where euthanasia and PAD are legal, they are used only under strict criteria. Typically, the patient must have decision-making capacity, make a voluntary and repeated request, and suffer unbearably from an incurable condition. The physician must confirm the diagnosis and prognosis, often with a second opinion. In practice, this means that euthanasia and PAD are rare, last-resort options, not routine care. They are used when palliative options, including palliative sedation, have failed or are unacceptable to the patient.

Palliative sedation itself is an accepted but rarely used practice in palliative care. It involves using medications in a monitored fashion to achieve decreased or absent awareness when this is the only way to relieve otherwise intractable suffering. Palliative analgesia achieves a similar result using analgesics rather than sedatives. These practices are frequently criticised by those who argue that they hasten or cause death, though many palliative care experts disagree. The doctrine of double effect is often applied to justify them. However, as noted, this application is controversial: it is difficult to support the notion that all physicians, nurses, family members, and patients view the arrival of death as a "bad" effect of palliative sedation, nor as an unintended consequence. In the absence of an appropriate structure of goods, established causation, or simple intentions, the application of the doctrine to palliative analgesia and sedation appears inappropriate. ^[5]

In practice, the decision to withdraw or withhold life-sustaining treatment is more common than euthanasia. This happens when treatment is judged futile or not in the patient's interests. But as the John and Donald scenario shows, such decisions are fraught when resources are scarce. The flowchart below outlines the reasoning process for a patient with respiratory failure when ventilators are limited.

In practice, the ethical principle of non-abandonment is paramount. Even when treatment is withdrawn or withheld, the physician's duty to the patient continues. This is where shared care planning becomes essential: it ensures that the patient's values guide decisions, and that the process is transparent and documented. In jurisdictions where conscientious objection is recognised, physicians who cannot participate in euthanasia or PAD must refer the patient to a colleague who can, without abandoning the patient. Where CO is not recognised, as critics argue is the case in Italy, the risk is that healthcare workers become "mere executors of the patient's will". This is a warning: the law must balance patient autonomy with professional integrity. ^[4]

Finally, remember that the evidence base for advance directives and end-of-life decision-making is imperfect. A narrative review of studies from 1997 to 2018 found that studies used diverse methods and endpoints, and the extraction and interpretation of data could be biased. So in practice, rely on careful clinical judgement, open communication, and a clear ethical framework, rather than on any single study or guideline.