Withholding and Withdrawing Treatment

Decisions to withhold or withdraw medical treatment are among the most challenging in clinical practice, yet they are an inevitable and essential component of good medical care. In Australia, approximately 54% of deaths occur in hospital, and a significant proportion of these involve the limitation or withdrawal of life-sustaining treatment. As the population ages and the prevalence of chronic, progressive illness increases, clinicians across all specialties — not only palliative care — must be competent in these conversations and decisions.

The principle that withholding (not initiating) and withdrawing (ceasing) a treatment are ethically and legally equivalent is well established in Australian and international bioethics. If an intervention is judged to be non-beneficial, or if its burdens outweigh its benefits in the context of a patient's goals, then neither initiating nor continuing it is obligatory. This principle is supported by the Australian Medical Association (AMA) Code of Ethics, the Medical Board of Australia's Good Medical Practice framework, and multiple Australian court decisions.

This article addresses the four core domains of treatment limitation: understanding and assessing treatment burden, structuring time-limited trials of therapy, undertaking resuscitation and advance care planning, and the processes for stopping non-beneficial treatments. It is relevant to all clinicians involved in the care of patients with serious, advanced, or life-limiting illness.

Australian Epidemiology of Treatment Limitation

• Approximately 54% of Australian deaths occur in acute hospitals; many involve decisions to limit or withdraw treatment in the final days of life.
• ICU studies in Australia show that treatment limitation (including withdrawal of life-sustaining therapy) occurs in 25–40% of ICU deaths.
• The Australian Commission on Safety and Quality in Health Care (ACSQHC) National Consensus Statement: Essential Elements for Safe and High-Quality End-of-Life Care (2015) mandates that all health service organisations have processes for advance care planning and treatment limitation.
• Palliative Care Australia's National Palliative Care Standards (5th edition, 2018) include shared decision-making and advance care planning as core standards.
• Australian Bureau of Statistics data (2022) shows ischaemic heart disease, dementia, cerebrovascular disease, and lung cancer as leading causes of death — all conditions where treatment limitation is frequently appropriate.

Treatment burden (also termed 'therapeutic burden') refers to the cumulative workload of healthcare and its impact on a patient's quality of life. It encompasses the direct physical effects of treatments (side effects, complications, pain), the logistical demands (hospital attendances, investigations, medication regimens), the financial costs, the psychological toll (anxiety, loss of autonomy), and the social consequences (separation from family, inability to work or participate in valued activities).

Conceptual Framework

The Cumulative Complexity Model posits that when treatment workload exceeds a patient's capacity to cope, outcomes deteriorate — regardless of whether the treatment itself is physiologically effective. This model is particularly relevant in advanced illness where patients may be simultaneously managing multiple comorbidities, polypharmacy, functional decline, and progressive symptom burden.

Domains of Treatment Burden

Assessing Treatment Burden in Practice

• Explicitly ask patients about the impact of their treatment on daily life: "What is the hardest part of your treatment for you?"
• Use validated tools where available — the Treatment Burden Questionnaire (TBQ-15) is a patient-reported measure of burden across 15 domains.
• Reassess burden at every clinical encounter, particularly after new diagnoses, hospitalisations, or treatment changes.
• Recognise that treatment burden is subjective — the same regimen may be acceptable to one patient and intolerable to another, depending on goals, values, and stage of illness.
• In advanced illness (e.g., metastatic cancer, end-stage organ failure, advanced dementia), the threshold for acceptable burden should be higher — treatments must offer meaningful benefit relative to the effort required.

A time-limited trial (TLT) of treatment is a structured agreement between the clinical team, the patient, and/or their surrogate decision-maker to initiate or continue an intensive treatment for a defined period, with pre-specified goals and a planned reassessment date. If the agreed goals are not met by the review date, treatment is de-escalated or withdrawn. TLTs are particularly valuable when prognosis is uncertain and there is equipoise about the likely benefit of an intervention.

Rationale

• TLTs reduce moral distress among clinicians by providing a transparent, pre-agreed stopping point for non-beneficial treatment.
• They reduce conflict with families by framing treatment limitation as the fulfilment of an agreed plan, not a unilateral decision to "give up."
• They allow prognostic uncertainty to be acknowledged honestly while still providing structure.
• Australian ICU studies have shown that structured TLT agreements reduce ICU length of stay and the duration of non-beneficial mechanical ventilation without increasing mortality.

Key Elements of a Time-Limited Trial

Resuscitation planning — the process of documenting decisions about cardiopulmonary resuscitation (CPR) and other emergency treatments — is a critical component of advance care planning. In Australia, this is formalised through jurisdiction-specific Advance Care Directives (ACDs) and, in some states, statutory health attorney frameworks.

Australian Legal Framework

• Each Australian state and territory has legislation governing ACDs. Key statutes include the Advance Care Directives Act 2013 (SA), Medical Treatment Planning and Decisions Act 2016 (Vic), Powers of Attorney Act 1998 (Qld), and equivalent legislation in other jurisdictions.
• A valid ACD that clearly covers the clinical situation is legally binding and must be followed by clinicians, even if they disagree with the content.
• Clinicians who override a valid ACD without lawful justification may face civil or professional consequences.
• In the absence of an ACD, the treating team makes decisions based on best interests, consulting with the patient's substitute decision-maker (SDM) or statutory health attorney.

The ReSPECT Process

The Recommended Summary Plan for Emergency Care and Treatment (ReSPECT) process, adapted for Australian use, provides a structured approach to resuscitation planning that goes beyond a simple "for" or "against" CPR decision. It documents:

• The patient's understanding of their condition and treatment options.
• Their preferences and values regarding quality of life, independence, and acceptable treatment intensity.
• Clinical recommendations about which treatments are likely to be beneficial and which are not.
• A personalised plan for emergency and non-emergency treatment, including CPR, ICU admission, antibiotics, fluid resuscitation, and other interventions.

When to Initiate Resuscitation Planning

Conducting the Resuscitation Planning Conversation

1. Prepare: Review the patient's medical record, identify prognosis, and consider what treatments may or may not be beneficial. Choose a private setting with adequate time.
2. Explore understanding: Ask the patient what they understand about their condition and what they have been told. Use open-ended questions: "What have the doctors told you about your illness?"
3. Identify values and goals: "What is most important to you in your life right now?" "Are there things you would not want to live without?" "What does a good day look like for you?"
4. Share medical information: Explain the likely trajectory, what CPR involves (including realistic survival rates for their condition — typically <10% for in-hospital cardiac arrest in patients with advanced illness), and what alternatives exist.
5. Make a recommendation: Based on clinical judgement, recommend whether CPR or other treatments would or would not be beneficial. Use clear, kind language: "Based on what you've told me and your medical situation, I would recommend that if your heart were to stop, we focus on keeping you comfortable rather than attempting chest compressions and electric shocks. This means we would not do CPR."
6. Document: Record the discussion, the patient's values and preferences, the clinical recommendation, and the agreed plan. Complete the jurisdiction-appropriate form (e.g., Resuscitation Plan in Victoria, ACD in SA).
7. Communicate: Ensure the plan is accessible across all care settings — hospital, GP, RACF, ambulance service. Provide copies to the patient, SDM, GP, and relevant services.

The decision to stop a treatment that is no longer providing benefit — or whose burden now clearly outweighs its benefit — is a core clinical responsibility. Australian law, ethical guidelines, and professional standards are clear: clinicians are not obliged to provide non-beneficial treatment, and continuing such treatment may itself constitute harm.

Defining Non-Beneficial Treatment

Treatment is non-beneficial when it can no longer achieve its intended physiological goal, or when the physiological goal itself is no longer aligned with the patient's overall goals of care. It is important to distinguish:

Specific Treatments Requiring Consideration

Mechanical Ventilation

• Withdrawal of mechanical ventilation is one of the most common forms of treatment limitation in Australian ICUs.
• Follow a structured protocol: pre-medication (opioids for dyspnoea, benzodiazepines for anxiety), family presence, extubation or terminal wean (reducing FiO₂ and PEEP incrementally), ongoing symptom assessment.
• Morphine 2.5–5 mg IV bolus PRN or infusion 1–5 mg/hr; midazolam 1–2.5 mg IV bolus PRN or infusion 0.5–2 mg/hr. Titrate to comfort — respiratory rate and blood pressure are not treatment targets at this point.

Dialysis

• Withdrawal of dialysis accounts for approximately 20% of deaths in Australian patients on maintenance haemodialysis.
• Patients with dialysis withdrawal typically survive 7–14 days (range 1–40+ days). Palliative care referral should be concurrent with the decision.
• Symptoms to anticipate: uraemic symptoms (pruritus, nausea, drowsiness — typically late), fluid overload (dyspnoea, oedema), and general terminal decline.
• Management: frusemide if residual renal function, fluid restriction, subcutaneous opioid for dyspnoea, ondansetron 4 mg SC/IV for nausea.

Artificial Nutrition and Hydration (ANH)

• There is no evidence that artificial nutrition prolongs survival in advanced dementia or advanced cancer cachexia. It may increase burden (aspiration, diarrhoea, tube complications).
• Tube feeding (PEG or NG) should not be initiated in advanced dementia when the patient can no longer take food safely; careful hand feeding is preferred.
• Withdrawing ANH is ethically equivalent to withholding it, though it remains emotionally challenging for families. Clear explanation of the natural dying process (including that hunger and thirst diminish in the dying phase, and that dry mouth is best managed with mouth care, not IV fluids) is essential.
• IV fluids at end of life may worsen symptoms (pulmonary oedema, peripheral oedema, urinary frequency) and are generally not recommended unless there is a reversible cause of dehydration.

Medications

• In the last days of life, discontinue all medications that are not providing comfort or symptom control (e.g., statins, antihypertensives, hypoglycaemics, anticoagulants where there is no active indication).
• Continue medications that are managing symptoms (opioids, anxiolytics, antiemetics, anticonvulsants).
• Convert all non-essential medications to subcutaneous routes where possible in the dying phase; discontinue oral medications the patient cannot swallow.

The Withdrawal Process

Symptom Management During Withdrawal of Life-Sustaining Treatment

The Doctrine of Double Effect

Australian law and medical ethics recognise the doctrine of double effect: a treatment that is given with the primary intention of relieving suffering (e.g., opioid titration for dyspnoea or pain) is ethical even if a foreseeable but unintended secondary effect is to hasten death. This is distinct from euthanasia, where the primary intention is to cause death. The key requirements are:

• The primary intention must be to relieve symptoms, not to cause death.
• The treatment must be proportionate to the symptom — doses should be titrated to symptom relief, not given in arbitrary large amounts.
• The symptom must be serious enough to warrant the treatment.
• There must be no reasonable alternative that does not carry the same risk.

Investigations in the context of treatment limitation serve a different purpose from diagnostic workup. They are used to inform prognosis, confirm futility, or guide symptom management — not to pursue further disease-modifying interventions unless consistent with the patient's goals.

Risk stratification in treatment limitation is less about scoring systems and more about recognising clinical trajectories and integrating patient values with prognostic information.

Clinical Trajectory Framework

Integrating Prognosis with Patient Values

Prognostic information alone is insufficient for treatment limitation decisions. A treatment that is "futile" in statistical terms may still align with a patient's deeply held values (e.g., religious beliefs about the sanctity of life requiring all possible interventions). Conversely, a treatment that offers a small statistical benefit may be clearly unwanted by a patient who prioritises comfort and time at home.

The decision-making framework must always involve: (1) an honest, compassionate sharing of prognostic information; (2) an exploration of the patient's values, goals, and fears; (3) a clinical recommendation based on the integration of (1) and (2); and (4) documentation of the shared decision.

Monitoring in the context of treatment limitation shifts from disease-directed surveillance to symptom-directed assessment and communication quality.

During Active Treatment (Pre-Limitation)

• Regularly reassess treatment burden and alignment with goals at every clinical encounter.
• Monitor for signs of clinical futility: progressive organ failure despite maximal therapy, worsening functional status, recurrent hospital admissions.
• Document discussions about treatment burden and goals in the medical record and communicate with the GP.

During a Time-Limited Trial

• Daily assessment of progress toward agreed goals.
• Clear communication with the patient/family about trajectory ("we are on track" or "we are not seeing the improvement we hoped for").
• Avoid moving the goalposts — if the agreed criteria for success are not met, the agreed plan should be enacted.

After Withdrawal of Life-Sustaining Treatment

• Ongoing symptom assessment every 15–30 minutes during the active withdrawal phase, then as clinically indicated.
• Assess for signs of distress (facial grimacing, restlessness, dyspnoea, autonomic signs) — treat proactively with opioids and benzodiazepines.
• Monitor for respiratory secretions — reposition and use anticholinergic agents as above.
• Assess family coping and provide updates on expected trajectory.

Clinician Monitoring (Post-Event)

• Conduct a clinical debrief after withdrawal of life-sustaining treatment, particularly in ICU and emergency settings.
• Monitor for moral distress among clinical team members and provide access to employee wellbeing services.
• Offer bereavement follow-up to the family (phone call, letter, or referral to bereavement services) at 4–6 weeks.

1. 1. Australian Medical Association. AMA Code of Ethics — 2004. Editorially Revised 2006. Revised 2016. Canberra: AMA; 2016.
2. 2. Australian Commission on Safety and Quality in Health Care. National Consensus Statement: Essential Elements for Safe and High-Quality End-of-Life Care. Sydney: ACSQHC; 2015.
3. 3. Palliative Care Australia. National Palliative Care Standards. 5th ed. Canberra: Palliative Care Australia; 2018.
4. 4. Willmott L, White B, Chipman M. Withholding and withdrawing life-sustaining treatment in a person's best interests: Australian legal and clinical perspectives. J Law Med. 2020;27(4):804–823.
5. 5. Medical Board of Australia. Good Medical Practice: A Code of Conduct for Doctors in Australia. Melbourne: Medical Board of Australia; 2020.
6. 6. Downar J, You JJ, Bagshaw SM, et al. Nonbeneficial treatment Canada: definitions, causes, and potential solutions from the perspective of healthcare workers. CMAJ. 2015;187(4):E159–E166.
7. 7. Cardona-Morrell M, Kim J, Turner RM, Anstey M, Mitchell IA, Hillman K. Non-beneficial treatments in hospital at the end of life: a systematic review on extent of the problem. Int J Qual Health Care. 2016;28(4):456–469.
8. 8. Quill TE, Holloway R. Time-limited trials near the end of life. JAMA. 2011;306(13):1483–1484.
9. 9. Vink R, Azoulay E, Hill A, Sprung CL, Benbenishty J. Time-limited trial of intensive care treatment: an overview of current literature. Intensive Care Med. 2018;44(9):1369–1377.
10. 10. Resuscitation Council (UK). Recommended Summary Plan for Emergency Care and Treatment (ReSPECT). London: Resuscitation Council (UK); 2020. [Adapted for Australian use by Resuscitation Council of Australia and New Zealand.]
11. 11. Australian Institute of Health and Welfare. Palliative Care Services in Australia. Cat. No. HWI 301. Canberra: AIHW; 2023.
12. 12. Douglas C, Murtagh FE, Blyth DM, et al. Symptom management during the last day of life: a national survey of Australian hospitals. J Palliat Med. 2017;20(7):751–758.
13. 13. Smith TA, Mendoza-Nassi T, Luckett T, et al. Palliative care for Aboriginal and Torres Strait Islander peoples: a framework for practice. Aust J Prim Health. 2022;28(3):185–192.
14. 14. Crozier F, Duckett SJ. Withdrawal of treatment and the doctrine of double effect in Australian law. J Bioeth Inq. 2019;16(1):51–59.
15. 15. Batchelor F, Hwang K, Haralambous B, et al. Advance care planning and Aboriginal and Torres Strait Islander peoples: a rapid review of the literature. Aust Health Rev. 2019;43(6):645–652.