Palliative medicine in the ICU

Considering how many patients in the ICU have a controlled end-of-life process, it is somewhat surprising that the only appeared in the exam once, as Question 24 from the second paper of 2022. This question asked mainly about the different approaches of managing patients with, or without, a palliative medicine consultation. Hard to know what prompted this, but there was certainly a flurry of papers about this in the 12-18 months prior to this written paper (Hill et al, 2022Einav et al, 2021Kościuczuk et al , 2022Stamper et al, 2022; Ito et al, 2022; and many others). Fortunately, this has made the process of writing an answer much easier; and it would appear that the trainees have been reading these papers as well, because the pass rate for this question was over 78%.

How ICU palliative care differs from ward or hospice

The 2008 ACCCM recommendations mentioned that (at the time of writing but probably also now) in America 22% of all deaths occur in ICU or following admission to the ICU, so you'd think the ICUs would be better designed to accommodate the needs of dying patients and their families. Unfortunately, they are also designed to provide intensive care. Apart from the barriers identified by studies like  Effendy et al (2022), there are some other differences to the whole process of dying in ICU which need to be taken into accou t:

  • Environment: the ICU is often a noisy, busy place. Not always are single rooms available (ie. the ICU may be a "open floor" design) and the constant alarms and lights are an intrusion into family time. This differs from a palliative medicine facility which would usually be designed to put the emphasis on comfort and rest. Moreover it becomes difficult to discuss sensitive issues with the family and answer their questions.
  • Commiunication with the patient is often impossible. Whereas under normal circumstances the patient's wishes and values would be central to the conversation, in the ICU the patient is often not in a position to participate in decisionmaking. The result is a reliance on the family to help guide their care, which is rather different to the way that palliative care is provided in the community.
  • Procedure of withdrawing treatment: whereas in the ward or community the withdrawal of therapy is usually just an inaction (eg. not going ahead with the next round of chemotherapy or not proceeding with a surgery), in the ICU the withdrawal is often of a treatment which is actively and constantly sustaining life, eg. mechanical ventilation, dialysis, ECMO or inotropes. Though the two are ethically identical, the latter may be far more confronting.
  • Grieving processes of the family are by necessity more abrupt because often the events that have led to the palliation are very sudden, often affecting young patients or those who had been previously well. There is often not enough time to gradually come to terms with the events and the inevitable outcome. It may even be a situation where the family have no direct access to the patient because of concerns about communicable disease.
  • Patient symptoms may be more unpredictable. There is the potential to grossly over or underestimate the severity of the distress the patient is going to experience when something like ventilator support is discontinued. There will also be less time during which the sedation or analgesia can be titrated, making it more likely that the intensivist will use an empirically determined higher dose of everything
  • Time course of palliative therapy is obviously going to be shorter if the patient was requiring a lot of support. This means long term side effects of palliative medications do not need to be considered to the same extent. For example, one may bravely use morphine in patients with renal failure knowing that the constipation and the accumulation of the dysphoria-inducing metabolite will not have a huge effect on the overall patient experience.
  • Invasive lines are usually already in place and so it is not especially burdensome to continue using these, whih means the IV route of administration remains open. Whereas in the ward or the community IV access is usually not expected, and palliative medicine physicians tend to rely on subcutaneous administration or oral medications.
  • Follow-up planning by the ICU team is essential for anybody expected to be discharged from the ICU for ongoing palliative care in a ward setting, and a multidisciplinary approach is probably essential so that the symptom control and lines of communication with the family are not interrupted.

Following from that last point, it would make sense to involve external palliative services early. So: let's discuss that:

Palliative medicine consultation in the ICU

Let's say the intensivist, upon achieving an understanding with the family of a dying patient regarding the goals of care, hands over the management to a palliative medicine specialist. 


  • Specialist care: though intensivists deal with death and dying a lot, the palliative medicine specialist service is literally immersed in this throughout their training and career, making this their unique area of expertise. It is difficult to argue that their approach to symptom control will be somehow inferior to that of the ICU specialist. More likely, they will be able to bring something novel to the table.
  • Resources and service provision: busy ICU staff will often appreciate being able to delegate some of their work to an external care provider, and this could be viewed as a way conserving ICU resources for patients that require that specific skill set. Speaking of which:
  • Skill mix for palliative care is not entirely overlapping with the ICU nursing and medical skill mix, and therefore it is possible that the patient will actually have better care and attention in the hands of dedicated palliative medicine staff. For example, it is possible that experienced palliative medicine nurses will be more attentive to subtle signs of distress, and will be more proactive with PRN medications.
  • Transition to dedicated palliative medicine facilities would be smoother, and potentially also faster, which would reduce overall healthcare costs
  • Communication with the family is said to improve, but this is often based on studies from overseas (eg. Martins et al, 2017) where the ICU team does not take as much of a leadership role in the end of life decisionmaking. Yes, in a culture where the intensivist never engages with the family on difficult topics, a palliative medicine consultation would be essential, as it would probably bring about some appropriate goal-setting.


  • ICU patient being managed by non-ICU staff: Grey old men of the ICU will often complain endlessly about the loss of "sovereignty" being disempowering. This is the supposed feeling of losing control which comes with surrendering the management decisions to another team of specialists while the patient is still ostensibly in the ICU. One might respond to this grumbling by pointing out that there is no such thing as  "ICU sovereignty" anyway, as (for example) we don't tell the surgeons how or when to operate, we don't give the oncologists advice about when to start or stop chemotherapy, and generally there are many examples where "closed" ICU models of care are by necessity collaborative. There are few legitimate patient-care-related reasons to resist collaboration. Still, one cannot rule out the distant possibility that the palliative medicine specialist will come along and say something preposterous to the family, or do something else unexpected that (for example) degrades their trust in the medical system. So, one should probably list this as a potential disadvantage.
  • Handover and interruption of care: transferring the management of the patient to another team may be a situation when critical pieces of information are lost
  • Interruption of rapport: the patient and their family may have already established a relationship with the intensivist and may actually prefer to continue dealing with them, particularly given the difficulty and seriousness of the situation. To be expected to explain potentially sensitive pieces of the patient's history to a new specialist may not be beneficial.
  • The perception of "failure" can develop in the intensive care team, as the result of the perceived inability to meet the palliative care needs of the dying patient
  • The perception of abandonment by the family can develop as the result of the intensivist disenaging and handing over care to the palliative medicine team. This was actually one of the findings of a study by Carson et al (2017), and it gets quoted a lot, except the study looked at the delegation of prognosis discussions which would be unheard of in  Australian Intensive Care culture. Still, it is possible the family may feel an abrupt withdrawal of attention, especially if they were having daily meetings with the intensivist and are now having only weekly meetings with the palliative medicine team.
  • No evidence for benefit in a closed model. All the evidence for benefit comes from studies based on "open model" ICUs, where - looking at the findings - there does not seem to be any governance over end-of-life matters in the ICU. For example, "proactive palliative care consultations led to significant reductions in ICU length of stay, increased DNR (“do not resuscitate”) orders, reduced nonbeneficial resource utilization after a DNR order, and increased transition to goals focused on comfort" according to Strand & Billings (2012), implying that in the studied environments the intensivist would otherwise never establish those "DNR" orders and would continue to utilise resources in a nonbeneficial manner. This is clearly not reflective of local practice.

So, all things considered, perhaps it is better to leave this to the intensivist?

Intensivist-led palliative care in the ICU


  • Rapport and familiarity: the intensivist would typically be the main source of information and decisionmaking during the period of the patient's active management, and the family would have become familiar with them. It would be sensible to use the strength of the connection they have formed to help the family though an even more difficult time.
  • Specialist training: Yes, the palliative medicine physicians have specialist training in symptom control, but then their training is usually carried out in the ward and community environment, and they may not be as familiar with the specific needs of ICU patients; whereas the intensivist may have more insight
  • Resource allocation should not be so important that it would be placed before the patient's or family's experience. Some might consider the allocation of ICU medical and nursing staff to dying patients an appropriate use of ICU resources, and the work as an essential part of wholistic critical care.
  • Increased satisfaction: some intensivists report having some positive experience of managing death. Trankle (2014) quoted intensive care specialists noting how their speciality is changing, and how more and more people are being referred to the ICU in the last days of life; but at the same time these staff reported feeling a sense of accomplishment from being able to relieve symptoms and to provide a "good death".


  • The ICU is not the best place to die; and to transition more of the palliation role to the intensivist would be an unreasonable move, as it would build the expectation that everyone needs to die in the ICU. There should be an understanding that the most effective management of ICU resources would involve preventing the admission of patients who are highly likely to die in spite of aggressive therapy.
  • Impact on the clinician: Even though it is well known that most intensive care specialists are robotic psychopaths, one cannot rule out the possibility that they might, over repeated exposure to human misery and grief,  develop what Tom Solano refers to as "cortical scarring", a defensive psychological callus that hardens the soul against injury but also erodes patience and decreases the amount of empathy and softness one can bring to an interaction. This is normal and it would in fact be unexpected if this did not happen, as nobody's reserves are infinite. On this basis, one could say that to share the burden of providing care to the terminally ill would be an appreciated relief for some intensive care staff.
  • Governance and ownership: the intensivists tend to change every few days, and the staff work in rotating shifts, which means there is still the handover of care from one team to another, and continuity is not preserved anyway (so why not give the handover to the palliative medicine team?)


Hill, Stephanie A., et al. "Palliative medicine in the intensive care unit: needs, delivery, quality." BMJ Supportive & Palliative Care 12.1 (2022): 38-41.

Einav, Sharon, Nathan I. Cherny, and J. Randall Curtis. "Palliative medicine in the intensive care unit." Oxford Textbook of Palliative Medicine (2021): 121.

Kościuczuk, Urszula, et al. "Aspects of palliative medicine in intensive care units: A narrative review." Palliative Medicine in Practice (2022).

Stamper, Tara Orgon, Renee Kerr, and Daniela Sporter. "The Evolution of Palliative Medicine in Intensive Care." Critical Care Nursing Quarterly 45.4 (2022): 332-338.

Tomko, Caitlin, et al. "The Practical Implementation of an Embedded Palliative Care Team in a Medical Intensive Care Unit (QI435)." Journal of Pain and Symptom Management 63.5 (2022): 904.

Bhan, Swati, et al. "Palliative Care in Intensive Care Unit." Onco-critical Care: An Evidence-based Approach. Singapore: Springer Nature Singapore, 2022. 515-524.

Ito, Kaori, et al. "Primary palliative care recommendations for critical care clinicians." Journal of Intensive Care 10.1 (2022): 20.

Effendy, Christantie, et al. "Barriers and facilitators in the provision of palliative care in adult intensive care units: a scoping review." Acute and Critical Care 37.4 (2022): 516-526.

Martins, Belmira DCPCC, Reinaldo A. Oliveira, and Antonio JM Cataneo. "Palliative care for terminally ill patients in the intensive care unit: Systematic review and metaanalysis." Palliative & Supportive Care 15.3 (2017): 376-383.

Carson, Shannon S., et al. "Effect of palliative care–led meetings for families of patients with chronic critical illness: a randomized clinical trial." Jama 316.1 (2016): 51-62.

Strand, Jacob J., and J. Andrew Billings. "Integrating palliative care in the intensive care unit." The journal of supportive oncology 10.5 (2012): 180-187.

Trankle, Steven A. "Is a good death possible in Australian critical and acute settings?: physician experiences with end-of-life care." BMC palliative care 13.1 (2014): 1-14.