Palliative care admissions I: Recognizing dying

Dedicated readers familiar with the course of critical illness understand that the admission of a critically ill patient is generally a time of great upheaval, when numerous investigations, interventions, and interested parties all converge upon the person in the bed. Scans are performed, meds given, tubes and lines placed, and the amplitude of medical care swells to a crescendo, hopefully in preparation for a corresponding deescalation in care in a few days as the patient begins to stabilize.

This is how we do our job, and how it always should be. Right?

In fact, it’s not. Sometimes, when a patient comes to the hospital, and even when the ICU is called to admit them, it’s not a sign that they need to be keelhauled by the machine of modern medicine. It’s a sign that they need palliative care.

Note what I didn’t say: I didn’t say they do not need to be admitted to the hospital, or even to the ICU. (Palliative measures sometimes need to be delivered aggressively, at least at first.) I also didn’t say that a palliative care service needs to be consulted. (While that may be true, any competent critical care provider should be able to start the ball rolling. Non-emergent consultations are often unavailable at the zero hour of admission.)

The specter of palliative care is never far from our work in the ICU. Often, though, it’s something that enters the picture days or weeks after presentation, when the clinical picture is beginning to become clear, and curative therapy has proven futile. Making this transition is an important art and science, but today I’d like to focus on a different one: palliative care admissions, the patients who should be managed with a palliative lens from the very beginning.

Who needs the palliative care admission?

Since by default, the majority of our admissions are fully “curative” in nature, you’ll need to be on the lookout for indicators that a particular patient would benefit from a different approach.

The most obvious is when it’s handed to you: a signed legal document stating that the patient should only receive certain measures, or a legally-empowered patient representative who tells you how to focus your care. Unfortunately, this usually isn’t the case. The most common example will be a paper asking you not to attempt CPR or intubate, but that gives little guidance when a patient requires neither.

Instead, what you’ll need to do is consider the context of the patient’s disease. With the exception of trauma, most critically ill adult patients in the developed world are not ill from purely acute diseases; they are demonstrating perturbations and exacerbations of chronic conditions. In addition to addressing the acute changes, the wise clinician will locate its latitude upon the predictable arc of the disease.

Patterns of functional decline at the end of life, from Lunney 2002, later validated in Lunney 2003

Consider the disease pathways above.

  • “True” sudden death. Sudden cardiac arrest, generally in an unlucky healthy patient or one with otherwise well-controlled comorbidities. Usually cardiac or traumatic in nature. The patient is fully functional until the moment of death, and if they survive, they can return to life at or near their previous baseline.
  • Terminal illness. Most classically cancer. There is little or no decline until a sudden drop-off near the end of life. Function is usually unimpaired and the outward manifestations of disease hidden until very late; once the patient exhibits frank decompensation, they are often in the last months of their life.
  • Organ failure. Most often heart failure, COPD, renal insufficiency, and liver disease. The patient exhibits intermittent “exacerbations,” potentially becoming extremely ill each time, but each time has the potential to demonstrate excellent recovery. Over time, however, the general trend is still downward, and eventually they either succumb to the baseline loss of function, or they fail to survive one of the exacerbations.
  • Frailty. Death by “old age,” best demonstrated in dementia. Decline is very prolonged and slow, almost imperceptible. Perturbations occur, but the hallmark pattern is a gradual, irreversible loss of function, not the dramatic episodes of organ failure. Once basic life functions and ADLs are lost, they are usually gone forever.

With an understanding of these patterns, we can begin to recognize where each patient falls—and how much they stand to benefit from aggressive critical care. We can estimate a patient’s chances of “bouncing back” from any particular admission, and if they manage it, what their landing zone might resemble.

The sudden death patient is the perfect candidate for our ICU toolkit. These patients (classically described by Peter Safar as “too young to die”) may or may not survive, but if they do, they can live long, healthy lives.

The terminal illness patient often has little to gain from aggressive curative therapies. Unless their complications are totally unrelated to their terminal diagnosis, by the time they are bedbound and anorexic, there is often little to be done. Palliative care may be the best pathway.

The organ failure patient may be the most common admission in most ICUs. Often they are reasonable admissions and can do well in the short term, but in the long term, their disease usually continues to progress. Understanding their overall trend is more important than their acute picture.

Finally, the frail or demented patient, although they may or may not survive acute complications, usually has no baseline to return to. For instance, if they develop aspiration pneumonia, it can be treated—but if they cannot manage secretions as a result of their cognitive deterioration, they will likely continue to aspirate until it eventually kills them. If their baseline quality of life has become unacceptable, it is usually appropriate to focus care on comfort rather than “cure” (which is impossible).

Your end-of-life GPS

This is all easy enough to say, but it’s often difficult to pinpoint exactly where a patient falls on these curves. Fortunately, there’s help. You can use the Gold Standards Framework, the SPICT tool, the NECPAL CCOMS-ICO, or the RADPAC tool (for cancer, CHF, or COPD). Using these rubrics to interview the patient, their family, or their caregivers will help focus your attention onto markers of poor prognosis.

You can pick the tool that suits you best. However, you’ll notice some common features that signal when a patient is approaching the end of their life:

  • The “Surprise” question: “Would you be surprised if the patient were to die in the next [days, months, up to a year]?” This accesses our gestalt impression of prognosis as clinicians, and probably has good specificity. However, its sensitivity is rather poor; we tend to overestimate survival. I find it useful only sometimes.
  • Increased frequency of complications: More frequent hospitalizations, an increase in falls, recurrent infections, etc.
  • General functional decline: Worsening and loss of basic human abilities. Losing weight and no longer eating, spending much of the time in bed, increasing care requirements (from independence at home up through skilled nursing), incontinence, inability to speak, and similar losses of function.
  • Worsening of disease burden: A progression of signs and symptoms, and an increasing failure of medical therapies to control them. Details are generally specific to disease type.

Using these tools, and with our new understanding of the progression of chronic diseases, we can start to understand the context for a patient’s acute admission.

A few examples

Sudden death

She was totally fine until the bus hit her. Just a little high blood pressure.

Terrific! Full steam ahead.

Terminal illness

He was diagnosed with the lung cancer last year. It’s stage IV now. He was doing okay for a while, but now he’s pretty much stopped eating; he must have lost 40 pounds the last couple of months. He’s hardly getting out of bed anymore. We finally had to come to the ER when he started having more and more trouble breathing.

This seems to be a patient in the last phase of the disease that will kill him. Acute, superimposed problems should be ruled out, easily managed issues addressed, and the plan of care discussed with the patient’s oncologist or PCP, but this is likely a person with little to benefit from intubation and the ICU gauntlet. They may be better served with palliative care and perhaps even a discharge to hospice.

Organ failure

I’ve had the kidney failure for ten years. Mostly I’m doing okay. I live at home and see my normal doc and my kidney guy regularly. I had a transplant once, but it failed a couple years ago. Every so often I do miss my dialysis, and sometimes I end up having trouble breathing and I end up in the ER. This time seems like it’s pretty bad.

This sounds like a typical kidney failure patient. Even if her acute presentation is very concerning, that is not prognostic of a poor outcome; with good critical care she may recover to her (fairly intact) baseline. Certainly, her disease is chronic, and she is undoubtedly getting a little sicker with each admission, but in the short-term she should do well (perhaps with some care management to help her keep her appointments). At the very least a trial of full curative measures is reasonable to see how she responds. However, consider her roommate…

I used to be doing pretty well with the kidney failure, but now it just seems like my body is falling apart. I’ve had two fistulas fail, and now I can only get these catheters, but I’ve already had two of them removed because of infections. I’m sick of going to dialysis. Even when I go, I always have problems with my blood pressure and often they can’t finish. I can’t walk much and I’ve had to move into assisted living. I even have these chronic ulcers on my legs and my butt now. And I feel like I’m constantly in the hospital; I was discharged only two weeks ago!

This tells a different tale. This patient has demonstrated a pattern of functional decline and a failure of available therapies to fully control her disease process. While she is not necessarily actively dying, her quality of life and disease burden have worsened to the point where she may no longer want the added challenge of recovering from a difficult acute illness or aggressive treatments. Reasonable expectations should be set, and an earnest conversation had about her goals of care, both for this admission and in general.


They diagnosed the Alzheimer’s almost ten years ago. At first it was just a little trouble with his memory. It got worse slowly, almost too slow to notice. Then he had a stroke four years ago and wasn’t able to talk afterwards. Since then he’s been doing less and less. He mostly just sits in a chair or in bed. When he does walk, it’s really slow, and he’s fallen four or five times this past year. He only eats a little, and I guess he’s not doing that properly either, because now they’re telling me he’s probably aspirating. He wouldn’t want to live like this.

This is a patient in the end stages of the neurodegenerative disease of dementia. It is possible that he will survive this admission, although his poor baseline health leaves him with very little reserve with which to battle acute illness. If he does survive to discharge, it will likely be with further dramatic decline in his function (for instance, with a new tracheostomy and PEG tube). Even in the best case, where he somehow returns to his full pre-admission baseline, that “baseline” was a very poor quality of life that may not be desirable. A discussion with his family about palliative care would be best.

That’s it for now. In Part II, we discuss the nuts and bolts of how to execute these admissions.

Thanks to the great work of Ashley Shreves for inspiring many of these ideas.

2 Replies to “Palliative care admissions I: Recognizing dying”

Leave a Reply

Your email address will not be published. Required fields are marked *