book cover

Palliative Care Perspectives : Chapter 2: Dying Trajectories and Prognostication : Dying Trajectories

Our prognoses will never be perfect. There will always be mystery regarding exactly when people die (and why). However, we can recognize certain patterns of dying that can help us better prognosticate and communicate about dying.

The concept of a dying trajectory was first suggested by Glaser and Strauss in 1965 and refers to the change in health status over time as a patient approaches death.13 Glaser and Strauss identified different "patterns" of dying - sudden death, lingering, certain to die on time, and the vacillating pattern. Their focus was on expectations about when a patient would die and communication among hospital staff, patients, and families regarding such expectations. The first to use the term dying trajectory may have been Gustafson in an article referring to the "careers" of nursing home patients in terms of such trajectories.10 Shneidman, in a book published shortly after this, explicitly uses the term death trajectories relative to the expected time until death.11 Formally plotting dying trajectories as health status over a longer period of time and considering systemic implications of different trajectories appears to have been the creation of Joanne Lynn and colleagues relative to the very famous SUPPORT study.12 It is usually plotted (retrospectively for an individual) with time on the X axis and health status on the Y axis. Sudden death has the simplest graph, a rectangle with a straight line down from a state of being healthy to death. The concept of dying trajectories has been helpful in understanding patterns of advanced illness and dying for different disease processes, which, in turn, have implications for care needs, decision making, and prognostication. Individuals will, of course, vary in their personal dying trajectories; however, it is remarkable how similar dying trajectories can be for patients with similar disease processes. Here I will discuss certain common trajectories and their implications.

Sudden Death

The person found suddenly and unpredictably dead is technically beyond the care of the physician. However, survivors are not. It is often the physician who must notify others and provide initial support. Physicians also must provide healthcare to those who have been so bereaved (see section on bereavement in Chapter 7). People may die suddenly and peacefully but more commonly die from accidents or violence, especially the young. The hallmark of this trajectory is lack of preparation for dying. Usually, those who die, unless already very ill or of advanced age, will not have prepared for their death. Even if details such as wills and funeral plans have been made, they have often not had a chance to settle matters in their minds or say goodbye, nor have their loved ones. In addition to the shock of sudden separation and bereavement, survivors may have regrets or feel guilty that they did something wrong or should have done or said something different. Bereavement needs are intense, especially for those bereaved of the young or through violence. Incredibly, many health care systems that deal with sudden death have not established even rudimentary support systems for those in need.

Cancer Deaths

Cancers of very different tumor types and locations often follow very similar dying trajectories. In fact, the dying trajectory of cancer is one of the most predictable trajectories. Most patients with metastatic cancer remain quite functional until approximately five to six months before their deaths. Their health statuses then tend to slowly decline until the rate of decline begins to accelerate rapidly two to three months before death.14 In Teno's study one month prior to death more than 50 percent of the 1655 cancer patients studied had difficulty getting out of a bed or chair. Prior to the period of decline in the last few months of life, cancer patients may have various symptom needs such as pain management but tend to remain at high functional levels. During the rapid decline phase patients start to "take to their beds." A good rule of thumb is that a patient with advanced cancer who has taken to bed without a correctable cause will usually die in a matter of weeks to a few months. (Usually means usually, as I have treated many patients who are exceptions to this rule of thumb.) The caveat about correctable causes is important. Treating certain very responsive cancers with chemo-therapy can get people out of bed. Treating complications such as certain infections sometimes makes a difference. This rapid decline in functional status in the last few weeks to months of life correlates highly with admission to and life expectancy in hospice. Christakis in 1996 identified a median survival of 36 days for 6451 patients followed on the Medicare hospice benefit. Eighty percent of these patients had cancer.7

Palliative Care Note
Patients with advanced cancer who have taken to bed without a correctable cause will usually die in a matter of weeks to months.

Hospice care, as initially envisioned by Cicely Saunders, was developed primarily with this dying trajectory in mind, and it works rather well for cancer patients, as their deaths are relatively predictable once this decline begins. Prior to the rapid decline phase in the last few weeks to months, it may be very hard to predict when patients with certain cancers will die. Patients with metastatic prostate, breast, and colon cancer, for example, may live for years prior to declining. Many cancer patients do not really need home hospice care in the months prior to the accelerated period of rapid decline; their symptoms can be adequately managed on office visits, and psychosocial concerns may be addressed in different ways such as through support groups.

What are the implications of this dying trajectory? When the rapid decline phase begins, prediction of death in a matter of weeks to months is considerably more certain than for most other diseases. With accurate prognostication by the clinician, patients really can know that they will likely die soon. This may come as a surprise, despite having known and even accepted their terminal diagnoses. Patients and families usually have not been taught about this trajectory, and thus, when rapid decline begins, they are often shocked and say, "But he was just fine a couple of weeks ago." Frequently, there is a push to do something to reverse the course of decline. Patients, families, and physicians may be drawn into a frantic dance, which does little but reassure everyone that "everything was tried." This dying trajectory, if recognized, offers the potential for getting one's house in order. Clear prognostication can grant patients and families the time and the permission to take care of unfinished business and to say their goodbyes. For many this is a valuable gift.


Congestive heart failure (CHF), chronic obstructive pulmonary disease (COPD), strokes, and many infirmities of advanced age follow a very different pattern.14 The overall health status of the patient is low 6 to 24 months prior to death. Acute exacerbations of illness (episodes of pulmonary edema, COPD flare-ups, aspiration pneumonias) occur intermittently and tend to increase in frequency until patients often appear to be "sine-waving" (oscillating from chronic ill health to acute crisis).15 Such patients, if being treated in the hospital, will usually have two to three or more admissions for such exacerbations in the year prior to death. Compared to patients with cancer, it is generally much more difficult to predict death within a matter of weeks to months.16-18

Clinicians often will not identify patients with serious, life-limiting illnesses as having terminal illnesses and thus may delay necessary discussions. If asked if such patients are dying, many clinicians will say "no." However, if asked, "Would you be surprised if such a patient were to die in the next two years," clinicians will tend to say they would not be surprised. This now famous question, "Would you be surprised if the patient died in the next two years," was brought to the attention of the palliative care community by Joanne Lynn in her publications.19 The first person to use this question as a means of identifying patients with serious, life-threatening illnesses was Dr. Marilyn Patterson, a colleague of Dr. Lynn (personal communication). While such patients may or may not qualify for hospice, they have special care needs, and their dying trajectory has important implications for decision making.

This trajectory is extremely common. Patients are at risk of bouncing back and forth from home or nursing home to the hospital, on and off antibiotics, into and out of crisis. House staff in a given hospital often know these patients by name. They are the "frequent fliers." They are particularly troublesome to physicians and the health care system because they do not behave as they should. Once "fixed," they do not stay fixed. Only with tongue in cheek can the discharge summary conclude, "The patient was stabilized and discharged to home." Patients and families entering this trajectory frequently live miserably on a roller coaster of decline and transient improvement.

Palliative Care Note
If you would not be surprised if a patient were to die in the next two years, that patient has a serious, life-limiting illness and would likely benefit from serious discussion and planning relative to end-of-life care.

This trajectory often can be recognized only after several such episodes of decline. Even so, there is no guarantee that the next dip on the ride will be the last one. Patients, families, and physicians, while despising the seemingly endless cycle of decline and improvement, often act as if they are addicted to it. Even knowing that one more transfusion will not make a difference or that one more round of antibiotics or one more trip to the intensive care unit (ICU) will not help, many have trouble breaking the cycle.

People have trouble accepting the inherent uncertainty in this trajectory. They want to know for sure when death will come - that is, be sure that one more round of therapy will not work. Patients may ask me, "But Doc, the antibiotics got me through the last pneumonia, how do you know they won't work the next time?" We do not, and that is the problem. To deal with this dying trajectory, patients, families, and clinicians must accept prognostic uncertainty and incorporate such uncertainty into their decision making. The question for such patients may be less about how long they will live and more about how and where they wish to live until they do die at some point in an uncertain future.

This trajectory also highlights a problem that can also occur on the cancer trajectory and subsequent trajectory. Will care be defined by what will be done or what will not be done? If clinicians approach patients and families and recommend that no antibiotics be given, blood no longer be transfused, no more admissions to the ICU be authorized, no more IVs or tube feeding be administered, addressing only what will not be done . . . , is it any wonder that they recoil? Here I offer a clinical pearl, nobody is going to love you for what you do not do. If only treatment limitation or withdrawal is discussed, patients and families will understandably feel abandoned. Reasonable and caring alternatives must be offered. The clinician must be able to say what will be done. A weak statement such as "We'll try to keep you comfortable," is not good enough. Unfortunately, many clinicians are ill-prepared to discuss alternatives.

Palliative Care Note
Nobody is going to love you for what you do not do.

Patients on this trajectory face complex choices: where to live (home, assisted living, nursing home, palliative care unit), under what care arrangements (family support, home care, hospice care, etc.), and with what goals of care (comfort only, intermediate life-prolonging efforts, or maximal life-prolonging -efforts). The complexity of such choices makes it tempting for patients, families, and physicians to ignore such choices and to hope that the problem will solve itself. It will eventually, but often at too high a price. One of the most difficult tasks in palliative care, requiring great communication skills in addition to a solid knowledge base about reasonable options, is helping patients and families sort through the complex maze of choices patients encounter on this trajectory.

Deaths Following Aggressive Life-Sustaining Treatment in Acute Care

This dying trajectory is common to many illnesses; patients may have chronic illnesses with sine-waving (see above) or may suffer acute, catastrophic events, such as stroke, overwhelming sepsis, or adverse outcomes of surgery.20 What characterizes this trajectory is that despite aggressive treatment, the patient does not improve and continues to decline. Many patients will die while such aggressive care continues. The probability of imminent death may not be recognized by patients, families, or health care workers. For others a point comes at which there is recognition that further aggressive, life-sustaining care is very unlikely to be successful, and death can be anticipated in minutes to days.

When impending death has not been recognized, the actuality of death often comes as a shock to families and health care workers. These patients do not die, they "code." The emotional impact of such deaths can be great. While families and physicians may experience some satisfaction in that "everything was tried," second thoughts and guilt also commonly arise. "Was there some mistake made?" "What else could have been done?" "Should we have seen this coming?" Patients who die in this way are usually heavily enmeshed in complex, high-tech medicine, and early grief reactions of patients, families, and clinicians may focus on the appropriateness or inappropriateness of specific medical therapies. In bereavement, grief may dwell on missed opportunities, especially if, in retrospect, it should have been clear that the patient was dying.

When death is predicted, a different challenge arises. In most such cases the patient's death will occur in the not-distant future, regardless of specific care decisions. Nevertheless, a myriad of care decisions present themselves to clinicians and families (and occasionally patients), usually relating to treatment withdrawal. Should ICU care, pressors, intubation, antibiotics, dialysis, and so on be withdrawn? Secondarily, parties may discuss where the patient should go following such withdrawal (to the floor, nursing home, home, hospice). In such situations families and clinicians frequently feel an overwhelming sense of responsibility. It is no small thing to discontinue life support and see a patient die soon thereafter. These difficult decisions beset families at a time when they are heavily grieving the impending loss of a loved one. Special communication skills are needed to help families at such times. (See Chapter 8 on Communication) At times families may have trouble experiencing this grief directly, as their -energy is consumed with one medical decision after another. Grief becomes walled-off and delayed, risking more intense bereavement later.

Patients on this trajectory often move abruptly from aggressive, life--prolonging efforts to care focused on comfort, a process often called treatment withdrawal. I hate the term. It suggests that care goes from something to nothing. I hate it more when this is, in fact, the case. Statements such as "There is nothing more we can do" reflect such a mindset. I prefer to think of it as a shift or transition in the focus of care. It may be perfectly reasonable to go full-bore trying to save a person's life and equally reasonable to shift that intensity and energy into -supporting the patient and the family when it is recognized that dying is -happening.21

Even when excellent palliative care is available, patients and families on this trajectory often do experience a form of withdrawal. "Detoxification" is required. They have grown accustomed to being cared for through high-tech medicine. Strange as it may seem, having one's blood pressure or O2 saturation measured may be a way of being nurtured. Suddenly, not having tests or vital signs taken can be experienced as abandonment, even when there is no rational reason for continuing such procedures. Thus, careful explanation is needed for why certain things are not being done and why others may be required, even when one might think the reasoning is obvious. New forms of nurturing must be identified.

Finally, a peculiar boomerang effect may be experienced with this trajectory. Having pushed very hard to prolong life and having heard that the patient will soon die, there may arise an expectation or even a desire that death happen very soon after a shift in the focus of care. Some deaths do soon follow, but not always. The reasons for a desire for a quick death are complex. First, having come to recognize dying, there is commonly an understandable desire not to draw out the process. There may also be a less conscious desire to confirm that a decision to withdraw treatment was, in fact, proper, and a quick death can validate such a decision. Physicians may encourage such thinking by stressing that the patient will certainly die soon without life-sustaining care and may mistakenly suggest that death will inevitably follow discontinuation in a very tidy manner, when in fact this does not always happen. Family members who have been following a patient during an acute illness are usually already exhausted when such a decision is made and out of their own suffering and personal time constraints may wish, consciously or unconsciously, for death to come sooner rather than later. Such feelings may prompt intense guilt in the bereaved. Finally, some patients and families deal with their living and dying like an on-off light switch. If they cannot have it all, they want it over as soon as possible. Physicians who do not understand this may be surprised when a patient or family that yesterday was asking for maximally aggressive care is today asking for a hastened death.

<<< Previous Next >>> [ Go Up ]

book cover

Palliative Care Perspectives

James L. Hallenbeck, M.D.

Order hardback from
Order paperback from

Copyright 2003 by Oxford University Press, Inc.

The online version of this book is used with permission of the publisher and author on web sites affiliated with the Inter-Institutional Collaborating Network on End-of-life Care (IICN), sponsored by Growth House, Inc.