Communication about cancer near the end of life



Cancer communication near the end of life has a growing evidence base, and requires clinicians to draw on a distinct set of communication skills. Patients with advanced and incurable cancers are dealing with the emotional impact of a life-limiting illness, treatment decisions that are complex and frequently involve consideration of clinical trials, and the challenges of sustaining hope while also having realistic goals. In this review, the authors sought to provide a guide to important evidence about communication for patients with advanced cancer regarding communication at diagnosis, discussing prognosis, decision making about palliative anticancer therapy and phase 1 trials, advance care planning, transitions in focus from anticancer to palliative care, and preparing patients and families for dying and death. Cancer 2008;113(7 suppl):1897–910. © 2008 American Cancer Society.

Cancer communication near the end of life has a growing evidence base, and requires clinicians to draw on a distinct set of communication skills. The recent National Cancer Institute monograph on Patient-Centered Communication in Cancer Care identified 6 functions of patient-clinician communication: exchanging information, making decisions, fostering healing relationships, enabling patient self-management, managing uncertainty, and responding to emotions.1 Whereas these communication functions shape communication for all patients with cancer, special skills are required for the subset of patients with advanced and incurable cancer who must face death more squarely.2

These patients, with advanced and incurable cancers, are dealing with the emotional impact of a life-limiting illness, treatment decisions that are complex and frequently involve consideration of clinical trials, and the challenges of sustaining hope while also having realistic goals. The clinicians must establish a therapeutic relationship based on trust and mutual respect with patients who often access a great deal of medical information, come from culturally diverse backgrounds, have various levels of social support, and confront the existential and spiritual aspects of dying, all while trying to access complicated healthcare systems.3, 4 Because new therapeutic technologies enable these patients to live longer with cancer than ever before, the communication challenges faced by oncology clinicians have become ever more complex, involving uncertainty, hope, and widely available anecdotes of patients who “beat the odds” available on the Internet and in books.

In this review, we sought to provide a guide to important evidence and best practices about communication for patients with advanced cancer who face life-limiting illness. Because the state of the science has produced few controlled trials, this narrative review also includes relevant observational, survey, and qualitative studies. We have framed the data from the point of view of oncology clinicians who often have relationships with patients that span the trajectory of the illness. Thus we have divided the review into sections relevant to practicing clinicians: core communication skills; communication at diagnosis, discussing prognosis, decision making about palliative anticancer therapy and phase 1 trials, advance care planning, transitions to palliative care, and preparing patients and families for dying and death.

Core Communication Skills

Patient concerns have both informational and emotional components, so patients need both information and emotional support to resolve their concerns. They look to their oncologist for both.5‒7 When these concerns are not resolved, they lead to psychological distress and affective disorders.8 Introductions help set the stage for a therapeutic alliance.9 Core communication skills enable clinicians to elicit concerns and consequently provide more effective and compassionate care (Table 1).

Table 1. Core Communication Skills
Recommended SkillExample
I. Identifying Concerns and Recognizing Cues 
 Elicit concerns 
  Open-ended questions“Is there anything you wanted to talk to me about today?”
  Active listeningAllowing patient to speak without interruption; allowing pauses to encourage patient to speak
 Recognize cues 
  Informational concernsPatient: “I'm not sure about the treatment options”
  Emotional concernsPatient: “I'm worried about that”
II. Responding to Informational Concerns
 “Ask-Tell-Ask”Topic: communicating information about cancer stage
  Ask“Have any of the other doctors talked about what stage this cancer is?”
  Tell“That's right, this is a Stage IV cancer, which is also called metastatic cancer…”
  Ask“Do you have questions about the staging?”
III. Responding to Emotional Concerns
 Nonverbal empathy: S-O-L-E-R 
  SFace the patient SQUARELY
  OAdopt an OPEN body posture
  LLean towards the patient
  EUse EYE contact
  RMaintain a RELAXED body posture
 Verbal empathy: N-U-R-S-E 
  NNAME the emotion: “You seem worried”
  UUNDERSTAND the emotion: “I see why you are concerned about this”
  RRESPECT the emotion: “You have shown a lot of strength”
  SSUPPORT the patient: “I want you to know that I will still be your doctor whether you have chemotherapy or not”
  EEXPLORE the emotion: “Tell me more about what is worrying you”

Identifying concerns: eliciting concerns and recognizing cues

For an oncologist to address a patient's information and emotional concerns, the concerns must first be disclosed to and recognized by the oncologist. This occurs in 2 ways: the provider eliciting the concerns from the patient, or the patient spontaneously giving cues about their concerns. Although patients want to talk about their concerns with their oncologist, whether they disclose their concerns depends on how their oncologists communicates with them.10, 11 Certain provider behaviors have been shown to elicit these concerns: open-ended questions, allowing more time for patients to speak, and empathy (Table 1).12‒15 Open questions such as “How are you doing with all of this?” or “Is there something else you'd like to talk about today?”16 let patients know that their physician will be willing to listen. Patients also give “cues” to providers about their concerns, for example, “I'm not sure what the treatment options are” (cue for information) or “That's 1 thing I have as a big fear” (cue for emotional exploration and support).17, 18 Physicians tend to focus communication on biomedical topics, and often do not elicit or recognize cues about emotional and psychosocial concerns.19‒21 Although other factors influence whether patients disclose,10 oncologist inquiry is an important first step.

Responding to informational concerns: Ask-Tell-Ask

Patients need information about their illness and how it will affect them.22 They use this information to plan for the future and to make medical and life decisions.23, 24 They want their physician to be honest and realistic in giving information while being sensitive to what information they are ready to hear and how it is affecting them.25, 26 They want to negotiate when and how information is given, but don't want a greater amount of detail than they are ready to hear.26 They want their oncologist to be open to discussing sensitive topics such as dying, but don't want to be forced to talk about it before they are ready.

The Ask-Tell-Ask technique ensures that the physician: 1) gives information slowly enough to ensure that the patient understands, and 2) remains sensitive to the effect the information is having on the patient and does not give too much information.27 This is accomplished by the oncologist bracketing each piece of information she gives with questions to check understanding and the impact on the patient (Table 1).

Responding to Emotional Concerns: Empathy

The stress of cancer and its treatment are often associated with intense negative emotions: sadness, fear, and anger. Although oncologists often cannot “fix” the causes of these emotions, empirical studies indicate that providing emotional support ameliorates distress. Patients are emotionally supported when their physician shows care for them as a person, by spending enough time with them, allowing them to ask questions, and listening to their concerns.25 Providers can show emotional support by listening and using specific language that expresses empathy, which is simply acknowledging the presence of a patient's emotion without trying to fix or alter it.18 For example, cancer patients whose physician addressed emotions and presented a “caring attitude” at diagnosis were less anxious a year later.28 Breast cancer patients were more likely to believe the physician cared about them and were less anxious when the physician expressed empathy.29 Empathy can be expressed without significantly prolonging the time spent talking with patients; as little as 40 seconds of empathic language results in significant changes is patient's anxiety level.30, 31

Physicians can respond empathically to patients' expressions of emotion nonverbally and verbally. The acronym S-O-L-E-R described nonverbal expressions such as using one's body posture and facial expression to convey care about the patient and relaxation (Table 1).32 Two techniques are useful guides for expressing verbal empathy: the N-U-R-S-E acronym33 and “I wish” statements34 (Table 1).

Communication at Diagnosis: Giving Bad News

Most of the cancer communication literature on bad news focuses on diagnosis. Over the past 50 years, physician practices for disclosing the life-limiting nature of cancer has changed dramatically. In 1961, Oken showed that 88% of physicians followed a policy of not telling patients about their cancer.35 The same questionnaire was repeated in 1979 by Novack and colleagues, and in a complete reversal it was found that 98% of physicians reported telling patients their diagnosis.36 This cultural change outpaced training for oncologists in how to give bad news. In a survey of 500 oncologists at the 1998 American Society of Clinical Oncology meeting, only 4.8% of participants reported formal training in breaking bad news.37 Training programs uncommonly use evidence-based methods to build communication skills.38 Given this, it is not surprising that physicians find disclosing bad news difficult. In 1 study, 1 third of physicians reported experiencing significant stress when preparing for and delivering the bad news, and for 86% of these physicians, the stress persisted after the interaction with the patient—for 20%, more than a day.39

What patients prefer when hearing bad news

Retrospective studies have identified what patients want when hearing bad news. They prefer to have bad news discussed in person, and in a private, quiet place; a physician who is able to communicate their diagnosis, prognosis, and treatment options clearly; full attention of the physician; time to ask questions; and to be given information about how the diagnosis will affect their life.29, 40, 41 Patients vary in their wish to have a family member or friend present, with some preferring to have the discussion alone. One study identified 2 factors that decreased patient satisfaction: “The physician rushed through the news,” and “The physician seemed to struggle to find the right words.”42 This same study correlated the following factors with greater patient satisfaction: a comfortable place, no interruptions, “physician sat close to me,” “let me know that bad news was forthcoming,” “took my personality and emotionality into account,” and “tried to empathize with what I was feeling.” Although treatment failure clearly represents another type of bad news, no studies have been performed to identify how this kind of bad news should be handled differently. However, patient preferences about disclosure vary considerably by cultural background, as well as level of assimilation.43, 44 For example, in a survey conducted in Los Angeles, 87% of European Americans felt that a patient with metastatic cancer should be told their diagnosis; only 35% of Korean Americans agreed.45 Worldwide, variation in disclosure of bad news is substantial.46, 47

Current practice in giving bad news

Physician education in bad news protocols has been widespread.48 Although physicians generally self-assess their communication about disclosure favorably, studies of patients show gaps in what physicians think they convey and what patients actually perceive. In a Canadian study, 98% of physicians thought they had accurately described the extent of disease to their patients, but almost 1 third of patients with metastases thought their disease was localized. Similarly, although 90% of physicians reported telling the patient the intent of therapy, almost 1 third of patients being treated palliatively thought that their therapy was curative.49

Recommendations for giving bad news

Evidence-based and best practice recommendations are summarized in Table 2. Although protocols for giving bad news have been designed and successfully taught in educational interventions,37, 50 there are no studies that evaluate the importance of individual steps of a protocol, compare the use of different bad news protocols, or link them to patient outcomes such as distress or anxiety after the bad news visit. In addition, despite attention to different cultural preferences in information sharing, there have been no studies to date that have examined disclosure approaches designed for patient populations with variable cultural preferences. Finally, many bad news conversations involve patients and family members, and there are no studies that have identified communication strategies for multiple participants in a bad news conversation.51

Table 2. Communication About Diagnosis: Giving Bad News
Find a comfortable and private place to talk
Ask whether the patient would like to have others present
Minimize interruptions
Assess the patient's understanding of the situation
Let the patient know explicitly that bad news is forthcoming
Provide information honestly and in simple language
Give time for questions
Encourage patient to express emotions and respond empathically
Check understanding
Arrange a clear follow-up plan

Discussing Prognostic Information

Discussing prognosis is challenging because oncologists face conflicting directives. The medical literature contains both studies that suggest that patients want and need a great deal of information disclosure, and other studies indicating that patients want to limit their discussions about poor prognoses. This mismatch has clinical consequences: a large study of patients indicated that patients with unrealistically optimistic views of their prognosis were more likely to die after intubation or cardiopulmonary resuscitation (CPR).52

Patient preferences in prognostic disclosure

In surveys, the majority of patients tell researchers that, in general, they want a great deal of prognostic information. A national survey revealed that 85% of Americans want a “realistic estimate” of how long they have to live if their cancer “leads to death in less than a year.”53 In a British survey of 2331 cancer patients, 87% wanted “all possible information.”54

However, studies also show that a significant minority of patients want their oncologist to limit discussion of prognosis, especially when survival is limited. In an Australian study, 40% of patients wanted to negotiate when to discuss prognostic information.22 In the British survey, 5% of patients did not want to hear bad news.54 In focus groups, patients and family members said that they did not want the oncologist to be “too blunt.”25

What patients want in prognostic discussions is influenced by their views on hope, cultural background, and beliefs about discussing or acknowledging how a poor prognosis impacts physiological processes. Sustaining and maintaining hope is a well-described function of the therapeutic relationship in oncology.55 Patients articulate their interest in maintaining hope and seem to be more flexible in their formulation of hope than oncologists, who interpret hope in terms of a biological cure. However, patients and families also derive hope from interactions with their oncologist even when prognostic information is unfavorable.56 Certain cultural beliefs lead patients to believe that an explicit discussion of a poor prognosis becomes a prediction that comes true.57 However, even when an explicit discussion is not approved, patients and families may acknowledge an impending death indirectly.58 A randomized study showed that providing patients with a prompt list of questions led to more discussion with oncologists about prognostic information.59

Physician practice in prognostic disclosure

Physicians describe facing a quandary in talking about prognosis with patients. Physicians are aware of their responsibility to talk about prognosis and the important role it plays in advance care planning and realistic goal setting.60 Yet physicians also believe they are “giving a death sentence,”61 are aware that a subset of patients do not want to discuss prognosis explicitly, and are aware that their prognostic predictions often prove to be inaccurate. A descriptive study asking oncologists to describe their emotions when describing these conversations includes: “unhappy,” “unpleasant,” “frustrating,” “bothersome,” “difficult,” “hard,” “exhausting,” and “draining.”62

Given these conflicting physician motivations, it is no surprise that physician behavior in disclosing prognosis is quite variable. In a study of cancer patients referred to hospice, 37% of physicians said that they would disclose a prognosis they thought was accurate, 40% would disclose a prognosis they thought was partially accurate (usually optimistic), and 23% would refrain from disclosing prognosis even if asked directly by the patient.63

Physicians and patients influence each other in how much explicit discussion of prognosis occurs. Some behaviors in limiting talk about prognosis involve tacit agreements by the physician and patient, which has been well documented. For example, in a qualitative study of patients with nonsmall cell lung cancer in the Netherlands, the phenomenon of collusion was identified in which physicians did not offer prognostic information and patients did not explicitly seek it—a tacit agreement to avoid an uncomfortable topic.64

Only limited studies examine the impact of specific communication practices on patient comprehension in the setting of advanced cancer. Interestingly, it seems to be the absence of pessimistic statements, not the presence of optimistic statements, that most significantly influences patients' perception of prognosis. In 1 study, advanced cancer patients were more likely to agree with their oncologists' estimate of prognosis if the oncologists gave a pessimistic statement during a consultation, whereas optimistic statements did not affect agreement.65

Recommendations for discussing prognosis

Existing evidence suggests that patients should be offered a discussion of prognosis and may be upset or disappointed on hearing the information, and that physicians should frame the information both positively and negatively. Although evidence indicates that these discussions can be difficult for physicians, there is little empirical data to indicate how physicians can prepare themselves or how they can manage the personal impact of these conversations. Similarly, little evidence exists to guide physicians in discussing prognosis with patients who do not wish to have explicit discussions, including those with cultural beliefs that discussion could be harmful. A “how much do you want to know” communication strategy has been proposed, although it has not been empirically tested (Table 3).66, 67

Table 3. Discussing Prognostic Information
Ask how much prognostic information the patient wants
For patients who want explicit information, ask what kind of information (eg survival, response to treatment, ability to attend a future event)
Frame statistical information positively and also negatively (eg, by 2 years, 50% will have died, 50% will still be alive)
Offer to describe survival in a range (eg, “months to a year”) in addition to a specific period (“median survival of 1 year”)
Consider a separate conversation with a family member who wants more information (as long as the patient gives permission)
Respond to patient emotions, but do not assume that intense emotion means you should not have disclosed the prognosis
For patients who do not want explicit information, try to understand their perspective

Communicating Evidence for Decision Making About Palliative Anticancer Therapy and Phase 1 Clinical Trials

Slightly different from the challenges of talking about prognosis are those challenges that oncologists face when talking about anticancer treatments with patients whose cancer has progressed despite evidence-based therapy.

Patient perceptions of palliative chemotherapy and phase 1 studies

Cancer patients have different perspectives on the benefits and burdens of chemotherapy than those without cancer. They are more willing to undergo treatments with small benefits and major toxicity, and little patient-level information is available.68 In addition, several studies suggest that a significant subset of patients receiving palliative chemotherapy hold misconceptions about the intent of the therapy. Specifically, these patients indicate in surveys that they believe that the therapy is intended to cure them. For example, in a study of 149 patients with incurable cancer, 45 (31%) believed their cancer was incurable, 61 (42%) were uncertain, and 39 (27%) believed their cancer was curable.69 The limitation of such surveys is that because the actual conversation was not analyzed, it is unclear what the physician said or whether the physician allowed a misconception to persist. Thus, although these studies suggest that communication could be improved, the reasons it failed are unclear.

Studies of patients participating in phase 1 studies also suggest that oncologist-patient communication could be improved. Although the overlying long-term objective of phase 1 experimentation is to improve care, the immediate goal of the specific study is to determine the toxicity profile and, often, the maximum tolerated dose. Historically, the benefits of phase 1 participation are low, and only 5% of these participants experience clinical benefit with a median survival of 6.5 months.70 Existing studies suggest that patients have an incomplete understanding of this reality. In a study of 328 patients surveyed after informed consent for a phase 1 trial, the median benefit expected was rated at 60%.71 In another study of 27 patients, only 33% could state the purpose of the study.72

Current practice in communication about palliative chemotherapy and phase 1

Existing evidence, albeit scant, indicates that oncologists spend more time talking about the option of active anticancer therapy than the option of focusing on supportive care. In a European study of audiotaped visits, medical oncologists talked to patients with metastatic cancers in all cases, but mentioned focusing on supportive care (called “watchful waiting” in this study) in only 50% of visits (and in nearly half of these visits, oncologists said only 1 sentence about this).73 In an Australian study of audiotaped consultations, most patients were told that the aim of anticancer treatment was palliative (84.7%), and that their cancer was incurable (74.6%). In this Australian study, alternatives to anticancer treatments was presented to 44.1% of patients, 36.4% were informed about how anticancer treatment would affect quality of life, and 29.7% were offered a management choice. Oncologists checked patient understanding in only 10.2% of consultations.23 No comparable audiotape studies exist that analyze what physicians say when they talk about phase 1 studies with patients. Thus, although existing studies indicate that patients consenting to participate in phase 1 trials have a limited understanding of trial purpose, an unrealistic expectation of the benefits and risks associated with trial participation, and a questionable appreciation of their right to abstain or withdraw,74 it is unclear what physicians have actually told them and how physician communication practices impact patient understanding in this specific area.

Recommendations for communication about palliative chemotherapy and phase 1 trials

Existing studies indicate that patients considering palliative chemotherapy and phase 1 trials often have an incomplete understanding in a variety of areas important to informed consent, and limited data from direct observation of physicians talking with patients suggest that physicians may not be discussing difficult issues explicitly enough for patients to completely understand. Thus we recommend that physicians use the core communication skills mentioned earlier to establish a clear idea of patient understanding of their situation, and offer to discuss key aspects including treatment options, anticipated impact on survival, and anticipated impact on quality of life (Table 4). For patients considering enrollment in a clinical trial, physicians need to prepare patients to engage in an explicit discussion of the purpose of the trial, the risks and benefits, and the patient's ability to withdraw from the trial74—and because of the requirement of informed consent, it does not seem reasonable to allow patients to opt out of these discussions.

Table 4. Communicating Evidence for Decision Making About Anticancer Therapy
Offer to discuss treatment options, impact on survival and quality of life with patients considering palliative chemotherapy
When discussing treatment options, explicitly mention the option of supportive care without anticancer therapy
For patients considering clinical trials, explicitly discuss the purpose of the trial, risks and benefits, and the patient's ability to withdraw at any time
After your discussion, check patient understanding

Advance Care Planning

Advance care planning was developed by bioethicists responding to high-profile cases of patients who wanted to discontinue life-sustaining treatments.75 Most advance care documents address 2 broad preferences: first, a surrogate decision maker who can represent the patient's wishes if the patient loses decision-making capacity (a durable power of attorney for healthcare); and second, patient preferences about future use of specific life-sustaining therapies, such as CPR.

Patient preferences about advance care planning

Although most patients want to be involved in decisions about care at the end of their lives, and a majority of patients want physicians to initiate the discussion with them,76 a minority of cancer patients have advance directives. Cancer patients may be reluctant to initiate discussion with oncologists because they are worried that the oncologist will “give up.” In 1 study only 9% of cancer patients had talked to their oncologist about advance care plans, and only 23% of the remaining patients wished to do so—although 69% had discussed advance care plans with someone.77 As illness progresses and functional status declines, some patients are more willing to accept life-sustaining treatments that would result in a diminished state of health.78 Patients who use spiritual coping (eg, belief in divine intervention) are somewhat less likely to engage in advance care planning.79 Also, patients who do not want to engage in advance care planning may be most likely to need it. In a study of 343 patients undergoing stem cell transplantation, 171 had some form of advance care planning, and failure to engage in advance care planning was associated with a significantly greater risk of death.80

Current practice in advance care planning

Studies of living wills suggest that they have a limited effect on outcomes for patients who ultimately die of cancer. Empirical studies done examining advance directive communication highlight several deficiencies; despite data indicating that patients want their physicians to raise the issue of advance directives, physicians rarely discuss the topic with even seriously ill patients.81 When they do talk about the topic, they tend to focus narrowly on technical decisions, fail to elicit patient values, and do not clearly discuss the likelihood that the life-sustaining intervention will meaningfully prolong life. In an audiotaped study of discussions about CPR in an outpatient setting that compared experts to nonexperts, it is clear that experts communicated differently. Although this study involved primary care physicians (18 experts and 56 nonexperts), it is the only study examining expert practice around CPR discussions. The experts spent longer conducting the discussion, were less verbally dominant, and spent more time on psychosocial issues and positive talk. In short, experts are more likely to discuss patient values and personal goals related to end-of-life care and to embed the CPR discussion in those discussions, rather than leading a discussion about dying with CPR.82

Recommendations for advance care planning discussions

In general, discussion about advance care planning should focus more on goals of care than on specific treatments, and clinicians should be especially careful to respond to the emotional content of the discussion (Table 5).75 The underlying principle is that the discussion should move back and forth from preferences to reasons and values to information and back, ensuring that the patient understands the implications of his or her stated preferences and that the physician understands the patients' values. Although little evidence exists to guide practice, it may be more effective for a physician to make a values-based recommendation, rather than offering a variety of choices without guidance.

Table 5. Advance Care Planning
  1. POLST indicates Physician Orders for Life Sustaining Treatment.

Elicit patient values and goals relevant to medical care
Explain why advance care planning is worth discussing
Ask about patient perceptions of life-sustaining treatments such as cardiopulmonary resuscitation and correct misinformation
Make recommendations about life-sustaining treatments based on the patient's goals and values in the context of the clinical situation
Ask the patient about their reaction to the recommendations
Note that any advance care plans can be changed in the future at the patient's discretion
Document the discussion (in chart, POLST, etc)

Discussing Transitions in Focus From Anticancer to Palliative Care

For the purposes of this article, we define these transitions as the point when it becomes clear that disease-modifying anticancer therapy is no longer effective. These can be difficult conversations, because they provoke intense emotions from patients and families, and a sense of failure from oncologists.

Patient preferences for discussing transitions

Most patients and families want to discuss the transition to palliative care with their oncologists and report frustration when the conversations are not timely83; the majority want to discuss dying and palliative care with their oncologist.22 Studies also document dissatisfaction with the transition to palliative care, and cite physician's lack of knowledge and communication skills and reluctance to make referrals as barriers to effective transitions to palliative care.27, 84‒86 Other studies demonstrate that the importance of the relationship with the oncologist increases as the patient's illness progresses.87 At the initial consult, patients want primarily cognitive information about their illness, but as early as the first follow-up visit, their preference shifts toward wanting support and reassurance, and discussing their fears and worries.88 Thus transition discussions may span more than 1 visit. The literature identifies 3 aspects of a good discussion about the transition from curative to palliative care: clarity of information, grieving loss, and sustaining hope.

Clarity of information

Transitions often begin with a bad news conversation. In transition discussions, patients and their families want their oncologist to be sensitive to how information is affecting them, do not want a greater amount of detail than they are ready to hear, and want to be able to ask questions.22, 26, 89, 90 Patients often ask about the availability and efficacy of further treatments, including alternative and complementary therapies, and prognosis when they are told that their cancer cannot be cured. Given that a subset of patients have misconceptions about the intent of palliative anticancer therapy as noted earlier, the oncologist will likely need to communicate that the cancer cannot be cured even if this has been done previously. Previously discussed recommendations for giving bad news are useful, particularly those for responding to emotion.

Grieving Loss

Qualitative studies about transitions identify loss as a central issue; when faced with a life-limiting illness, patients face the loss of family, hopes for the future, and life as they knew it.26, 89, 91‒93 In survey studies, the concept of loss is captured in the category of psychosocial issues. Patients' fears and worries often focus on psychosocial and emotional issues, and in 1 study, 80% of patients wanted to discuss their family and social life with their physician.11 Yet psychosocial issues were discussed 2.5 times less frequently than biomedical issues in 1 study of oncology visits.14 A focus group study indicated that patients at transitions want their oncologist to be open to talking about sensitive topics such as about dying.25 However, patients often may not raise these issues unless invited to by their oncologist.13, 94

Because patients spontaneously express cues about their concerns (eg, “I'm worried about how this will affect my family”), oncologist responsiveness to cues using empathic communication skills is critical to building a therapeutic relationship. Experts suggest that attention to emotional cues may enable patients to more clearly understand cognitive issues such as incurability, and make clearer decisions. However, physicians often do not express empathy in response to expressions of emotion, and this blocks further discussion of the concern, possibly “training” patients not to bring these issues up in the future.65, 90 The oncologist's commitment to face the future with the patient, sometimes called nonabandonment,95 is especially important to express in transition discussions. Patients want to know that their oncologist will still be their physician even if they do not continue chemotherapy.88, 92

Physicians, as well as patients, often experience loss when a patient's disease progresses.96 These feelings of sadness, loss, and impotence may be part of the reason that oncologists report that issues related to death and dying are so stressful for them. In addition, physicians' distress over their perceived impotence may lead them to redouble chemotherapeutic efforts. It is important for the clinician to recognize and deal with one's emotional reaction to disease progression before seeing the patient.97

Supporting hope

Maintaining hope is essential to patients and their families at the end of life.26, 84 Although many clinicians equate honesty about a poor prognosis with destroying hope, healthy coping continually generates hope, even in difficult life situations. Even under situations of severe stress, positive emotions are prominent, and are an integral part of the coping process.98 Oncologists can help patients to redefine hope by exploring and reinforcing patients' hopes and talking about what can be done.26

Patients and families often remain hopeful, even after they have been given a life-threatening diagnosis.56, 99 Although many continue to hope for a miracle, or to be cured, studies show that patients also have other concurrent hopes. In an Australian study, cancer patients and healthcare providers described a spectrum of hopes: being cured, living longer, finding meaning in life, having special time with loved ones, finding spiritual meaning, and having a peaceful death.26 An American study described patients' hopes for a good death: freedom from pain and other symptoms, clear decision making, preparation for death, having a sense of completion, contributing to others, affirmation of the whole person, being at peace with God, being in the presence of family, being kept clean, and trusting one's physician.100

Patients want their oncologists to explore new hopes with them, to help them set realistic goals for the future, and to discuss the practicalities of day-to-day living.26 Within that relationship, patients say their oncologists can nurture hope by emphasizing what can be done: pain and symptom control, emotional and practical support, having dignity, and being in a caring environment.26 Common pitfalls when discussing the transition from curative to palliative care are focusing on hope for cure to the exclusion of other hopes and viewing anticancer therapies as the only thing that an oncologist can offer a patient. One empirical study focusing on the transition to hospice care, involving a survey of bereaved family members of cancer patients who had died in an inpatient hospice unit, indicated that a substantial proportion felt that communication needed to be improved and that family distress correlated with the family being told that “nothing more could be done.”101

Recommendations for transition discussions

Although existing literature on transition discussions is largely descriptive, these studies suggest that improving communication about palliative care for cancer patients should involve a discussion of patient values, and a clear statement about nonabandonment with specific care plans (rather than a statement that “nothing more can be done”)(Table 6).102 Knowledge of the elements of a good death may help guide a productive conversation. An expert approach described as “Hope for the Best, Prepare for the Worst” is a useful framework for discussing planning for the future while still embracing a patient's hope to be cured.102

Table 6. Discussing Transitions From Curative to Palliative Care
Discuss anticancer treatment failure in a timely manner
Use bad news recommendations to establish shared understanding of the clinical situation
Be prepared to acknowledge and grieve losses
Explore hopes other than cure, support realistic hopes
Allow patients to hold private hopes and avoid attacking these hopes
Emphasize what can be done for the patient and family
Reaffirm commitment to the patient and family

Preparing for Dying and Death

Many cancer clinicians observe their patients through the trajectory of illness, yet preparing patients for dying and death requires a shift in the oncologist's mindset. Discussions about prognosis, anticancer treatment choices, and treatment side effects, in general, give way to discussions about the physical, psychosocial, spiritual, and practical aspects of preparing for death. The discussions, however, are often asymmetric; the literature describes patients and caregivers' desire to talk about psychosocial concerns at the end of life, and physicians' struggle to talk about CPR.

Patient and caregiver views on end-of-life discussions

In descriptive studies, the majority of patients and caregivers report wanting information about the disease process, likely future symptoms and how these symptoms could be managed, and life expectancy.25, 103 Patients want to talk about these issues with a health professional with whom they have an established level of trust, and who have shown empathy, honesty, and sensitivity to different levels of understanding.104 Patients are also concerned about pain control, nonabandonment, and information about hospices.105 Interestingly, patient and caregiver information needs showed a tendency to diverge as the illness progressed, with caregivers needing more and patients wanting less information.104 In 1 study, patients and caregivers agreed that the following communication practices were desirable: 1) consistency among different health professionals and openness to questions and discussion, 2) provision of specific information needed to care for the patient, and 3) separate discussions with patient and caregiver.106 Qualitative studies have identified the following content issues to be discussed: good pain and symptom control, avoiding inappropriate prolongation of dying, strengthening important relationships, achieving a sense of control, and finding a sense of closure.107‒109

Current practice in end-of-life discussions

Interestingly, the empirical literature describing end-of-life discussions from the clinician' viewpoint does not mirror the patient studies described above and is largely centered around discussions of CPR. In a large study of seriously ill hospitalized patients that included over 800 patients with cancer, physicians misunderstood patients' wishes about CPR nearly half the time, suggesting that communication about CPR had been inadequate.110 Only 23% of patients reported discussing CPR with their physician, and importantly, of those who had not discussed CPR, 58% were uninterested in doing so; this group did not consider CPR to be a pressing concern.111 In a more recent Canadian study of hospitalized patients with advanced disease including cancer, only 3% reported discussing CPR with their physician, and patients who felt that end-of-life issues were relevant to them were 5.5 times more likely to want a discussion with their physician.112 These studies suggest the difficulty of raising CPR for patients who are not prepared to talk about dying, and raise a question about whether CPR discussions might be more successful if they occur after better discussions of prognosis and goals of care.

The only observational study of CPR discussions that studied actual conversations involves medical house staff and hospitalized patients (including cancer and noncancer diagnoses). These conversations demonstrated that physicians often provided incomplete information about what CPR involved, how often it was successful, and likely patient outcomes. For example, only 13% of physicians mentioned the patient's likelihood of survival after CPR, and no physician used a numerical estimate.113 Given that other studies indicate that the likelihood of a patient with metastatic cancer surviving CPR is between 0% and 20%,114 this study raises troubling questions about what physicians do or do not communicate in these situations. House staff initiated discussions about values or goals of care in <10% of the conversations, yet 90% rated their own performance highly and 77% reported feeling comfortable with the discussion. One third of these house staff had never been observed by a more senior physician while conducting a CPR discussion, suggesting little effective training in communication skills.113

Discussions about hastening death

Although a considerable proportion of patients tend to avoid discussions about end-of-life care, a small proportion of patients present a different issue for clinicians—they want to hasten death, and try to initiate discussions about assisted suicide or euthanasia. A longitudinal qualitative study of patients pursuing hastening death found that for these patients, talking about assisted suicide was a gateway to talking about dying, and patients were frustrated at being rebuffed by physicians. Importantly, physicians who presented scenarios of future suffering without effective treatment may have inadvertently encouraged patients to seek hastened death.115

Preparing caregivers for the patient's death

Preparing caregivers for the patient's death is often not a focus of communication.116 In a national sample, 23% of informants contacted reported that the death of their loved one was “extremely” unexpected.117 Communication between caregivers and healthcare providers is likely a major predictor of preparedness.11, 118 Observational studies suggest that inadequate information and unpredictable situations contribute to caregiver uncertainty, and also that caregiver uncertainty is associated with poorer caregiver health outcomes.119, 120 Studies of caregivers demonstrate that when they perceive that their questions have been answered, they experience fewer depressive symptoms, fewer economic and other burdens, and improved satisfaction and quality of life.121 The components of communication that could contribute to death preparedness have been outlined but not tested in intervention studies.120, 122 However, a French study that randomized caregivers of dying patients in intensive care units (including cancer and noncancer diagnoses) to a communication intervention demonstrated lower rates of postdeath post-traumatic stress disorder.123

Finally, communication before death continues to have an impact on family caregivers after death. Studies indicate that preparedness impacts the caregiver's bereavement in physical health, depression, substance abuse, and even the caregiver's own death.121, 124, 125

Recommendations for preparing patients and family caregivers for dying and death

Communication designed to prepare patients and family caregivers should start with a broad framework of patient goals and values before addressing specific interventions such as CPR (Table 7). If consistent with patient goals, physicians may recommend withholding CPR as part of this discussion. Because caregivers have distinct informational and emotional needs, they need conversations separate from the patient. Although there are suggestions about how to talk about a clinician-patient relationship that is ending because of death, these have not been empirically tested.126

Table 7. Preparing for Dying and Death
For patient:
 Arrange discussion with a clinician who has an established relationship with the patient, if possible
 Ask about concerns related to disease progression or dying
 Offer to talk about life expectancy and last hours of life
 Be prepared to discuss components of a good death:
  Pain and symptom control
  Avoiding inappropriate prolongation of dying
  Strengthening important relationships
  Achieving a sense of control
  Minimizing burden on family
  Finding a sense of closure or peace
 If consistent with patient goals, physicians may recommend withholding cardiopulmonary resuscitation as part of this discussion
For caregiver:
 Provide specific instructions for caring for patient
 Consider a separate discussion for caregiver if patient gives permission
 Recognize work caregiver does in caring for patient
 Acknowledge and empathize with caregiver grief
 Prepare caregiver when death is imminent


Talking with cancer patients who are approaching death represents a significant clinical challenge. Empirical research on how patients and oncologists discuss these issues has defined a few best practices, but further research should continue to link identified needs to communication practices that influence patient outcomes such as quality of life, quality of dying, and decision making, and to training that will enable oncologists to use these practices effectively.