Intended for healthcare professionals

Spotlight Spotlight: Palliative Care Beyond Cancer

Having the difficult conversations about the end of life

BMJ 2010; 341 doi: (Published 16 September 2010) Cite this as: BMJ 2010;341:c4862
  1. Stephen Barclay, general practitioner and Macmillan Postdoctoral Research Fellow1,
  2. Jane Maher, consultant clinical oncologist and chief medical officer, Macmillan Cancer Support2
  1. 1General Practice and Primary Care Research Unit, Institute of Public Health, Cambridge CB2 2SR
  2. 2Mount Vernon Cancer Centre, London
  1. Correspondence to: S Barclay sigb2{at}
  • Accepted 15 May 2010

Clinicians need to create repeated opportunities for patients to talk about their future and end of life care, guided by the patient as to timing, pace, and content of such talks, and respecting the wishes of those who do not want to discuss such matters

More than half a million people die each year in Britain—36% from cardiovascular disease, 27% from cancer, and 14% from respiratory disease; and 58% of all deaths occur in hospital,1 a proportion that has increased in recent years. While some deaths are sudden and unpredictable, many patients go through a period of illness when death becomes increasingly probable.

Recent General Medical Council guidance on good practice in decision making in treatment and care towards the end of life states that “patients whose death from their current condition is a foreseeable possibility are likely to want the opportunity to decide what arrangements should be made to manage their final illness” but also cautions that “you must approach all such discussions sensitively, as some patients may not be ready to think about their future care or may find the prospect of doing so too distressing.”

Some may not wish to talk with their clinicians or their family about the end of life, but others may greatly benefit from such conversations. The right conversations with the right people at the right time can enable patients and their loved ones to make the best use of the time that is left and prepare for what lies ahead.

In this article, some of our comments arise from our experience as clinicians in general practice and oncology, and others from the research evidence in this area, which is limited. We seek to stimulate discussion and debate: we focus mainly on issues that make these conversations difficult for patients and clinicians, and invite readers to expand on our suggestions of practical ways forward.

The difficulty of knowing what lies ahead

Uncertainty about prognosis creates anxieties for doctors when discussing end of life care, with patients and their families often expecting greater prognostic certainty than is possible. Cancer patients have traditionally been viewed as having an identifiable dying trajectory,2 but health professionals’ estimates of their prognoses are frequently inaccurate and over-optimistic,3 4 with deterioration and death coming sooner than expected by all concerned. In illnesses other than cancer, recurrent hospital admissions and interventions give rise to an unpredictable dying trajectory and a “prognostic paralysis,”5 in which the difficulty of prognostication results in failure to consider or raise end of life issues until death is very close and the patient too unwell for meaningful conversations. End of life discussions are particularly challenging with patients who have heart failure: up to half of deaths are sudden, particularly in the less severe stages.6 Many older patients have multiple comorbidities, each of which is potentially life limiting.

Changing illness trajectories

Therapeutic and healthcare advances are changing care at the end of life. Patients with cancer are increasingly receiving active treatment into their last weeks of life and their dying trajectories are becoming more akin to those of patients with non-malignant chronic illnesses. In exacerbations of non-cancer illnesses, patients and clinicians often see acute admission and active treatment as appropriate: “you never know what they might be able to do in the hospital.” Public and professional attitudes have not kept up with this increasing medical activism: end of life discussions are still often linked in their thinking with the stopping of active treatment and the close proximity of death. In modern health care, such cessation of treatment often takes place far too late for effective end of life care planning to happen, if it takes place at all.

Keeping in the frame of “curative change agent”

The communication of a poor prognosis is a most difficult conversation for doctor and patient and is a source of considerable physician stress.7 8 Doctors are often reluctant to discuss poor prognosis and treatment options,9 10 and when such conversations do occur, they frequently avoid the words “death” or “dying,” preferring euphemisms such as “time is getting short” that are intended to soften the shock but may also confuse or mislead.11 Patients with cancer frequently misunderstand the aim of their treatment, seeing therapy aimed to palliate disease as having curative potential.12 Patients view the option of supportive care without continued disease modifying treatment as the clinical team “giving up”: they value their doctors’ expertise in up to date active interventions and prefer them to remain in the role of curative agent.13

Coping with uncertainty and maintaining hope

Professionals often prefer to wait for patients to approach them to talk about the end of life, whereas patients often wait for the doctor to broach the subject.14 Conversations are thus avoided until disease is advanced and prognosis is more certain, and this delay is a common cause of late referrals to palliative care, unplanned hospital admissions, and inappropriate interventions when crises develop.15 Doctors are often uncomfortable with the inherently uncertain nature of prognostic estimates and find patients’ expectations of clarity and certainty impossible to meet.8 They struggle to bring that uncertainty into the open for themselves, the clinical team, and the patient.16

Maintaining hope during and after difficult conversations is challenging. Some patients would like open communication about their illness and its progress: others are more ambivalent, wanting to be told but not wanting to know, or having a compartmentalised awareness in which they acknowledge that their illness is terminal while retaining a sense of hope.17 Evidence suggests that open discussion is beneficial for those who desire it, with less inappropriate medical treatment, lower risk of depression, and better adjustment of caregivers to bereavement.18 However, to impose such open discussions on all patients, irrespective of their wishes, may destroy hope and cause considerable harm. Denial is an important ego defence mechanism that must not be broken down.

Understanding patients and carers’ perspectives

Patients’ fears may underlie their reluctance to discuss the end of life: fear of treatment withdrawal, of loss of the managing team, of uncontrolled symptoms, to name but a few. They may have cognitive impairment or low health literacy, and misunderstand or selectively retain information given. They may be protecting their families, using coping strategies such as denial, or they may simply not wish to address the issues at this time. Many, however, have information needs that could be addressed by sensitive, patient led conversations.

The financial impact of failure to start end of life conversations

Failure to discuss the end of life may have a substantial financial impact. In the UK, patients with a terminal prognosis (defined as six months or less to live) are entitled to both the higher rate disability living and attendance allowances, which are fast tracked on completion of form DS1500: over half of people who die from cancer receive neither allowance.19 In the United States, Medicare funded patients have to make a choice between home hospice care and hospital active treatment: in the absence of early end of life discussions, most continue with active treatment and are referred for hospice care very late in their illness.15

What are the organisational incentives?

Studies of the Gold Standards Framework for Palliative Care in primary care suggest that timely end of life conversations can trigger the introduction of processes that are associated with improvements in care.20 The current details of palliative care indicators for primary care in the Quality and Outcomes Framework are insufficient, and there are no organisational incentives in secondary care to encourage the appropriate initiation of end of life discussions. Tariffs for chemotherapy and radiotherapy do not include auditable communication or support elements, and end of life needs are rarely addressed in multidisciplinary team meetings. Hospitals do not routinely identify patients approaching the end of life, other than when very close to death when the Liverpool Care Pathway for the Dying is used. Nor do they have codes for end of life assessment and care planning.

Possible triggers for starting the conversation

Many triggers have been suggested for clinicians to consider opening up conversations about the end of life: poor control of symptoms, changing care needs, deteriorating function, withdrawal of active cancer treatment, diagnosis of incurable advanced disease, admission to hospital, or entry into a nursing or care home, among others. Recognition is growing that prognostic precision is rarely achievable and it may be better to identify patients who are “sick enough that dying within the next year would not be a surprise.”21 Those identified by this “surprise question” might be sensitively approached for end of life conversations and be put onto general practice palliative care registers. However, for many patients the proximity of death is not clear until very close to the end of life. For them, an approach of “hoping for the best and planning for the worst” may be the best way forward.

Initiating and holding the conversation

Hospital specialists, including oncologists, rarely initiate discussions about the end of life during active treatment, and hospital team care rarely permits the personal continuity that facilitates these difficult conversations. Primary care may be a better setting, where patients and families may have established and trusting relationships with their general practitioner, although personal continuity has declined in general practice over recent years. However, general practitioners may feel that they lack the specialist knowledge required and wait for a signal from the specialist team before opening up conversations. Patients may expect such information to come from their specialist, but disease specific specialist nurses often do not see these discussions as part of their role, and hospital palliative care teams are involved with a minority of dying patients.22 The consequence is that no professional takes responsibility for starting these conversations and the patient and family are left uncertain and confused about the future and their care options.

Discussions about the end of life require good communication skills and great sensitivity and respect for individual wishes.23 24 25 The crucial task is to ascertain which patient wants what information at this time—a judgment that doctors find very difficult to make.26 27 Some patients do want their doctors to talk in a straightforward and sensitive way when they judge the patient to be ready, listening and encouraging questions, and striking a balance between honesty and hope.28 Others may not want to talk now, not with that particular health professional, or not at all. Each patient’s care needs to be handled in the way they prefer, even if to do so creates untidiness and uncertainty for care providers. A patient’s preference for silence should be respected.29


In 1769 Samuel Bard wrote that “To buoy up a dying man with groundless expectations of recovery is really cruel” and could lead to “overlooking the important concerns of futurity, and involve families in confusion and distress.” Such practice is still very familiar 250 years on. In response, a conventional wisdom is developing that open awareness and communication about death and dying is the best option for everyone. Since patients’ preferences are varied and complex, such a “one size fits all” approach needs to be questioned: a patient’s preference not to hold such conversations must be respected.

Death is not a medical failure: it comes to all of our patients, and to all of us. It is part of our duty as doctors to provide optimal end of life care for our patients, a key part of which is the offer of timely, sensitive, patient-led conversations about the end of life.


Cite this as: BMJ 2010;341:c4862


  • doi:10.1136/bmj.c4863
  • doi:10.1136/bmj.c4861
  • doi:10.1136/bmj.c4860
  • doi:10.1136/bmj.c4859
  • Part of the costs of producing the BMJ supplement in which this article appeared were met by the British Heart Foundation. The article was commissioned and peer reviewed according to the BMJ’s usual process.

  • Contributors and sources: SB has been a general practitioner for 25 years. He is honorary consultant in palliative medicine at the Cambridge Hospice and Macmillan post-doctoral research fellow in the University of Cambridge General Practice and Primary Care Research Unit. He leads the End of Life research team of the NIHR CLAHRC for Cambridgeshire and Peterborough (Collaborations for Applied Health Research and Care). JM has been a consultant clinical oncologist at Mount Vernon Cancer Centre and Hillingdon Hospital for 20 years. She is senior clinical lecturer at University College London, visiting professor in cancer and supportive care at the Centre for Complexity Management at Hertfordshire University, and chief medical officer for Macmillan Cancer Support. Both authors are members of the Macmillan Palliative and Cancer Care Research Collaborative (MacPaCC), a group of clinical academics whose research has informed this paper, and whose contribution and support is acknowledged with gratitude. SB is guarantor.

  • Competing interests: All authors have completed the Unified Competing Interest form at (available on request from the corresponding author) and declare (1) no support from any company for the submitted work; (2) no relationships with companies that might have an interest in the submitted work in the previous 3 years; (3) no spouses, partners, or children with financial relationships that may be relevant to the submitted work; and (4) no non-financial interests that may be relevant to the submitted work.