View commentaries on this research

This is a plain English summary of an original research article

In the UK, end of life care is typically put in place for people who are likely to die within the next 12 months. But predictions of the length of time remaining are difficult to make. New research suggests that these predictions are often less accurate than chance alone. The study calls for palliative and end of life care services to be made available without waiting for a prediction of approaching death.

The study pooled previous research into the accuracy of predictions of death, and the problems that arise when predictions are incorrect. The researchers also investigated the types of end of life services available through primary care and in the community.

A lack of warning that someone is dying can have devastating consequences, the study found. Services may not be put in place in time. People may die in pain and away from loved ones; families may feel robbed of time. Those left behind may experience complicated grief for years to come.

Inaccurate predictions caused problems. If someone survived much longer than expected, families could struggle to look after them at home for long periods. People told they only have months to live could lose the hope that sustains them.

The study also described 8 types of end of life services available in primary care and the community. The researchers hope this will help professionals to make the most appropriate referrals.

Further information about end of life care is available on the NHS website.

What’s the issue?

End of life care is generally given in someone’s final weeks or months. This care aims to help people live as well as possible in the last part of their life, and to die with dignity. It does not aim for cure, but to reduce pain and suffering and to ensure the best possible quality of life in the time remaining.

Care to manage distressing symptoms (palliative), can be provided alongside care aimed at cure (curative). It is often wrongly believed, including by health professionals, that palliative care only starts when curative treatment stops. In the UK, end of life care is not usually started until a clinician judges that someone is likely to die in the next few days, weeks, or months.

This study explored how accurate predictions of death are, and whether access to end of life care should depend on a prediction of death. The team also looked at the palliative and end of life care services available in primary care and community settings.

What’s new?

The study pooled evidence from 21 studies, reviews, policy reports, and discussion papers, and from another 317 abstracts. Researchers investigated the accuracy of death predictions and the consequences of inaccurate predictions. They looked at factors that influenced referrals to palliative and end of life care services, and they categorised those available in primary care and in the community.

Predictions of death

The study found that predictions of death varied broadly in their accuracy and were often worse than chance. 

A lack of warning that someone is dying could have far-reaching consequences, the study showed. People could feel robbed of time to set their affairs in order, to say their goodbyes, or to arrange the care they prefer. Health and social care services could not always respond effectively in time. This could lead to poor quality care with insufficient pain control and symptom relief to increase the quality of life remaining. People were more likely to be admitted to hospital and receive inappropriate invasive treatments aimed at cure.

But inaccurate predictions could also be harmful. People who survived much longer than predicted could live for years with the constant threat of approaching death. Their family members could come under strain if they were only able to look after their loved one at home for a short time.

Once they had a prediction of death, some people chose to try new curative medicines. Few benefitted from new or experimental treatments; those who did often had small gains only. For some, however, this could mean they died without receiving good palliative care.

Referrals to end of life care services were influenced by whether someone could be looked after at home (family and carers’ needs must be considered) and what systems were available locally. Occasionally, professionals claimed that people had less time left than they believed, to allow access to a ‘better’ service. Some professionals were not aware of all available services.

Carers’ and family members’ observations about changes in a person’s condition were not always taken seriously, meaning that critical information was not included in decisions about end of life care.

Primary care and community services

As a new resource, the researchers grouped primary care and community services (delivered fully or mainly outside of hospitals or hospices) according to how or why they were expected to work.

Different services had different roles. They aimed to:

  • increase availability, such as out of hours and home-based services
  • address personal needs and provide holistic support, for example through social interactions, financial support, counselling, and complementary therapies
  • help people move from one type of care to another (transition care), for instance to manage hospital discharge
  • keep the person at the centre of their care, by ensuring their preferences are documented (using both legally binding tools and more flexible decision aids) and by supporting self-management
  • provide care according to the phase of someone’s illness, for instance around the time of a diagnosis of serious illness, or in the final weeks or days of life
  • take a proactive, ‘whole-system’ approach, such as through advance care planning discussions and electronic systems for recording preferences and sharing data across health services
  • improve outcomes, and fulfil someone’s final wishes, through support and training for staff and volunteers, and expanding staff roles
  • provide support to informal carers.

Why is this important?

Recent research and policy has a strong focus on supporting dying at home. But there has been little specific discussion about the timing of end of life care.

Palliative and end of life care services are not usually offered until people are identified as being at the end of their lives. Many people do not receive pain relief, symptom control and other support when they need it.

Findings about the inaccuracy of predictions should not discourage professionals from talking about death or timelines, the researchers say. They need to find ways of approaching uncertainty and to give people warning while not robbing them of hope.

The researchers recommend that palliative and end of life care services should not depend so heavily on expectations and conversations about death and dying. They should be offered according to people’s needs. It should not matter how long this care is needed, or whether someone wishes to continue, in parallel, with curative treatments.

Professionals also need more training and awareness of the many kinds of palliative and end of life care services that are available in the community and through primary care. This could help them make more appropriate referrals and put in place more timely care.  

What’s next?

The researchers suggest that professionals should have discussions about palliative care far earlier, and not only when someone is dying. Palliative care now includes management of any suffering related to illness. It is needed in many healthcare settings other than in life-limiting illness, death and dying.

Discussions about palliative and end of life services should include:

  • patients' physical, psychological, social, spiritual, and practical support needs
  • patients' preferences for home care and home care support needs
  • patients' preferences for life-saving treatments, including in future when mental capacity might be lacking
  • carer needs
  • the maximum level of intervention, including whether resuscitation is appropriate
  • pain relief and symptom control
  • data-sharing preferences which could allow or limit different health professionals’ access to people's notes.

Service providers need a new approach to palliative and end of life care. Many more people need access to holistic care to meet their physical, psychological, social, spiritual, and practical needs. Many aspects of palliative and end of life care represent good care that should be available regardless of whether someone has been judged to be approaching the end of life.

You may be interested to read

This NIHR Alert is based on: Petrova M, Wong G, Kuhn I, Wellwood I and Barclay S. Timely community palliative and end-of-life care: a realist synthesis. BMJ Supportive & Palliative Care 2021; doi: 10.1136/bmjspcare-2021-003066

The Palliative and End of Life Care Group at Cambridge website.

A report from the World Health Organisation (WHO) on the need for universal access to palliative care: Why palliative care is an essential function of primary health care

A report on palliative care and pain relief from the Lancet Commission on Palliative Care and Pain Relief Study Group: Knaul FM, and others. Alleviating the access abyss in palliative care and pain relief-an imperative of universal health coverage: the Lancet Commission report. The Lancet Commissions 2018;391:10128

A national framework from the National Palliative and End of Life Care Partnership: Ambitions for Palliative and End of Life Care.

Funding: This study was funded by the NIHR School for Primary Care Research.

Conflicts of Interest: One of the authors, Geoff Wong, is deputy chair of the NIHR Health Technology Assessment Primary Care Panel.

Disclaimer: NIHR Alerts are not a substitute for professional medical advice. They provide information about research which is funded or supported by the NIHR. Please note that views expressed in NIHR Alerts are those of the author(s) and reviewer(s) and not necessarily those of the NHS, the NIHR or the Department of Health and Social Care.

  • Share via:
  • Print article


Study authors

Current policy and practice assume that we need to identify a patient as being at the end of their life and/or not likely to benefit from curative treatments before palliative and end of life care services are discussed. Our results challenge the foundation of this sequence of steps. Since the identification of patients at the end of life is so often inaccurate, the basic assumption is problematic.

The shift to palliative and end of life care services in the community and primary care is positive, but complex. We offered detailed examples of the kinds of services that are available so that healthcare professionals, patients and carers have better awareness of the numerous sources of support they can access. Policy and decision makers, commissioners and service designers can also choose from a broader range of established and innovative programmes.

Mila Petrova, Visiting Fellow, PELiCam (Palliative and End of Life Care in Cambridge) Research Group, Primary Care Unit, Department of Public Health and Primary Care, University of Cambridge

Stephen Barclay, Associate Professor in General Practice and Palliative Care, Honorary Consultant Physician in Palliative Care, Lead of PELiCam, University of Cambridge 

Marie Curie 

Access to high quality palliative care is essential to ensure that quality-of-life, for however long that might last is achieved. This timely research is a reminder of the importance of properly funded, person-centred palliative and end of life care options.

Everyone should be supported to have the best possible end of life experience, reflecting what is important to them. But this depends on the right support and care being available in the community and on wishes and preferences being identified and supported as early as possible.

Briony Hudson, Senior Research Manager, Knowledge Translation and Research Network, Marie Curie 

Research Nurse

This review supports what we already know in practice – that predicting death is difficult. This evidence will support services to define how and when they might assess a patient’s prognosis and when support will be needed. As an organisation, we need to be able to say that we know a prediction is vague and rely less on that prediction as a gateway to services and further support.

Some of the findings support changes to organisational pathways that rely on an ‘accurate’ estimation of time of death. This can mean the difference between access to services or not, and have a real impact on patients and their families and carers.

Sarah Stirrup, Senior Research Nurse, William Harvey Hospital Research Delivery Team, East Kent Hospitals University NHS Foundation Trust 

Lived Experience

“This is relevant to me as someone living with a progressive neurological condition. I think it will interest anyone who is likely to need palliative or end of life care at some point, as well as their families or informal carers. People may expect a definitive timescale so that they can plan and prepare psychologically, emotionally, spiritually and practically for the end of life.

The research shows a need to revise guidance and change mindset, particularly in health professionals but also patients, to move away from trying to identify a timescale. Less emphasis on death and dying could lead to new ways of working that improve the delivery of end of life care and may make it more widely available.

Revised guidance is needed but commissioners and professionals may need to assess the possible impacts of a new approach which includes access to different forms of care across a range of specialties and services. There will be an impact on financial resources, location of care provision and staffing.”

ME, Public Contributor, North of England 


I work on supportive care models that integrate supportive/palliative care for non-malignant conditions. This research seems to validate some of our service design choices, such as parallel planning by which service users can receive optimal medical management for their life limiting condition alongside supportive care.

This research recognises the uncertainty among service users and professionals regarding palliative and end-of-life care. I hope it will encourage transformational change in which all specialties learn from each other and services benefit from service user-led engagement.

This paper is relevant for service reviews and theory development, and in exploring ways that different specialties can learn and work together. Services need to be flexible and based on individual needs rather than the estimated time to death.

More debate is needed to refine ideas such as that prognostic indicators can be helpful in some instances but not as sole indicator for service referral. We need to continue to develop innovation in services that emphasise value-based care, and improvements that matter to service users.

Sian Hughes, Supportive and Palliative Care Clinical Nurse Specialist, Cardiff and Vale University Health Board 


“This is a useful piece of research, drawing together important issues in palliative care. It highlights the limitations of advance care planning, and it emphasises the need for an informed and thoughtful approach.

This work challenges professionals in service commissioning and provision to consider broader issues and the patient perspective. Services need to work together in a joined up way. Clinicians need to consider the impact including, benefits and limitations, of giving a prognosis. And communication between professionals, patients and families must be realistic. ”

Susi Lund, Visiting Researcher, University of Southampton


We need further discussion about how to identify palliative care patients and how to adapt care for each individual.

This paper raises awareness of the need to consider access to palliative care services at the appropriate time, but it also demonstrates the challenges of doing this. The paper could be used to suggest to oncologists that, while treatment is ongoing, active palliative care services could be used in a complementary and supportive way.

The recommendations may be useful in joint training for clinicians in acute care and palliative care teams.

Samantha Russell, Highly Specialised Palliative Care Physiotherapist

Member of the public

“The issues around dying will interest members of the public. Some people assume chronic conditions such as cardiac disease can be arrested rather than be progressive and terminal. Providing early information will give patients the opportunity to engage with the topic, and form views earlier in their lifetimes.

The honesty of this review in stating difficulties such as the accuracy of timing in prognosis will be appreciated by those wrestling with the human consequences of uncertainty in care.

I have taken an active interest in healthcare for many years and, now in my 70s, am encouraged to make some more informed choices about the issues involved in the end of life.”

Bill McPate, Public Contributor, York


Back to top