Guide for commissioners on end of life care for adults

NICE commissioning guides [CMG42] Published date:

3 Assessing service levels for adults needing end of life care services

3.1 Information and data to support the provision of end of life care

Around 455,000 adults died in England in 2010, two-thirds of which were 75 years of age or older[12]. The majority of these deaths, 53%[12], occured in an acute hospital. 75% of NHS hospital deaths are from diseases of circulatory system, respiratory system or cancer[13]. The number of deaths each year in the United Kingdom is expected to rise by 17 per cent between 2012 and 2030[14]. In addition, the average age at death is predicted to increase[13], and those dying are likely to have increasingly complex co-morbidities.

Commissioners and their partners should conduct a local needs assessment in order to estimate local service need, to plan capacity and develop plans to improve the accessibility and inclusivity of services. A local needs assessment for end of life care for adults is likely to include:

  • number of people who may need end of life care including annual number of deaths by condition and place of death (see section 3.1)

  • additional factors that influence local service need (see section 3.2)

  • existing current and optimal local practice (see section 3.3).

3.1.1 Identification of the number of adults at end of life

In England, approximately 455,000 adults died in 2010[12] which equates to approximately 1.12% of the adult population[15].

A proportion of this population will not need end of life care before death because the cause of death is sudden, either through unpredictable onset of disease or an external cause. Recent analysis of all deaths in England by the National End of Life Care Intelligence Network estimate that at least 25% of all deaths in England are unexpected deaths from sudden causes[16]. This is in line with the palliative care funding review[17], which estimated that between 70% and 80% of all deaths are likely to need end of life care.

All people whose deaths are not sudden or unexpected (approximately 75% of all deaths) should have their end of life care needs recognised and provided for in the last year of life. Therefore, the number of adults needing end of life care services is 0.83% or 830 per 100,000 population aged 18 years or over; this is around 341,000 people in England[18].

3.1.2 Number of deaths by underlying cause

Figure 4 illustrates that three-quarters of people who died in 2010 died from diseases of the circulatory system, respiratory system or cancer. The cause of death can affect the type of care needed in the last year of life and the appropriate place of death. It should be noted that the cause of death does not necessarily reflect the main condition needing care in the last year of life, and many people will have comorbidities.

Figure 4: Underlying cause of death England and Wales 2010[19]

Commissioners may wish to pay particular attention to people with respiratory disease. The report 'Deaths from respiratory diseases: implications for end of life care in England' shows that, of the three main causes of death, respiratory diseases (excluding lung cancer) is the category for which there is the highest proportion of emergency hospital admissions, the highest proportion of people who die in hospital (69%) and the lowest proportion who die in their own residence (13%). Also:

  • there are relatively few respiratory (excluding lung cancer) deaths in hospices;

  • there is a higher proportion of respiratory (excluding lung cancer) deaths in the most deprived quintile

  • the proportion of deaths due to respiratory diseases increases with age.

3.1.3 Place of death

The Office for National Statistics data 'Mortality Statistics: Deaths registered in England and Wales (Series DR), 2010', sets out the place of death. The most common place of death is NHS hospitals where over 53% of deaths occur (table 2) for all causes.

Table 2 Place of death: all causes [20] , [21]



Care homes

Local authority


Non-local authority







Hospitals (acute or community, not psychiatric)





Other communal establishment




Place of death for just cancers (table 3) shows a higher number of deaths in hospice and at home with lower numbers of deaths in NHS hospitals. The national audit office report 'End of life care' suggests that a large proportion of people (40%) who currently die in hospital have no medical reason to do so.

The majority of people said that they would prefer to die at home if circumstances allowed (section 3.1.4) but the number of people who are able to do so varies with age, geographical area and, most significantly, by condition[22].

Table 3 Place of death: cancer only [17]



Care homes

Local authority


Non-local authority







Hospitals (acute or community, not psychiatric)





Other communal establishment




3.1.4 Preferred place of death

Commissioners should ensure that they engage with the public and service users when developing their needs assessment to understand service user experience and local barriers to accessing end of life care. This information should be used to inform commissioning plans and to improve the accessibility of local services.

The actual place of death (section 3.1.3) contrasts sharply with the report, 'Local preferences and place of death in regions within England'[23] which showed public preferences for place of death in the nine English Government Office regions.

The majority of participants in all regions said that they would prefer to die at home if circumstances allowed, this ranged from 60% of participants in the West Midlands to 67% in the north east. This preference was lower (45%) for the older age group (75+) overall. The older age group (75+) also expressed the greatest preference to die in hospice care, despite being the least likely to do so. Hospital was the least preferred place of death in all regions except for the north east, where 34% reported 'care home' as their least preferred place (against 31% for hospital).

It should be noted that this was a telephone survey of participants (not necessarily people who were dying) in their own home, thus excluding care home residents. People's preferences for where they are cared for and die are not yet currently systematically captured or routinely reviewed as their illness progresses and their needs change, it is not possible to get a clear picture of how many people would prefer to die somewhere other than their actual place of death. For the future, this gap may be filled nationally through the 'Views of informal carers – evaluation of services' (VOICES) survey of carers and bereaved relatives and locally through the growing use of electronic palliative care coordination systems which will record this information.

3.2 Additional factors that influence local service needs

Commissioners should conduct a local needs assessment for health and social care to determine local service levels for people with end of life care needs. To determine local commissioning priorities, end of life care commissioners should conduct or review their baseline assessment of local practice in line with the 'End of life care strategy', NICE cancer service guidance and the 'Operating framework'.

Commissioners will need to use intelligence on local demography to inform their needs analysis. The following are in addition to the areas covered in section 3.1 and can have a significant impact on the local need and uptake of end of life care:

  • age distribution and sex of the person

  • regional and local variation

  • non-malignant diseases

  • deprivation

  • cognitive impairment (dementia)

  • religious or cultural groups within the local community

  • lesbian, gay, transgender and transsexual people

  • children and young adults.

3.2.1 Age and sex

The trend for life expectancy to increase has been observed for many years in the United Kingdom, as the age at death increases those dying are likely to have increasingly complex comorbidities. The World Health Organization (WHO) 'European health for all database' 2009 data shows life expectance in the UK (table 4) has increased dramatically in the years since 1970.

Table 4 Life expectancy at birth, in years






United Kingdom






Commissioners should be aware that the number of deaths by age and sex varies significantly from region to region. The office of national statistics publication 'Deaths registered in England and Wales by area of usual residence' splits the number of deaths by Government Office region. Around two-thirds of all deaths occur in the 75 and older age group (74% of females and 59% of males[12]). However, older people may be at risk of over-identification if assumptions about their prognosis and end of life care needs are based on age alone.

Figure 5, Deaths by age and sex [12]

3.2.2 Regional and local variation

Identifying groups of people at the end of life is complex and the profile of these groups will differ regionally. The 'Health profile of England' 2009 shows a distinct 'north/south' divide for both male and female life expectancy at birth. In all northern regions, as well as both the East Midlands and West Midlands, life expectancy is significantly shorter than in the regions to the south. For both sexes, those living in the north east or north west live approximately 2 years less than those in the South East or South West.

In England two-thirds of all deaths are in people aged 75 years of age or older[12]. Around 17% of deaths are in people aged 65 or under. However, this varies with some inner city primary care trusts recording over 30% of deaths in the 65 or under age group.

3.2.3 Comorbid conditions

In the last year of life many people will have comorbid conditions. The cause of death is not necessarily the main condition that will be treated in the last year. Dementia has a relatively low number of deaths. However, studies have shown that a large proportion of older people have dementia.

The NICE commissioning guide 'End of life care for people with dementia' reported the number of deaths in hospital for people with dementia. The prevalence of dementia increases with age, from around 1.6% in people aged 65–69 to over 30% in those people aged 90 or over. 66,000 people who die each year with dementia are likely to need end of life care before death. This is equivalent to around 800 deaths per 100,000 people aged 65 and older, per year.

Some groups (such as people with non-malignant disease) are currently under-identified when providing end of life care. The association of public health observatories disease prevalence models[24], the health needs of a population derive from the prevalence of diseases, that is, the numbers of people with different types of illness. This disease prevalence model can help commissioners to identify and address those needs.

3.2.4 Deprivation

The National End of Life Care Intelligence Network report 'Variations in place of death in England' examines the complexity and appropriateness of different end of life care settings for different groups, depending on age, sex and socioeconomic status.

The 'Health profile of England' 2009 shows inequalities in the determinants of health across England; for example, approximately 34% of people living in the north-east live in the most deprived fifth of neighborhoods in England. This compares with approximately 6% of people in the south-east who live in the most deprived fifth of neighborhoods.

Risk factors for diseases such as cancer and circulatory diseases, including smoking, physical inactivity and obesity, increase with greater levels of deprivation. The Marmot report 'Fair society, healthy lives: a strategic review of health inequalities in England post-2010' identified that the burden of disease falls disproportionately on people living in deprived conditions.

3.2.5 Household composition

Household composition will impact on the level of care needed for a person; household composition differs for men and women. Data from the Office for National Statistics showed that in 2010, 7.5 million people in UK households lived alone, of which 3.4 million were aged 65 and over, of these, 70% were women.

The National End of Life Care Intelligence Network report 'Deaths in older adults in England' again shows the proportion of older men living alone is lower than for older women, with more men living as a married couple and fewer who are widowed compared with women. Similarly, north west public health observatory reports the proportion of older men living in a communal establishment is lower than for women.

Factors that are thought to contribute to differences in living arrangements include a greater life expectancy[12] for women than men and a tendency for women to marry men who are older than themselves. Living alone will affect the level of care needed for a person and their ability to continue to live in their own home. This may also affect the person's aspirations to die in their own home and the levels of care needed.

3.2.6 Religious or cultural groups in the local community

Local communities may contain small or large pockets of people with particular religious or cultural beliefs, in which discussions about death may be considered to be inappropriate and the management of death is handled in a particular manner. These communities may also be from particular black and minority ethnic groups (but this is not necessarily the case). Differing language, religious and cultural needs should be considered when identifying and providing end of life care services.

3.2.7 Lesbian, gay, transgender and transsexual people

Any service provided for people at the end of their lives must be equitably and sensitively provided to lesbian, gay, transgender and transsexual people and their partners and families.

3.2.8 Children and young people

Around 2,500[12] children and young people aged 15–24 years old died in 2010. However, as a result of advances in medical care, many children with life-limiting, long-term conditions live into adulthood. 'Better care better lives' (Department of Health 2008) identifies the need to ensure that transition should be 'a planned and purposeful process'.

3.2.9 Sources of data for the needs analysis

A comprehensive compendium of data sources, publications and research is available on the National End of Life Care Intelligence Network website. The National End of Life Care Intelligence Network has further developed a set of tools to help to identify and understand variation in end of life care across England. The indicators focus on place and cause of death, broken down by age and sex, by local authority area in England.

3.3 Current practice

3.3.1 Hospital episode statistics

The 'Hospital episode statistics (HES)' is a data warehouse that contains information about hospital admissions and outpatient attendances in England. The data in HES comes from the Secondary Uses Service, which collects data that have passed between healthcare providers and commissioners. It includes private patients treated in NHS hospitals, people resident outside England who were treated in an NHS hospital and care delivered by treatment centres (including those in the independent sector) funded by the NHS.

In 2010/11 HES data show that over 200,000 in people died in hospital. Over 40% of these people had a final length of stay of 8 days or more. Figure 6 below illustrates the length of stay and primary chapter (diagnosis area) for people who died in hospital in 2010/11.

The length of stay may be appropriate for many of these people; however, it is possible that the length of stay could be reduced for some people at the end of life through providing alternative care options, if appropriate.

Further, detailed analysis of the HES data, split by primary care trust, can be found in the commissioning and benchmarking tool. Figure 7 illustrates the English regions that have the largest number of episodes per 100,000 population of long length of stay (8 days or more). Commissioners should use the commissioning and benchmarking tool to model service capacity at each step of care.

Figure 6 Length of stay of hospital admissions which result in death

Figure 7 Episodes per 100,000 people who died in hospital, with a length of stay of 8 days or more

Figure 7 categorises strategic health authority (SHA) regions based on the number of hospital episodes ending in death, with a length of stay of 8 days or more per 100,000 people ('Hospital episode statistics [HES]' data). Of the 10 SHA regions, the highest three are marked as high in relation to the other regions. The next three are marked as medium and the last four are marked as low. Further analysis and more detail, including a split by primary care trust is available in the commissioning and benchmarking tool.

3.3.2 Palliative care registers

One source of data for people who are currently identified as needing end of life care is palliative care registers. Quality and outcomes framework (QOF) data for 2009/10 show the number of people on a palliative care register (people in need of palliative care or support) is 0.1% of the population or 75,000 people in England. This is approximately 16% of the number of people who die every year.

People with cancer tend to be over-represented on the QOF palliative care register. Data extracted from IMS Disease Analyzer, a database that holds data from a sample of GP practice systems, show that 73% of people available in their practice for the 12 months from 1 April 2010 to 31 March 2011 new to the QOF palliative care register during the study year, had a diagnosis of cancer anywhere in the record (International Classification of Disease 10 cancer definitions[25]). This compares with less than 30% who had cancer as their primary cause of death.[26]

Electronic palliative care coordination systems that link together all the organisations involved in a person's care (currently being piloted) have been considered a potentially useful tool for improving end of life care and enabling people to be supported in their place of choice and to die in their usual place of residence.

The process of placing an individual on a palliative care register or electronic palliative care coordination system should be sensitive to the possibility that not all people will want to be recognised as approaching the end of life. Some people choose not to acknowledge their prognosis and others, such as those with sensitive conditions, may worry about potential stigma if they are identified.

3.4 Conclusion

Based on the epidemiological data and other information outlined above, based on the following, it was concluded that:

  • around 455,000 adults died in England in 2010

  • 75% of all deaths were not sudden or unexpected

Therefore, the indicative benchmark rate for the number of adults needing end of life care services is 0.83% or 830 per 100,000 population aged 18 years or over. This equates to 341,000 adults in England.

Commissioners should use their local needs assessment to determine optimum levels for local service provision. Commissioners should note that the benchmark rates do not represent NICE's view of desirable, or maximum or minimum, service levels.

Commissioners should use this benchmark and local data to facilitate local discussion on optimum service levels. There is considerable variation in the identification of people in the last year of life. This is influenced by the social, economic and demographic profile of the local population, therefore commissioners are encouraged to consider local assumptions.

Use the end of life care commissioning and benchmarking tool to determine the level of service that might be needed locally and to calculate the cost of commissioning the service using the indicative benchmark and/or your own local data.

[12] Death registrations in England and Wales, selected data tables 2010, Office for National Statistics (ONS), London , 2011

[13] HESONLINE, Provisional linked 2009/10 HES-ONS mortality data aggregated by diagnosis/procedure, 2010, Leeds

[14] Gomes B, Higginson IJ. Where people die (1974-2030): Past trends, future projections and implications for care. Palliative Medicine, 2008; 22(1):33-41

[15] Applied to Office for National Statistics population estimates for people in England

[16] Blackmore S, Pring A, Verne J (2011) Predicting death: estimating the proportion of deaths that are 'unexpected.' National End of Life Care Intelligence Network

[18] This figure is for end of life care prior to death. Care after death is examined in section 4.5. See also section 3.2, which examines the areas that will have the greatest influence locally on the above figure, principally the age profile of the population and the expected number of deaths.

[19] Death registrations in England and Wales, selected data tables 2010, Office for National Statistics, London , 2011

[21] Place of death is defined using a revised place of death classification. deaths at home are those at the usual residence of the deceased (according to the informant)‚ where this is not a communal establishment. Care homes includes homes for the chronic sick; nursing homes; homes for people with mental health problems and non-nhs multi function sites. Hospices include Sue Ryder Homes; Marie Curie Centres; oncology centres; voluntary hospice units; and palliative care centres. Other communal establishments include schools for people with learning disabilities; holiday homes and hotels; common lodging houses; aged persons' accommodation; assessment centres; schools; convents and monasteries; nurses' homes; university and college halls of residence; young offender institutions; secure training centres; detention centres; prisons and remand homes. Elsewhere includes all places not covered above such as deaths on a motorway; at the beach; climbing a mountain; walking down the street; at the cinema; at a football match; while out shopping; or in someone else's home. This category also includes people who are pronounced dead on arrival at hospital.

[22] Deaths: area of usual residence and sex, by place of occurrence 2009, Office for National Statistics, 2010, London

[23] Gomes B, Calanzani N, Higginson IJ Local preferences and place of death in regions within England. London, Cicely Saunders International, 2011

[24] Prevalence models provide estimates of underlying prevalence derived from population statistics and scientific research on the risk factors for each disease. This enables commissioners to assess the true needs of their community, calculate the level of services needed and invest the appropriate level of resources for prevention, early detection, treatment and end of life care.

[25] ICD-10 Chapter II: Neoplasms (C00- D48)

[26] Death registrations in England and Wales, selected data tables 2010, Office for National Statistics (ONS); London 2011

NICE has accredited the process used by NICE to produce guides for commissioners. Accreditation is valid for 5 years from November 2011 and applies to guides produced since November 2008 using the processes described in 'Process manual for developing guides from NICE for commissioners: Information for internal NICE teams' (2011). More information on accreditation can be viewed at
Get involved