Article Text

Download PDFPDF
General practice nurses and physicians and end of life: a systematic review of models of care
  1. Geoffrey Mitchell1,
  2. Michèle Aubin2,
  3. Hugh Senior1,3,
  4. Claire Johnson4,5,
  5. Julia Fallon-Ferguson6,7,
  6. Briony Williams6,7,
  7. Leanne Monterosso8,9,
  8. Joel J Rhee10,
  9. Peta McVey11,
  10. Matthew Grant12,
  11. Harriet Nwachukwu1 and
  12. Patsy Yates13
  1. 1Primary Care Clinical Unit, University of Queensland, Herston, Queensland, Australia
  2. 2Département de médecine familiale et de médecine d'urgence, Faculté de médecine, Université Laval, Quebec City, Quebec, Canada
  3. 3College of Health, Massey University, Auckland, New Zealand
  4. 4Nursing and Midwifery, Monash University, Clayton, Victoria, Australia
  5. 5Medical School, The University of Western Australia, Crawley, Western Australia, Australia
  6. 6Primary Care Collaborative Cancer Clinical Trials Group, The University of Melbourne, Melbourne, Victoria, Australia
  7. 7General Practice, University of Western Australia, Perth, Western Australia, Australia
  8. 8School of Nursing, Notre Dame University, Perth, Western Australia, Australia
  9. 9School of Nursing, Edith Cowan University, Joondalup, Western Australia, Australia
  10. 10General Practice Unit, University of Wollongong, Wollongong, New South Wales, Australia
  11. 11Susan Wakil School of Nursing, University of Sydney, Sydney, New South Wales, Australia
  12. 12School of Medicine, Monash University, Melbourne, Victoria, Australia
  13. 13Nursing, Queensland University of Technology, Brisbane, Queensland, Australia
  1. Correspondence to Professor Geoffrey Mitchell, Primary Care Clinical Unit, University of Queensland, Herston, QLD 4006, Australia; g.mitchell{at}


Background General practitioners (GPs) and general practice nurses (GPNs) face increasing demands to provide palliative care (PC) or end-of-life care (EoLC) as the population ages. In order to maximise the impact of GPs and GPNs, the impact of different models of care that have been developed to support their practice of EoLC needs to be understood.

Objective To examine published models of EoLC that incorporate or support GP and GPN practice, and their impact on patients, families and the health system.

Method Systematic literature review. Data included papers (2000 to 2017) sought from Medline, Psychinfo, Embase, Joanna Briggs Institute and Cochrane databases.

Results From 6209 journal articles, 13 papers reported models of care supporting the GP and GPN’s role in EoLC or PC practice. Services and guidelines for clinical issues have mixed impact on improving symptoms, but improved adherence to clinical guidelines. National Frameworks facilitated patients being able to die in their preferred place. A single specialist PC-GP case conference reduced hospitalisations, better maintained functional capacity and improved quality of life parameters in both patients with cancer and without cancer. No studies examined models of care aimed at supporting GPNs.

Conclusions Primary care practitioners have a natural role to play in EoLC, and most patient and health system outcomes are substantially improved with their involvement. Successful integrative models need to be tested, particularly in non-malignant diseases. Such models need to be explored further. More work is required on the role of GPNs and how to support them in this role.

  • service evaluation
  • supportive care
  • terminal care
  • symptoms and symptom management
  • home care

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Primary care practitioners—general practitioners (GPs) or family physicians (hereafter termed GPs) and general practice nurses (GPNs)—are central to the provision of person-centred end-of-life care (EoLC) to improve the quality of life of patients, and to prevent and relieve suffering.1 Primary care is defined as first-contact, accessible, continued, comprehensive and co-ordinated care… accessible at the time of need; focusing on the long-term health of a person rather than the short duration of the disease; providing the range of services appropriate to the common problems in the respective population and co-ordinating other specialists that the patient may need.2 GPs should therefore be adept at general medicine, and develop clinical relationships with patients and carers which allows an understanding of their needs.3 They should also knowledgeable of the health and social services available in the community.3 Most people visit a GP regularly,4 5 and GPs feel that caring for palliative patients is a key role.6 In Australia, over 80% of GPs report providing EoLC for at least one person in the past year.7

Primary care is the foundation of healthcare in most developed health systems. With the increasing ageing of the population, a paradigm shift is occurring in the provision of EoLC, whereby the numbers of people approaching death will increase rapidly, and the way EoLC is approached has changed significantly.8 While specialist palliative care (PC) services predominantly treat patients with cancer, most people will die of non-malignant diseases, either specific organ failure or multiple conditions.9 The most common causes of death with a predictable EoL are heart failure, cerebrovascular disease, end-stage respiratory conditions,10 multimorbidity, frailty and the dementias.11 12 Their EoLC is delivered by primary care and various medical specialists. Most spend the majority of the last year of life at home, either cared for by spouses of similar age, often with their own health problems, and/or by adult children with their own work and family.13 14As a result of this paradigm shift, we decided to refer to this care as EoLC rather than PC, to indicate that it is more than care in the last weeks or days, which PC can sometimes be perceived to be. Formal definitions of PC, EoLC and specialist PC are presented in online supplementary appendix 1.

The role of GPNs has grown substantially in recent years as the value of a multidisciplinary team approach has been more widely recognised. The WHO definition of primary care highlights the role of first point of contact and comprehensive general care of all people within a community.2 A primary care team at its heart has a GP and a GPN.2 Murray et al’s view that PC should be available to all people across all diseases, all dimensions of the person, in all settings and all countries15 accords with the WHO view that high-quality EoLC is a basic human right.1 16 This can only be achieved with active involvement of primary care worldwide.

To deliver quality EoLC, GPs have to identify the patient with future EoLC needs, then provide skilled management of comorbidities while reducing the risk of complications, address psychosocial issues, and liaise with family and other health professionals as well as ensure the patient’s end-of-life wishes and caregiver needs are considered.15 Often the care requires a multidisciplinary approach17 involving hospital-based consultants, inpatient services and community services. The GP and/or GPN may be leading or be involved in the co-ordination of this care.

To ensure GPs and GPNs continue to build capacity in providing PC within the community,18 we have sought to improve our understanding of the role of these professionals in the delivery of EoLC. In 2002, a systematic review was published on how well GPs provide EoLC.3 To date, this is the only attempt to bring together the world literature on GP performance on EoLC.3 However, extensive work on the role of primary care at the end of life has been done in many settings worldwide since that time. While national health systems dictate the nature and role of general practice and primary PC to some extent, there are central tasks and roles that are common worldwide.

To facilitate GPs and GPNs to build capacity in providing PC within the community, we have sought to integrate the literature on all aspects of general practice PC that has been generated since the 2002 systematic review. To this end, we have conducted a systematic review of literature published from 2000 to October 2017. The review sought to answer two major questions: (1) How well do GPs and GPNs deliver EoLC; and (2) what are the facilitators and barriers to the involvement of GPs and GPNs in providing EoLC? This publication is the final part of a five-part series.6 19–21 It seeks the models of care that have been developed to facilitate GP EoLC and evidence of their effectiveness.


We conducted a systematic review to critically appraise all aspects of the care provided by GPs and GPNs to patients with life-limiting illnesses and their carers. We followed the PRISMA method of identifying and assessing the quality of studies, and evaluating outcomes.

Inclusion criteria

Papers that reported on the practice of general practitioners or family physicians (henceforth termed GPs) and general practice nurses (GPNs) in EoLC. Patients were aged ≥18 years, suffering from advanced malignant or non-malignant illness, no longer responding to curative or maintenance treatment, and who required treatment with a palliative intent.

Types of intervention(s)/phenomena of interest included models of care that were developed to enhance GP or GPN practice in EoLC.

Types of outcomes

This study sought studies that included one or more of the following outcomes:

  1. Patient and carer outcomes relating to symptom control, psychosocial well-being or quality of life.

  2. Economic outcomes of models of care.

  3. Changes in GP, GPN or specialist team clinical practice.

Types of studies reviewed included

  1. Randomised individual or cluster controlled trials.

  2. Non-randomised controlled trials (RCTs).

  3. Controlled before-and-after studies.

  4. Qualitative studies (using semi-structured interviewing or focus groups).

  5. Other (discussion papers, position papers).

We excluded papers not reporting research findings, including editorials and opinion pieces.

Search strategy

We searched Medline, Psychinfo, Embase, Joanna Briggs Institute and Cochrane databases from 2000 to October 2017. The search strategy was based on that used in Mitchell’s 2002 systematic review3 and team discussion. The full search strategy is presented in online supplementary appendix 2.

The EndNote V.8.0 reference package (Clarivate Analytics, USA) was used to manage references. The initial database search was by single review of Titles and Abstracts in these databases, and hand-searching references in systematic reviews was conducted by JF-F, BW and HN. This initial search yielded 6209 articles after duplicates were removed. Titles and abstracts were then reviewed by both JF-F and BW to 2014, and GM and HN to 2017: 5732 articles were excluded, leaving 474 articles for full-text review. This included articles with a relevant title but no abstract. Two authors conducted independent assessment of each article, following the protocol. Any disagreements were resolved by discussion between the two authors or by arbitration by a third author if necessary. A further 209 articles were excluded after this process, leaving 265 articles for analysis (figure 1). The EndNote library was downloaded into EPPI Reviewer4 (EPPI-Centre, University of London) a multi-user web-based application for managing and analysing data for use in research synthesis.

Figure 1

PRISMA diagram of included studies.

Quality assessment

Each article was assessed for quality using a tool relevant to the study type: JADAD-RCT22 for RCTs; the Critical Appraisal Skills Programme for qualitative research23; the Newcastle-Ottawa Quality Assessment Scale (NOS) for cohort studies24; the NOS for cross-sectional studies24; the National Institute for Heart Lung and Blood (NIHLB) assessment for before-and-after studies.25


Included studies

Fourteen papers reporting 13 studies were included in this part of the review (figure 1). The studies were conducted in Australia (three studies, one of which has two publications17 26–28), UK,29–32 the Netherlands,33 34 USA,35 36 France37 and Japan.38 Several programmes ranging from national comprehensive approaches to EoLC, through programme-level organised care through to informal arrangements were identified. Two RCTs tested single case conferences between GPs and specialist PC units,17 26 and one RCT, a joint geriatrician/PC/primary care programme,36 while there were two before-and-after studies,18 27 35 one historical control study of a local pain management programme38 and one non-random comparator group evaluation of a national programme30 of initiatives to provide clinical guidance to GPs. Two observational studies described self-reported outcomes or practices,32 33 37 and there were evaluations of one district-wide service,29 one telephone advisory service34 and a national primary PC programme.31 Quality assessment was performed on each study. Most studies scored moderately to highly on their respective scales, but one of the RCTs was of low quality. The details of the included studies are presented in online supplementary appendix 3.

Impact of organised programmes supporting primary PC

The detailed findings are presented in table 1.

Table 1

Summary of findings

Interventions providing clinical support

A Dutch specialist PC telephone advice service fielded questions from GPs most commonly for pain (49%), delirium (20%), nausea and vomiting (16%), and dyspnoea (12%). Fourteen per cent of calls were for advice on end-of-life issues (terminal sedation 11% and euthanasia 3%).34 One-third of French GPs were actively engaged in end-of-life decisions taken within a hospital, and 60% reported being consulted by phone before a decision on withdrawal of treatment at the end of life.37

Pain and symptom control

There were mixed results for pain and symptom control. A home shared programme in the USA showed improvements in pain control,35 as did a prescribed pain pathway in rural Japan.38 However, a GP education programme showed no improvement.17 Similarly, there were mixed results for non-pain symptom control. Case conferences led to functional capacity being maintained for longer,17 and improved physical well-being and mental function in the last month of life.26 The home shared care programme between GPs and PC reduced patient anxiety, depression, tiredness and appetite,35 but a shared geriatric/PC service showed no persistent improvements in a range of parameters.36 Regular telephone contact between GPs and head and neck cancer team did not stop almost half of patients experiencing unresolved symptoms, despite the GPs reporting that they were satisfied with the standard of care they delivered.33 A GP facilitation service in the UK did not improve GP confidence in symptom management.29

Home deaths

Sixty-nine per cent of patients whose GP was involved in a system-wide palliative care framework (UK’s Gold Standards Framework (GSF)) achieved their preferred place of death,31 of which almost half of deaths were home deaths.31 Half of Dutch GPs participating in a local head and neck support programme achieved a home death.33

Adherence to clinical practice guidelines

Systematic programmes like the GSF and other end-of-life pathways led to improved adherence to PC guidelines,30 as did a prescribed pain pathway in rural Japan.38 The underlying assumption of GSF is that identifying patients at risk of dying early is fundamental to better EoLC for all. However, the highest proportion of patients on UK PC registers (a key component of GSF) was still people with cancer, despite them being a small proportion of general practice patients who died.32 Further, GPs used intuition more than formal tools developed to facilitate identification of people nearing the end of life.32

Health service utilisation

Case conferences in a specialist PC population (90% cancer) reduced hospital admissions by about 30% in Australia.17 A before-and-after study reported substantial reductions in emergency department visits, admission rates, length of stay and cost.27 28 However, there was no change in service utilisation noted in patients cared for by a USA-based combined PC/geriatrician service.36

No studies reported outcomes related to GPNs.


This review has described interventions to facilitate primary care involvement at the end of life that have been reported since 2002. Most were programmes that involved either some form of information sharing, or shared care between specialist services and primary care practitioners. There were some clear gains in clinical outcomes for some studies, but not all. It appears that patient outcomes were better the more actively the GP was involved in the decision-making process as well and/or the care itself. Most reported interventions did not use RCT methodology: formal comparative trial evidence of such interventions is difficult to achieve. Hence, results and their generalisability have to be interpreted with caution. In most cases, the comparator group used in the report was the best that circumstances allowed.

Of the three RCTs that were carried out, only one17 reached its proposed sample size target of 461, making it the largest PC trial ever carried out successfully. Its results should be accepted as robust. The reduction in hospitalisations and the improved functional status of patients over time the study demonstrated are results that should encourage changes in practice and policy. They reflect in some respects of the work of Temel et al,39 who showed that early introduction of specialist PC for oncology patients with advanced lung cancer improved quality of life and increased survival compared with oncology care alone. Here, integration of primary care and specialist PC in a robust way improved outcomes for PC patients, compared with PC alone.

There are three lessons from this review. First, the more comprehensive the engagement with GPs in the integrated planning and care of these patients, the better the outcome for patients, and the more likely there will be attendant economic benefits. If the interventions tested did not include active engagement of the GP in negotiating clinical plans, patients did no better than usual care.

This observation is also the finding of a systematic review of integration of primary care into the management of people hospitalised with advanced stroke.40 It was clear that in studies where the primary care involvement was collaborative and active (eg, the Trondheim study),41 patients were almost twice as likely to be alive and living at home than in programmes that simply told GPs what the plan was through routine correspondence.

Second, much of the work done focuses on the care of people with cancer. This is not a surprise since most specialist PC services treat mainly patients with cancer. However, as most people who die do not have cancer, it is important to establish whether the findings reported here are reproducible in non-malignant conditions. This may not necessarily be extrapolated from cancer-based populations—the needs of people suffering end-stage non-malignant disease, like frail elderly with multimorbidity and possibly dementia, are very different and occur over a much longer and more uncertain time course. The usual palliative care programme assumes a predictable and mostly time-limited time course.

While much work has been done to advance our understanding of the nature of the dying process and the needs these people have at the end of their lives, reliable translation of this work to new or improved community-based models of care is yet to be achieved. What is certain is that GPs who manage frail older people as a matter of course have and will continue to have a major role to play in the care of these people at the end of their lives. Given the ageing of the population and the anticipated growing needs in PC, GPs should be considered as important resources in the provision of EoLC, and they should be trained accordingly.

Third, the nature of national health systems has a major bearing on the nature of interventions that are possible. Some health systems are heavily slanted towards specialist care, and in particular system-based specialist care. It is possible to have this form of care entirely delivered by specialists, but there are significant risks of duplication of care, problems in determining treatment priorities, and patient confusion and uncertainty, particularly in frail and multimorbid patients. This form of care may be all that practitioners know, so introducing primary care into such systems will challenge both the specialists and GPs. Specialists may have insufficient knowledge of or confidence in the work of their GP colleagues. GPs may feel highly uncertain of being involved, due to reduced exposure to EoLC teaching at undergraduate or vocational training levels and lack of confidence. In health systems where primary care acts as the gatekeeper to the system, there is more likelihood that they will have basic skills and experience required to manage care at the end of life as their clinical responsibilities are likely to be broader and they are often integral parts of the community they serve. Furthermore, if the system in operation includes capitation payments, such as the UK, it appears that national programmes are feasible and may improve the proportion of people who die in their place of choice.

This outcome is entirely in line with the work of Starfield, who showed a direct relationship between a nation’s investment in primary care and health outcomes.42 43 The more primary care takes the majority of responsibility for health, the better the health of the population. It appears this finding extrapolates to care at the end of life as well.

As with other topics within this large systematic review, there was no research examining or enhancing the role of GPNs. This role needs to be investigated, and the potential of nurses to improve patient outcomes strongly encouraged. GPNs currently do have important roles in the management of long-term conditions. Given the change in the nature of dying as the population ages, management of long-term conditions in older people actually merges imperceptibly into EoLC. GPNs (and GPs) are already dealing with EoLC on a day-to-day basis. A change in focus which brings this perspective to the fore, alongside routine chronic disease management, is really what is needed to enhance the GPN’s role in EoLC. Adding appropriate anticipatory planning appropriate to the risk of deterioration and dying, and providing the tools to achieve this is what is required in further research.

Summation of the findings of the full systematic review

This is the final of five papers detailing research done on the role of primary care in the end of life since the previously published review.3 Much has changed in the 17 years of reported literature, and much remains the same. The comprehensive whole person role of primary care still defines primary care,2 but it is delivered by a team that now comprises the GP and others including GPNs. The relationship between the patient and their doctors and nurses remains the bedrock of that care and offers great opportunities to deliver high-quality EoLC. Indeed patients and their carers expect it from their GP and GPN.20 GPs can be effective at symptom control towards the end of life, but require exposure to these patients and support when delivering that care early in their career so they can become confident in time. New challenges, especially terminal sedation and the legalisation of euthanasia/medically assisted dying (in some countries), have developed in the time frame under review.

GP and GPN self-reported conduct of EoLC involves multiple roles.20 New studies reported include a renewed focus on sharing the journey with the patient, from identifying patients at risk of dying in a timely manner, through breaking the bad news, to providing bereavement care. The nature of patients requiring EoLC has widened, as has the ethnic diversity of populations that primary care deals with.19 The role of spirituality has been highlighted but remains a challenge to fulfil. Disturbingly, a loss of confidence in GP’s ability to perform appropriate EoLC compared with 2000 was observed in the literature, at the very time the death rate is rising throughout the world.20

Patients and carers expected GPs to be competent in all aspects of EoLC.6 They expected and valued availability, empathy and good communication skills, as they did in 2000. The GPs’ involvement contributed positively to the use of Advance Health Directives (which were not in existence in 2000) and the possibility of patients dying in their preferred place.

A range of facilitators and barriers to GPs and GPNs conducting EoLC was identified in the last 17 years.19 There is a much better understanding of the needs of patients and carers, and of the characteristics that facilitate or hinder the conscious decision of practitioners to participate in EoLC. Much care is reactive to patient need and not enough thought is given to what will happen to the patient in the future. This is critical to EoLC in non-malignant disease. Impressive national frameworks of models of care have arisen, particularly in the UK, in an attempt to facilitate identification of palliative needs early. Similarly, the co-ordination of EoLC, between SPC and primary care has been examined.


Interventions that aim to improve the knowledge and skill of general practitioners are likely to improve outcomes of patients with cancer. More studies are needed to confirm that similar interventions will achieve the same results in non-malignant conditions. Strategies to enhance the role of GPNs in EoLC are also required.

Ethics statements

Patient consent for publication


We would like to acknowledge generous research support from the Primary Care Collaborative Cancer Clinical Trials Group, which is funded by Cancer Australia. Dr May-Lill Johansen joined the group in 2016 as part of her sabbatical leave, and her input has been invaluable.



  • Twitter @GeoffM_GP

  • Correction notice This article has been updated since it was first published. The article type has been changed to Systematic review.

  • Contributors GM initiated this study, oversaw the design and conducted all parts of the review, and wrote the initial draft of the manuscript. JF-F and BW did the initial search and screen of ineligible papers. GM and HN searched data and screened ineligible papers from 2014 to 1017. All other authors (MG, HS, LM, PM, CJ, MA, PY) contributed to the design, the development of the separate research questions, read abstracts, determined eligibility, read full manuscripts, did quality assessments and data extraction, and approved the final manuscript.

  • Funding This paper was funded by Royal Australian College of General Practitioner/HCF grant in 2013.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.