bloomer mj, moss c & cross wm (2011)Journal of Nursing and Healthcare of Chronic Illness 3, 165–173
End-of-life care in acute hospitals: an integrative literature review
Aim. To explore and define the current status of published literature related to end-of-life care in the acute hospital setting.
Background. Chronic illness is now the leading cause of death in developed countries because the population is and the advances in health care delivery. In Australia, two-third of people with chronic illness who die is inpatient in an acute hospital. Whilst most of the attention regarding end-of-life care and dying has focused on palliative care settings, there is a need to consider those dying outside of specialist palliative care settings.
Design. Integrative literature review.
Methods. Using an integrative approach, electronic searches of selected databases were undertaken, using comprehensive search terms. Secondary sources were also identified through the reference lists of retrieved literature.
Results. The literature review identified a lack of high-quality scientific evidence to direct end-of-life care as delivered in the acute (non-palliative care) sector. From the literature available, five core themes/issues were generated to reflect the literature.
Conclusions. There is a need for research to investigate end-of-life care in the acute hospital setting. Care pathways are in use to varying degrees without validation, but implementation alone does not ensure quality end-of-life care. Cultural factors can impede/impact practice and the ability of clinicians to recognise ‘active dying’ has been shown to be problematic, delaying the transition to end-of-life care.
Relevance to clinical practice. Given that the majority of deaths in Australia occur in acute hospital settings, clinicians should be delivering expert life-sustaining acute care and also quality end-of-life care.
Care of the person at end-of-life (EOL) has changed dramatically over the last century (Thompson et al. 2006). In Australia, there was over 140 000 deaths in 2008 (Australian Bureau of Statistics 2009) and the 2006 figures tell us that around two-third of deaths occur in hospital (Australian Institute of Health and Welfare 2010). Moreover, it is estimated that almost three quarters of these deaths could have been anticipated (Palliative Care Australia 2008).
With the population of developed countries ageing, longer life expectancy also comes as a result of advances in medical treatment and health care delivery (World Health Organization 2004). These advances are increasing longevity, the pattern of disease worldwide is changing (World Health Organization 2004) and chronic illness is now the leading cause of death (Thompson et al. 2006), where chronic illness is defined as illness that persists over time with the person’s health status fluctuating between maximal functioning and serious, even life-threatening health relapses (Crisp & Taylor 2009). Treatment advances mean that many people are living longer with a chronic illness and in some cases they are living with multiple chronic illnesses (O’Grady 2005). Whilst most of the attention regarding EOL care and dying has focused on people with cancer, many others, such as those with chronic illness and chronic organ failure are similarly in need of EOL care (Department of Health 2009).
In Australia, acute care hospitals still remain as the most common place of death (Heyland et al. 2005) with figures showing that 50–60% of those who die in Australia, actually die in a hospital setting (Jennings 2005, Foreman et al. 2006, Tabor et al. 2007, Department of Health, 2009). This is despite findings from a study by Murphy et al. (2007), which reported that when asked the majority of people stated that their preference was to die at home, rather than in hospital. Society’s expectations of care at EOL impacts significantly on how healthcare is viewed and the current reality of service delivery may not meet the wishes of society.
Given the diverse origins of data related to EOL care and the lack of higher level evidence in this area, an integrative literature review was considered to be the most appropriate method to enquire, explore and create new meaning from the available literature which comes from a variety of sources, including scientific, political, humanitarian and anecdotal sources. An integrative literature review is used to determine the current state of knowledge and identify any knowledge gaps through the investigation of literature (Whittemore 2005) generated previously from diverse methodologies, where both experimental and non-experimental research and theoretical papers are included (Whittemore & Knafl 2005, Schneider et al. 2007). By re-interpreting and re-analysing text data (Pope et al. 2007), this integration has the potential to build nursing science, inform future research, care delivery and policy (Whittemore & Knafl 2005).
Well-defined literature searches are critical in enhancing the rigour of the integrative review (Whittemore & Knafl 2005), because incomplete and biased searches can result in an inadequate search and inaccurate findings (Conn et al. 2003). A search was carried out using CINAHL, Medline, Google Scholar and Google, for literature published in English between 2005–2010, however, if other publications were deemed particularly relevant and important or were regularly cited in other retrieved publications, but published before 2005, they may have been included also. The rationale for a Google search was to ensure that non-research publications, such as government/private reports/guidelines/position statements etc. were identified, as these may make direct recommendations for EOL care and not be identified via an academic database search.
The search terms were limited to end-of-life/end-of-life/EOL/death/dying/terminal care combined with hospital and articles retrieved from these searches were selected based on their relevance to care provided in acute care hospital environments. Any literature that more specifically related to specialist palliative care or palliative care or hospice environments was excluded. Further, associated searches were undertaken for literature on ‘chronic illness’ and statistics related to ‘place of dying’ to assist to understand the magnitude of this situation under review. A total of 73 publications were included in this integrative review, of which 41 were research publications and 32 were non-research publications.
After these initial retrievals were undertaken and from reading the literature retrieved, it became clear it was impossible to explore EOL care in acute hospitals without giving consideration to the Liverpool care pathway (LCP). The LCP was developed in the UK from a collaboration between a hospital-based palliative care team and the Marie Curie Hospice Liverpool to assist in transfer of the best practice model of care of the dying patient from the hospice setting to the acute setting (Murphy et al. 2007). Fourteen publications relating to LCP use were included.
Assessing the quality of sources in an integrative review is complex and challenging due to the diversity of studies and publications that can be included in this type of review (Whittemore & Knafl 2005); however, each publication source included was evaluated on its own merit, with consideration given to study design or conclusions drawn, or the public integrity of the organisation who produced the publication. Any publications that were deemed to be questionable or lacking credibility were also excluded.
This integrative literature review identified five core themes/issues that are reported to impact directly on the care of the dying in the acute hospital environment.
The terminology used to refer to care of the dying
Common and clearly defined terminology related to EOL care is a crucial foundation for ensuring consistent quality care (Goodridge 2010). The terminology applied to care of people dying in the clinical setting creates confusion and ambiguity, which may contribute to inequity in care delivery. Most notable is the relationship between the dying and ‘palliative care’. Many scholarly articles refer to ‘end-of-life’ and ‘palliative care’ interchangeably, as well as other terms such as ‘terminal care’, ‘hospice care’ and ‘care of the dying’ (Lorenz et al. 2005, Thompson et al. 2006, Murphy et al. 2007). There is a need for differentiation between palliative care and EOL care to ensure clinicians are clear about the intent (Goodridge 2010). EOL care in the acute setting may take place over a very short time frame, where death is recognisably imminent (Lorenz et al. 2005, Pugh et al. 2010, Norton et al. in press), for example where the patient is dying from an unexpected event or illness (Hardy et al. 2007) or where little time passes between diagnosis and death (Norton et al. in press). In this way, a planned palliative care approach to dying is not possible, thus highlighting the clear difference between palliative care and EOL care delivery. The ambiguity surrounding this terminology hinders development and delivery of EOL care (O’Grady 2005).
Although the World Health Organization (WHO) (World Health Organization, 2009) defines palliative care, there is no WHO definition available for EOL care. Lunney et al. (2003) define EOL care in a way that differentiates it from palliative care. EOL care is that which focuses on the last days and hours-of-life and is not preceded by a planned coordinated team approach to symptom control and disease progression (Lunney et al. 2003). This definition more appropriately describes care for dying, where previous plans about the dying process had not been made, or where the person is dying from an unexpected sudden event. Irrespective of the definition, what is explicit is the need for the dying person to have access to the right care at the right time.
How fear of death and dying influences actions
In 2009, a policy brief was published (Tilden & Thompson 2009) which reports that individuals are not only afraid of death, but they are also concerned about how they will die and the possibility of ‘suffering’, more than the death itself. Further, a significant issue impeding quality EOL care for the dying worldwide stems from the discordance between the goal of medicine ‘to cure’ and the acceptance that death is the most likely and sometimes the best option (Tilden & Thompson 2009). Whilst the United States leads the way in the medicalisation of death (Tilden & Thompson 2009), the Australian health system is similarly designed to treat and cure disease and to provide short-term episodic care for acute illness.
Similarly, although all people are mortal and will die, the beliefs and culture surrounding death needs to be acknowledged. An interesting finding from the literature is that there is a widespread reluctance to talk about death (Department of Health 2009), where death is considered a ‘taboo’ topic (Grady 2000, Wasserman 2008) and as a result talking about death and dying is not done enough (Ingleton et al. 2009).
Death of the aged and frail may be more acceptable than death of the young. Likewise, death as a result of a cancer diagnosis is perhaps more acknowledged or accepted as a possible outcome; however, the same acceptance has not been afforded to other chronic illnesses or disease states which are also likely to result in death. Nevertheless, the likely outcome for people with chronic illness is death either at home or in acute settings. Some may receive palliative care in specialist palliative care settings, but the majority will not and will subsequently die in acute settings. But when death is not expected, recognising or acceptance that death is the likely outcome is a difficult process for individuals, fraught with uncertainty and emotions (Thompson et al. 2006), in particular, where a person is facing their own imminent death or that of a close loved one.
According to Murray et al. (2005), acknowledging and openly discussing illness trajectories with patients and their families enables them to understand how they may die and make appropriate plans. The implications for service planning and development means that different models of care will be required for different illness trajectories and that planning may assist to avoid hospitalisation. ‘Now-a-days, we can cure some diseases and manage others effectively. Where we cannot alter the course of events we must at least (when the patient so wishes) predict sensitively and together plan care for better or for worse’ (Murray et al. 2005, p. 1011).
In tertiary care settings, the overall momentum in care delivery is towards life extension (Chen 2008). The focus is on resuscitation and assertive treatment to restore health, increase survival and avoid death (Aleksandric & Hanson 2010a). For those patients who are in the dying phase, aiming for cure is not only unrealistic, but may result in over-inflated expectations for patients and family, situations of conflict and inappropriate or inadequate care (Aleksandric & Hanson 2010a). This failure to recognise or acknowledge treatment futility (Chan & Webster 2010) results in a system, where clinicians miss the opportunity to deliver timely and appropriate EOL care (Murphy et al. 2007).
Transitioning to an EOL care focus
The essential first step in planning and providing EOL care is acceptance that death is the likely outcome by the patient, family and clinicians (Thompson et al. 2006). This acknowledgement in itself can be a difficult process for all; the treating team, patients, their family and significant others (Duggleby & Wright 2004, Jackson et al. 2006, Thompson et al. 2006). Furthermore, the difficulty in accurate prognostication by physicians and ‘waffling’ mean that patients and families are often given a broad range of survival estimates that create more uncertainty than certainty (Thompson et al. 2006).
When symptoms are out of control (Buckley & Herth 2004), when the patient has pain (Appelin & Berterop 2004), when there is a fear of addiction to strong drugs (Tilden & Thompson 2009) and when patients and their families fear a hastening of death (Arnold 2004), acceptance of a change in care direction can be traumatic and frightening for patients and also their families.
Tilden and Thompson (2009) describe how families and often patients themselves resist the option of transfer to a hospice setting, which implies acceptance of death and is tantamount to ‘giving up’; a situation similarly described in other literature, where changing focus for curative treatment to EOL care results in patients and family losing ‘hope’ (Jackson et al. 2006, Kylma et al. 2009). Thompson et al. (2006) also explains that the uncertainty and emotions experienced at this time make acceptance of death complicated and a clear cut transition from curative care to an EOL care is rare.
Notably, Thompson et al. (2006) found that clinicians also described losing hope once the active treatment was discontinued for their patient. The recognition by clinicians that treatment is futile is not easy (Chan & Webster 2010) and the decision to stop a life-sustaining treatment, or withhold a treatment in the first place can be very difficult and a source of moral distress (Yam et al. 2001). In Western societies, where death has become medicalised and curative procedures are often prioritised ahead of palliation (Economist Intelligence Unit 2010), many clinicians still report regarding patient death as their ‘failure’ or as a result of their inadequacy (Boroujeni et al. 2008), rather than the inevitable outcome of disease (General Medical Council 2010).
Regardless of the difficulties and emotions experienced when transitioning from curative care to EOL care, what should be of primary concern is that ‘Good care at the end-of-life and a dignified death should be regarded as basic human rights …’ (Clark 2007, p. 108), where the dying patient and their family are supported in the respectful completion of the life cycle (Wasserman 2008) and, where care is person-centred (Murphy et al. 2007) irrespective of the illness timeline, diagnosis, age or any other variable.
Knox and Vereb (2005) suggest a change to the curative focus of medical care as a strategy that may improve acceptance and understanding of EOL care. Replacing the term ‘Do not Resuscitate’ with the term ‘Allow Natural Death’, which reflects that death is the likely outcome, may assist in enabling clinicians, their patients and families, to see EOL care as an empowered choice, rather than the end result of failed curative care.
For physicians, even when the decision is made to provide EOL care in some parts of the US health system, certification that the patient’s life expectancy is less than six months is necessary and agreement by the patient, physician and family to forego any further curative care is essential before a patient is eligible (Weiner & Tilly 2003, Higginson 2005).
Multiple organisations, national and international, government and non-government have developed position statements, standards for care, care pathways, guidelines and recommendations designed to inform the provision of EOL care. A 2010 review from the Cochrane Collaboration reported, however, that there was insufficient data at present to make recommendations regarding the development and evaluation of EOL care pathways, due to the lack of evidence from randomised control trials (RCTs), quasi-randomised trials or high quality controlled before and after studies (Chan & Webster 2010). Studies of this type that deal with the dying are difficult and fraught with ethical and methodological issues (Fowell et al. 2004). As a result, current guidelines, recommendations, standards and position statements are not supported by a strong evidence base (Goodridge 2010) and instead, are reliant on lower levels of evidence to inform practice and guide care.
These documents describe best available practice for care of people at EOL. They are used as a resource and contribute to the existing knowledge base for clinical competence. Moreover, they provide a framework that supports clinical decision-making (Taylor & Randall 2007) and promotes practice change (Ellershaw 2002, Curtiss 2004). Importantly, they promote practice change in mainstream acute hospitals, where many dying patients and their families (Jack et al. 2003) would benefit from guided EOL care, not just in specialist palliative care facilities.
In the United States, the National Consensus Project for Quality Palliative Care released the second edition of Clinical Practice Guidelines for Quality Palliative Care (2004), which aims to promote quality and reduce the variation in new and existing programs and encourage continuity and consistency of care across all healthcare delivery settings. Whilst there is an obvious synergy between palliative care and EOL care, the differences between palliative and EOL care means that palliative care clinical practice guidelines are not entirely appropriate for the patient in the acute care setting who is imminently about to die, with no known/anticipated illness trajectory leading to death, prohibiting a more planned response. Having said that Domain 7 of the Guidelines relates more specifically to ‘Care of the Imminently Dying Patient’ and suggests that the signs and symptoms of impending death are recognised and communicated, that post death care is delivered in a respectful manner and that a post death bereavement plan is activated (National Consensus Project for Quality Palliative 2004), all of which is relevant EOL care in the acute care setting.
Recognising the lack of high-quality evidence, The American College of Physicians has also released recommendations for care delivery at EOL (Huntzinger 2008). These recommendations focus on symptom management for common symptoms, such as pain, dyspnoea and depression, ensuring regular assessment and treatment with therapies of proven effectiveness. They also encourage clinicians to ensure advance care plans (ACP) are initiated for all patients as early as possible and re-assessed when significant clinical changes occur (Huntzinger 2008). Nevertheless, the unpredictability of acute illness means that some patients will be near dying and not have considered or used an ACP.
In Australia, the Respecting Patient Choices Program was launched to promote the use of ACPs and ensure that the patients have the ability to make their wishes known about future care in the event that they lose capacity to communicate (Seal 2007), however the program has not resulted in significant uptake, with fewer than 1% of acute hospital case notes in South Australia containing an advanced directive (AD) (Harrison Health Research 2004 as cited by Seal 2007). Other studies, which looked at the prevalence of ADs in Australia also found that the presence of an AD was the exception, rather than the rule (Bezzina 2009) with very low levels of formal planning (Nair et al. 2000). This may be related to cultural issues (Johnstone & Kanitsaki 2009) that make planning for death unpopular or that many people die without luxury of advanced warning (Norton et al. in press). But also, for those patients not afforded advance warning of death, time is not available to consider AD as a part of planning for their death.
In the absence of an ACP or AD, or when decision-making is complex, NSW (New South Wales) Health, Australia, released Guidelines for EOL care decision-making (NSW Health 2005), where the goal is to assist clinicians to reach consensus in decision-making, where the patient and family are included in the process. Given that family inclusion is promoted by some (American Association of Colleges of Nursing 2004, Cook et al. 2006, Truog et al. 2008), it must be acknowledged that family/significant others do not always wish to be included or involved in decision-making, where the negative impact of sharing in these processes is significant and results in symptoms of anxiety and depression (Pochard et al. 2005) and post-traumatic stress disorder (Azoulay et al. 2005). Although the guidelines are clear and supportive of collaborative decision-making; they do not guide care.
Care/clinical pathways have also become a common source of guidance for clinicians, but the terms used to describe a clinical pathway number more than 80 (Kinsman et al. 2010), where the goal of such a pathway is to base the care of the patient on the best available evidence and establish an institutional commitment of care for a particular diagnosis (Verdu et al. 2009). In the United Kingdom, the Department of Health developed an EOL care pathway, which identifies five steps to care of those approaching EOL. The first step in this care pathway is to identify those who are approaching EOL so that ‘good care’ can be made available to all persons approaching EOL regardless of their diagnosis, rather than just those with cancer who are most commonly the recipients of palliative and EOL care (Department of Health 2009). Again, the luxury of forewarning or advance notice is not afforded to all dying patients and their families and the recognition of condition changes that may indicate active dying has been reported as a difficult step for clinicians (Jack et al. 2003, Alsop 2010).
Care pathways encourage forward planning, effective coordination of care and central monitoring (Taylor & Randall 2007, Department of Health, 2009) and whilst ideal, may have little to no relevance to those persons rapidly approaching death in the acute healthcare system. For those at their last days of life, the LCP is recommended for guiding care (Department of Health 2009), as it was designed to assist in transfer of the best practice model of care of the dying patient from the hospice setting to the acute setting (Murphy et al. 2007), thus ensuring that the patient and family are provided with the best possible care at this time (Ellershaw & Ward 2003, Jack et al. 2003). Ellershaw explains that previously, care of the dying was the goal and mainstay of hospice care, which resulted in feeling that the only place patients can die well is in a hospice and a de-skilling of other staff. This creates an affirmation of the sense of ‘care failure’ for clinicians working outside of the hospice sector (Ellershaw 2007), so the need for urgent integration of such a pathway into acute hospital environments becomes an even higher priority.
Many iterations, adaptations and modifications of the LCP have been developed and tailored for various hospital and healthcare settings worldwide with a variety of challenges and successes. What is common, however, is that the LCP focuses on the physical, psychological, spiritual, social and religious aspects of care of the dying, the information needs of the patient and their carers (Ellershaw 2007), as well as care of the family before and after patient death (Ellershaw & Ward 2003).
Although the focus of this integrative review is on EOL care in the acute hospital sector, it is interesting to note that in the residential aged care setting the introduction and implementation of the LCP to guide care delivery to the dying is reported as a positive step (Duffy & Woodland 2006). Given that the median length of stay in these facilities ranged from 1–2 years (Bebbington et al. 1998) and that the population was already aged and requiring nursing support, the acceptance of a pathway for the dying may not have been considered such a vast change from attitudes and care goals already in place in the facility. Improvement in cross-boundary communication was reported as another positive outcome, leading to improved care (Hockley et al. 2005, Duffy & Woodland 2006), where staff reported that the LCP empowered them to speak more openly and proactively with relatives and promote EOL care, resulting in an overall reduction in the number of residents being transferred to acute facilities only to die in the Emergency Department (ED) (Duffy & Woodland 2006). The introduction of the LCP into residential aged care facilities also resulted in staff feeling empowered and prepared to enter into conversations residents, families and other clinicians about death and dying (Hockley et al. 2005).
Now to focus on the acute hospital sector, the results are similar. The LCP has been implemented and its use evaluated, where it was used to direct care for the acutely dying in the ED, such as those patients who were postcardiac arrest. This study highlighted increased staff satisfaction when the LCP was used to guide care for the dying and more importantly, it showed that patient care improved (Paterson et al. 2009).
In Australia, an adaptation of the LCP was also implemented and evaluated in medical wards with the aim to facilitate ‘good deaths’ by acknowledging the inevitability of a patient’s death and guiding care. The study showed marked improvements in care for the dying patients in terms of involvement of support service personnel, such as social work and pastoral care and the facilitation of communication (Jackson et al. 2009).
The LCP was also implemented and evaluated in a large Dutch cancer hospital (Verbeek et al. 2006). The findings were largely positive however Verbeek et al. study report that in 30% of cases some care goals/recommendations for the after death period were marked as ‘not applicable’ and a relatively high number of missing assessments were reported (Verbeek et al. 2006), suggesting that clinicians chose to disregard recommendations on the pathway for alternate/absent care. The authors suggest that ongoing education is a necessary requirement for the LCP to maintain its function as a tool that facilitates quality care of the dying (Verbeek et al. 2006).
Improvements in the quality and completeness of documentation were reported in another study following implementation and evaluation of the LCP into hospital, nursing home and home settings. Whilst this does not necessarily translate to an improvement in care delivery, improvements in care documentation is still a positive outcome from the LCP (Verbeek et al. 2008).
Clinical judgment in dying
Whilst evaluations of the LCP are positive, what they also show is the continued need for clinical judgment. Active dying manifests in each and every patient differently and even when clear criteria apply, not all dying patients will fit. How does a dying patient present or look differently than a chronically ill patient who is not dying? Pugh et al. (2010) suggest the commencement of the LCP when death is expected within one to two days, based upon the patient meeting at least two of the following four criteria: (1) bed bound, (2) semi-comatose, (3) only able to take sips of fluids and (4) is no longer able to take tablets. Some dying patients will meet at least two of these criteria, many will not. Conversely, some patients may also have met two or more of these criteria for a long time, but not be dying, such as progressive long-term neurological conditions. Therefore, even though the criteria may be helpful if dying is not recognised, others may still miss the opportunity to receive quality care at the EOL. Verbeek et al. (2006) report that clinicians have difficulty in diagnosing impending death and thus commencing EOL care. For some clinicians, acceptance that death is likely difficult to accept, so their focus remains on curative treatment (Thompson et al. 2006).
Implications for clinical practice
This integrative literature review has highlighted several factors that are imperative if care for people who are dying in the acute sector of health care is to improve. First, there is a need for further research to investigate the care outcomes for those receiving current standard care vs. those whose care is guided by a care pathway, however this type of study may be fraught with ethical issues, given that current care delivery is determined by lower levels of evidence.
Second, care pathways such as the LCP, aim to assist clinicians to provide better care for the dying patient and aim to ensure dignified death is as important as saving lives. However, the implementation of care pathways/care guidelines alone is not enough to ensure ongoing success in mainstream acute hospital wards. This literature review showed the degree of fluidity in results evident in evaluations of the LCP across a variety of settings and that the nursing/medical and healthcare cultures and workplace systems/process can impede practice change in this area. As a result, persons dying from an unexpected illness, or where palliative care services were not already in place, will not receive EOL care comparable to that delivered in palliative care settings.
Third, despite the introduction of ACP and ADs, their prevalence in practice is negligible. Notwithstanding their promotion, individuals are not using them, resulting in little more patient self-determination than previously. Given that the population is ageing and chronic illness is the leading cause of death, the need for individuals to use ADs is now more urgent than ever. As previously stated, unless the prognosis is acknowledged as terminal, both patients and clinicians are unlikely to consider death as inevitable.
Fourth, the ability of clinicians to recognise ‘active dying’ has also shown to be problematic, particularly if the clinical presentation of the chronically ill patient, sometimes with multiple diagnoses, may not be too different from the dying. This creates uncertainty for clinicians about when to transition from life-saving treatment to dignified and quality care of the dying.