The Liverpool Care Pathway for the Dying Patient: a critical analysis of its rise, demise and legacy in England

Background: The Liverpool Care Pathway for the Dying Patient (‘LCP’) was an integrated care pathway (ICP) recommended by successive governments in England and Wales to improve end-of-life care. It was discontinued in 2014 following mounting criticism and a national review. Understanding the problems encountered in the roll out of the LCP has crucial importance for future policy making in end of life care. We provide an in-depth account of LCP development and implementation with explanatory theoretical perspectives. We address three critical questions: 1) why and how did the LCP come to prominence as a vehicle of policy and practice? 2) what factors contributed to its demise? 3) what immediate implications and lessons resulted from its withdrawal? Methods: We use primary and secondary sources in the public domain to assemble a critical and historical review. We also draw on the ‘boundary object’ concept and on wider analyses of the use of ICPs. Results: The rapidity of transfer and translation of the LCP reflected uncritical enthusiasm for ICPs in the early 2000s. While the LCP had some weaknesses in its formulation and implementation, it became the bearer of responsibility for all aspects of NHS end-of-life care. It exposed fault lines in the NHS, provided a platform for debates about the ‘evidence’ required to underpin innovations in palliative care and became a conduit of discord about ‘good’ or ‘bad’ practice in care of the dying. It also fostered a previously unseen critique of assumptions within palliative care. Conclusions: In contrast to most observers of the LCP story who refer to the dangers of scaling up clinical interventions without an evidence base, we call for greater assessment of the wider risks and more careful consideration of the unintended consequences that might result from the roll out of new end-of-life interventions.


Introduction
Major policy innovations covering a whole jurisdiction are rare in palliative care. When one such intervention is introduced with government support and with a vigorous programme of implementation, but then runs into significant difficulties, it is vital to make sense of the factors at work. This paper draws on the concept of 'boundary objects' 1 to provide an in-depth analysis of an unsuccessful attempt to provide palliative care at scale. The Liverpool Care Pathway for the Dying Patient (LCP) was an intervention based on an integrated care pathway. It grew out of the hospice context and over more than a decade was promoted across the health care system in the United Kingdom before it was suddenly withdrawn from use. It has become a critical learning point for those charged with delivering end of life care to large populations and there is continued interest in trying to make sense of its rise and demise.
There is a wider context to such concerns. The importance of palliative care interventions to relieve suffering in the face of advanced disease and of death, has gained policy, clinical and academic endorsement worldwide 2,3 . Beginning in the 1960s with the emergence of new hospice programmes, the modern field of palliative and end of life care grew rapidly in the later decades of the twentieth century and has continued to make progress, attracting wide interest and support, and extending its reach 4 . The World Health Organization first defined palliative care in 1990 5 and in 2014 the World Health Assembly issued a resolution calling on all governments to adopt policies to support the delivery of palliative care across the life course and in all relevant institutional and community settings 6 . Nevertheless, palliative care faces many challenges. It is still weakly developed in many low and middle income countries, and at the same time advanced and well-resourced health care systems are struggling to deliver its benefits, that have been demonstrated in specialist settings, across the wider spectrum of health and social care services.
The epidemiology of dying is also changing. Increasingly for many, death will follow an extended period of uncertainty, frailty and multiple morbidity in advanced old age. In contrast, the rapidly progressing downward trajectory of dying with its clear point of entry to the dying phase, which was central to the original hospice model of cancer palliative care, will become less common 7 . Already a series of complex challenges is emerging around prognostication, communication and the planning of care for the gravely ill or dying person, especially in hospitals 8 . The scenario has given rise to numerous attempts to take specialist palliative care knowledge and apply it 'at scale' within the mainstream of the health and social care system -in hospitals, care homes, and in the community.
The LCP was a specific and concerted attempt to respond to these challenges. It was first described in a publication in 1997 9 as a means of transferring key principles derived from hospice care into general health care settings, such as hospitals and care homes. It was endorsed by successive governments and in particular adopted as a key initiative within a National Programme and Strategy for End of Life Care in England and Wales 10 . Then, in the face of mounting criticism and following a government requested national review, led by Baroness Julia Neuberger in 2013 11 , it was abruptly discontinued. An extensive debate ensued, seeking to understand how this had happened and why the LCP had 'failed'.
Our purpose in subjecting the LCP story to detailed scrutiny is to move beyond the positions so far seen in the literature. These polarise between simplistic and retrospective 'blaming' of its limitations, often from sources that had hitherto been silent on the matter; or on the other hand, regret for its demise. Both these tendencies are evident in the many published commentaries, particularly in the clinical literature after 2013 12,13 . In contrast, we seek to provide a theoretically informed analysis of the rise and fall of the LCP and to conclude with comments that might be relevant to future policy and practice in the context of complex health systems in which the character of terminal illness is undergoing significant change. We aim to answer the following questions: 1) why and how did the LCP come to prominence as a vehicle of policy and practice 2) what factors contributed to its demise? 3) what immediate implications and lessons resulted from its withdrawal?

Background
Integrated care pathways Integrated care pathways (Box 1) are complex interventions to enable the organization of health care for specific groups of patients, often in the context of time limited decision-making 14 . Their use commonly involves structured documents outlining essential steps in care to be followed by members of multidisciplinary teams involved with particular groups of patients. They can also be used to introduce clinical guidelines and to provide a

Amendments from Version 1
In our revised version 2, we have: • Added some text to the Abstract to bring out the policy implications of the work.
• Added some new text to the Introduction to clarify the purpose of the work, and re-ordered some sentences in the introduction. • Deleted a 'note to self' in the references.

See referee reports
REVISED framework for audit 15 . A key intent behind their use is to standardise or 'rationalise' care for a particular issue, problem or clinical care episode 16,17 , although their potential to enhance 'person' centred care is also often emphasised, creating a degree of ambivalence about their core purpose 18,19 . Integrated pathways can go by various names: critical pathways, clinical pathways, and case management plans 14 . Originating in the USA in the 1980s in the context of a 'payment for service' health care system, they were part of a wider movement to manage concerns about spiralling health care costs, whilst sustaining care quality and improvement. Their appearance in the United Kingdom (UK) in the 1990s, mirrored a 'modernisation' emphasis on clinical efficiency and a drive to ensure the application in everyday practice of standardised national guidelines 20 . The integrative focus of pathways refers to the intent to use them as a means of formalising multidisciplinary channels of communication and of enabling professionals to work together across disciplinary or care setting boundaries (for example in health and social care) 21 .
In a review of the evidence base behind integrated care pathways in stroke care, Allen and Rixson 22 note the widely cited aims of integrated pathways within NHS discourse: '…the right people, doing the right things, in the right order, at the right time, in the right place, with the right outcome'(p81). The reviewers conclude that integrated pathways in stroke care are potentially successful in the acute context, where the patient's illness and care trajectory are reasonably predictable; however, their value in rehabilitative care, where recovery pathways are variable, is less clear. In a similar vein, Pinder et al. 19 describe how pathways 'abstract the patient and reify the condition' (p765), creating a tendency to '….omit the plasticity of patients' personal circumstances and lived experience, providing no map of the terrain that the ill person has to traverse' (p775). In this sense, pathways can have a powerful influence on shaping practice and the way in which practitioners understand clinical issues, and the most appropriate response to them.
An early review of integrated care pathways published in the British Medical Journal in 1998 23 summarises them as 'task oriented care plans' offering not only the essential steps in patient care but also a structured means of implementing local protocols of care based on evidence based guidelines and analysing why care may fall short of, or vary from, any adopted standards.
In this context, the LCP was a clear example of an integrated care pathway which reflected all of these features, although as we will see, debate later emerged about the quality of evidence on which its key features were based and whether it was effective in achieving its goals. Such issues were foreshadowed in an observation made from the 1998 review: '…despite the sound principles which underlie care pathways, few evaluations have been done of the cost of developing and implementing them and their effectiveness in changing practice and improving outcomes' 23 (p133).

'Boundary objects'
A boundary object is an artefact that provides a means of sharing ideas, technologies and practices across and between organisational settings, cultures and communities 24 . The interest in boundary objects reflects a concern to articulate the meanings and perspectives of actors from a variety of 'social worlds' or 'sites of difference' 25 . Originally conceptualised in 1989 by Star 26 as '…objects which both inhabit several intersecting social worlds … and satisfy the informational requirements of each of them' (p393), boundary objects were soon categorised along four dimensions: repositories, ideal types, maps and standardised forms 1 . Integrated care pathways have been described by Allen 27 as 'classic examples' (p305) of boundary objects because they straddle clinical, managerial and user interests. They are thus potentially associated with differing meanings across the particular groups involved.
Following Carlile's 28 definition of a successful boundary object as one which provides a shared language, allows concerns to be expressed and enhances knowledge, Fox 29 has made the distinction between 'positive' and 'negative' boundary objects, each of which have consequences for the transformation of knowledge and practice within a community. This highlights examples where implementation is able to proceed or to where it is blocked in some way. Both are important.
Boundary objects are therefore subjects of and for reinterpretation and renegotiation, often through a process of conflict, although their dynamic aspects have been somewhat neglected 30 . While boundary objects are capable of bridging different perspectives, they can be associated with negative unintended consequences resulting in inhibition of the very improvements they were intended to facilitate 30 . For Carlile, the production of knowledge across boundaries entails several activities: transfer (involving information and knowledge processing), translation (involving interpretation and the use of new information and knowledge), and transformation, which occurs when the interests of actors diverge, leading to power struggles over the legitimacy of the 'object' at hand 28 .

Methodological orientation
Building on the twin concepts of integrated care pathways and boundary objects, we set out to write a critical and historical account of the LCP, informed by relevant theory and thereby generating understanding that may inform future end of life interventions. We have made use solely of sources available in the public domain, namely: 1) Content emerging from the LCP programme -this includes guidance materials, LCP documentation, and writings concerned to support the use of the LCP.
2) Evidence from a spectrum of published studies on the use of the LCP -qualitative research and improvement project findings, surveys of practice and attitudes, clinical evaluations, and randomised trials.
3) Letters, articles, broadcasts, and online content found in the mass media and on social media. 4) Professional commentaries on the LCP, particularly those written in the aftermath of its withdrawal and published in clinical journals. 5) Content from and associated with the Neuberger report 11 into the use and efficacy of the LCP.
We contend that these sources are sufficient to answer our three research questions. We acknowledge that other lines of enquiry about the LCP are still to be pursued, and that these would require other methods. For example, oral history interviews might be useful in assessing the rise and demise of the LCP from the personal perspectives of key actors involved in the wider processes we describe in this paper. We suggest that it may still be too early for such work to be conducted, as personal investments in the LCP, individual and organisational reputations, questions of anonymity, confidentiality, and other sensitivities may be sufficiently marked as to inhibit the successful conduct of such work. Primary and secondary sources in the public domain still offer much material for analysis however, and allow the short to medium term barriers that would arise from research involving human subjects, to be overcome.

History and development
The LCP was originally formulated during the 1990s at the Royal Liverpool University Hospitals NHS Trust and the Marie Curie Hospice in Liverpool and underwent incremental development and revision over the next 15 years 31 . The LCP was subject to annual review by a multi-disciplinary steering group that included carer representation. Utilising the quality improvement methodology of 'Plan, do, study, act' (PDSA) 32 , the LCP programme and documentation was updated and revised over time in response to feedback, with 12 versions published in total.
The LCP was focused on patients in the 'dying phase' (described as the last 48 hours of life in the original LCP documentation). Up to version 11, recognition of the dying phase was recommended to be the responsibility of the multi-professional team agreeing that the patient was dying and when at least two of the following features were present: the patient is bedridden; the patient is semi-comatose; the patient is able to take only sips of fluids; the patient is no longer able to take tablets 9 . Following review of feedback by the multi-disciplinary steering group, version 12 removed these features and replaced them with an algorithm to guide recognition of the dying phase.
Once established that the patient was entering the dying phase, the use of the LCP entailed the following four steps: initial assessment; care planning against suggested 'goals' of care; ongoing assessment and care after death 33  Perhaps in anticipation of the risk that the LCP could be seen as a set of instructions for care rather than a guide to individualised decision-making, the importance of 'variance recording' was emphasised (see Box 1). For example, in their text book about the LCP, Ellershaw and Wilkinson noted that variance recording in the context of integrated care pathways is: …a mechanism by which a seemingly process driven approach to care can be tempered in line with individual patient need. The potential to use clinical skill and judgement to deviate from the suggested plan of care in response to an individual patient's needs makes the LCP a more flexible and practical document. Variance recording tells the story of the patient's journey and current condition 35 (p17).
The context of the development of the LCP was twofold: first, recognition that NHS hospital care of the dying fell short of best practice as understood at the time, and second, a recognition that the shortfall in specialist palliative care resource in hospitals meant it was unrealistic to expect specialist palliative care teams to be involved with every patient 9 . The LCP was therefore about 'going to scale' with an approach that had to date been limited in its availability.

Transfer
The Marie Curie Palliative Care Institute in Liverpool took the lead role in disseminating information about the LCP, through series of publications 34,36-38 and a process of networking nationally and later internationally. The work was co-ordinated by a central LCP team and organised in a series of programmes or work streams. In a 2005 publication Ellershaw and Murphy describe the extent of this enterprise and how it worked: The LCP framework is now based at the Marie Curie Palliative Care Institute Liverpool and is focused on four programmes; these are non-cancer, bench marking, education and international. These four programmes are supported by an audit and a research team. The education programme for 2005 will incorporate 2500 health-care professionals. Five clinical champions and 13 clinical facilitators now support the central team in the education-spread programme. An annual national conference based on the theme of care of the dying acts as a focus for new developments and research in this field 39 (p133).
The LCP had become a major enterprise with its own dissemination and transfer needs. There was a great deal to be done to inculcate LCP knowledge and capacity at a local level. The central team encouraged a six to 12 month local implementation process for the LCP, described by Murphy in 2003 as entailing ten steps (see Box 2). This later became known as the 'Ten step continuous quality improvement programme' or CQUIP and was described as occurring in four phases: induction, implementation, dissemination and sustainability 40 . By this point, the LCP as boundary object seems to have become firmly established: strategically, clinically, organisationally.

Translation
The LCP was not a fixed entity. It went through a series of revisions as its nationwide use began to quicken. Its first iteration was reported in 1997, in the European Journal of Palliative Care 9 and the final 'generic' version was issued (version 12) in December 2009. The latter followed two years of consultation across the sector by the Marie Curie Palliative Care Institute, examination of two rounds of national audit data a 51,52 and consideration of criticisms that were by now emerging in the media about risks associated with the use of the LCP 53 . An associated 'data dictionary', perhaps seeking to reassure, provided detailed instructions about the use of the LCP to enable: …explicit and robust understanding of the core meaning of each of the goals of care and the rationale, required behaviour and correct coding of information 31 (p4).
National and locally produced documents associated with version 12 also affirmed that: …the responsibility for the use of the LCP generic document as part of a continuous quality improvement programme sits within the governance of an organisation and must be underpinned by a robust education and training programme 54 (p4). Peer reviewed evidence and a developing critique As the LCP became 'high profile' in the policy arena of end of life care in the UK, and especially so in England, it began to receive more attention from the academic research community and to be studied in other countries (notably the Netherlands). A key research criticism was that the LCP lacked an under-pinning 'gold standard' of evidence in the form of supporting data from randomised controlled trials (RCTs). The critique missed an important nuance: the LCP was styled quite explicitly as a quality improvement programme; something not usually associated with the more formalised approach of the RCT. This type of criticism was voiced openly in 2005, with publication of a letter in the journal Age and Ageing by a hospice doctor from Kent, Shah 57 , lamenting the absence of large scale RCT data to support the LCP. Shah was writing in response to an earlier research letter published in 2004 by Ellershaw's team 58 that reported descriptive data on use of the LCP on an acute stroke unit from a 'before and after' study of 20 clinical cases, which showed: …marked improvement in levels of documentation, including a change in the prescribing of medication 58 (p625).
In response, Shah raised the following criticisms: The LCP appears to have a potential to improve some of the patient and carer-centred outcomes of good death, but it would be helpful to have some evidence. ... Current circumstances almost resemble a setting of a multi-centre randomised trial in regards to LCP implementation. In hospitals of different sizes and variable set-ups there are wards implementing LCP, and others that are not. It will be useful to compare the patient-, carer-and staff-related outcomes of care of dying in intervention (LCP) and control (conventional care of the dying in hospitals) groups before the LCP is rolled out to more wards and hospitals. If we miss this opportunity, we will end up with an untested LCP accepted as a gold standard everywhere. We will then be unable to test its efficacy, as ethical approval will almost be impossible 57 (p197-8).
There is a reply to Shah by the LCP team in the same issue of Age and Ageing 59 , taking the line that the whole field of palliative care is evidence poor and also observing that the LCP is a template of care that: … promotes the spread of the palliative care approach in the dying phase to members of the generic team. In this way, it has the potential to impact on the 'culture' of the delivery of care to dying patients in a way that a relatively small team of palliative care professionals … could probably never do (p198).
Over the next four years, Ellershaw  No randomized controlled trials or meta-analyses were identified, although 26 studies of other types d were included: 15 from the UK; four from the Netherlands; three from the USA; two from Australia and one each in Ireland and China. Phillips et al. 83 concluded that there was some low level evidence suggesting that: pathways promote good practice in end of life care; increase accessibility in palliative care and promote better management of patient comfort. However, they observed a range of weaknesses of end of life care pathways from their review, including: a lack of association with palliative care approaches that are ideally in place before the dying phase and dependence of use upon timely recognition and diagnosis of dying in a context where only around 50% of deaths in acute care settings are predicted 83 .
Phillips et al. 83 also provided evidence from several studies of pathway implementation, indicating that key factors for successful implementation include: clinical education sessions and professionals with the necessary competences to use the pathway; strong clinical leadership; and the use of pathway 'facilitators'. The criticality of facilitators to the safe use of the pathway for patients and the development of palliative care capacity amongst health care professionals was especially emphasised, though the review noted that there had been a decrease in the numbers of pathway facilitators in the UK over time: The palliative care capabilities of the pathway facilitator appear to be central to ensuring that the dying patient's transition onto the pathway is appropriately negotiated and safely managed. The pathway facilitator also plays a key role in building the palliative care capacity of health professionals. Despite this positive relationship, the number of U.K. pathway facilitators actually decreased over time, reflecting a trend to use facilitators for a defined period during the path-way implementation phase 83 (p952).
In response to the growing disquiet over the LCP, Chan and Webster repeated their Cochrane review in 2013 84 but still found no evidence from randomised controlled trials, concluding that: …with sustained concerns about the safety of the pathway implementation and the lack of available evidence on important patient and relative outcomes, recommendations for the use of end-of-life pathways in caring for the dying cannot be made 84 (p1).

Transformation
The absence of 'gold standard' research evidence made it difficult to counter criticisms of the LCP that began to emerge beyond academia from 2009 onwards. Now more mainstream print media and the emerging power of social media were seen to combined effect as a range of 'stakeholders' colonised the LCP debate for a variety of political ends, ranging from 'pro-life' campaigners, to anti 'Obama care' activists in the USA. In between, the voices of people who had been bereaved also found a space. Many of these cited the LCP as a key factor in poor care, although others reported positive experiences of it.
A gathering 'storm' Thunder clouds associated with the LCP appeared in the wider public domain as early as September 2009 with the publication of a short but highly critical letter in the Daily Telegraph 85  In the case of the LCP, a health correspondent in the Daily Telegraph quickly picked up on the significance of the story and subsequently ran a series of short reports, the first of which had the memorable title: 'Sentenced to death on the NHS' 88 . It was sufficient to obscure the inclusion in the article of observations from a range of agencies in support of the LCP, including Marie Curie and the Department of Health. Other pieces quickly followed, usually taking the familiar form of personal storiesthe first of these from a woman who claimed her father had been wrongly placed on the LCP 89  My overriding objection to this paper centres on the use of the word euthanasia. It is suggested that widespread use of the LCP equates to institutionalised euthanasia, and implicit in this is an accusation that individual practitioners have killed their patients. To read this, as a doctor who has used the LCP, is very difficult. The accusation is made in the conclusion without any supporting evidence. The 'evidence' that is reviewed in the paper does not touch upon intentional killing. If we are regularly making inaccurate predictions (or prognoses), that is of course unacceptable and must be addressed, but the term euthanasia suggests that we are intentionally killing our patients. There is absolutely no evidence for this. This paper, and the thoughts behind it, sparked a huge controversy over end of life care in this country. I think it is methodologically weak and structurally flawed. I think it contains baseless conclusions, and is excessively liberal with emotive, hurtful accusations of intentional killing.
of life care began to be noticed. The essence of these centred on awareness of their relative's dying, lack of involvement in decision-making and poor quality of communication. The use at that time of comparatively new social media such as Facebook and Twitter enabled what might be seen as an essentially 'grass roots' movement to come to wider attention, and provided a means of networking between individuals and organisations. This in turn furnished the mainstream media with more material, including reports or 'think pieces' published in the Guardian Society 95 , the Daily Mail 92 and most other UK newspapers at regular intervals between 2010-2012. Most were critical and used extreme language. For example, the Daily Mail used the term 'Liverpool Killing Pathway' and went on to erroneously claim that the LCP was implicated in the deaths of young babies and children: the latter report was the subject of a Press Complaints Inquiry and was reported in the BMJ 96 . Perhaps reflecting either an inability to gain newspaper publication or a fear of being misreported, most commentators seeking to lend support to the LCP did so in the traditional medical and health care press, although many were also active on Twitter.
Some correspondents on social media apparently sought not only to air their complaints but also to 'bring down' the LCP. For example, one Twitter hashtag @NHSNaziHunters j echoed the language of the anti-government activist Lyndon Larouche reported earlier.
As the wide ranging media activity peaked, it culminated in a petition by the organisation Change.org for a national enquiry, which took the form of an open letter to the Health Secretary of the United Kingdom government k .
A rising sense of risk Controversy and criticism continued unabated in spite of efforts to lend support to the LCP by a range of organisations across the sector l , and reiteration in a variety of policy and practice documents that the LCP: …requires senior clinical decision making, communication, a management plan and regular reassessment. It is not a treatment in itself but a framework for managing treatment. It aims to support, but does not replace, clinical judgment. Communication, care and compassion must come from all the healthcare workers caring for an individual and their family 97 (p17).
In 2012 a report that half of all NHS trusts had received, or were due to receive, financial awards for demonstrating use of the pathway under a system known as 'Commissioning for Quality and Innovation' (CQUIN), was cited in the Daily Telegraph as evidence of doubtful use of the LCP, and claiming undue financial and bureaucratic influence of individual clinical practice 98 .
In January 2013, two stories were run in the Daily Telegraph reporting that a national enquiry into the LCP has been commissioned by the Norman Lamb, MP, Minister of State for Care Support. The first cited an interview with Dr Bee Wee (then president of the Association for Palliative Medicine): …some hospitals appeared to be treating the pathway as just another "thing to be done" rather than something to be handled with extreme care. She added that the cases which have come to light suggested that "packaging up" principles used in hospices for hospitals had caused difficulties. "There's a very big difference in the culture of hospitals," she said. "So the environment and the attention to training and support ongoing is an important point" 99 .
In the same article, Lamb, apparently capitulating to the growing storm, was reported as saying: I have committed to appoint an independent chair to review how end of life care is working and oversee the reviews into the LCP. …This review will also consider the value of locally set incentives, and whether they are leading to bad decisions or practice 99 .
As a national independent review panel was convened, key elements in the controversy were distilled into the content of a Channel 4 television programme in February 2013 called 'Death on the Wards', in the investigative series 'Dispatches'. This broadcast sought to establish the truth or otherwise about accusations in the media about the LCP. A month later a survey was published of 647 UK hospital doctors, jointly conducted by the BMJ and Channel 4 100 , which found that critical reports of the LCP were negatively affecting its use. Many clinicians responding to the survey expressed concerns about distress caused to relatives of dying people by 'scaremongering' in the press about the LCP. Many were also anxious about how to respond. One respondent wrote this: Negative press regarding LCP [the pathway] has caused additional distress for relatives at an already distressing time when their loved one is dying. This has caused a dilemma in judging if discussing the LCP will cause more distress than the benefit of being on the LCP for coordination of care in the dying phase 100 (p18).
The survey enquired into doctors' views about the accusation in the press that the LCP was used to 'save money' 101 . The vast majority of respondents (98%) did not think that resource considerations played a part in decisions to use the LCP, although most (58%) disagreed with the principle of using financial incentives provided to NHS Trusts for adoption of the LCP and other similar initiatives 100 .

The National Independent Review and its recommendations
The national independent review panel was chaired by Baroness Julia Neuberger and was given the brief of examining in detail …where the LCP is used properly patients die a peaceful and dignified death. But the review panel is also convinced, from what it has heard and read, that implementation of the LCP is not infrequently associated with poor care 11 (p7).
Overall, the panel made 44 recommendations, organised under 25 themes p , of which the very large majority provide an agenda to improve the quality of care for dying people and their relatives that extends far beyond the scope and remit of the LCP as originally conceived, and which at the time of writing this paper were still being developed, discussed and disseminated. As well as proposals to cease reference to the LCP because of confusion about the term 'pathway', the recommendations addressed deficiencies in: documentation; prognostication and diagnosis of dying; communication and clinical decision-making at the bedside (especially in relation to nursing practice surrounding clinically assisted and oral hydration and nutrition, provision of pain relief and sedatives), consent; care planning; cardio-pulmonary resuscitation and other ethical issues. Wider aspects were also considered, including the quality of the environment of care and staff resources and equipment. Likewise, a proposal was made for a system-wide strategic approach to be adopted to improve care of the dying, building on the foundations of the End of Life Strategy published five years earlier in 2008 10 , with end of life care incorporated into hospital inspections.
In the press release q accompanying the review, Baroness Neuberger stated that: All the major players in the health and care system, including the Government, need to do their part in reforming care for the dying, so that people everywhere can be sure they will be treated with respect and compassion, supported to die a peaceful, dignified death.
The panel concluded that use of the LCP should be phased out by July 2014, with the intention that it be replaced by a personalised 'end of life care plan' backed up by good practice guidance specific to disease groups. Interim guidance to this effect was published for doctors and nurses on 16 th July 2013 by NHS England 102 . On 15 th August, and after the publication of the Neuberger report, BBC Radio 4 broadcast a 30 minute programme assessing the rise and demise of the LCP and its future implications. There were contributions from family members of patients who had been put on the LCP, as well as from palliative care specialists, including Professor Ellershaw 103 .

New guidance develops
The independent review panel primarily directed its recommendations about the LCP to a range of national organisations with statutory or regulatory roles in health care (for example: NHS England, Care Quality Commission, Department of Health, NICE; General Medical Council). In England, quite quickly after the publication of the independent review, representatives from these organisations together those from a range of others, including charities, formed a coalition of 21 known as the 'Leadership alliance for the care of dying people' (LACDP). The stated purpose of the alliance was '…to take collective action to secure improvements in the consistency of care given in England to everyone in the last few hours or days of life and their families' 104 para 3, with the following objectives: first, to support all those involved in the care of dying people to respond to the findings of the review; and second, to be the focal point for the The Priorities for Care are that, when it is thought that a person may die within the next few days or hours: 1. This possibility is recognised and communicated clearly, decisions made and actions taken in accordance with the person's needs and wishes, and these are regularly reviewed and decisions revised accordingly. 2. Sensitive communication takes place between staff and the dying person, and those identified as important to them. 3. The dying person, and those identified as important to them, are involved in decisions about treatment and care to the extent that the dying person wants. 4. The needs of families and others identified as important to the dying person are actively explored, respected and met as far as possible.
5. An individual plan of care, which includes food and drink, symptom control and psychological, social and spiritual support, is agreed, co-ordinated and delivered with compassion.
The priorities were worded to direct attention to the importance of individualised assessment and care planning, and away from the standardised approach that the review panel found had sometimes been adopted in the use of the LCP. The alliance emphasised that '…where change is needed, it is in the practice of particular local organisations and staff' 104 (p7), with the role of national organisations to: '…require, encourage and support that change' 104 (p7). In addition, Alliance members agreed and published a 'commitment statement', setting out their individual and collective approaches to improved care in the last hours and days of life. The following year an editorial published in the British Medical Bulletin by a leading palliative care doctor, Dr. Nigel Sykes, communicated the step change in shared responsibility for end of life care that he believed the Alliance's report introduced: …. doing the right clinical thing is no longer the sole responsibility of care providers. In addition, the role of contracting and resources is recognized through an explicit expectation that Commissioners of care will share the responsibility for effective end-of-life care, while previous training deficiencies are acknowledged through placing this responsibility also with Commissioners of education and training as well as the medical and nursing Royal Colleges 105 (p45). Although the LCP was designed to bring values of good end of life care from the hospice movement to mainstream hospitals and elsewhere, it met with increasing criticism from the public, health care professionals and the media. There were 3 main areas of concern: • recognising that a person was dying was not always supported by an experienced clinician and not reliably reviewed, even if the person may have had the potential to improve • the dying person may have been unduly sedated as a result of injudiciously prescribed symptom control medicines • the perception that hydration and some essential medicines may have been withheld or withdrawn, resulting in a negative effect on the dying person.
These were not necessarily a direct consequence of following the LCP, but often happened because of poor or indiscriminate implementation or a lack of staff training and supervision 110 (p5).
As we will see below, the review of the LCP and the subsequent questions of implementation and quality raised by NICE were much debated in the health care press, especially in the academic research journals related to palliative care.
Published commentaries and editorials: key themes Following its withdrawal, a large number of editorials and comment pieces were published on the theme of 'what went wrong with the LCP'? Here three major themes were in evidence: 1) the LCP as a iconic example of what happens when 'gold standard' evidence from randomised controlled trials is not gathered before adoption of a complex intervention; 2) the LCP as a failure of implementation rather than any fundamental problem in its formulation; and 3) the LCP as something misunderstood and abandoned without due cause. Straddling the first two of these is an editorial in the BMJ by Sleeman and Collis about the LCP 111 : Was this a failure of the paperwork or of its implementation? If implementation and training are key, would investment in these areas-rather than developing guidelines from scratchbe a more efficient use of resources? The lack of strong evidence of the LCP's benefits undermines this argument. However, the converse is also true: the absence of prospective evidence of harm should caution us against the assumption that simply withdrawing the LCP will improve end of life care…. Ultimately, the decision to phase out the LCP was made on the basis of little more than an accumulation of anecdotal evidence. Without independent prospective evidence from controlled trials, the LCP became unusable. This should serve to warn us of the dangers of the national implementation of tools that are not properly evidence based 111 (p7).
A year later and in similar voice, an editorial in the Lancet 12 by Currow and Abernethy accompanying the publication 112 of the first RCT of the LCP, asserts: …the LCP was taken up by bureaucrats who did not understand the implications of widespread implementation of an initiative for which the net effects were poorly defined … As demonstrated by the results of Costantini and colleagues' study, a government, when introducing such initiatives, should properly assess them in rigorous trials of health services, preferably randomised; if this cannot be achieved then a formal prospective assessment of new interventions as they are implemented must be the minimum standard. Either assessment should be done in a health-care environment where new interventions are thoughtfully introduced, corresponding data are routinely gathered for the interventions, and analyses inform understanding of the net benefit and opportunities for iterative enhancement-namely, a learning health system framework, as described by the US Institute of Medicine 12 (p192).
The multicentre cluster randomized trial 112 to which Currow and Abernethy's editorial 12 refers was carried out in Italian hospitals across 16 general medicine wards with at least 25 deaths from cancer annually, and involved 308 patients who died from cancer and their families. Although some commentators have argued that the study was underpowered 15,111 and subject to bias for methodological reasons 15 , it was well designed and rigourous. It found no significant difference in overall rating of quality of care between patients who died in wards in which the LCP had been implemented when compared with those in which it had not, although improvements were observed in two out of 9 secondary outcomes (respect and control of breathlessness). No differences in survival or medication use were observed. The 'take home' message from the cluster trial was that any benefit derived from the LCP depends on the quality of its implementation. Since then, a cluster randomized trial of the 'Care Programme in the Last Days of Life' (CAREFul) has been conducted in the Netherlands, based on the LCP and taking account of its critique 113 , finding that the progamme improved nurse assessed comfort but not satisfaction with care among relatives. The authors conclude that further qualitative research is required to gain an understanding of this apparent discrepancy and recommend 'controlled implementation' with due regard for ongoing training of clinicians especially in communication skills 113 (p132).
A similar theme of implementation is picked up by the American authors Billings and Block in a review in the Journal of Palliative Medicine 114 which noted that: …the tragedy of the LCP lies in the gap between the apparent value of its guidance for clinical care in the last few days of life and its performance in daily practice 114 (p1493).
These commentators also draw four wider conclusions: first, that the Neuberger report itself may be criticized for not being a scientific study of the LCP and for relying too much on anecdote; second, that the LCP story has revealed serious deficiencies in the perception of palliative care in the UK; third, that it reveals serious deficiencies in care in NHS hospitals which go far beyond the scope of the LCP; and fourth, that engagement of anti-euthanasia groups clearly impacted on the reputation of the LCP.
Nor did political and media interest in the LCP disappear. When in late 2015 evidence was given to a Health Select Committee of MPs, confirming that thinly disguised versions of the LCP were in use in some settings despite the Neuberger recommendations, the Daily Mail once again returned to the issue, this time condemning the 'arrogance' of doctors who were flying in the face of best practice 115 . Prominent figures in the palliative care world, including Professor Sam Ahmedzai, who chaired the NICE guidelines committee, were equally vocal in the condemnation.

Discussion
Following Carlile's 28 description of boundary object implementation as a process involving transfer, translation and transformation, we have shown how the LCP quickly assumed national prominence as a key means to deliver the goals of a National End of Life Care Strategy 10 . On entering the phase of transformation, it became a site of expression and a signifier of difference around a range of domains that extended far beyond its original conceptualization as a guide for clinical staff caring for people at the end of life. It was some five years into the national roll-out of the LCP that a sense of public scandal began to break. As Butler and Drakeford (2005)  The transfer and translation of the LCP took placed in the context of a widespread and uncritical enthusiasm for integrated care pathways that was gathering momentum in many modern health systems in the early 2000s. The LCP caught the wave of this, but in her description of integrated pathways as classical examples of boundary objects, Allen 27 (p355) notes that while they have a 'strong cohesive power' to appeal to a range of stakeholder group, their breadth of appeal also disguises tensions between different agendas and frames of reference. To this extent, the development and evolution of boundary objects such as the LCP must always be understood as a political process, with the resultant tools functioning as 'embodied practices for routing patients through the system' 17 not neutral mechanisms 19 . As we have illustrated, before the LCP came to the awareness of the wider media, it encapsulated an ambiguity shared with all other 'pathways', between standardisation and person centred care 16 . This ultimately undermined its core purpose of improving decisions for each individual patient's care in spite of efforts by its promoters to build in safeguards against poor practice and the publication of version 12 in response to criticism. The problem of ambiguity was compounded by an inadequate focus in the wider NHS on clinicians' underlying knowledge in palliative care principles and ethical approaches to end of life decision-making 11 . In a cogent discussion of this latter issue, an Australian palliative medicine specialist, Mackintosh, has observed: …it was not so much about what was said in the LCP documentation as what was not said and brings to light the difficulties of end of life decision-making. Ticking two out of four boxes about symptoms that were not specific to the dying patient now seems a rather naive approach; diagnosing dying can be a complex process filled with uncertainties ... specialist palliative care teams make their decisions very carefully following all the safeguards contained within the LCP. Much of what is implicit in the practice of specialist palliative care was never made explicit; the V12 algorithm came too late ... the LCP has been accused of adopting a 'one size fits all' approach. However, this was not the case as a careful examination of the documentation will reveal. Instead the LCP was interpreted as being a 'one size fits all' tool to practitioners without the required skills to read between the lines 90 (p651).
The LCP was understood by its designers to be a complex intervention and, as we have seen, for its implementation they recommended ongoing training and support of clinicians at an organisational level to be provided by expert facilitators. As the LCP was rolled out at scale, this aspect was weakened and the LCP began to be represented as a stand-alone document s , rather than a broad approach to care in which the documentation was merely one aspect. Lack of training, together with ambiguities and misunderstandings surrounding the LCP, were combined with rapidly increasing pressures on the workforce. Indeed the Review Panel observed 'constant pressures on staff and that some find the workload unmanageable' 11 (para 2.25), with hospitals in England running at 90% capacity rather than the 80% recommended for safety. The LCP story thereby highlighted an unpalatable truth of much broader significance: when healthcare workers are exhausted and overstretched, quality of care, communication and decision-making are compromised 116 . As other commentators have concluded: '…services that provide poor quality general care will undoubtedly provide very poor end-of-life care 117 (p510).
As the LCP story unfolded, it became a conduit for debate about the proper epistemic foundations of innovation in health care (quality improvement versus evidence based medicine). The LCP was not originally underpinned by research evidence from RCTs, but was far from alone in this as an intervention in palliative care or in health care more generally. For example, there was no LCP-type reaction to the 'Gold Standards Framework' 118 or 'Preferred Priorities of Care' 119 , which were similarly high profile boundary objects promoted by the National End of Life Care Programme, each lacking supporting evidence from RCTs. Moreover, Mackintosh 90 has argued that the calls for better evidence in the LCP debate are misplaced '….since what is being measured is not the performance of the tool but the performance of the user' (p651) . The RCT evidence that has since emerged from Italy 120 and from the Netherlands 113 powerfully draws attention to this point.
An essential characteristic of boundary objects described by Star is their interpretive flexibility 121 . Where over standardization occurs this works against the reflective use and local tailoring that enables boundary objects to be used effectively. In her prescient case study of the ways in which integrated care pathways develop, Allen 27 concludes that diversity rather than standardization should be encouraged, thereby ensuring their utility for different purposes and contexts: They should not be treated as pre-fabricated tools that supposedly can be easily transferred from one context to another. Furthermore, it is not sufficient to introduce a boundary object and wait for it to do its magic. The local process of making, introducing and using care pathways is crucial for making them work as boundary objects and continuous work is needed to sustain them 27 (p 648). It is well known that boundary objects need to be accompanied by the involvement of 'knowledge brokers' to enable translation into practice 122 . This was something recommended by the Marie Curie Palliative Care Team in their recommendations for implementation (albeit using different terminology) 123 , and highlighted in the conclusions drawn from the results of the randomised controlled trial of the LCP in Italy 112 . Knowledge brokers were also found to be critical to the success of the wider National End of Life Care Programme in an evaluation of its first phase 48 . Wenger describes such brokerage as a '… process of translation, co-ordination and alignment between perspectives. It also requires the ability to link practices by facilitating transactions between them' 124 (p38). Following its withdrawal in England, a study in the Netherlands 125 126 examining what hindered the implementation of the LCP showed that successful interventions are configured to address and influence the understandings of staff about end of life care and to increase their motivation and self-efficacy, and highlighted the support of senior managers as essential, both to release necessary resources and to enable culture change in organisations.
Integrated care pathways are simultaneously work flow models and records of care 27 . They are both management and clinical tools. As Allen 27 demonstrates, they also appeal to two distinct but related logics within the health care system. Pathways are about evidence based practice as well as quality improvement. If the former suggests scientific knowledge that is slow to accumulate and problematic to implement, the latter promotes local initiatives, quick implementation, adaptation and rapid feedback into practice. The LCP embodied these tensions. The formal evidence from a randomised control trial that brought its efficacy into question did not come until after major concerns had led to recommendations for its withdrawal. At the local level however, LCP was subject to annual scrutiny by a multi-disciplinary steering group that included carer representation and over time it moved through a series of revisions and numbered versions. Its 'roll out' was therefore the product of quality improvement enthusiasms at local level which were quickly scaled up. In this LCP can be seen as a positive 'boundary object' with the potential to enrol clinicians, managers, service users and indeed wider publics in the common aim of improving care of the dying. When this broke down however, LCP was transformed into a negative boundary object, which served to highlight not shared enrolment, but rather fundamental conflicts of view. The LCP is not alone in this. As Allen's studies show 22,27 , other pathways have undergone this transition -and have been quietly abandoned in the process before a crisis occurred. The LCP however was unusual in becoming high profile. It was not obscured within the complexities of daily clinical practice, but rather was championed as the diffusion of hospice principles to meet the needs of the majority rather than the few. It was heralded as a transformative intervention to improve system-wide terminal care. Although its protagonists were always careful to emphasise that the LCP was only as good as the skills of those using it, the LCP, as it quickly rolled out, ran ahead of these skills 126,127 .
Hendy and Barlow 128 have shown that organisational champions can be highly effective in the initial phases of health system innovation, when change is contained within distinct sub-sets of practice. But they caution against change being positioned in the hands of too few individuals, which may prove detrimental to wider implementation. The LCP may have suffered from this problem. It might also be argued that the relevant champions lacked foresight. In environmental areas it has become routine to practice 'responsible innovation' 129 , acknowledging that innovation can raise questions and dilemmas, is often ambiguous in its purposes and motivations and can be unpredictable in its effects, beneficial or otherwise. The approach operates on four principles of: 1) anticipating impacts, intended or not 2) reflecting on potential motivations, uncertainties, framings, dilemmas and transformations that might ensue 3) engaging in dialogue and debate about these issues in an inclusive way 4) acting to influence the direction and trajectory of the innovation process. Such principles could map fairly easily onto the LCP narrative, but were largely absent in practice. In a related way, Greenhalgh and colleagues 130 have published a framework for understanding abandonment and challenges to the scale up of health and care technologies. They show that innovations that fail to address complexity across 7 specific domains are unlikely to be sustainable: (i) nature of illness and co-morbidities; ii) material features of the technology and knowledge needed to use it; iii) values (including efficacy, safety and cost-effectiveness); iv) adopters (including staff roles and practices); v) organisation (organisational readiness to adopt and implement); vi) wider system (consideration of professional and stakeholder perspectives) and vii) scope for embedding and adaptation over time. As our account shows, several of these factors are relevant to the rise and fall of the LCP.
The sequence of events surrounding the LCP brought to an end the era of 'unconditional regard' 4 for palliative care. No longer could it be assumed that palliative care was universally welcomed. The LCP exposed 'sites of difference' upon which the goals of care and methods of care at the end of life were opened up to wide scrutiny. The ramifications seem not yet to have been fully acknowledged by those involved.

Conclusion
We set out to answer three sets of questions: 1) why and how did the LCP come to prominence as a vehicle of policy and practice 2) what factors contributed to its demise? 3) what immediate implications and lessons resulted from its withdrawal?
The LCP emerged in response to a clear need for 'scaled up' approaches to care interventions at the end of life. The case in the late 1990s was compelling. The available provision of palliative care through hospices and specialist palliative care units in hospitals was clearly incommensurate with the prevailing level of need for such services and the number of people who could benefit from them. This was therefore addressed through the development of a mechanism that fitted with then current enthusiasms for integrated care pathways and which could bring some of the essentials of palliative care practice to a wide range of beneficiaries. The initial success of the approach, what might be called the period of the LCP as successful boundary object, can be measured by the traction that it gained in its roll out and in the early results from local and service improvement studies. This was based on successful involvement from a number of stakeholders -across health care disciplines, hospital management and service users. Such an approach resonated strongly at that time with a predisposing climate of policy interest in the provision of multi-disciplinary end of life care and a strategic commitment to support the transfer of successes seen in the British hospice context, into the wider health and social care system.
The factors that contributed to the demise of the LCP in turn relate to how it then broke down as a boundary object. The LCP needed constant cultivation by knowledge brokers if it was to hold together and sustain the engagement of disparate stakeholders. Where this faltered, the default position was one of mechanistic and potentially insensitive implementation of the LCP as a protocol or checklist. It then became the site of conflicts provoked by ideologically inspired interest groups, both within and beyond the healthcare environment, which fostered a previously unseen critique of certain assumptions within the palliative care paradigm. This found its correlates in the media and wider expressions of public concern. As these grew, the LCP became the bearer of responsibility for all aspects of NHS end of life care, well beyond its original remit, goals and aspirations.
Meanwhile, palliative care experts were slow to respond to these concerns, or to develop evidence about the LCP in relation to which they could be judged. Instead their contributions only emerged with any significance in the aftermath of withdrawal when some hastened to publish their thoughts. The contributions polarised between those who had 'always' had reservations about the LCP and those who saw its demise as a matter of regret and who in turn challenged the critics to present a better alternative. The five 'priorities of care for the dying person' developed by the LACDP and the subsequent NICE guidance, with its flimsy evidence base, were long on values and aspirations, but short on a practical course of action. They replaced one set of deficiencies with another.
On the available evidence and within the limits of our chosen method here, we judge that the LCP boundary object was well conceived, but not matched by sufficient subtlety and foresight in its implementation and dissemination. The dramatic demise, when it came, resulted from a combination of media-fuelled public criticism and long-delayed professional judgement, hitherto never experienced in the developing field of palliative and end of life care. The most important lesson that can be learned however is not about the dangers of scaling up clinical interventions that lack an 'evidence base'. Rather it is about the need for greater assessment of the wider risks involved and more careful consideration of the unintended consequences that might result from a given course of action -especially in the politically and morally charged arena of end of life care interventions.

Data availability
All data underlying the results are available as part of the article and no additional source data are required. The paper is based on an analysis of documents in the public domain. In this thoughtful and illuminating article, Professors Seymour and Clark critically analyse the rise, demise and legacy of the Liverpool Care Pathway (LCP), using the sociological concept of "boundary objects". A boundary object occupies multiple social worlds and enables communication between them, making this a good conceptual fit to the LCP, which was an integrated care pathway that sought to straddle various clinical, managerial and "user" interests.

Open Peer Review
The use of the boundary objects metaphor is instructive and the authors (commendably) clearly marshal and convey a diverse range of evidence from the literature. The authors take a literature-based approach, in part because qualitative work (e.g. interviewing key participants and stakeholders) would likely be premature. This seems plausible, although future such work would no doubt be welcome. It might have been useful to hear a little more about how the authors designed their (systematic?) review, but this certainly appears to be a thorough review, which successfully traces the history of the LCP and highlights the key themes surrounding the LCP that one might expect to encounter.
Numerous overlapping themes emerge: striking a balance between individualised (person-centred) care and standardised care; the nature and use of evidence in supporting or discrediting health interventions (and, indeed, in policy-making); translating theory into practice, through appropriate education, support and dissemination; and the respective roles of process and judgment when designing and delivering care. Many of the themes will resonate beyond the LCP and the authors do engage with some (important) broader themes, such as the role of care pathways, funding of and in the UK's National Health Service, and perceptions of palliative and end of life care more generally. Some such themes might even have been pushed further e.g. in spurring national discussions about death and dying, might the LCP's demise be associated with an oft-reported reluctance to openly confront these difficult matters? However, the authors' reflections appear to be carefully evidence-based. Indeed, the evidential issues surrounding the LCP were notable throughout their article: I was particularly struck by the sense that the sort of evidence that was judged problematically to underpin the LCP -this being "soft", less than "gold standard", evidence -was precisely the sort of evidence then used to discredit it, and to underpin its review and subsequent removal. The sort of hard evidence sought by some critics was later collected, but I was left pondering the sort of evidence best suited to this complex, uncertain context (indeed, uncertainty might be another theme associated with the LCP, as the authors acknowledge early in their article).
In discussing the themes that they found, the authors display a sure grasp of context and history. Some elements could have been explained a little further, especially for non-UK audiences and for those less familiar with navigating the (potentially bewildering) health policy landscape. As other reviewers also comment, examples include the references to the Gold Standard Framework, Preferred Priorities of Care, familiar with navigating the (potentially bewildering) health policy landscape. As other reviewers also comment, examples include the references to the Gold Standard Framework, Preferred Priorities of Care, End of Life Care Strategy, and to bodies like the National Council for Palliative Care (which has since merged with Hospice UK). There are also very occasional -and very mild -typographical errors. However, none of this should detract from the value of this important article. The authors display insight and balance throughout and this article should become a go-to piece for anyone wishing to explore the LCP and, more generally, the promise and pitfalls of using pathways and making policy in healthcare.

If applicable, is the statistical analysis and its interpretation appropriate? Not applicable
Are all the source data underlying the results available to ensure full reproducibility? Yes We thank Professor Huxtable for his insightful review. Work is underway to collect interview data from key stakeholders involved in the international application of the LCP in a follow on project. The latter will provide us with an opportunity to consider some of the important issues raised in this review, such as the relationship between palliative care policy and wider societal perceptions of end of life care, and the type of 'evidence' required to underpin innovations in palliative care. We have made some small clarifications to the paper, which hopefully eradicate the typographical errors identified and make aspects of the paper a little more accessible to international readers.
No competing interests were disclosed. Competing Interests: Nevertheless, having discussed the presentation of the LCP conflict by the authors in the background, we find the introductory perspective to be somewhat less distanced than the rest of the article.
The study is contextualized to the global spread of palliative care and how it has evolved as a response to challenges in the population, while the analysis focuses on LCP on the UK scene. International readers might benefit from having British palliative care put in context, within the responsibility of the NHS on the one hand and wider society -and maybe communities -on the other. This particularly surfaces in the analysis, where the authors clarify that, at the same time as the LCP, the 'Gold Standards Framework' and the 'Preferred Priorities for Care' were also implemented. How these three tools were related is not explicated, however. All three are mentioned, but it remains unclear to what extent, if at all, these latter tools were part of the LCP discourse, and if so, in what way.
The theoretical lens of 'boundary objects' which is used in the study appears to have a practical fit. In our interpretation, the lens articulates -and demarcates -an understanding of the study object that could range along opposites poles in a continuum from considered negative to positive, which is related to negotiations, renegotiations and conflicts as reflected in the swings of the pendulum appearing in the history of the LCP. Moreover, the perspective could help shed light on how different perspectives could be bridged and also on unintended consequences, which were clearly related to the implementation of the LCP. However, the article does not make clear what aspects might have been left out in the analysis. This would have been useful, especially for international readers who are not knowledgeable of the details in the British context. Moreover, the discussion section does little to go back or problematize from the view point of this framework.
A fairly unsurprising but still important result of the analysis is the fact that the tool itself cannot be blamed for poor quality of care. Thus, the analysis explicates the importance of considering what a practice tool involves (e.g. as related to underpinnings, reliability, usefulness and so forth) and the driving forces for implementation and diffusion of the tool. After independently reading the article and then discussing it, we became unsure as to how the whole controversy were handled by the media -could this have been more explicated and problematized in the analysis? This also raises the question of whether the NHS makes room for self-criticism (i.e. allows sufficient implementation and follow-up procedure, and a workable structure, resources and competence of care etc.).
On the whole, we find this analysis complete and important, both for practice and research as related to palliative care practice and policy development and, of course, especially for knowledge translation or 1 2 1 2 palliative care practice and policy development and, of course, especially for knowledge translation or dissemination. In our view, the significance of the analysis is that it shows how critical it is to have an informed understanding of the not only the development of innovative practice support tools, but moreover of complexities and the diversity of stakeholders involved in their dissemination, and that both innovative development processes and implementation processes do not only happen in discourses of evidenced-based practice. Over simplified conclusions of problems related to the use of several practice tools in palliative care, such as the LCP, are either that the tool has sufficient evidence or that the tool in itself is dangerous or disconcerts wellbeing for patients and/or their informal carers. Furthermore, the analysis of the LCP case shows that a tool can be both recommended and promoted despite lacking support from significant evidence gatekeepers such as Cochrane reviews. Rather, the analysis by Seymour and Clark clarifies the need for practitioners to take action based on their practical wisdom and not only rely on tools, as well as the merits of considering unintended consequences related to the use of tools. At least from our Scandinavian perspective we could learn from the LCP lesson, and without doubts Seymour & Clark's analysis contributes to a more in-depth and critical understanding than previous national reports from the UK on the use and recommendations of LCP.

Are sufficient details of methods and analysis provided to allow replication by others? Yes
If applicable, is the statistical analysis and its interpretation appropriate?

Not applicable
Are all the source data underlying the results available to ensure full reproducibility? Yes

Are the conclusions drawn adequately supported by the results? Yes
No competing interests were disclosed.

Competing Interests:
We have read this submission. We believe that we have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.
Author Response 20 Apr 2018 , University of Glasgow, UK David Clark We thank Professor Öhlen and Dr Håkanson for their thoughtful and helpful review. We accept their view that the some aspects of the paper (especially the relationship between the three 'tools' ) may be challenging for an international audience unfamiliar with the UK context, but hope that the small clarifications we have now made to the paper go some way to addressing this. We anticipate making further use of the boundary object concept in a follow on study examining the international spread of the LCP, and so their observation that we could perhaps have made further interpretive use of the latter in this paper is valuable to us as we embark on that new project. The observations about what may have been left out of the analysis and whether we could have provided an use of the latter in this paper is valuable to us as we embark on that new project. The observations about what may have been left out of the analysis and whether we could have provided an expanded account to the role of the media are well made. It was challenging to weave together a coherent narrative about a very complex affair and we have tried to be balanced in our use of sources. We hope that the paper will encourage others to take a similarly 'forensic' approach to examining the many aspects of this important phenomenon in palliative care history.
No competing interests were disclosed. Competing Interests: I've never seen such a comprehensive or dispassionate account of the affair. You have definitely moved the debate up several notches. The boundary object concept is very helpful and has implications for service design programmes.

Discuss this Article
I am thinking about how the industrial origins of ICPs, following something like the Toyota method may have laid the foundations for the failure of the LCP as a boundary object in the world of NHS palliative care. When the object of the original method is standardisation and its users are healthcare professionals, it would have required additional work to allow a common language with families. The criticisms of the LCP system seem to have come mostly from families who do not share the professional's view of the illness as terminal, as well as some who see the care the LCP as an aberration when compared to previous care. It is possible that the lack of continuity from previous care, privileging care of the "dying phase" over continuing care of the individual was its downfall.
The GSF had different, perhaps scriptural origins, with its list of 6 commandments, later incorporating the LCP as a 7th. It was designed within general practice, following consultation with primary care teams and has always chimed well with patients and families. It apparently changed organisation and practice within primary care and did not attract the problems of the hospital version of the LCP. Although the paperwork made it rather cumbersome, it was adopted much faster than the LCP, and without the institutional coercion, at least this side of the Scottish border. Work on the GSF always identified the need for facilitation and support.
Also, your call for consideration of wider risks etc is well put. We might ask why this did not happen as it should have done when such a clinical practice replaces another. The complex system of clinical governance and risk registration in NHS hospitals, that should have overseen its application, must have been circumvented by the way it was imposed on NHS Trusts.
Another aspect of the story might be that it represents a failure, in professional terms, of my medical specialty. I include myself in this as a naysayer, too much influenced by the politics of influence and decisions taken further up the NHS. At the time, I lacked the evidence or the motivation to engage properly with local implementation of the LCP, commenting that effort should be expended further upstream of the with local implementation of the LCP, commenting that effort should be expended further upstream of the last 48 hours. I would only take issue with you that your call contrasts with other observers. Your cll for consideration of risk before implementation would have been informed by interventional studies. A useful evidence base will always elucidate unintended consequences, including those that none predicted before implementation. RCTs might be slow, but might have precluded the LCP's mass implementation. Another analysis might conclude that the PDSA cycle enabled the LCP to dodge its own coffin for 20 years! Thank you Professors Seymour and Clark for this highly accessible, analytical history and critical analysis of the Liverpool Care Pathway. I found it thoughtful and thought provoking. At times uncomfortable reading, at other moments providing a different perspective. I believe this will be a seminal reference for the future.
I guess my main observation and question is making sure we move from rhetoric to practice in terms of reflections and lessons learnt, pragmatically respond rather than react to the ever changing communications and technological landscape as well as combine an attitude to care and practice that embraces the generation of robust evidence and translation of knowledge into the real world practice which has every day meaning for patients, families and the community within which we live (and die) in.
Thank you for providing a useful and significant contribution to the debate, analysis and knowledge arena.
I am a palliative care nurse working in a national hospice charity Competing Interests: