Home care and end-of-life hospital admissions: a retrospective interview study in English primary and secondary care

Background Enabling death at home remains an important priority in end-of-life care policy. However, hospital continues to be a more prevalent place of death than home in the UK, with admissions at the end-of-life often negatively labelled. Admissions are frequently attributed to an unsuitable home environment, associated with inadequate family care provision and insufficient professional care delivery. Aim To understand problems in professional and lay care provision that discourage death at home and lead to hospital admissions at the end of life. Design and setting A qualitative study of admission to a large English hospital of patients close to the end of their life. Method Retrospective in-depth semi-structured interviews with healthcare professionals (n = 30) and next-of-kin (n = 3) involved in an admission. Interviews addressed why older patients (>65 years) close to the end of life are admitted to hospital. Interviews were transcribed and analysed thematically. Results Home-based end-of-life care appeared precarious. Hospital admission was considered by healthcare staff when there was insufficient nursing provision, or where family support, which was often extensive but under supported, was challenged. In these circumstances, home was not recognised to be a suitable place of care or death, justifying seeking care provision elsewhere. Conclusion Challenges in home care provision led to hospital admissions. Home end-of-life care depended on substantial input from family and professional carers, both of which were under-resourced. Where either care was insufficient to meet the needs of patients, home was no longer deemed to be desirable by healthcare staff and hospital care was sought.


INTRODUCTION
Supporting patients at the end of life to die at home and preventing their admission to hospital is a significant task for GPs and community nursing teams. Place of death is a key policy marker of end-of-life care success in both the UK 1,2 and internationally, 3 though recent literature challenges the priority given to location in end-of-life care. 4 Research has demonstrated that patient preferences for place of death are complex, 5 that notions of home are malleable and include other settings which may feel 'home like', [6][7][8] and that the hospital is important in end-of-life care provision. [9][10][11] However, death at home remains a policy concern. Reflecting this, end-of-life hospital admissions are often defined as 'avoidable', 'preventable' or 'inappropriate', 5,12,13 particularly compared to community-based care.
It is not clear why patients at the end of their lives are admitted to, and die, in hospital 14,15 though many reasons have been suggested. It seems likely that the provision of end-of-life care in the community from both professional and lay carers is an important factor, recognised in both the limited literature exploring end-of-life hospital admissions 11,14,[16][17][18] and in related research on community-based end-oflife care. For example, the availability of professional home-based end-of-life care is associated with the likelihood of patients with cancer dying at home rather than in hospital. 19 Notably however, the relationship between professional end-of-life care in the community and hospital admissions has yet to be explored robustly. 20 The challenge for family carers in providing care at home and the association between problems in this care and end-of-life hospital admissions is frequently identified in policy, and recognised by GPs. 11,18,21 Problems are often attributed to family members being 'panicked' by changes to a patient's symptoms in the dying phase, 22 or because family members felt unable to cope with incessant care provision, and their own distress in watching a relative die. 23,24 Similar to professional care however, the circumstances that lead families to seek hospital care have not been considered in detail.
Hospital continues to be a more prevalent place of death than home in England. 25 Understanding why these admissions occur will help GPs and community nurses to support patients in their end-of-life care, whether at home or elsewhere. This article explores empirically the relationship between the care requirements of home-based endof-life provision and hospital admissions using healthcare staff, and family carer perspectives of why patients previously cared for in the community are admitted to, and die, in the acute hospital setting.

METHOD Design and participants
The data presented are from a larger study exploring end-of-life hospital admissions. 26,27 Research Abstract Background Enabling death at home remains an important priority in end-of-life care policy. However, hospital continues to be a more prevalent place of death than home in the UK, with admissions at the end-of-life often negatively labelled. Admissions are frequently attributed to an unsuitable home environment, associated with inadequate family care provision and insufficient professional care delivery.

Aim
To understand problems in professional and lay care provision that discourage death at home and lead to hospital admissions at the end of life.

Design and setting
A qualitative study of admission to a large English hospital of patients close to the end of their life.

Method
Retrospective in-depth semi-structured interviews with healthcare professionals (n = 30) and next-of-kin (n = 3) involved in an admission. Interviews addressed why older patients (>65 years) close to the end of life are admitted to hospital. Interviews were transcribed and analysed thematically.

Results
Home-based end-of-life care appeared precarious. Hospital admission was considered by healthcare staff when there was insufficient nursing provision, or where family support, which was often extensive but under supported, was challenged. In these circumstances, home was not recognised to be a suitable place of care or death, justifying seeking care provision elsewhere.

Conclusion
Challenges in home care provision led to hospital admissions. Home end-of-life care depended on substantial input from family and professional carers, both of which were under-resourced. Where either care was insufficient to meet the needs of patients, home was no longer deemed to be desirable by healthcare staff and hospital care was sought.
Keywords end of life care; home palliative care; hospitalisation; primary care.
The focus of this article is on why patients close to the end of life who are cared for at home do not die there but are instead admitted to hospital. Interviews (n = 33) were conducted with healthcare staff and next of kin involved in the admission to a large English hospital of a patient who subsequently died there within 3 days of admission (a 'case-patient').
Interviewees included community healthcare staff (GPs, a community specialist nurse, and ambulance staff) hospital staff, and the case-patients' next of kin. This study centres on the accounts of interviewees involved in the care of a case-patient who had been admitted to hospital from home. Additionally, the accounts of interviewees involved in the admission of case-patients from other community settings have been included as they usefully reflected on other patients admitted from home. These quotations are indicated by (a).

Data collection
All interviews were semi-structured, in depth, and were conducted between 2012 and 2013. Healthcare staff interviews addressed the participants' involvement in the case-patients' admission, their views on the reasons for the admission, and end-of-life admissions generally. These interviews were conducted within a month of the interviewee's involvement in the case-patient's care, lasted approximately 30 minutes, and occurred at the participants' place of work or nearby. Next-of-kin interviews considered the casepatient's care in the last months before death, including their involvement in care provision. These interviews took place 4-7 months after the case-patient's death, lasted approximately 1 hour, and occurred at their home.

Analysis
Interviews with staff and next of kin provided highly detailed accounts of the circumstances that led to the hospital admission of case-patients and for staff interviewees, similar patients they had cared for previously. Interviews were audiorecorded and professionally transcribed verbatim. The transcripts were analysed thematically to understand the reasons that facilitated admission to hospital. Transcripts were coded by hand and then using the software NVivo (version 10), with sections of the text tagged using both a priori and emergent data categories. Codes were then sorted and charted to assess both the breadth and depth of data. Data were synthesised to understand both the individual casepatients' admission, and the admissions collectively, with synthesis aided by reference to both end-of-life care and sociological literature. Extended details about the study methodology 26,27 and related findings 10 can be found elsewhere.
Quotations in the text are followed by the participants' role, study number, and interview page number. Identifying details have been omitted to protect the anonymity of the interviewees and the deceased casepatients.

RESULTS
Appendix 1 shows the characteristics of the case-patients and Table 1 provides a summary of all participant roles. Hospital admissions were instigated by patients,

How this fits in
It is known that dying at home is an end-of-life care policy priority, with reducing end-of-life hospital admissions targeted in end-of-life care improvement measures. Many people currently die in hospital. Understanding why deaths occur in hospital rather than at home is underexplored but is often assumed to be associated with problems in care provision. This study shows that homebased end-of-life care can be precarious and an absence of nursing care or family support may facilitate hospital admission; deaths in hospitals can occur where home is not considered a suitable place of care and death, and offers benefits to patients, family carers, and staff. To understand why hospital admissions may have been sought because family members did not 'want' to continue to provide care it is useful to explore the family caring role. In the next section, the demands on family carers' time, physicality, and experience are examined in relation to end-of-life hospital admissions.

Experience to provide care
Family care provision at home could be compromised by the limited experience some family carers had of death and endof-life care. For healthcare staff, naïve family carer expectations of the dying process were thought to hinder the likelihood of death at home, as unfamiliar but typical end-oflife care symptoms were reported to lead family carers to seek reassurance and care. Consequently, hospital admissions were thought to occur, either directly, as carers sought help from ambulance or out-ofhours services, or indirectly, if they requested additional help that could realistically only be fulfilled in hospital: ' [...] So it does seem that we had talked about end-of-life care but we probably didn't talk about preferred place of care or maybe the family weren't prepared well enough for the last stages of [the case-patient's] dying to manage with the support that we have.' (Ward doctor, 14, 1) For family carers, inexperience was often expressed as concern about the quality of their care provision, particularly compared to professional carers. While this was not cited by family carers as a reason for admission -family carers who had facilitated casepatient admissions said they did so to access necessary medical support -they acknowledged the hospital to be a place for professional attention and a place where their relative would receive appropriate care: 'I was pleased [they] died in hospital, because I always used to think "If you're ill, hospital is the best place to be, because there you can have all the correct attention and everything that you want". At home, I mean, I don't know all that much about the medical service really, [...] I couldn't be a nurse, I would do what I could for [my spouse] but I don't know, you know, how well I would be doing it compared with a qualified nurse.' (Next of kin, 6, 13)

All-encompassing care
Supporting a dying relative often involved significant investment of time and resource by family carers. The consequences of this to the family carer was infrequently recognised by the family carer themselves, who typically dismissed their own care provision as just part of their family role. However, providing care could have negative effects on the family carer's wellbeing, and healthcare staff described how a patient's care need could exceed the family carer's capacity to provide care. In these instances, additional care was often recognised to be necessary, whether sought from hospital providers or elsewhere:

Physical care
The family caring role was recognised by professional staff to be physically demanding. This could make family care provision difficult, particularly if a carer was a frail older individual and the patients' health had deteriorated such that they required substantial hands-on support. In these instances, both community and hospital interviewees recognised that the continued maintenance of patients at home was impractical without additional support. Where this was not thought possible, staff recognised the helpfulness of care elsewhere, including hospital: 'And there is, you know, this practical side of nursing people, you know, mobilising, even moving them safely when they can no longer move themselves. [...] I went to visit an elderly couple and found both of them on the floor where they had been, God help them, all that night, because she had been trying to get him out of bed to the loo and he had fallen on top of her, and there they had lain all night long.' (aGP, 28

Caring for patients at home
Ensuring a patient could die at home in a manner that met staff expectations of an appropriate death was not easy. The factors identified above were often considered to be interdependent issues, which could be challenging for GPs to meet and therefore could precipitate hospital admissions:

DISCUSSION Summary
The present research identifies that hospital admissions at the end of life can be a consequence of the challenges in delivering end-of-life care at home. Insufficient available nursing provision and family carers who had exceeded their capacity to care countered the desirability of home as a place to deliver end-of-life care. These circumstances often caused GPs, together with family carers and other healthcare staff to seek care for patients through hospital admissions.

Strengths and limitations
This article has explored the consequences of insufficient community-based care as a precursor to end-of-life hospital admissions. This research usefully contributes to the limited evidence base on end-of-life hospital admissions. In a policy and clinical context where the desirability of dying at home is presumed, the present findings also add insight to the practicalities of home-based end-of-life care.
The findings of this study are from an in-depth qualitative study of 33 healthcare professionals and next of kin involved in the admission to one English hospital of nine patients close to the end of life. The extent to which interviewee accounts of the case-patients', and similar patients', community-based care, and the reasons for their hospital admission, can be extrapolated to end-of-life care in general is therefore limited. For instance, it is likely that some other patients with problems in the provision of professional or family care would continue to remain at home and not seek hospital care. However, the challenges faced by healthcare staff and family carers in providing end-of-life care accords with existing research, as outlined later on, and contributes to an understanding of why admissions occur.
Formal care provision that occurs at the interface of health and social care, such as supporting personal hygiene, or eating and drinking, has not been addressed in this article. This care is pertinent to understanding how end-of-life care can be facilitated at home, 28 and paid carers were known to have supported some of the case-patients. Interviewing these carers was outside the study scope, and their role was not sufficiently elaborated by other interviewees to be considered.

Comparison with existing literature
The limited availability and capacity of domiciliary nursing care was cited as a reason why interviewees instigated hospital admissions for patients close to the end of life. The challenge for GPs in arranging community nursing end-oflife care is also reported in a UK-wide survey, 29 and seems likely to be due to the restricted capacity of community nursing, where rising demand for services coincides with staff shortages. 30 Internationally, the importance of community-based end-oflife care provision for facilitating death at home accords with review evidence, which identifies receipt of community-based nursing, home palliative care, and GP home visits as key factors. 31,32 The importance of family care provision in facilitating death at home is wellestablished. 31,33 The present research shows that where lay care is recognised to be no longer tenable, healthcare providers may seek hospital care for patients, supporting GP accounts about the reasons for endof-life admissions 18 and pejorative rhetoric about family carers. 6,18 However, it would be inaccurate to assert on the basis of this that problems in lay care provision instigate these admissions, 23,34 without acknowledging simultaneously the contribution of family care provision in maintaining patients at home. Interviewees demonstrated that family members delivering hands-on care provided vital support, supplementing professional care provision. 35,36 Therefore, it is plausible that hands-on care from family members can defer admission to hospital for some patients while resident at home. This is pertinent given the context of lay care provision: family carers were understood to be often ill-prepared for their role 1,18,34,37 with often substantial, if typical, psychological 23,34 and physical 6,36,38 demands made of them, particularly for those with limited mobility such as frail older people. It seems likely therefore that family carers' ability to sustain their care provision, and prevent hospital admissions, is compromised without supplementary support. 39 Facilitating end-of-life care at home was important for professional carers, who echoed prevalent notions about the desirability of death at home. 8,37 However, interviewees also recognised that the safety of end-of-life care at home could be compromised. 8,10,40 Where the cause of this could not be addressed adequately, healthcare staff and family carers sought care for the patients elsewhere. While hospice and residential nursing homes were considered, challenges in accessing these institutions at short notice meant hospital care was prioritised 41 highlighting the importance of hospital as a provider for end-of-life care. 10

Implications for research and practice
The present analysis was informed by a sociological interest in dying at home, focusing on the physical and technical infrastructure of home and hospital, the skills of lay carers and professionals, and the way dying is understood by those involved. 42 This perspective exposed the vulnerability of home care for some patients, which can be dependent on stretched professional and lay provision, and builds on over a decade of sociological scholarship on the role of home as a place of care. 6 The case-patients' hospital admissions demonstrate the difficulty of providing and maintaining care at home. The authors' findings challenge the prevalent discourse that home is an inherently better place to receive end-of-life care than hospital and highlight the dissonance between policy rhetoric and the everyday reality of caring for patients close to the end of life. The present research suggests that if policy and practice maintain an emphasis on facilitating deaths at home, there must also be a concurrent focus on ensuring that patients can die there safely. Identifying how best to achieve this will require further research, and is likely to require investment to ensure that community nursing provision is adequately staffed, responsive, and available throughout the day and night. 43 Supplementing this essential provision with rapid and reliable specialist clinical services that offer support for patients at home for extended periods of time, such as hospice at home and Marie Curie nursing services, would also be pertinent. [44][45][46] To facilitate care delivery, primary care clinicians must be supported by adequate information-sharing practices, 47 and be confident in delivering palliative and endof-life care. 48 Family carers must also be supported, including greater recognition of their needs. 49 Pejorative assessments of the association between family care and end-of-life hospital admissions obscures the significant undertaking of lay carers to support patients at home. Future research could productively address how families persist in providing care, and how GPs and others can best support them.

Funding
This study presents independent research part funded by the National Institute for Health Research Collaborations for Leadership in Applied Health Research and Care (NIHR CLAHRC) Cambridge and Peterborough (grant number: RG55908), and NIHR CLAHRC East of England (grant number: RG74481). The writing of the article was funded by the NIHR School for Primary Care Research (NIHR SPCR) (grant number: RG85520 [NIHR SPCR-2016-S67]). The views expressed are those of the authors and not necessarily those of the NHS, the NIHR or the Department of Health. The Marie Curie Design to Care programme is a service improvement programme. This research forms part of the design phase of this programme (grant number: RG87570), which is funded by Marie Curie, the UK's leading charity caring for people living with any terminal illness and their families.

Ethical approval
The study was approved by the Hertfordshire Research Ethics Committee (England) (ethical approval number: 11/EE/0491. Approval to view the patients' medical records to identify potential interviewees was granted by the then National Information Governance Board (ethical approval number: ECC 1-5 (G)/2012). All interviewees gave consent to take part in the study.

Provenance
Freely submitted; externally peer reviewed.