Out-of-hours services and end-of-life hospital admissions: a complex intervention systematic review and narrative synthesis

Background Out-of-hours (OOH) hospital admissions for patients receiving end-of-life care are a common cause of concern for patients, families, clinicians, and policymakers. It is unclear what issues, or combinations of issues, lead OOH clinicians to initiate hospital care for these patients. Aim To investigate the circumstances, processes, and mechanisms of UK OOH services-initiated end-of-life care hospital admissions. Design and setting Systematic literature review and narrative synthesis. Method Eight electronic databases were searched from inception to December 2019 supplemented by hand-searching of the British Journal of General Practice. Key search terms included: ‘out-of-hours services’, ‘hospital admissions’, and ‘end-of-life care’. Two reviewers independently screened and selected articles, and undertook quality appraisal using Gough’s Weight of Evidence framework. Data was analysed using narrative synthesis and reported following PRISMA Complex Intervention guidance. Results Searches identified 20 727 unique citations, 25 of which met the inclusion criteria. Few studies had a primary focus on the review questions. Admissions were instigated primarily to address clinical needs, caregiver and/or patient distress, and discontinuity or unavailability of care provision, and they were arranged by a range of OOH providers. Reported frequencies of patients receiving end-of-life care being admitted to hospital varied greatly; most evidence related to cancer patients. Conclusion Although OOH end-of-life care can often be readily resolved by hospital admissions, it comes with multiple challenges that seem to be widespread and systemic. Further research is therefore necessary to understand the complexities of OOH services-initiated end-of-life care hospital admissions and how the challenges underpinning such admissions might best be addressed.


INTRODUCTION
The out-of-hours (OOH) period (nights, weekends, and bank holidays) comprises 63% of the week in the UK, when normal in-hours primary care services are not available. End-of-life care in the community is an important and challenging aspect of OOH provision, and hospital admissions for patients at the end of life are controversial. [1][2][3][4][5] Recent research has challenged perceptions of hospital admissions for patients close to the end of life as inappropriate, preventable, or avoidable, [6][7][8] and has highlighted that at times hospital is the only place where care is reliably, safely, and urgently available. 9,10 The challenges facing GPs, ambulance staff, nurses, and other OOH providers in the delivery of high-quality end-of-life care are significant and multifaceted. These include access to patient information; 11 meeting the clinical needs of patients often very close to death; 12 lack of confidence in providing end-of-life care; 13,14 a potentially awkward fit between end-of-life care and services' wider remit of urgent care; [15][16][17] uncertainties of prognostication; and decisions about whether the patient's condition is potentially reversible with hospital treatment or if they are best kept at home for symptomatic relief and care. 15 Hospital care is at times the best option for patients at the end of life to reliably obtain urgently needed care OOH. 15 Only some end-of-life-care-related OOH calls, however, lead to hospital admissions. 12 It is unclear which issues or combination of issues lead OOH clinicians to initiate hospital care and when patients are best kept at home. The aim of this study was to review the literature concerning the mechanisms (the components of the system that initiate admissions), the circumstances (the context of and reasons for admissions), and the processes (the actions and steps through which admissions are instigated) involved.
For the purposes of this review end-oflife care was defined as the care of patients with advanced incurable disease and an anticipated prognosis of ≤12 months of life. The term OOH providers is used to refer to all services providing access to health care for patients at night, weekends, or bank holidays. In this review, all end-of-life care hospital admissions occurring OOH, whatever the outcome of the admission, were included, and these were not limited to admissions in which patients necessarily died in hospital after admission.
The following six questions were addressed: • Which patients are admitted?
• What are the mechanisms of these admissions?
• Which OOH providers arrange these admissions?

Abstract Background
Out-of-hours (OOH) hospital admissions for patients receiving end-of-life care are a common cause of concern for patients, families, clinicians, and policymakers. It is unclear what issues, or combinations of issues, lead OOH clinicians to initiate hospital care for these patients.

Aim
To investigate the circumstances, processes, and mechanisms of UK OOH services-initiated end-of-life care hospital admissions.

Design and setting
Systematic literature review and narrative synthesis.

Method
Eight electronic databases were searched from inception to December 2019 supplemented by hand-searching of the British Journal of General Practice. Key search terms included: 'out-of-hours services', 'hospital admissions', and 'end-of-life care'. Two reviewers independently screened and selected articles, and undertook quality appraisal using Gough's Weight of Evidence framework. Data was analysed using narrative synthesis and reported following PRISMA Complex Intervention guidance.

Results
Searches identified 20 727 unique citations, 25 of which met the inclusion criteria. Few studies had a primary focus on the review questions. Admissions were instigated primarily to address clinical needs, caregiver and/or patient distress, and discontinuity or unavailability of care provision, and they were arranged by a range of OOH providers. Reported frequencies of patients receiving end-of-life care being admitted to hospital varied greatly; most evidence related to cancer patients.
• How frequently are these admissions arranged?
• Why are these admissions arranged?
• What are the processes of these admissions? Search results were imported to EndNote (version X9) and de-duplicated. The PRISMA flow diagram for study selection is shown in Figure 1.

Study selection and data extraction
Title screening was undertaken by the first author and abstract screening by the first author and one other author independently, with disagreements resolved by discussion. Full texts of potentially eligible publications were assessed by the first author and reviewed by two or more authors if there was uncertainty. Data were extracted into a review-specific data extraction form provided in Supplementary Table S1.

Quality appraisal
Gough's Weight of Evidence (WoE) framework was used to assess the quality and relevance of included publications. 19 This framework rates the internal validity of the study (WoE A), appropriateness of study design to the review aims (WoE B), and the focus or relevance of the study to the review aims (WoE C). Each of the three domains is individually scored on a scale from 1 (low) to 3 (high), and combined to provide an overall judgement (mean score of A, B, and C) to generate an overall assessment of study quality and relevance (WoE D). WoE appraisal was carried out independently by three authors, with disagreements resolved by discussion. Publications assessed as being of high WoE were considered more credible and relevant and were given priority during data synthesis. 19,20 Given the scarcity and diffuse nature of the evidence identified, publications assessed as low WoE were also included.

Data synthesis
Data synthesis used a narrative approach, 20,21 selected for its potential to assess and synthesise heterogeneous and complex evidence in a rigorous and replicable way. 22,23 A thematic approach was employed for review questions where sufficient data was available whereas a descriptive approach was adopted for questions with limited available evidence. Reporting followed the PRISMA Complex

How this fits in
Out-of-hours (OOH) end-of-life care hospital admissions are a concern for patients, families, clinicians, and policymakers. Little is known about the mechanisms, processes, and circumstances under which such admissions occur. This review found that admissions occur to address clinical needs, unavailability or discontinuity of care, and patient or carer distress. They are initiated by a variety of OOH providers and services, indicating the issues identified are widespread and systemic. Existing evidence, however, is scarce, and further research is required to understand why these admissions occur and how the issues identified might best be addressed.
Interventions extension statement and checklist. 24,25 Patient and public involvement A patient and public involvement advisory group of six people with experience of hospital admissions towards the end of life met twice, initially advising on refinement of review questions and later commenting on emerging findings. The review protocol was registered with PROSPERO (registration number: 42019156827).

RESULTS
Database searches retrieved 30 033 records. Following de-duplication and title and abstract screening, 22 articles were assessed in full text for eligibility. Four were excluded and seven additional articles were identified from hand-searching and reference lists of eligible articles, leading to a final total of 25 included publications ( Figure 1). 1,10,12,15, Characteristics of included studies The included publications comprised 23 original articles and 2 project reports, using qualitative (n = 9), quantitative (n = 10), and mixed-methods designs (n = 6). Studies were conducted in the UK: UK-wide (n = 2), England (n = 13), Scotland (n = 7), England and Scotland (n = 2), and Northern Ireland (n = 1). Articles were published between 2002 and 2020. Only one article was dated before the 2004 UK introduction of the new contract through which responsibility for commissioning and providing OOH care passed from GPs to primary care trusts. 26 Supplementary Table S2 summarises the included publications.

Focus and nature of available evidence
Publications varied considerably in terms of their focus and the nature of the evidence reported with respect to the three dimensions of this review. Most presented either 'indirect evidence' (data extracted was interpreted, for example '68% of hospital admissions occurred within working hours' implies that 32% occurred OOH) and/or 'generalised evidence' (data extracted was implied, for example '68% of advanced cancer patients were admitted to hospital' implies some were admitted OOH). See Supplementary Table S3 for detailed mapping of included publications based on focus and nature of available evidence. Results are presented with priority being given to those review questions for which the most evidence is available. were identified, grouped under thematic areas of participants, structures and processes, and clinical factors (see Box 2).

Reasons for OOH hospital admissions at the end of life
In terms of participants, reported circumstances related predominantly to informal caregivers, associated with their preferences, inexperience of death, and feelings of uncertainty, burnout, or emotional breakdown. 1,10,36,37,41 Less frequently, admissions were arranged in response to patients' needs or wishes, [36][37][38] or patient and informal caregiver distress and anxiety. 39,42 Circumstances pertaining to structures and processes included lack of systems to ensure continuity of care and/or limited access to patient information; 1,10,15,28,32,36,38 lack of flexibility in OOH service provision; 28,35,36 unavailability or limited capacity of alternative community services; 15,38 lack of time to assess and address needs; 38,42 and complexities of workforce management. 1 Provider-related circumstances included poor communication and lack of training; 1 unfamiliarity with patients; 28,35,37,42 and feelings of professional underperformance or fears of disciplinary repercussions if hospital admissions were not instigated. 15 Policy-related circumstances were associated with lack of advance care planning 15,28,42 or concerns about timely access to anticipatory medications. 28 Clinical factors related to assessment or diagnosis, symptom management, and precautionary admissions. Admissions were predominantly associated with treatment of pain or other physical symptoms including rapid deterioration in a patient's condition [27][28][29][30][31]33,34,36,37,40 and complications or failure of initial treatment. 29,30,34 Other reported circumstances related to: clinical or diagnostic uncertainty; 15,30 complexity of the situation; 39 need for further investigation; 37 abnormal laboratory results; 29 or as a precautionary measure. 31

Processes of OOH hospital admissions at the end of life
Limited evidence is available for the processes review question from seven publications, 1,12,26,35,39,43,44 with considerable variation in the level of detail provided. Processes were described as a chain of communication including a series of steps (triaging) and involving a number of different providers, at times using NHS pathways algorithms 12 or agreed protocols. 44 These processes varied considerably between services. Calls were answered by nonclinical call handlers, 1,12,44 experienced specialist nursing staff, 39 or other healthcare professionals. 1 Triaging was by initial telephone consultation with a clinician, 12 by forwarding to specialist advice 39 or call back from a healthcare professional. 44 Hospital admissions were arranged at all triage stages: on initial call to a call handler, after clinician phone consultation, or following a home visit. 44

DISCUSSION Summary
The literature addressing OOH end-oflife care hospital admissions in the UK is largely unfocused and limited by the heterogeneity of the evidence presented. Available data indicates that admissions are initiated in relation to informal caregiver and patient distress, discontinuity (or unavailability) of services and/or access to patient information, and symptom management issues. Hospital admissions are arranged by a variety of OOH services and providers, most prominently OOH GPs. The limited evidence focuses largely on cancer populations and reported admission rates varied greatly between studies.

Strengths and limitations
To the authors' knowledge, this is the first systematic review of the factors leading UK OOH services to admit patients receiving end-of-life care to hospital. Initial database searches were adapted to ensure conciseness and concreteness. Clearly defined criteria for study selection and explicit methods for data extraction and synthesis reduced biases and offered a transparent and replicable process. The patient and public involvement advisory group assisted in refining the research questions and interpretation of findings.
The review was hindered by inconsistent definitions of patients receiving palliative care and end-of-life care, and often heterogeneous patient populations in the included studies. The depth of narrative synthesis was limited by the focus of the studies identified and the nature of available evidence, which addressed diverse aspects of OOH provision either generically or indirectly. The focus of many studies on patients with cancer may not be generalisable to non-cancer populations that have been little studied to date.

Comparison with existing literature
The circumstances identified in this review as leading OOH providers to

Implications for research and practice
This review provides evidence as to why issues experienced during OOH may lead to end-of-life hospital admissions (circumstances), how such admissions occur (processes), and by whom they are instigated (mechanisms). Importantly, although the findings may be unsurprising to many clinicians and end-of-life care researchers, this review highlights significant gaps in the evidence. Knowledge on how the identified factors interact with each other (for example, how circumstances may affect processes or how different mechanisms may respond to different circumstances) is currently lacking. Also lacking is evidence of effective interventions to improve care to prevent potentially avoidable end-of-life hospital admissions. 54 The issues highlighted are pertinent to end-of-life care provision at all times of the day and night, although they seem to be particularly acute when they occur OOH. What the current review suggests is that, although OOH end-of-life care can often be readily resolved by hospital admissions, it comes with multiple challenges that appear to be widespread and systemic. Some of these challenges might be prevented by action in-hours 1 or better management of unscheduled care episodes within the community leading to reduced hospital admissions, which is what most recent empirical evidence seems to suggest. 55 Bearing in mind, however, that the OOH period comprise the majority of the week, service managers, commissioners, and policymakers need to continue to strive for integrated and comprehensive approaches to end-of-life care, 24 hours a day, 7 days a week. 56 Ethical approval Not applicable.

Provenance
Freely submitted; externally peer reviewed.

Competing interests
The authors have declared no competing interests.