Empowering Better End‐of‐Life Dementia Care (EMBED‐Care): A mixed methods protocol to achieve integrated person‐centred care across settings

Globally, the number of people with dementia who have palliative care needs will increase fourfold over the next 40 years. The Empowering Better End‐of‐Life Dementia Care (EMBED‐Care) Programme aims to deliver a step change in care through a large sequential study, spanning multiple work streams.


| Dying with dementia
One-third of people aged over 65 will die with dementia, 1 the commonest cause of death in the United Kingdom. By 2040, it is estimated that 220 000 people will die each year with dementia in England and Wales, and many will have palliative care needs. 2 Although people may die in the early or moderate stages from diseases other than dementia, about one-third of people with dementia live until advanced stages. 3 Globally, the number of people dying with dementia who experience serious health-related suffering will increase fourfold over the next 40 years. 4 Most live and die with dementia and other conditions that are becoming increasingly common with advancing age (eg, diabetes and heart disease).
As dementia and multi-morbidities progress, symptoms such as pain and breathlessness 3,5 are common and cause significant distress for the person and their carers. 6 People with dementia and their families may struggle to express their needs and concerns, leaving problems under detected and undertreated over the last months and years of life. Death can be unpredictable, and uncertainty is inherent, 7 with inequity of access to palliative care. 5 A shift is required, from delivering care based on an often unreliable prognosis to care that is responsive and based on need, whatever the stage of dementia.

| The role of palliative care for people with dementia
Dementia is the most feared disease for people aged over 60 years, 8 but there is less awareness of dementia as a life limiting illness where a "good death" is possible. The concept that a person can "live well" with dementia is promoted widely and is beginning to gain widespread understanding by the public. That people can "die well" with dementia is less accepted.
Most people with dementia die in care homes (58%). 9 England has one of the highest rates of hospital death in dementia (40%), 10 and few people die at home. 11,12 Avoiding care transitions is an important goal, but policy has focussed mainly on promoting death in usual place of residence rather than on quality of dying. 12 Carers and people with dementia report comfort and quality of life (QoL) should be the main goals of care. 13 Recent NICE Dementia Guidance states people with dementia should receive flexible, needs-based palliative care that takes into account the unpredictable progression. 14 Palliative care is patient-centred care in advanced disease (see Box 1). 15 The cornerstone is a focus on detailed detection and assessment of physical, psychosocial, and spiritual needs, with regular review and careful management. This needs-based approach allows the person, their carers, and staff to focus on living well, maximising comfort, QoL, and managing uncertainty. Palliative care models for other life-limiting conditions reduce symptom burden and are cost effective. 16 A palliative care approach for people with dementia is recommended 17

Key Points
• Dementia palliative care is an under developed priority area.
• We lack cost-effective interventions to empower people with dementia, carers and staff, to identify and meet physical, psychosocial and spiritual needs, across care settings and transitions.
• We will undertake policy and literature reviews, large data and cohort studies, knowledge synthesis, develop and evaluate interventions, underpinned by a new conceptual framework for dementia palliative care.
• Research is guided by patient and public involvement, innovative work with artists, policymakers and third sector organisations amend accordingly.
widely used in practice. Most care for people with dementia is delivered via "generalist" staff who frequently deal with palliative care in their role (ie, general practitioner [GP], community and mental health nurses, and ambulance and care home staff) without specialist training or support. A palliative care approach is often started in a crisis or when dying is imminent 5 but should be integrated earlier into routine practice by generalist staff and be "everybody's business." 18 Significant deficits in knowledge and service provision remain; palliative care is a neglected area of dementia research. 19 Prior studies have focussed on the advanced stages of dementia in care home residents. 3,5,13 There are no UK studies of palliative care need in those with early-stage or moderate dementia who may be dying from other life limiting diseases or co-morbidities, which have different profiles of symptoms and carer experiences, compared with those dying from advanced dementia, or in people with young onset and rapidly progressive dementias (an archetype of which is prion disease). The latter conditions pose special challenges, with rapid deterioration, a lack of time to prepare for death and are highly distressing for carers who may also be managing employment and younger families. Services are patchy, inequitable, and not always age appropriate. 20,21 Given the increasing numbers of people who will die with dementia, the distress this can cause carers, the pressures on staff to manage dying in the context of an unpredictable trajectory and funding pressures on the health system, we need to develop new knowledge and innovative cost-effective care. New interventions will need to empower people with dementia of all ages, carers, and staff, to identify and act upon changing physical, psychosocial, and spiritual needs, addressing unmet needs across care settings and facilitating the transfer of information across transitions.

| Previous work informing EMBED-Care
In previous studies, we found high levels of persistent pain and distressing symptoms in care home residents with severe dementia 5 and increasing distress in carers. 6,22 Our routine data analyses show place of death is moving from hospital to the community 10 but increasing numbers of people with dementia attend emergency departments in the last year of life. 23 We have developed decision aids, using a codesign approach, based on heuristics 24,25 (rules of thumb) to manage difficult decisions towards the end of life. 25 We developed, validated, and refined the Palliative Care Outcome Scale (POS, http://pos-pal. org/), a tool to improve comprehensive assessment, identification of unmet needs, screening, communication, care, and outcomes in lifelimiting disease in routine care. 26 Systematic reviews 27 and empirical, quantitative, and qualitative work 28,29 have informed development of the Integrated-POS for Dementia (IPOS-Dem). 30,31 This multifaceted tool takes a person-centred approach, 27 which asks about the person's most important concerns, and then assesses symptoms and unmet needs, enabling carers and nonclinical staff to identify needs and plan and improve care. 31 Using realist methods, we developed an understanding of integration, 32 facilitators and barriers to good dementia palliative care, 33 and guidance for evaluation of regional health care strategies and policies. 34 The EMBED-Care Programme will add to our existing knowledge and build on this work, developing a complex intervention based on our research to identify unmet needs using IPOS-Dem and providing practitioners with resources to inform clinical decision making, communication, and management of these needs (ie, heuristics or clinical "rules of thumb").

| AIMS AND OBJECTIVES
Our overall aim is to promote a "step change" in care for people with dementia and their carers to enable delivery of timely person-centred care, to identify unmet needs and improve outcomes, including comfort and QoL, towards end of life. Each of six-interdependent work streams (WSs) will deliver new knowledge and sequentially build a complex intervention to improve palliative dementia care. Our objectives are to: 1. Synthesise health and social care policy, guidance, and evidence for the effectiveness of interventions to improve QoL, to understand context, inequalities, workforce, and levers for change. We will use this data to develop our programme theory to underpin our research (WS 1);

BOX 1 Definitions of palliative and end of life care
Palliative care is the active, total care of the patient whose disease is not responsive to curative treatment. Control of pain, of other symptoms, and of social, psychological, and spiritual problems is paramount. Palliative care is interdisciplinary in its approach and encompasses the patient, the family, and the community in its scope. In a sense, palliative care is to offer the most basic concept of care-that of providing for the needs of the patient wherever he or she is with UK and international partners, to leverage a sustained step change in care. We will involve people affected by dementia, carers, staff, researchers, third sector, and policymakers to stimulate a public discourse on dying with dementia and provide a platform for workforce development, research, education, implementation, policy, and public engagement.

| METHODS
We first describe the key perspectives and theoretical models that underpin the programme followed by aspects of the work that run through the programme, including patient and public involvement and engagement activities. We then give an overview of the individual WSs and how these will work together to build knowledge and develop interventions to improve care. The overall structure of the programme is illustrated in Figure 1

| Conceptual framework and implementation
A conceptual framework of integrated palliative dementia care will underpin the programme. A theory of change (ToC) model 35 will be used to understand how the EMBED-Care intervention can facilitate integrated palliative dementia care using a theory-driven approach. This will describe organisational context, the challenges facing integrated palliative dementia care, implementation requirements, and linkages between processes and intended outcomes.
The framework will provide theoretical cohesion, inform the choice of research methods, for example the selection of tools and outcomes, and be refined across the programme as we build knowledge.
Implementation is considered from the beginning of the programme. The Knowledge-to-Action (KTA) Implementation Framework 36

| Involvement of people with dementia and their carers
Our programme is built on research priorities identified by people affected by dementia, the Alzheimer's Society and James Lind Alliance. 37 Active and collaborative involvement is integrated throughout.
Experts by experience, one of whom is a co-applicant, helped write the proposal. We held a workshop funded by the NIHR Research Design Service "Enabling Involvement Fund," which led to changes including developing thinking around comfort as a key outcome and widening to include QoL, endorsing the use of the term "carer" for family and friend carers and strengthening our social media strategy.
F I G U R E 1 Overview of the EMBED-Care Programme. EMBED-Care, Empowering Better End-of-Life Dementia Care; WS, work stream Our study reference panel will comprise eight experts by experience: carers and people with dementia, including young onset and prion diseases. It will meet every 4 to 6 months with programme and WS leads and their teams to inform important aspects of the work.
The panel will contribute to study materials such as information sheets to improve accessibility, develop qualitative topic guides, and support data analysis. They will inform broader public and policy engagement, for example, designing of our YouGov social attitudes survey and contributing personal narratives. We will work with people living with dementia holding four workshops across the United Kingdom with the Dementia Engagement and Empowerment Project (DEEP) at key stages of the programme.

| Engagement
We will deliver a novel programme of public and policy engagement targeting different sections of society to maximise knowledge exchange.
We will work with Science Gallery London (https://london.sciencegallery. com/engage), a space hosting participatory exhibitions, linking art, science and health, the university, and the local community, particularly young people. Our objective is to start conversations among groups who may not have been previously engaged with this issue but are likely to experience dementia in the future as relatives or carers. Key beneficiaries are people with dementia and their families and friends, with whom we hope to build awareness and understanding; service commissioners, voluntary sector organisations, and policymakers who we need to equip with information to drive change; and health and social care practitioners to ensure the intervention is embedded in day-to-day practice.
Our impact plan is based around three phases.
1. Early programme: a co-creation model to shape our research and engagement strategy. Activities include a co-creation workshop with young people, which will inform an art installation by commissioned artists from the Science Gallery London network. We will hold public workshops to identify questions for a nationally representative YouGov survey to understand knowledge, attitudes, priorities, and concerns around dying with dementia. A policy lab, which is an interactive event bringing together researchers and policymakers using co-design techniques, will define the information policymakers require and how we can best work with them. Monitoring and evaluation of impact will be led by a specialist evaluation consultant from SGL's network. Working with the research team, they will employ mixed methods, pre-and post-participation questionnaires deployed at each activity, semi-structured interviews, and focus groups.
The Network for Excellence in Palliative Dementia Care will bring together the third sector, academics, policymakers, advisors, and international partner organisations from across the specialities of dementia and palliative care, including the Worldwide Hospice Palliative Care Alliance, Marie Curie, Hospice UK, Dementia UK, Alzheimer's Society, CJD Support Network, and YoungDementia UK. The NEPDC facilitates input from this broad range of stakeholders into design and delivery of this and future research and supports dissemination and implementation of research outputs. The Network for Excellence in Palliative Dementia Care will be facilitated through our website, regular newsletters, and knowledge exchange events and will facilitate dissemination of our findings and learning to stakeholders including people with dementia, carers, and staff.

| Capacity building
We have included four PhD studentships, linked to local ESRC graduate schools, in the programme to build capacity at doctoral level:

| Research methods
Methods will change and develop during the programme as knowledge is built at each stage. Below, we give an overview of each WS and how they will work together. Indicative details of methods, populations, and deliverables for each WS are given in Table 1. Synthesise evidence and knowledge from WSs 1 to 3, to provide a foundation and inform recommendations and components for our new intervention.
Data synthesis using a matrix approach that draws on framework analysis and systematically evaluates evidence for strategies to address unmet needs.
Focus groups will appraise the matrix to help understand and organise the data and introduce conflicting evidence and ideas that provide a rich "real world" discussion.
WS 1 policy reviews, WS 2 linked data sets analysis, and WS 3 cohort studies (over 400 000 patients), enriched with a range of external data linkages and natural language-processing algorithms to extract structured data from free text. 39 CRIS has been previously used to investigate end-of-life issues in dementia, 23 and this WS builds on WS 1 reviews and will inform the case for new models of dementia palliative care.
We will specifically investigate key factors, including inequalities and multi-morbidity, associated with health service use in the last 2 years of life for people with a previous diagnosis of dementia of any severity and use this information to project future need. Analysis will be informed by our prior work and the programme conceptual framework. We will hold a second DEEP workshop to consider the implications of findings for people with dementia.

| WS 3: Cohort studies on unmet palliative care needs
The aim of this WS is to understand unmet palliative care needs in people at all stages of dementia, including those in the early and moderate stages who may be dying from other illnesses. We will include sub-cohorts with young onset and rapidly progressive (prion) dementias. This will be a prospective longitudinal mixed method cohort study involving people with dementia (all subtypes) over 9 months. Inclusion criteria will be informed by WS 2 and our prior studies. Study measures will be informed by WSs 1 and 2 and our patient and public involvement (PPI) Study Reference Panel. We will define unmet needs using tools such as IPOS-Dem. 30 Outcomes will include validated scales for comfort and symptom management and care transitions. We will measure carer unmet needs and outcomes such as anxiety, depression, and anticipatory grief, to explore how these are associated with unmet palliative care need. If participants die, we will interview carers postbereavement to understand final care transitions and care quality.

Health economics
Our aims in the economic evaluation are to assess participants' QoL and estimate costs of care from health and social care (NHS) cost perspectives.
This will include social services and community support, day care services, outreach, symptom control, and respite services. Unit costs will be obtained from Personal Social Services Research Unit and National End of Life Care Intelligence Network.
Unpaid carers often provide essential support and care to patients with dementia. Their contribution to care needs to be recognised and valued. A failure to do so risks undervaluing the total cost of care.
Thus, in addition to health and social care costs, our analysis will take into account the contribution of unpaid carers as if they were paid at the same rate as a paid carer. Total cost estimates will also be expressed as increments per 1 point change in our primary outcome, IPOS-Dem. Quality-adjusted life years (QALYs) will be calculated using the EQ-5D-5L values from each follow-up point from baseline to 9 months. In the population nearing the end of life, it is unclear how well the EQ-5D-5L will function, and so a dementia-specific QoL measure, DEMQOL, will also be utilised.

Sample size
We estimate that care transition occurs in 50% of people with demen-

Analysis
We will describe participant characteristics and quantitative outcome measures using mean/median (SD, IQ range) and counts (percentage) as appropriate. We will investigate bias from missing data and use multiple imputation for missing values of explanatory variables if considered appropriate. Logistic regression accounting for whether patients are recruited from care homes or the community will be used to estimate the associations between pre-specified explanatory variables such as unmet needs and outcomes, ie, first care transition during the study. To explore trends in IPOS-Dem, comfort, symptom management, and care transitions over time, we will use regression models that can incorporate repeated measurements.
We will explore further unmet needs and care transitions using individual interviews comprising dyads of a person with dementia (when possible) and a carer (ie, family). We will interview separately the main practitioner involved in providing care identified by the family to understand their perspective. Interview data will undergo thematic analysis, triangulated with our quantitative data. A third DEEP workshop will consider implications of findings for people with dementia, key messages for policymakers and commissioners, and how they should inform care.

| WS 4: Knowledge synthesis for intervention development
This critical step of the EMBED-Care programme will synthesise knowledge into a product (the intervention) to enable translation of research evidence into practice as part of the KTA Framework. 41 This innovative approach moves beyond traditional evidence-based guidelines, by incorporating information about existing practice and policy and the resources needed in different contexts. Our complex intervention will require synthesis of diverse types of knowledge, in differing formats, with varying degrees of value, to be robust, reliable, and practical for implementation in real world settings. 42 We will synthesise data from WSs 1 to 3, our work on IPOS-Dem 30 and the heuristic decision aid. 25 We will systematically evaluate evidence for strategies to address unmet needs, incorporating them into a knowledge synthesis matrix (for an example of this, see Table 2). This pivotal step of the KTA links knowledge creation with action to support effective uptake by potential end users.
The matrix will be based on framework analysis, 43  3. An intervention manual and training and support package for staff.
We will consider the use of telehealth as a key mode of delivery.
We will use the WS 4 knowledge synthesis to develop prototype decision aids. We will work with co-design 45 groups involving practitioners, current and former carers, and people with mild dementia.
This allows collaboration on intervention design and delivery between researchers and end users to promote successful implementation and translation into practice 46 and links knowledge creation with action in our KTA framework. 36 Co-design groups will iteratively refine the heuristic resource. A training and support package will be designed and reviewed in a final workshop. The intervention will be field tested in practice (home and care home settings) for 4 weeks, and after this, we will hold a focus group to explore use barriers and facilitators, content and design, and modifications to implementation or delivery.

| WS 6 Feasibility study
We will conduct a feasibility study of intervention implementation and recruitment and then pilot the intervention in a cluster RCT. Clusters will be individual care homes or communities of practice. We will refine the intervention, inform the design of future studies, and establish whether a future definitive trial is warranted. To ensure methodological rigour, we are supported by an MRC-registered clinical trials unit (PRIMENT-CTU), who will be involved at all stages: planning, approvals, randomisation, data collection, storage, and analysis. We will need to be flexible with feasibility and pilot trial design to allow for likely changes in health and social care systems during the programme. Detailed methods will be informed by WSs 1 to 5. For example, potential outcome measures will likely include care transitions, comfort and service use, but selection will be informed by programme The feasibility study will run the intervention for 3 months, supported by a champion in each care setting. We will test the feasibility of data collection at baseline and 3 months. We will evaluate acceptability of the intervention and conduct a process evaluation to understand risks of harm, implementation requirements, and contextspecific adaptations. 50 We will use data from a range of organisational levels and informants. For example, clinical commissioning group service data to examine unplanned care transitions to hospital and a utility questionnaire for staff to explore practical use of the tools. Semistructured interviews/focus groups in each cluster, detailed case studies, observations of multidisciplinary meetings, and case note review will inform adaptations. Final consultation with potential end users will employ consensus methods to make recommendations on intervention components and implementation requirements. 51 The pilot cluster RC will run for 6 months in six clusters (three intervention and three control). Outcome measures will be identified from the feasibility study. Data will be collected on people with dementia and their carers at baseline and monthly. Levels of missing data will be assessed and evaluated. Recruitment and attrition rates will be estimated using proportions with associated 95% CI. We will estimate the difference (with 95% CI) in our main clinical outcomes between trial arms, using appropriate regression models adjusted for baseline values. This will inform sample size calculations for a full cluster RCT. We will assess intervention fidelity and hold focus groups and semi-structured interviews with people with dementia, carers, and health and social care staff. We will test recruitment processes, whether outcome data can be collected by researchers blind to allocation, implementation, and identify stop-go criteria for a full trial.

| Key outputs from EMBED-Care
We will ensure that EMBED-Care produces outputs that benefit people with dementia, their family carers, practitioners, and wider society. Data sets and methodological developments will build research capacity. The main outputs include the following: • Conceptual framework for integrated palliative dementia care: including key components and mechanisms of change, to underpin all aspects of the programme. Theoretical work will expand notions of "comfort," "person-centredness," and good death with dementia to inform both theory and practice of care in society (All WSs).
• EMBED-Care intervention: which may be delivered via telehealth technologies, with associated training materials and implementation guidance (WSs 5-6).
• Methodological development: innovation in data linkage (WS 2), economic methods in palliative care (WS 3), synthesis of qualitative and quantitative data and underlying theory (WS 4), and development of intervention implementation theory and framework (WS 6).
• Data sets: new quantitative and qualitative data on palliative care needs and outcomes in people with moderate dementia and their carers, and the first data on young onset and prion dementias, will be anonymised and made freely available. This will provide significant scientific benefit to sociological, economic, gerontological, and health researchers and a basis to develop and evaluate future interventions, instrument design, and services (WS 3).
• Future projections of need for palliative dementia care: to inform service planning and commissioning (WS 2).
• Policy briefings/labs: to inform and brief policymakers and commissioners together with practitioners and service users to enable evidence-based policymaking.
• Knowledge exchange events: provide pathways to policy and academic impact (all WSs).
• Art installation: co-created in workshops with an artist to increase outreach to groups who may not have engaged with the issue of dying with dementia.
• YouGov poll: snapshot measure of societal attitudes towards dying with dementia.
• Newsletters and blogs: co-developed with patient groups at key milestones to contribute their perspective to policy briefings and dissemination.
• Network for Excellence in Palliative Dementia Care: to boost the impact of the research by uniting contributing practitioners, people with dementia, carers, researchers, other stakeholders and existing communities of practice, and international partners.

| DISCUSSION
The multidisciplinary team working on our programme will sequentially build new epidemiological and social science knowledge on dying with dementia and a new model of integrated palliative dementia care.
This will address unmet needs regardless of the severity of dementia across a range of care settings. We will apply a person-centred approach maximising comfort and QoL. Our multi-component intervention will be carefully developed with attention to contextual factors such as adaptability to local models of service provision, workforce capability, and economic factors. Our innovative combination of underpinning theory, implementation science, and co-design methods will ensure adoption in practice and improved palliative care for people with dementia and their carers.
This step change in care provision requires public support and engagement, and therefore, changing societal attitudes towards dying with dementia is key to our work, as demonstrated by our pathways to impact that use a range of methods, from academic dissemination to policy work and novel projects with young people who will be affected by dementia in their future. Active and meaningful engagement of people with dementia and their carers will shape all aspects of our work. We will deliver new knowledge and interventions so that people dying with dementia and their carers receive good quality palliative care in the right place at the right time, delivered by a work force with the skills necessary to achieve this.