Using Palliative Care Needs Rounds in the UK for care home staff and residents: an implementation science study

Size and impact of the problem

Care homes (CHs) now operate as subacute units with high levels of morbidity and mortality. 1 Care homes will be the leading place of death by 2040, and mortality can be as high as 56% within a year of admission. 2 Consequently, there is an urgent need to find methods and approaches to optimise end-of-life care for this population.

Although some specialist palliative care (SPC) teams offer services to CH residents, recent surveys in England, Wales and Scotland3–5 demonstrate that referrals are usually reactive rather than anticipatory, but there is considerable goodwill from SPC to support residents in CH. 6 The National Audit of Care at the End of Life has prioritised systems and processes that support CH residents to receive personalised end-of-life care. 7

Some CH residents (hereafter ‘residents’) experience multiple admissions to hospital prior to their death,6 despite some admissions being preventable. 8 Hospital admissions are costly and may prompt futile or burdensome interventions that can cause distress to residents and family members. 9 Many residents will require SPC to enable CH staff to manage complex symptoms1 to avoid hospitalisation at end of life. Well-managed death and dying is contingent on high-quality interdisciplinary care,10 anticipatory care11,12 and resident-centred planning. 13

Care homes can be homely, warm and supportive environments. Supporting these establishments to improve the care they give at end of life can make them even better places for people to spend their final months and weeks of life, especially when compared with clinical settings like hospitals. CHs are an important location in the nexus of service provision for older people and are the de facto hospice. 14

Supporting residents to die in the care home

Care home staff wish to reduce preventable hospitalisations, yet often lack clear methods of doing so. 15 Palliative care interventions for residents in CHs report positive outcomes including reducing hospitalisations16 and improving advance/anticipatory care planning (ACP),17 while advocating for the role of specialist senior nursing practitioners. 18 Increasing anticipatory planning (including Advance Care Plans and anticipatory prescribing) improves the confidence of care staff to discuss goals of care with residents, their families and friends and supporting healthcare professionals and can lead to a reduction in hospitalisations. 19 Some studies issue caution on the impact on hospitalisations, but report improvements in preferred place of death. 20

Nurses in CHs who are supported to administer anticipatory medications can reduce hospital admissions and facilitate faster symptom management. 21 Provision of support to CH staff has been shown to improve end-of-life care for residents. 22 Needs Rounds have clear benefits for residents’ health outcomes: reducing admissions to acute care and improving quality of dying, and increasing workforce knowledge and confidence. 23–25

Telehealth is increasingly used to support CHs engage in early palliative care discussions and is acceptable to CH staff and relatives. 26 Such evidence supports changes to practice due to COVID-19,27 and telehealth has been used in Australian Needs Rounds adaptations, with positive outcomes reported. 28

Supporting the palliative care education of CH staff (despite it not being part of statutory training) needs to be seen as a priority to improve outcomes both for staff and for residents requiring a palliative approach. 29,30 Education in ACP, as part of palliative care provision, has led to increasing rates of completed plans and advance directives, improving consistency of clinical decision with resident preferences. 11,31 Advance care planning interventions mitigate distress and improve communication with relatives. 32 Yet, in a recent systematic review of 16 trials with a staff education component,33 many interventions did not result in improved quality of life, quality of death or reduced hospitalisations. Educational interventions can lack sufficient power to result in different clinical behaviour, programmes can be inconsistent and the necessary steps for sustainable change are often lacking. 29,34 Calls for standardisations in palliative care education in CHs may be helpful,35 yet this risks education not being tailored to specific contexts and circumstances. 36 Furthermore, the wider workforce context means there are inherent difficulties in providing coherent training and education where there is high turnover of staff. 35

Workforce issues continue to hamper many initiatives in improving palliative care in CHs,37 though some studies report that retention and burnout were not related, and indeed that burnout rates were unexpectedly low. 38 Maintaining few staffing vacancies is associated with greater likelihood of CHs being rated good or outstanding by the regulatory body. 39

Providing end-of-life support to CHs is an increasingly important area of service development. UK service delivery innovations such as ECHO,40 Gold Standards Framework (GSF),41 Macmillan’s education for carers ‘Foundations in Palliative Care’, Six Steps to Success,42 the EU funded PACE work43 and person-centred dementia care with the Namaste programme44 offer staff training and development, but rarely provide facilitation of evidence-based clinical input for people who are dying. Currently, only the PACE study has been fully tested in a randomised controlled trial across six European countries. The English Enhanced Health in Care Homes (EHCH) approach aims to improve links between CHs and primary care, yet it had patchy implementation due to multiple barriers such as communication, appropriate outcomes, trust and complexity. 44

Hospital prevention studies form an important backdrop to CH research. 45,46 Care Home Innovation Partnerships (CHIP) in Liverpool, for example, paired community matrons with CHs to review residents, develop ACP and treat minor ailments. The study reported a 15% reduction in emergency calls and a 19% reduction of hospital transfers. 47 Integrated care approaches, with health and social care working together, show promising impact on hospitalisations,48,49 alongside active management of long-term conditions. 50 Determining what constitutes an inappropriate admission, however, is not straightforward. 51

Advance care planning is an important tool in ensuring tailored and appropriate care. Studies suggest that while staff knowledge is important,52 their self-efficacy also plays an important role in engaging in ACP conversations. 53 However, training CH staff in conducting ACP conversations is a recognised deficit. 54 Involvement of family members can be important in ACP. 55 It is likely that ACP is one of several tools which need to be used to avoid unnecessary hospitalisations. 56

COVID-19 in care homes

Since commencing this study, publications on CH research have focused on responses to, and experiences of, the COVID-19 pandemic. In England, there were already 29,542 excess deaths in CHs by 7 August 2020. 57 In Scotland, estimates of years of life lost by CH residents were calculated at 5600 years. 58 Data from before COVID-19 suggest that relative risk of dying in CHs had already increased, with shorter intervals from admission to death. 59 Recommendations from a mixed-methods study on COVID-19 on CHs in England suggested a need for greater training and support for CH staff providing palliative care. 30 Evidence suggests an increase in ACP in CHs during the pandemic, which was helped by education/training for staff to promote their confidence and competence in having these conversations. 60,61 The increase in resident mortality alongside increased responsibility for end-of-life care took a considerable toll on staff mental health. 62 COVID-19 also brought about new practice developments63 and research examining ACP in the broader health and social care community, including the role of GPs64 and SPC teams. 65

Despite these studies, there is a need for greater understanding of the delivery of palliative care during the pandemic to CH residents. 66

National Health Service policy and practice

There is currently no statutory training requirement and no robust approach to delivering optimal palliative care to CH residents. NHS England wants to improve care in all settings and has committed to ‘explore improvements’ (p. 13)67 for residents in CHs but recognises that there are substantial difficulties in providing adequate care in these settings. The Care Quality Commission (CQC) articulates a requirement for ‘a clear focus on end of life care and applies in all services where end of life is delivered. The approach includes […] care homes’ (p. 21). 67 The CQC further states that there is more outstanding care by hospices than any other service, yet their data separate hospice care from nursing/CH care. This underlines a fundamental problem that CHs are not yet considered to be providing effective palliative care, despite the evident morbidity and mortality of residents.

Facilitating improved end-of-life care in CHs is an explicit driver for NHS England. 68 The ‘Ambitions Framework’ for palliative and end-of-life care has yet to be fully realised, but includes important elements such as fair access to care and staff/communities able to provide care and talk about death/dying. Clinical commissioning of palliative care clinical and education services across England is variable. 69

In the UK, while some SPC teams are based in hospitals and generalist palliative care is provided by primary care, it is the hospice teams who provide at-home support and care to people living in the community with advanced disease.

The Scottish Government’s 2015 Strategic Framework for Action on Palliative and End of Life Care sets out a vision of universal access to palliative care by 2021. This includes individuals, families and carers having timely, focused, conversations with appropriately skilled professionals to plan end-of-life care, in accordance with their needs and preferences. The vision sought to widen the range of health and social care staff providing palliative care, delivering appropriate training and supporting clinical and health economic evaluations of palliative and end-of-life care models. Furthermore, Healthcare Improvement Scotland commits to testing and implementing improvements to identify those who can benefit from palliative and end-of-life care,70 yet at present, there is no delivery model for this in CHs.

Wales set out a priority action in Health Boards providing access, support and education from SPC to CHs,71 but with no dominant model being offered beyond each community clinical nurse specialist linking with one or more CH. Northern Ireland guidelines recommend a designated nurse within the SPC team for CHs. 72

CH culture inevitably impacts working practices and resident care/experiences. 73 There is a need for high-quality leadership and facilitation alongside innovation41 and consideration of the nursing workforce and culture of education. 74 CH context, including the level of anticipatory and scheduled contact with primary care, is an important facet in effective healthcare in residential care, alongside joint working where residents’ needs are discussed, prioritised and responded to with a common purpose and shared values. 75

Understanding context and adjusting implementation in response to CH culture, including local priorities, readiness for change and facilitation champions are all important. 76 CH culture, and the culture change movement, has become a research focus seeking to identify characteristics which lead to care improvements. 77–80 Indeed, CHs which embrace culture change are more likely to provide better resident care. 81

The context in which CHs operate is complex, with stretched social care funding budgets from years of austerity, residents with complex needs and high levels of multiple morbidity, alongside suboptimal continuity of care within primary care services. 82 While approaches such as EHCH and My Health My Care My Home83 seek to address some of the multisector issues, this continues to be a complex space for providers, commissioners, service users and families. 84

Why is this research needed now?

Recognising the scale of people dying in CHs, the clinical risks associated with hospitalisation, and hence the drive to enable people to remain in their CHs with appropriate advance planning and support for staff, creates a context of urgent clinical innovation. Without a clear national model with robust evidence base, suboptimal care is a risk. However, there is a model of providing SPC in CHs in Australia which provided compelling data to improve care, called Palliative Care Needs Rounds (hereafter referred to as Needs Rounds). The definitive stepped-wedge trial of Needs Rounds85 provided robust evidence from 1700 residents that the new approach to care can substantially improve outcomes for residents, staff and the acute sector.

In order to deliver the Needs Rounds model in the UK, adaptation is required due to the different service organisation and delivery models and the need to engage with key stakeholders such as residents, carers, palliative care services, CHs and the acute care sector. 86

Needs Rounds outcomes

The pilot Needs Rounds intervention outcomes in Australia support UK strategic priorities, such as decreasing length of hospitalisations [p < 0.01; confidence interval (CI) −5.05 to −1.41 days], improving rates of residents dying in their preferred place23 and enabling staff to normalise death and dying25 by adopting an outreach model of SPC. A further Australian stepped-wedge trial of Needs Rounds with 1700 CH residents achieved reduced acute length of stay (p = 0.048) and evidenced improvements in residents dying with dignity, compassion and comfort (p = 0.019) and improved workforce confidence (p = 0.09). 85

Goals of the elements essential to the intervention

Needs Rounds draw together requirements for looking after older people in care, including case management and specialist outreach services,87 increasing ACP88 and staff education. 49,89 Needs Rounds strengthen current models by widening the beneficiary beyond those with advanced dementia90 and meeting core strategic targets such as improving symptom management,91 increasing preferred place of death92 and providing a framework for person-centred care to residents. 93

Creating a basic organisational structure that promotes palliative care collaboration through monthly multidisciplinary meetings (both internal and external health and social care professionals) is an important first step to build a solid foundation to provide palliative care. Such a foundation helps to break the isolation of CHs and can promote greater sustainability of further initiatives.

The context of Australian and United Kingdom care homes using Needs Rounds

While Australian CHs and UK CHs vary, many of the practicalities are similar. Both countries face similar tensions in service delivery for example high turnover of staff and residents with complex multiple morbidities. Usual care in the area of Australia where the stepped-wedge trial was conducted had reactive provision of direct clinical care from the local SPC team. This is very similar to most UK provision. 3,4

The four core elements by which the two countries’ CH contexts differ are as follows:

1. The sites involved in the Australian study all employed a registered nurse, which is an important difference in adapting and implementing Needs Rounds for the UK. This may mean that UK Needs Rounds require greater links with primary care (not just SPC). In the UK, SPC will require primary care to facilitate prescribing and, in CHs with no on-site nurses, administering of medication. It may also mean that the Needs Round ‘case-based education’ component of the model will include greater emphasis on core information and skills.

2. Australian CHs tend to be larger than the average size in the UK (average Australian size in the stepped-wedge study was 90, whereas the average in England is 29.5 and 38 in Scotland). 94 This means that the delivery of Needs Rounds will be to fewer residents and smaller staff groups in the UK.

3. Australia has a larger proportion of adults over the age of 65 residing in CHs (5.2%)95 compared with England/Wales (3.2%). 96 It is not clear, however, whether the acuity of residents’ symptoms is comparable across countries.

4. Both countries operate CHs without mandatory training for their staff on palliative or end-of-life care. In Australia, however, CH staff are able to access a national education programme (PEPA – Programme of Experience in the Palliative Approach https://pepaeducation.com/) which enables CH staff to attend a workshop on palliative care, and some days shadowing staff from SPC (e.g. in an inpatient unit). By contrast, in the UK, education is provided via initiatives such as ECHO, Six Steps or the GSF.

Overview of Phase 1

Purpose: Phase 1 sought to generate an initial programme theory of ‘what elements of Needs Rounds would work for whom in what circumstances and why, in the UK context’.

1. Qualitative interviews were conducted with key stakeholders to understand how Needs Rounds could be used.

2. The findings from the interviews were presented to participants during two online workshops. Workshop discussion focused on whether the data reflected participants’ experiences, any identified gaps and the anticipated mechanisms of change required for Needs Rounds to be successfully implemented.

3. Team analysis of interview and workshop data led to the development of five theories and an initial programme theory.

4. A third workshop was held to generate feedback on the initial programme theory and the context, mechanism outcome (CMO) configurations. The five theories’ CMOs were communicated within the workshop using fictionalised vignettes – to help illustrate the interconnections by describing example residents (see Report Supplementary Material 7–11 for vignettes). Discussions arising from the presentation of CMOs focused on overarching themes and connectors between the theories. Discussions were captured by the investigator team (including academics and PPIE).

5. A fourth Phase 1 workshop was run, focused on presenting the initial programme theory and providing training on the UK Needs Round approach.

Workshop participants included Phase 1 interviewees and a wider pool of stakeholders. The workshops created the collaborative space to co-design the UK approach to Needs Rounds. All workshop discussions were audio-recorded. Phase 1 ran for 6 months, from January 2021 to June 2021.

Recruitment to the study

Specialist palliative care sites were recruited to the study first. Existing networks of the principal investigator (PI) were approached with study materials outlining the aims, objectives and methods of the study. These sites were named in the protocol and hence agreed prior to funding. One hospice withdrew after the study was funded, and consequently, further recruitment activity took place to identify an additional site. This site was confirmed and in place prior to commencement of the study.

Subsequently, hospices that agreed to participate in the study identified local CHs to invite to participate.

Recruitment was augmented by ENRICH, a UK network focused on enabling research in CHs. ENRICH holds databases of CHs who have demonstrated an interest in participating in studies, and they contacted CHs in some of the case study areas to invite eligible homes to participate.

Interviewees were recruited through the collaborating CHs and SPC sites, using snowballing and personal networks to identify relevant stakeholders in the local communities.

Recruitment was sought across England and Scotland, with four sites in England (west Midlands, Midlands, East Anglia and South) and two in Scotland (central and northern). Purposive maximum variability sampling of SPC services focused on recruiting a heterogeneous and information-rich sample, including urban/rural, service size, deprivation, cultural demographics, use of ECHO or other SPC input models and funding models. CH recruitment sought diversity by being part of a national chain or independently run and size. These variables reflect the dominant contextual influences which were likely to impact how Needs Rounds are used in the UK.

Due to COVID-19 restrictions, recruitment was all conducted remotely, using e-mail, phone calls and video calls, rather than a combination of these approaches alongside in-person meetings.

Inclusion criteria

Stakeholders (for interviews in Phase 1 and workshop participants)

1. Work for SPC or a CH in one of the six cases, are a resident in one of the CHs, are a relative of a CH resident in one of the six cases or work in acute care impacted by hospitalised CH residents.

2. Willing to provide informed consent.

3. Have capacity to provide their own consent to participate.

4. Not engaged in any current safeguarding investigations.

Care homes (to use UK Needs Rounds)

1. Located within the service area of the SPC service.

2. Provide care to residents who have high clinical nursing/medical needs.

3. Willing to sign a memorandum of understanding with the research team, outlining provision of hospitalisation data, facilitate access to staff for interviews and engagement in delivering Needs Rounds.

4. A range of medium and large sizes (focusing primarily on larger CHs), sole traders and large corporate provider and with a range of funding models (social care and self-funded residents).

Residents who are discussed at Needs Rounds

1. Resident in a collaborating CH in one of the six case study locations.

2. An anticipated life expectancy of < 6 months or a deteriorating condition where they are at risk from dying without an adequate care plan in place.

3. Experiencing suboptimal biopsychosocial symptoms.

Relatives completing the CANHELP Lite (family’s perceptions of care) questionnaire

1. Relative of a resident who was discussed in Needs Rounds.

2. Able to provide their own informed consent.

Interviewees for the patient and public involvement and engagement evaluation

1. Coinvestigator or staff at one of the case study sites.

2. Able to provide their own informed consent.

Phase 1

Data were generated from two sources: (1) interviews and (2) a sequence of four workshop discussions.

Twenty-eight interviews (individual or paired) were conducted with participants from six case study sites. Interviews involved relatives (n = 2), clinicians/managers in CHs (n = 12), clinicians in SPC (n = 7) and their related acute/primary care staff (n = 5), allied health practitioners (n = 1) and one CH staff member who did not specify their role. Formal respondent checking of transcripts was not used, since the workshops provided opportunity to clarify, check accuracy and validate ideas/opinions with participants.

Staff participants were asked about their local context, such as services’ geography, policy, structure, funding and practice elements. Relatives were asked about their experiences of their family member’s care including symptom management, goals of care and ACP and communication and responsiveness.

These data were analysed to develop realist theories regarding how implementation would work in practice, what might influence implementation in each case study site, to identify CMOs. Interview data were collected between February 2021 and April 2021. Interviews ranged from 37 to 119 minutes, with an average of 56 minutes.

The analysis from these interviews was then presented to participants during two online workshops. Workshop discussion focused on whether the data reflected participants’ experiences, any identified gaps and the anticipated mechanisms of change required for Needs Rounds to be successfully implemented.

Team analysis of both the interview data and the two workshop discussions led to the development of five theories and an initial programme theory. These five theories and programme theory were then presented for discussion and examination at a third workshop, to generate feedback on (1) the initial programme theory, (2) the CMO configurations and (3) overarching themes and connectors between the theories.

A fourth workshop was then held to present the initial programme theory and use feedback from participants to make revisions to it. This fourth workshop also included training on the UK Needs Round approach for the CH staff and SPC staff who would be delivering the intervention.

Forty-three unique participants attended the workshops in Phase 1. These participants were from 23 organisations including CHs, SPC organisations, and people with PPIE experience. The workshops were conducted in April and June 2021 and lasted for approximately 3.5 hours each.

Throughout the four workshops, the research team, including PPIE members, made notes capturing key parts of the discussion. Transcripts of the workshop discussion were also used as data sources and hence were subject to analysis.

It was intended to conduct interviews and workshops face to face. However, due to ongoing COVID-19 restrictions, reduced willingness/ability to travel and greater familiarity with online platforms, all data collection was conducted online.

Phase 2

Qualitative data

Transcripts of audio data and documentary evidence were stored and organised using Nvivo 11, 12 and 20.

Critical realism informed the thematic analysis used with all qualitative data, following the five-step process outlined by Braun and Clarke. 111 Stage 1 involved familiarisation with the data set through repeated rereadings. Stage 2 involved identifying an initial thematic framework, which was used in Stage 3 where data were indexed with reference to the thematic framework. In Stage 4, data were synthesised from across respondents into consolidated themes. Stage 5 focused on interpretation and finalisation of key themes generated from the data.

Within- and between-case analysis was conducted inductively (inferring from specific data to draw conclusions and develop theory), drawing on process tracing and constant comparative methods, respectively. Deductive analysis (testing the theories with reference to data) was also used to refine the CMO theories. Following realist approaches, data analysis also drew on a ‘retroductive’ approach, to move beyond induction and deduction to allow for ‘the identification of hidden causal forces that lie behind identified patterns or changes in those patterns’ (p. 1). 112 Retroduction describes the team’s interpretation of generative causation in identifying and naming contexts and mechanisms which lead to the intervention outcomes.

Phase 1 (pre-implementation) data were analysed to create an initial programme theory to be tested during implementation. Analysis focused on identifying the contexts (in which Needs Rounds would run), mechanisms of change and outcomes. Contexts were analysed as either inner/micro/meso context (individual or organisational level) or outer/macro (such as the wider policy/cultural context). Mechanisms of change were organised into categories to include facilitators (people), facilitation (process), resources and reasoning. Innovation focused on understandings of Needs Rounds, perceptions of value and degree of fit. CMO configurations and their chains of inference were generated for five theories from these data.

Qualitative data collected in Phase 2 (at months 4, 8 and 11 of implementation) were used to refine the programme theory. The coding was organised in relation to the five theories (see Theory 1: confidence and competence and Theory 5: supporting families), identifying CMOs. CMO configurations were updated based on what worked or did not work during implementation. Analysis was an iterative process, examining differences between the initial programme theories and Phase 2 implementation data.

Further coding was developed to allow analysis of engagement of sites in the data collection by CH staff. Tables were created identifying the main CMOs under each theory with illustrative data.

The PPIE evaluation analysis was conducted by the independent researcher to maintain the confidentiality of respondents. All other qualitative data were analysed by three experienced female qualitative researchers (all of whom hold PhDs) who had ongoing contact with staff at all study sites. Codes and themes were discussed between those three team members to develop consensus on data interpretation.

The guiding principle for analysis and sampling was theoretical sufficiency113 with triangulation used for all data sources to aid development of the theory.

Survey data

The quantitative analysis was conducted by social scientists with expertise in statistics, who were not involved in data collection. Therefore, the researchers responsible for analysing the survey data had no interaction with either CH staff or hospice staff. During the analysis, ID numbers were used (e.g. ST00335) in place of organisations’ names, creating a quasi-blinded state whereby the researchers analysing the survey data were not connecting results to specific CHs or case study identity during analysis.

Fidelity

Fidelity is concerned with the extent to which an intervention or programme is carried out in the way intended. It encapsulates the delivery, receipt and enactment of an innovation. 115 Various tools have been created to measure fidelity, focusing on both content and/or processes that impact implementation outcomes. Content focuses on the specific information to be conveyed, while process focuses on facilitation, including the skills of the practitioner, and the quality of delivery and engagement. 116 Given the centrality of i-PARIHS to this study, facilitation is reported in more detail (see Facilitation and facilitators).

Assessment of fidelity to the agreed approach to Needs Rounds was conducted based on the singular UK model developed and agreed in workshops. Fidelity was assessed against the published checklist,24 since Phase 1 determined that the UK approach would closely model the approach outlined in the checklist.

Fidelity was assessed through analysis of a random sample of 20% of all audio-recorded Needs Rounds. Fidelity is most often associated with positivist research paradigms; within this study, however, it was approached as a realist endeavour to map engagement with the agreed model. A three-tier scoring system, of 1 (high adherence), 2 (moderate) and 3 (low), was adopted. Operational definitions for these scores were developed to aid coding. For example, case-based learning is a core element of Needs Rounds meetings and was coded as high (relevant learning included for the majority of residents discussed), moderate (some relevant learning provided during the Needs Rounds, but not for most residents) and low (little or no relevant learning provided during the meeting). The fidelity schema was collaboratively developed by two researchers, with reference to the literature on fidelity, and coding was conducted by one researcher. The fidelity criteria are provided in Report Supplementary Material 14.

Estimating the treatment effect of the intervention on health service outcomes

Data on hospitalisations were requested from CHs across 5 months at baseline, between March and July 2021, and across a further 3 months at follow-up, from March to May 2022 (months 9–12 of the intervention).

At baseline, 11 CHs returned data for at least one of the months requested; however, there were a substantial number of missing and incomplete data. A minimum of 1 month’s complete data were obtained from nine of the 29 participating CHs; the average response rate was 3 months of data across those nine CHs, with only one site providing complete data for all five baseline months. At follow-up, six CHs returned at least 1 month of complete data; however, one had not provided adequate baseline data, and thus comparison could not be made.

Exposure was calculated by examining the reported data on the number of beds within each of the five CHs, the number of empty beds reported each month and the number of days of data that were reported (i.e. if data were returned for March and April, the number of baseline days equalled 61).

Using data reported from the CHs about the number of hospital admissions over the baseline period and the resulting number of nights spent in hospital, calculations were made to create ratios for admissions per 1000 exposure and nights in hospital per 1000 exposure for analysis.

Only five CHs returned data at both baseline and follow-up, which severely restricted the possibilities for analysis and interpretation. Our analysis therefore reports descriptive statistics in Chapter 5 for the estimated treatment effect on the two outcomes above (number of admissions and number of hospital days) for each of the CHs, as well as an average treatment effect. The sample size is too small to conduct an inferential test for statistical significance on the average treatment effect.

Care must be taken in interpreting both the site-specific treatment effects and the average treatment effect both due to the low power of the analysis and the confounding effect of COVID-19 on estimating a casual effect using a pre/post research design.

Estimating the cost effectiveness of the intervention on health service outcomes

A cost–benefit analysis of the intervention was planned from a health and social services perspective. The health economic analysis plan was described in the original funding bid. The intervention cost was to be derived from both direct and indirect costs to both NHS and CHs of delivering the intervention. Calculations were to include the change in NHS costs incurred following the intervention, including both primary and secondary care, by valuing the reduction in hospital stays, hospital days as a result of the intervention, ambulance usage, GP callouts and visits by specialists.

As described above, due to insufficient data, this analysis was not possible, both as a result of the small samples of data returned by CHs and the compromised validity of the primary treatment effect estimation due to COVID-19. There was insufficient detail in the data collected to accurately estimate the costs of hospitalisations or the wider health service costs. To augment the sparse quantitative data, in Chapter 5 we also provide qualitative data on costs gathered in the study.

Care home participation

To examine which, if any, factors related to the involved CHs were associated with greater participation in the study, analyses were conducted to examine different areas of participation.

Four binary variables were created: whether the CHs provided baseline data, whether the CHs provided follow-up data, whether the CHs returned any QODDI surveys and whether the CHs returned any CAPA surveys. These were scored as zero for no data returned and one for some data returned. A fifth overall participation score was calculated by summing four variables representing participation in different areas of the study (Table 3), including the number of QODDI and CAPA surveys returned and the number of months CHs returned data on hospitalisations and assessments and interventions triggered by Needs Rounds (see Resident data).

Descriptive data were compared for all aforementioned participation variables across the following predictor variables: CH size, ownership, country, location and types of care provided (see Descriptive analysis).

Some variables were derived using data that CHs provided. For example, CHs were categorised as either medium or large based on the number of beds they reported against CQC categories (see Staff capability: CAPA), and each site reported their ownership as either private or not for profit. The CHs were located either in England or Scotland, so a simple binary variable was created to reflect this and draw comparison between the two countries.

To determine whether location impacted CH participation, a binary rural–urban variable was created. Using the Office for National Statistics postcode directory,117 the postcode for each CH was used to find the 2011 output area (OA11) code. For the English CHs, the OA11 codes were cross-referenced in the Department for Environment, Food and Rural Affairs, 2011 Rural–Urban Classification lookup tables for small area geographies118 to determine which CHs were in rural areas and which were urban. This process differed slightly for the Scottish CHs, in that the OA11 codes were used to obtain sixfold urban or rural codes from the Scottish Postcode Directory Files. 119 The sixfold codes were collapsed into the twofold (binary) rural urban classification using the Scottish Government Urban Rural Classification. 120

Finally, using the carehome.co.uk profile for each site, data were collected on the types of care provided at each home. Types of care included nursing, residential, convalescent, respite, day or lunch clubs, specialist care, dementia care and care to adults under 65 years.

A binary logistic regression was conducted to model participation, by means of completing baseline data, as a function of potential predictors of participation. The same models were applied in a linear regression analysis with overall participation score as the outcome. In model 1, location variables were included to identify how participation varied in England compared to Scotland and in urban versus rural CHs. In model 2, factors relating to the CH such as the size and the ownership were added. Ownership was dropped from this model for the baseline data analysis because, as shown in Descriptive analysis, 100% of the voluntary, not-for-profit CHs in the sample provided baseline data. In model 3, we controlled for the type of care provided. Residential was selected as it is one of the two main types of CH (i.e. CH with nursing – referred to in this section as nursing, or CH without nursing – referred to in this section as residential). When considering types of care provided, 24 (82.8%) provided some form of nursing care and 11 provided residential care (37.95). All CHs that provided nursing care also provided residential care, but not all CHs that provided residential care also provided nursing care (Table 4). Therefore, by including a dichotomous variable for the provision of residential care, we are comparing CHs that provide residential or residential and nursing care with those that solely provide nursing care. Due to small cell counts, it was not possible to include all care type variables in the analyses.

Training and skills of people facilitating Needs Rounds

Training was provided on Needs Round for all CH staff and hospice staff who would be delivering the intervention located in six geographically disparate areas.

The fourth videoconference-based workshop session in Phase 1 encompassed several facilitated examples of a Needs Round case discussion with some participants taking the role of SPC clinician and others in the role of CH staff. Other clinicians adopted the role of observer, with a task of reflecting on key learning regarding how to participate in Needs Rounds and actions they would need to undertake to be ready at their site for implementation. Observers therefore had the opportunity to act as a reflecting team to identify strengths and areas for development.

The Needs Round simulation exercise made use of the videoconference platform’s ‘chat’ function, allowing the verbal exchanges in the Needs Round to be accompanied by synchronous written commentary on the techniques used. Supplying this written commentary within the chat facility allowed the trainers to demonstrate links between theory and practice without the need to stop the simulation.

The platform’s breakout room structure was then used to form small groups to discuss the learning and run further bespoke simulations to allow participants to change roles and engage as both observers and active members. These smaller groups enabled deepening relationships between participants who would be delivering the intervention in each of the case study sites.

Action learning sets

To aid the SPC facilitators and gain coherence around the process of facilitation, action learning sets126 were adopted for monthly discussions with the SPC nurses delivering the Needs Rounds. The action learning sets offered a regular time for online video meetings of peer-to-peer learning, to share difficulties and solutions, in a collaborative relationship, facilitated by a member of the research team. Action learning sets provided opportunity to reflect on issues prospectively and develop learning and insights into the implementation across Scotland and England and enabled SPC staff to:

• focus on their role within the Needs Round project

• provide an opportunity to discuss emergent issues with each other and engage in problem-solving

• give space for each SPC clinician to describe their situation and learning

• provide a context which values being open, reflexive, practical and shared

• not be overly structured, to allow for maximum scope of issues and topics to be within the remit of the action learning set

• be non-hierarchical

• be attentive to individuals’ contexts and hence complementary to the importance of context in implementation science.

Reflections within these monthly meetings fed into development of the programme theory. Due to the geographic spread of the sites, these action learning sets were always anticipated to run online. However, due to pandemic restrictions, the SPC clinicians did not get the opportunity to augment their online relationships with face-to-face meetings at workshops.

Theory 1: confidence and competence

This theory is provided in Box 2.

The skills and knowledge of people involved in implementation is a key micro (inner) contextual factor and a focus of the recipients’ category within i-PARIHS. 128 Participants identified a need for better understanding of death and dying in CHs without registered nursing staff and for staff members who lacked experience of caring for people who are dying:

She [SPC clinician] gave the staff more confidence really. Because some of them haven’t dealt with death and then when they did, they thought people were choking when they weren’t choking, and to realise how it happened and stuff that’s just give them confidence.

CH manager, ST00329, Phase 2

Confidence was gained through the Needs Rounds, which helped to provide a better understanding of how various individual components must be attended to and integrated for good end-of-life care to happen:

They know that some person, some residents are dying, but they didn’t have this bigger picture, what it takes to deliver the best care at the time, they didn’t know what the nurses need to do, what the GP needs to do, how you need to approach the family, how you need to approach the residents. And it was all discussed, everyone has a like, you know, there was an open atmosphere that everyone could ask questions, there was no, you know, stupid questions, and they saw a huge value in this.

CH deputy clinical manager, ST00348, Phase 2

Although there were gaps in knowledge and practice, many CH staff believed they were already delivering high-quality care, and this view was echoed by SPC clinicians. Many had previous experience of innovation and change, including GSF, ECHO and regular contact with primary care, which meant that good practice was often in place on ACP and identifying if residents needed a palliative approach: ‘We already followed the Gold Standards Framework. So, that was already embedded in the culture in the home, and the surgeries, and all the other professionals’ (CH manager, ST00395, Phase 2). This highlights how there is a good degree of fit between Needs Rounds and existing practice, a key facet of the innovation component within i-PARIHS.

SPC clinicians provided positive feedback to CH staff when they were doing well and having the validation of an expert helped to build confidence:

Because they knew I had many years of hospice experience and stuff like that, so you could genuinely say, yeah, well, you’ve got a handle on that, well done, that kind of thing. So, I would hope that that was a good kind of sort of confidence boost.

SPC, S25614, Phase 2

Needs Rounds helped CH staff feel supported. Instead of making decisions in isolation, CH staff valued having an extra person to discuss residents with, especially given that staff in CHs lacked face-to-face support from outside health services for extended periods of time during COVID-19. 129 This was reinforced by one CH manager:

I think it gave the nurses a bit of confidence that someone is there … the way we felt during the past years is like you’re on your own, you’re locked in here, you’re on your own … And you sort of feel like you’re left on you own so you have to make very important decisions in regard to someone’s life, which you’re not a specialist in, but having [hospice] involvement has, you know, you’ve got that peace of mind, you’re not left on your own.

CH manager, ST00348, Phase 2

Participants who attended Needs Rounds highlighted the positive impact that discussions had on their knowledge and confidence in supporting CH residents, specifically on disease trajectories and the dying process, symptom management and communicating with families and other professionals:

So, certain factors that normally may not get picked up. It kind of comes up [in the Needs Rounds] in terms of the families, the contact points, whether we covered certain emotional as well as certain sort of care documents that we need to look at. So, these kind of things really help going into detail of someone’s pain management and also, in terms of their hopes and concerns which are sometimes difficult things to discuss with family but with the discussions we have as a group discussion, it helps our senior nurses and end-of-life champions to go back and then answer the questions asked and then at the end, it overall helps us and the family.

CH staff, ST00403, Phase 2

However, there were concerns raised about the extent to which information was disseminated across wider teams. Although some staff in CHs described how information was shared during handovers and meetings, for others, this was limited to those participating in the Needs Rounds discussions:

What I doubt would happen as that that disseminates beyond the four walls of that room. So, in terms of, you know, oh well I learned this today about how to approach this situation, let’s share this with staff because this is not something—these are not conversations that we routinely have or think about having. So yeah, that would be my guess that they’ve benefitted but how wide that goes.

SPC, S25614, Phase 2

Some SPC clinicians offered separate education sessions outside of Needs Rounds to meet CH staff needs, and while this was not part of the protocol, Needs Rounds acted as a catalyst for this. These sessions focused on the dying process, verification of death and ACP.

Some SPC clinicians provided direct clinical support to CHs, and this helped to build confidence. While this was part of the original intended model of care, adherence to this varied across the sites. There were multiple reasons for this: (1) the ability to provide direct clinical input was restricted by COVID-19, (2) some CH staff were unaware that this was an option and (3) some SPC clinicians stated that they had offered to do direct clinical work but that this had been declined. Yet when it did happen, clinical work was perceived as valuable in developing knowledge and understanding. This included support with pain management, as one SPC clinician explained:

I was able to assess her swallow and give her a little drink and realised that there’s no way she could take medication and she definitely needed to be on a syringe pump and then talk that through with the staff, and that was I think really helpful for them and just understanding syringe pumps and understand the situation.

SPC, S55997, Phase 2

Integrated Promoting Action on Research Implementation in Health Services emphasises the importance of facilitation, and the approach adopted by the SPC clinicians helped to build confidence. Adopting a collaborative, non-judgemental partnership approach, grounded in trust, helped CH staff to feel at ease and confident to participate in the Needs Rounds sessions.

Theory 2: place of care

This theory is provided in Box 3.

The context of residents’ place of death has shifted significantly due to COVID-19. Participants talked about how hospital transfers had reduced because of the pandemic, highlighting the overwhelming pressure on hospitals, residents and families not wanting to risk infection and CH staff only transferring residents in an emergency. COVID-19 resulted in ACP becoming embedded in CH practice:

I think COVID pushed them into having to know exactly what would a person’s wishes be because they were having so many deaths … it’s just almost standard that they’ve got all these things in place.

SPC S25614, Phase 2

After COVID, it did change because we did have a period where everybody was in isolation, the ambulances were not coming out, so we had to have those conversation with the families about the DNAR, about the preferred place to care, for care. So, we are pretty good at discussing the preferred place to.

CH manager, ST00322, Phase 1

By the start of implementation (July 2021), proactive discussion and documentation of resident and family wishes as part of admissions and review processes were already happening in many of the CHs. Alongside COVID-19, this was shaped by previous experience of innovation and change, including use of the GSF and regular GP rounds, and using tools such as ReSPECT (ACP documentation) and National Early Warning Score (NEWS2, a standardised assessment and response to acute illness tool to document wishes and assess residents’ needs): ‘We normally, from that admission, we’ll plan whether they need a resuscitation, or they want to go to hospital, their preferred place of last days, so things like that, we will plan anyway’ (CH manager/registered nurse, ST00335, Phase 2).

This again reinforces a good degree of fit between Needs Rounds processes and existing practices and procedures within CHs, which maps onto the innovation component of i-PARIHS.

Consequently, many CH staff already knew the wishes of their residents, and for most, their preferred place of death was the CH: ‘The majority of the time unless it’s an acute occurrence then they, their families, tend to want them to stay with us anyway, particularly around the end of life’ (CH manager, ST00435, Phase 2).

Despite the macro context altering practice on ACP, there were examples within the data of Needs Rounds instigating a proactive approach and contributing to residents dying in their preferred place. SPC clinicians discussed the nuances surrounding ACP, including what residents understood by preferences for active treatment, reasons for and against acute care and what outcomes residents and families would be hoping to achieve through hospitalisations. One SPC clinician reflected on the role of Needs Rounds in prompting these discussions:

I think there’s far more clarity about what you have to go to hospital for, and I can probably think of a number of discussions that we’ve had about residents who’ve not been sure about whether to go or not in one month and then they had the discussions and they’ve made the choices. And by the time we review them at the next Needs Round, they’ve made their choice about not going. So, I don’t know statistically but I would, I get the sense that yeah, there will have been avoided hospitalisations.

SPC, 8CL93, Phase 2

There was a perceived need amongst some participants that CHs without onsite nurses required more support to make decisions about hospitalisations. One GP, for example, noted: ‘They’ve got a lower threshold for contacting me. Whereas the one with the nursing, that’s just not necessary. They can handle a lot more’ (GP, 8JD12, Phase 1).

Key contextual factors influencing this, highlighted within the interviews, included there being fewer deaths in CHs without nurses, fear around making the wrong decisions, staff not having the same level of training due to an absence of in-house nurses and poor pay and conditions. These factors interact to shape confidence and competence around the referral of residents to hospital teams. Participants from a CH without onsite nursing reflected on the benefits of having support from a specialist with a nursing background to help improve their knowledge:

Yeah. It’s good speaking to [hospice nurse] and getting a different approach to things really because obviously she’s got a nursing background that helps us as well really into areas that we probably don’t know a lot about.

Deputy CH manager, ST00327, Phase 2

One SPC clinician talked about the lack of support from GPs and the progress they had made through case-based education:

I do think the residential homes need more support from a GP. Whereas now at the minute, a GP will say to send them into hospital. So, I think it’s quite difficult for them to make that call whether to send them or not. But I’ve seen a reduction in that home where they’re not blue lightening them, willy-nilly if you like because they was … I still think it must be scary for them to make that call.

SPC, 8C928, Phase 2

Low confidence was discussed by a CH participant without onsite nursing who said that their practice around hospital transfers had not changed because they did not want to take responsibility for decision-making about hospitalisation. Power, responsibility and clinical hierarchies are visible here (discussed further in Theory 3: collaboration and interagency working) as well as evident gaps in learning, with a need to build confidence on decision-making about hospital transfers:

We’re calling the ambulance because, obviously, it’s not down to us, obviously. So, we call the ambulance and they’ll ring us back … But that’ll go back on them then because they have been called. It’s not our responsibility, then, if anything happens.

CH manager, ST00329, Phase 2

The fear of being viewed as negligent by families was also perceived as impacting hospitalisations: ‘I think, it’s around that fear of the fact that if they didn’t do something [sending to hospital], they might be seen as being negligent’ (SPC 8JD12, Phase 1).

One of the key areas where Needs Rounds were able to add value was in relation to communication. Providing education on communicating with families was integral to the discussions, including discussion of hospitalisation. One SPC clinician reflected on how they incorporated this into the Needs Rounds:

I just gently tried to challenge that a little bit around what is going to be achieved by hospitalisation in a 95-year-old with Alzheimer’s. And sometimes rather than phrasing the question ‘do you want hospitalisation?’ it’s talking about is it going to change outcomes? Is it going to change what happens? Is it going to increase distress? So just started to really push that a little bit in terms of trying to give them confidence.

SPC, 8JD12, Phase 2

Although CH staff were having conversations about hospitalisation prior to implementation, the learning provided through Needs Rounds helped to improve the depth, content and approach towards the discussion so that families and residents could think through the implications of their choices. This helped build CH staff competence to provide high-quality proactive anticipatory care, as well as their confidence in communicating this with families, so that families felt at ease for their loved ones to be cared for in the home. This was reinforced by one participant who said:

I think this has given us the confidence to give education to support the next of kin of the families; that we can provide a certain level of care here in the care home. So, again, it gave myself and the staff a bit of encouragement to say, well, you don’t necessarily need to go to the hospital. And we haven’t, the residents that we have discussed, we have managed to nurse them here and the families, I believe, have got that confidence that we can do it here.

CH manager, ST00418, Phase 2

The learning component of Needs Rounds also helped to improve communication between some CH staff and GPs. Care home staff became more assertive with GPs in communicating their assessments of resident needs, to prevent a transfer to hospital. As one SPC clinician said:

She phoned up and she said she’s really anxious about this resident, she wants someone to come see them now and the GP has said, ‘If you need someone to see them now, you need to call 999’. And she said, ‘But I didn’t need to call 999, I know I didn’t need to call 999’. She’s great. And then she said, ‘So [the GP]’s coming in 45 minutes’.

SPC, S55997, Phase 2

Theory 3: collaboration and interagency working

This theory is provided in Box 4.

Needs Rounds supported interagency working by improving communication skills and confidence. Needs Rounds strengthened relationships between SPC and CH staff, which required reciprocity and a sharing of different but complementary expertise.

Theory 4: better-quality lives and deaths

This theory is provided in Box 5.

SPC clinicians praised the high-quality care being practised pre implementation, and Needs Rounds focused on augmenting this. CH staff had strong relationships with their residents: ‘Because they [CH staff] just know the residents so well, they care so much … I know it’s cliché but they are … staff are their family’ (SPC, S25614, Phase 2).

Knowing residents well enables CH staff to identify subtle changes that may not be detected by other professionals:

Senior NHS manager, Phase 1, workshop 3: ‘[CH staff] are the specialists actually because the nuances they know can make such a difference. They might not know the actual [...] clinical terms, but the amount of detail that they know can make such a difference to somebody’s life and the noticing that goes on that we don’t actually capture’.

Needs Rounds sought to triage residents to ensure that the most appropriate people were discussed. Triggers included those who were at the end of life, conflict, deterioration and residents who were difficult to manage. While residents who were perceived as being closer to death were frequently selected, the SPC clinicians encouraged CH staff to think more broadly about anyone who was in the last 12 months of life:

We [usually] pick a resident at quite late stage, really the last month maybe of their life. But last time [SPC clinician] was so really good, she was suggesting that we might talk about somebody who is not actually imminent, he’s not coming towards the very end.

Registered nurse, ST00348, Phase 2

Care home staff were thus informally learning that SPC could be appropriate to people with a life-limiting illness even if they were not in the final weeks or days of life. The SPC clinicians highlighted the need for CH staff to have more confidence in their knowledge of residents and identifying deterioration and dying:

I think the issue … around their confidence around perhaps that I think they … I said, ‘Do you think you recognise dying? Do the people here recognise dying?’ They said that, in the main, they feel that they do. But I think it’s perhaps having the confidence then to go with what their gut instinct is.

SPC, 8JD12, Phase 2

Drawing on ‘gut instinct’, or their knowledge of residents and their assessment skills, alongside additional support, case-based learning about signs of deterioration and being able to communicate this to other professionals, is integral to high-quality care provision.

One of the key areas that Needs Rounds aided was symptom management. One CH interviewee reflected that their approach to assessing, documenting and treating pain had changed because of Needs Rounds: ‘We are getting better at utilising [PRN medication], documenting it, putting stress and distress care plans in, and having pain management care plans now of what we’re doing to try and manage pain’ (CH manager, ST00419, Phase 2).

Staff were more proactive in their pain management, but also able to expand their approach, considering the last 12 months of life rather than the last few days. Having these plans in place also enabled CH staff to be more assertive about the needs of residents who experienced fluctuating levels of pain that was undetected by external health professionals:

The CPNs are very much looking for hard evidence rather than anecdotal evidence … I understand that, you know. I don’t want every resident within the care home sedated, you know. That’s not what we’re looking for. But we are wanting every resident in the care home to have the best quality of life for their last six months, a year, whatever, you know, or even weeks that they’ve got.

CH manager, ST00419, Phase 2

As identified in theory 3, relationship building between CH staff and SPC clinicians is a vital mechanism in improving palliative and end-of-life care. When SPC clinicians recognise and value the knowledge that CH staff have of residents, listen to their concerns and collectively create action plans; this helps to improve resident care, making sure that residents’ wishes are met and that their pain is managed so they have a comfortable peaceful death:

Other professionals aren’t always keen on the uptake to do anything different or hugely different, but since doing the Needs Rounds, we have been able to titrate her medication up and it’s been of huge benefit to this particular resident in managing her pain … I think that’s because this lady has been listened to and the staff have been listened to.

CH nursing manager, ST00418, Phase 2

As mentioned in theories 1, 2 and 3, Needs Rounds also helped to develop CH staff confidence in communicating with primary care clinicians, with requests being met and more timely actions being carried out. This timeliness was identified as being particularly important when a resident is at end of life.

Direct clinical work also assisted CH staff to recognise dying earlier in the process and ensure that anticipatory medications were in place so that resident pain was managed. As mentioned in theory 1, clinical work did not always take place because of COVID-19 restrictions and managers not wishing for this to happen; however, when it did happen, it worked well:

[The SPC clinician] really assisted, and that death was a lot easier than what it could’ve been, if that hadn’t got the [medication] patch sorted and things like that and [the resident] was probably a lot more comfortable having that on.

CH nurse manager, ST00435, Phase 2

Having time to reflect and proactively plan for death enabled CH staff to ensure that care was resident and family focused; symptoms were also well-managed, and residents and families were prepared:

[Needs Rounds] contributed because … everything was in place … and just to stop a minute and think of the resident’s experience … He did not suffer, the family was aware. He was ready. He got a little bit more religious towards the end, and it was just so peaceful.

Deputy CH manager, ST00327, Phase 2

The above example highlights the importance of spiritual care. While medical components such as pain and symptom management are integral to a peaceful death, so too are meeting residents’ psychosocial and spiritual needs. During the action learning sets, SPC clinicians reflected that there was often a lack of attention to residents’ non-medical requirements, so these gaps were focused on in the subsequent Needs Rounds. SPC clinicians, for example, asked questions about religion during the discussions: ‘I also asked the [hospice] chaplain to go in and see her … So I used [the hospice]’s resources to go into that nursing home and support that family and the patient and the staff as well’ (SPC, 8JD12, Phase 2).

Other services such as complementary therapy were also used to improve the quality of residents’ lives.

She [complementary therapist] came out with me to quite a few of the care homes on one of the Needs Rounds. I was able to talk to the activities coordinators, and was teaching them just touch massage. We use non-verbal communication and things like that … It was really worthwhile.

SPC, S55997, Phase 2

Thus, although the CH was often already providing excellent care, Needs Rounds identified gaps and helped CH staff to build on their existing practice to ensure that a holistic range of needs were met to improve quality of life and death for residents.

Theory 5: supporting families

This theory is provided in Box 6.

With limited primary data from family members, the theory on families is derived from what CH staff and SPC clinicians told us about what worked or did not work with families, rather than being based on their direct feedback.

Some CH teams were confident in communicating with and supporting families prior to Needs Rounds being delivered. Such sites already had good processes in place for this, for example ascertaining resident and family wishes at point of admission. During Phase 1, a family member reflected on the CH support they had received pre implementation:

I’ve tried to have my moments at home but there was a few times where I was distressed, and there was a couple of times where the senior nurse in the, in the building said, ‘come in and have a chat, we’re worried about you’.

Family member, S55997, Phase 1

During implementation, where CH staff were effectively supporting families; praise, support and positive reassurance were used to enhance confidence (see Theories 1 and 3).

In contrast, some CH staff felt nervous, found it difficult to talk about death and dying with relatives and did not always have the time to support them. As demonstrated in theory 1, Needs Rounds helped to develop CH staff confidence, including the confidence to talk to families about death and dying and how palliative and end-of-life care is managed, including care planning, symptom control and anticipatory medication. When CH staff recognise that residents are dying, they are better able to plan and care for residents, ensuring that the family are fully informed and involved in decisions about care, anticipate what care and treatment will be required to keep residents comfortable and work with other health professionals to ensure these are in place. Needs Rounds prompt action earlier in the process:

[Before Needs Rounds] we would wait for something to go wrong rather than being proactive and say to the family that this person is possibly in the last days of their life and we need to consider plans for them ahead. So, I think that’s the major change, so we are able to recognise that [dying process] in a much better way and able to communicate that with the GP and the family and make plans for them well ahead of time, yeah.

CH Registered nurse, ST00335, Phase 2

Being proactive and recognising deterioration, combined with timely communication and support, helps to prepare families so that they can process and plan for their relative dying:

What has helped us, is actually doing anticipatory care and preparing families for that. And, you know, and then being able to accommodate families for them to be better prepared for those last days and having that time to come in, and, you know, make peace with their families and that kind of thing.

CH manager, ST00419, Phase 2

During the Needs Rounds, the SPC clinicians assisted CH staff by talking through strategies for how best to communicate, what terminology to use and when to approach families. Sometimes, the SPC clinicians spoke directly with families, but the CH staff were also encouraged to do this themselves and use the Needs Rounds sessions for reflexive practice:

[The CH has] almost practiced the conversation before they’ve had it with them [the family] … they’ve kind of taken that, ‘give it a go and we’ll talk about it at the Round and see you know whether there’s anything that was really hard to do and we’ll kind of unpick that a bit and reflect on it’.

SPC, 8CL93, Phase 2

Having more confident and competent staff improves the quality of residents’ lives and deaths (see Theory 4), and this can build family confidence in the care being provided. CH staff communicated with families to aid understanding of what care can be offered as an alternative to hospitalisation. Having an extra layer of specialist support from the hospice was viewed positively:

They chose here, and we were able to provide this man a dignified death and he was able to, be able to have the syringe driver, his pain was managed well, we didn’t need to transfer him to hospital … so we were able to give them [family] the confidence … [Needs Rounds] had given myself and the staff the confidence to put our abilities out there.

Deputy CH manager, ST00418, Phase 2

Strong communication skills can also be used to resolve conflict. On the Needs Rounds checklist, family conflict is a trigger to discuss residents. Often conflict or disagreements with or within families relates to different perspectives on care and treatment decisions: ‘You can appreciate when residents’ conditions are declining and really, it’s not in their best interest for them continuing to go back and forth to hospital. We can explain that. We can’t really enforce it’ (Clinical lead, ST00333, Phase 1).

Families can find it difficult to accept that their relative is dying and seek burdensome and futile treatments which can cause tensions with CH staff. Care home staff talked about forming close relationships with residents and wanting them to have a peaceful death in the home surrounded by people who care about them. CH teams developed confidence through Needs Rounds to have conversations with families about the potential risks of hospitalisation:

The family, one son in particular was in denial … he talked about having [a treatment] giving her hope, and I’d said, you know, I can probably understand that and who wouldn’t want hope. I said but what they did explain to you was … it’s a terminal condition she’s got and we have to think of what is best for her at this moment in time and for her comfort and for her pain levels.

CH manager, ST00340, Phase 2

In the above example, the CH staff communicated to the family that their mother might not survive if transferred to hospital. Despite being advised of the potential risks, the family decided to proceed with a hospital admission, where their mother died. Although the CH staff’s communications did not prevent hospitalisation, they delivered the required information to allow the family to make an informed decision.

Tensions can also arise due to CH staff not listening to family concerns. In one site, a daughter had recognised changes in her mother and believed that she was dying. Although the CH staff did not share this view, the SPC clinician was able to identify signs of dying and arrange necessary end-of-life medications. Through direct clinical input (assessment and medication management), the resident’s symptoms were managed so that they were comfortable in their final days. Support was also provided to the daughter who had been distressed by witnessing her mother in such anguish:

I’ve never seen somebody so distressed with dementia, and she was, nobody wanted to diagnose that she was dying in this home … I’ve never seen that level of distress and her poor daughter was beside herself. And within … we got her really settled with just five milligrams of midazolam. So we just put a tiny amount in and supporting that daughter who’d been through a very traumatic experience, I think, it was very traumatic seeing mum like that for days on end.

SPC, 8A784, Phase 2

The SPC clinicians also supported CH staff to resolve conflict within families. In one site, a son was strongly advocating for his mother to have active treatment in hospital, despite his mother having capacity and not wanting to go. The SPC clinician offered to attend a family meeting, and although the CH felt confident enough to manage the communications themselves, they took advice about how best to approach this:

[The CH] said, ‘it went really, really well, I spoke to them [the family], used what you said you’d do, and it went well, and the son was fine’. So, yeah, I just think they’d rather ask me but then do it themselves.

SPC, 8C928, Phase 2

An approach based on doing alongside, with advice and support from the SPC clinician about how to communicate the potential risks and distress from hospitalisation, and not doing directly for, is important in enabling the CH team to build their confidence and have the skills to continue supporting families’ post implementation. This reflects the centrality of honing the facilitation approach, as advocated by i-PARIHS.

Time

Time was a key connector within the theories and is also central to the ‘recipients’ characteristic in i-PARIHS. Needs Rounds are intended to provide protected time to discuss residents’ holistic needs, and while they did sometimes, they were not always prioritised or were cut short. Available time was impacted by wider macro workforce issues, including high turnover and unfilled vacancies; COVID-19 exacerbated these problems. If CH managers are unable to allow their staff time to attend Needs Rounds, then they do not benefit from the improved outcomes that the approach can engender.

Having protected time to discuss residents resulted in more proactive care, with Needs Rounds prompting proactive management of resident needs to manage symptoms more effectively. Staff having time to continue learning about recognising deterioration and the dying process, and how to communicate with families, improves the quality of residents’ lives and deaths and prepares family members for their relative dying.

Time also shaped which residents were discussed, with triggers predominantly focusing on those closer to dying. Specialist palliative care clinicians did encourage CH staff to discuss residents at an earlier stage, to improve their quality of life.

Effectively planning Needs Rounds around CH routines so that they occurred at a quieter time of the day was helpful in enabling CH staff to attend the sessions. However, a lack of time was detrimental to implementation in several ways. First, it restricted who could attend the sessions, which impacted the overall scale of change. Carers who were identified in Phase 1 as being integral to the success of Needs Rounds were often excluded because there was no time or availability to participate. Second, time acted as a barrier to management and wider staff members engaging with materials about what Needs Rounds are, which disrupted the early implementation period. Third, time prevented MDT meetings from happening as aligning schedules between busy professionals was too difficult.

Relationships

Pervasive public discourses regarding CH staff being less skilled than other healthcare professionals were recognised and challenged during the Phase 1 interviews and workshops. Valuing the important work of CH staff, using their skills and knowledge of their residents, and adopting a facilitation approach that fostered these assets was integral to the programme theory. It was anticipated that a collaborative, partnership approach was needed where SPC and CHs learn together, to enable CH staff to provide high-quality palliative and end-of-life care beyond the duration of the project. It was evident that this had been achieved in many of the sites, with a non-judgemental partnership approach underpinned by trust.

Demonstrable shifts in power were evident in Phase 2 interviews, with clear acknowledgement of each party holding expertise. Care home staff were celebrated for providing excellent care, for knowing their residents, anticipating their needs and undertaking proactive care planning to facilitate good deaths. Indeed, rather than being viewed as inferior, CH staff were conceived as experts in the provision of palliative and end-of-life care and respected for working in an extremely challenging context.

Specialist palliative care clinicians were also praised for their skills and knowledge, being supportive and being present. However, for some, respect had to be earned rather than given by virtue of their role as senior nurses from specialist services. This occurred against a backdrop of COVID-19, where many CHs were left with little face-to-face external support, resulting in feelings of abandonment. One way for SPC staff to build trust was to demonstrate skills and knowledge by combining case-based learning with direct clinical work to enhance the value that they offered. This was part of the original implementation plan, though did not happen consistently in the sites. This was due to a combination of COVID-19 visiting restrictions, a lack of a perceived need and direct clinical intervention not being offered by the sites.

Some SPC clinicians had not readily identified their added value and expertise due to the existing high-quality work being conducted by the CH teams’ pre implementation. However, this does not accord with the feedback from CH staff who valued the support they had received, particularly in building staff confidence and improving communications with families and GPs. While the initial programme theory focused on developing the confidence of the CH staff, the final programme theory has been broadened out to include a feedback loop so that SPC clinicians also know what value they contribute. This feedback loop creates a more equal relationship where different expertise is acknowledged and collectively harnessed to improve resident and family care and support.

Despite there being a shift in power between the CHs and SPC, Needs Rounds did not address power dynamics in the wider health and social care context. Externally, it was evident that SPC clinicians were still conceived as experts and CH staff used this status to influence their interactions with GPs so that their requests were met with responsiveness. While this helped to improve communication, it was often by referring to the SPC clinicians’ recommendations that actions were carried out swiftly.

During the Phase 1 interviews and workshops it was made clear that for Needs Rounds to be effective, a range of staff members should attend the monthly meetings, including care staff. Care staff spend a significant amount of time with residents, resulting in the development of knowledge about their needs, which enables the detection of changes that may signal deterioration and death. However, the wider macro context impacted care staff attendance, with workforce issues restricting the ability of carers to participate, which often resulted in knowledge generated within the Needs Round being held by nursing and management staff. There is a lack of consensus in the data about the extent to which learning from the Needs Rounds was disseminated to staff not attending; therefore, the programme theory reflects that the benefits arise for those who participate in the Needs Rounds sessions.

Communication

Communication is a key theme that connects the five theories. The Phase 2 interview data demonstrate that improved communication was one of the main outcomes resulting from Needs Rounds. Needs Rounds provided a direct communication mechanism between CHs and SPC, in some cases circumventing the need for referral through the GP.

Needs Rounds also helped CH teams to improve their confidence and assertiveness when communicating resident needs to GPs. This augmented existing good practice on documenting, communicating, directly providing and working with GPs to ensure medications are in place to enable residents to die in the CH. Enhanced communication contributes to better-quality dying, with residents’ wishes regarding place of death being fulfilled.

CH staff also improved their communication with families. While for many of the CHs, there was already excellent practice around ACP, the SPC clinicians supported CH staff to improve both the content and process of communication to prepare families for their loved one dying and to resolve conflict between and within CHs and families. This included discussing staff skills to bolster family confidence in their ability to provide high-quality palliative and end-of-life care.

Despite this, three communication gaps were also evident. First, the full understanding what Needs Rounds are and how they function was missing for some participants. While information was provided through meetings with the research team, e-mail correspondence, an implementation pack, a memorandum of understanding and workshops, it was evident that these were often not fully engaged with. Furthermore, understanding and knowledge were also lost through staff changes. This resulted in gaps in understanding amongst CH staff members who lacked clarity about Needs Rounds.

Second, there was a lack of awareness amongst SPC clinicians about their value and how they were positively perceived by the CH staff. The action learning sets with the six SPC clinicians were an important peer support mechanism which enabled discussions and reflections on their learning and problem-solving, including this absence of a feedback loop. However, an additional feedback mechanism is required to aid communication between CHs and SPC clinicians.

Finally, the multidisciplinary team and family meetings did not happen in practice, largely due to COVID-19. These were intended to act as formal mechanisms to improve communication between CHs, external agencies and families.

Aspects of care that family members were less satisfied with

Family members were more likely to be satisfied that their relatives’ physical symptoms, such as pain, were being adequately assessed and controlled in comparison to their emotional problems, such as anxiety. A larger number (23.1%) indicated they were not sure about their satisfaction regarding their relatives’ emotional needs being met, and a small percentage (7.7%) indicated they were ‘not very satisfied’ (Figure 6).

FIGURE 6.

Bar chart: family member’s satisfaction that their relatives’ physical symptoms and emotional problems are being adequately assessed and controlled.

The lowest levels of satisfaction were observed for the questions regarding doctors (Figure 7), though the questionnaire does not elicit feedback on which doctors this relates to (e.g. primary or secondary care). While most family members indicated some level of satisfaction, there were higher percentages of people indicating they were ‘not at all satisfied’ or ‘not very satisfied’ than for other questions in the survey.

FIGURE 7.

Bar chart: family members’ satisfaction with doctors taking an interest in their relatives, being available when needed and having trust and confidence in those doctors.

Improving family members’ perception of care

Communication appears to be a key area to improve family members’ perception of their relatives’ care. A minority of family members indicated they were ‘not very satisfied’ with the consistency with which they received information (15.4%) or that staff were explaining things to them in an honest or straightforward manner (7.7%). Family members want to understand their relatives’ condition and receive regular updates. Finally, only 46.2% of family members were ‘completely satisfied’ that their relatives’ care environment was calm and restful, indicating that most family members saw room for improvement in this area.

Overall, findings suggest that family members are mostly satisfied with aspects of their relatives’ care. They considered relationships with doctors to be the least satisfactory, and areas for improvement were identified for communication, supporting residents’ emotional well-being and providing a calm care environment.

Priorities and floor time

Interviewees highlighted competing priorities as a significant factor impacting upon their availability to participate in Needs Rounds, often describing the challenge of finding time to complete the research tasks associated with the Needs Rounds study. For implementation, however, the primary concern was how Needs Rounds competed with clinically urgent priorities on the ‘floor’.

I feel the nurse should be in [the Needs Round] because she knows the resident well … So, I need a nurse to be present on that day and that’s very important. And sometimes […] there is GP rounds or there could be three admissions that day, and it’s just one nurse on one floor. So, it’s very difficult for the nurse to pick out that time […] she really needs at least an hour of time.

CH Nurse Manager, CHNM_ST00456, mid-implementation

The choices made to prioritise tasks, often those related to immediate benefit to patient care, meant that Needs Rounds were positioned as a lower clinical priority. The intention to use Needs Rounds to anticipate care needs and reduce unplanned admissions takes a backseat to urgent care. While no cost is attached to this element, there are opportunity costs in how staff use their time.

Organising efficient use of staff time to ensure optimal staff engagement with the monthly Needs Rounds was seen as possible, even if not always managed effectively:

I think these things are all doable with … it’s … they take, you know, it takes a chunk out of your time […] they’ve got that organised. So, it’s part of that nursing day to factor that in. So, yeah, that would put a pressure on resources if it was to be done … if they could do it in that proper way. But I think it really just depends on, you know, how proactive they are about that.

SPC clinician SPCS25614, mid implementation

Another interviewee described feeling that they were able to provide additional resources and opportunities by planning ahead and organising their caseload/workload in advance: ‘I think that’s what’s allowed me to get up these extra things like with the complimentary therapist and with health and wellbeing and with working with [Training Staff] to look at education, and things like that’ (SPC clinician, SPC8JD12, early implementation).

There are also costs associated with SPC clinician time. For this study, SPC staff worked 2 days per week, supporting the research tasks as well as the monthly clinical meetings. SPC clinicians reported that it would be feasible to complete 10 Needs Rounds per week if working on a full-time basis, though one noted that up to 20 meetings per week would be possible. Thus, the SPC staff costs could be distributed across a greater number of CHs.

Opportunity costs, particularly with last-minute cancellations of Needs Rounds, were noted. Such changes resulted in wasted time, notably if a Needs Round was cancelled after the SPC clinician had already travelled to the CH. In addition to the time/travel costs (described further below), there is a subsequent implicit cost in CH staff and residents not benefitting from the Needs Rounds embedded education and guidance.

Staff wages

Interviewees highlighted that a strategy to balance urgent resident care with attending Needs Rounds was staff using their own time to attend the meetings:

We are quite a busy care home. Even in the meetings, normally we’re trying to do them as quick as possible and then back on the floor. […] I used to have the meetings when I wasn’t on duty so I was coming on my day off, and that was the best thing.

CH Nurse Manager, ST00402, late implementation

Several CH staff describe financial costs incurred in paying staff to work additional hours to cover Needs Rounds, as well as the opportunity costs of staff being unable to attend to direct patient care during the Needs Rounds. Other data suggest that Needs Rounds might increase wage costs, but have longer-term positive impacts on the sector:

People are coming in on their off days. They’re having to squeeze things around last minute to make sure that care delivery isn’t affected and so you know while people can carry on for so long and they have done for the past kind of two years [during the COVID-19 pandemic], there will still be a point at which individually people won’t be able to and I think you know you’re going to start seeing people leaving jobs, changing jobs, just not being to carry on. Maybe the main aim of Needs Rounds wasn’t supporting staff’s well-being. I think where it has been able to do that hopefully that’s given some people enough support that they’ve considered that it’s worth carrying on doing the job that they are doing.

SPC Clinician, 83L93, late implementation

The use of overtime has both explicit and implicit cost implications. Paid overtime has more apparent financial costs to the organisation paying staff. However, both paid and unpaid overtime holds cost in terms of staff missing out on their free time and leisure activities.

Travel costs

Travel time was a cost incurred by SPC clinicians as they travelled to and from different CH locations. Journey times varied from as little as 5 minutes to as much as 3.5 hours each way. Most involved car journeys between 20 and 60 minutes each way, but some CHs were reached by foot or cycling. Opportunity costs were also incurred in travel time; to mitigate this, one CH located 3.5-hour drive from the hospice agreed to meet online some months to increase efficiencies.

Some SPC clinicians spoke about organising their diaries to manage combining journeys efficiently, as one said ‘If you’re organized, you can see, you can do a Round and see three or four people on the same day’ (SPC clinician 8A784, early implementation).

Travel time also has both explicit and implicit cost implications. Explicit implications are largely associated with the cost of staff wages (for organisations) and fuel costs:

I think just mileage is a huge one. Obviously for us as I put in my mileage sheet for the first month and then I can’t remember like a couple of hundred pounds or something because I had to ‘cause I was driving so far, but so that is a big one. Also, yeah, travel time.

SPC, ST00339, early implementation

Participants described a work culture of not claiming mileage costs for shorter distances. This therefore operates as displacing costs from service provider to SPC clinician:

Petrol […] I think I think the pay is good, so I wouldn’t say … ‘Actually I think we need to claim this back’. ‘cause I think the pay is good for the hours that I do. There’s a couple of times where I went to a nursing home which is quite away from me and [the manager] wasn’t there again. That was my fault. Should have double checked. I do that now I’ll ring on the day. ‘I’m just leaving. Are you still there?’ ‘Yeah, I’m fine’, but I did do it twice.

SPC, 8C928, early implementation

Medications

Forty-seven medicines across all six sites were recorded as having been deprescribed during the intervention period. One interviewee raised a financial cost to the NHS through the disposal of unused medications, specifically unused ‘Just In Case’ (JIC) medications:

The cost obviously to the health service having all these medications supplied and sometimes that person passed away and the medication must be destroyed, they are not cost directly to us but these are costs to the NHS because with COVID, we were not allowed to send any medication back to the pharmacy so we had lots of medications that had to be disposed of.

CH nurse, ST00403, mid implementation

In this case, the CH staff describe that the palliative care team made recommendations for all JIC medications to be in place, rather than seeking new prescriptions once symptoms have already escalated. While these are needed in some circumstances, in cases where they remain unused, the medication is wasted. One interviewee said: ‘The whole scale of the medication that would have been necessary should things having to turn to the worse, and if that person dies, suddenly all that would just had to be destroyed’ (CH staff, ST00403, mid implementation).

Care home staff explain that while this type of waste occurred prior to the intervention, it has increased with the pre-emptive approach of Needs Rounds. JIC medicines are considered good practice in the community, so standard practice (CH residents with complex symptoms not having JIC prescribed) is problematic, despite the cost. However, a SPC clinician described using a considered approach which bases these anticipatory prescription decisions upon assessment and individualised care:

There is a strong argument, of course, for Just in Case medications, which is why we have them. But equally that what the care home was saying is that often they’re not used and there is there is an economic cost to that that perhaps we I think is something that could be looked at. You know, Just in Case medications I think need to be arranged in a timely way. […] it is about individualised care … I think assessment is key in this situation. You’ve made a decision that you don’t want to prescribe Just in Case medications. You need to be assessing this resident for pain, agitation, secretions so that these medications can be got in a timely way. They had a good relationship with the pharmacy and I did point out coming up to a weekend and out of hours, these things are difficult. So I think it is I think it is a balance, but they were thinking about that. I suspect there probably is a cost in terms of just in case medications, but I think that that’s broader actually than just the Needs Round.

SPC, 8JD12, mid implementation

Emergency legislation introduced in the UK lockdowns to facilitate reuse of dispensed medicines was noted as an enabler of higher quality, anticipatory care in Phase 1 interviews:

[CHs] don’t really have stock but they can use the ‘medication reuse’ national policy. And ours have not needed to use that but we have seen that being used over [area] during the COVID pandemic, so if somebody dies, they consented for the medications to be used by somebody else so it’s that rather than a stock. […] that’s enabled access to timely anticipatory medication’.

GP, 8A784, Phase 1

The financial savings are implicit both in terms of the costs of medicines as well as dispensing time from pharmacies and the symptom burden of the resident.

Office costs

There were financial costs to CHs, SPCs and individuals in relation to practical and administrative elements of the study. These costs were research-related (e.g. printing and postage of questionnaires, and photocopying of posters) and hence reflect the study context rather than wider implementation. One SPC clinician (site 8A784, early implementation) said ‘I photocopy things at work […] I think if I didn’t photocopy the posters and bring them, I don’t think they’d be up on the walls’. Another said:

One cost is that some of the care homes do quite value some of the hard copies of information so you know printing costs. It’s small, but, you know, if you, so I’m working mainly from home. So printing costs. And you know if you are giving an envelope, a family questionnaire in an envelope, put a stamp on it, these are small costs, but that’s the only thing I’ve really come across.

SPC Clinician, 8JD12, mid implementation

A further SPC clinician noted her desire to purchase additional office supplies (files) to aid her admin processes. Such costs may endure with implementation.

Additional interventions and equipment

Two interviewees referred to occasions where referrals were made for complementary therapy for residents which would not ordinarily have occured. There were also explicit costs in relation to the purchase of items recommended by SPC clinicians:

The way she helps is coming up with some new ideas, things we don’t think about like for [a resident] she came up with the idea about having a red blanket in his room in case the cancer burst and there’d be a lot of blood. So like the red blanket and the family will not be too alarmed seeing a lot of blood […] So that was the thing that we put in plan and that was very helpful.

CH Nurse, ST00404, mid implementation

In this instance, the staff were able to source the blankets without purchasing externally. However, the CH staff explained that the manager had agreed to purchase any further items deemed necessary.

Summary

Qualitative data suggest the major cost of the Needs Rounds intervention was staff time. This had explicit costs of staff wages (existing working hours, overtime, and travel time) and implicit costs including opportunity costs, primarily due to the late cancellation of Needs Rounds. However, when CH staff cancel Needs Rounds due to competing priorities, they inadvertently create an implicit cost in that neither the staff nor the residents’ benefit from the intervention. Time-related issues such as prioritising floor time over engaging in Needs Rounds pose a dilemma which is not easily resolvable without a changed funding model for staffing.

While some of the costs highlighted by interviewees were spoken about within the context of specific situations, given the detail provided, it is reasonable to accept that many of the economic costs highlighted could apply to other individuals and organisations within the study. However, if Needs Rounds are implemented in practice, costs associated with the research element of the study would be avoidable, particularly the time required to fulfil the quantitative data requirements of the study.

Patient and public involvement and engagement was integral to the team

Patient and public involvement and engagement members were explicitly positioned as core to the team and described as ‘co-facilitators’ and ‘co-applicant’ (academic 9 and PPIE 3, respectively). Indeed, PPIE members were ‘involved from the word go’ and considered a ‘key part’ of the team (PPIE member 1 and academic 7, respectively). Another PPIE member stated:

We are members of the steering group that had oversaw the project, but also part of the operational group that met once a month just to double check how things were ticking on along. So it wasn’t like ‘go to PPIE when you want to do an easy read version of the leaflet’. I mean, that’s easy, isn’t it? It was more of, you know, we are there when they were talking about the funding, how much the funding was going to be, you know, what people’s roles are, what will the research look like, how often we would meet, what next, what next, what next, we’re looking at the phase of the project, what it would include, the model, the theory, the engagement part, like the workshops. All of it.

PPIE 3

Similarly, PPIE 2 stated that the study’s user involvement and engagement was ‘cutting edge’, stating that her involvement at the Needs Rounds Study had ‘exceeded [her] expectations’. The notion that this was ‘top level PPIE’ was related to the ‘complete autonomy’ that had been extended to PPIE, as well the responsibility she was allowed to assume in the project. Patient and public involvement and engagement members, for example, had control over the activities they were involved in and in shaping how the PPIE budget was spent. Further, she described the significance of being involved in an interview panel and the importance of her involvement in this given how the researcher being appointed ‘was pivotal for the project’.

Both academics and PPIE members described how the dedication of most PPIE members was consistent throughout the study. Patient and public involvement and engagement members were ‘involved from the outset’ (academic 1), which was in distinct contrast to some academic interviewees’ prior experiences of PPIE being ‘tacked on at the end’ (academic 1) and being treated as ‘tokenistic’ (academic; 1, 4, 5 and 9). Given her past experience, academic 5 describes her expectation of PPIE in the Needs Round study:

When I came into the study, because of my previous experience with PPIEs, I thought, ‘Yes, they would be there and contribute a bit.’ But I thought they would be in the background, but they’ve been very much more in the foreground of the study.

Academic 5

Patient and public involvement and engagement members were described as ‘active voices’ in discussions regarding research, funding and impact (academic 2). Patient and public involvement and engagement members' central roles in the research meant that the language used by the team shifted to facilitate their inclusion.

Language in the team – ‘bearing the audience in mind’

Patient and public involvement and engagement members framed part of their role as posing questions for academics to consider in relation to their research and dissemination practices:

The thing is, [academics] get a bit too professionalized sometimes and then you sort of forget about basic questions that might get asked and I know there was one time, you know, when you think you sort of covered everything. There’s always something that you’ve forgotten about.

PPIE 1

Patient and public involvement and engagement members described how they asked questions to clarify what academics shared in public. Patient and public involvement and engagement members were asked to act in the capacity of ‘guardian angels’ to ‘watch out for language becoming too technical’ (PPIE 2) in workshops which involved practitioners and members of the public. This was designed to make the workshops more accessible to participants, thereby encouraging greater participation.

Overall, academic interviewees considered the questions and comments from PPIE regarding language and the content of the study to be helpful and constructive. Several academics made note of how the research was communicated:

[It is] always valuable to get feedback when you’re presenting results, and particularly not just a sense check, but also, if people don’t understand it, then I’m not explaining it clearly enough. And it’s easy for me to start using jargon or make assumptions.

Academic 2

Others described adapting how they communicated findings. In this way, PPIE influenced the study’s engagement with practitioners and the public, foregrounding accessibility. Hence, the questions asked by PPIE members ‘just grounds people a bit more’ and ‘challenge[d] [academics] to be better at articulating what [they] want to achieve from the projects, in ways that make sense beyond academic circles’ (academic 6).

Academics described how the presence of the PPIE members provided them a helpful reminder to modify their communication; they were more likely to ‘bear the audience in mind’ or ‘pay […] attention to language’ when PPIE members were present (academic 6 and 9, respectively):

I’ve just enjoyed it more than meeting with sort of academics all the time. It’s made me think a lot about communicating to non-academic audiences […] I’m thinking about okay how am I going to make this easier to understand for somebody who isn’t a [area of expertise] researcher.

Academic 3

Relationship-centred research team

Patient and public involvement and engagement members appreciated the relational investment from the academics in the team. Monthly meetings, post-meeting debriefs, regular e-mail exchanges and staff availability meant that there were considerable points of contact in place. One PPIE continually referred to different moments in which a member of the team had taken time to explain something to her. This kind of support was described as both significant and continual: ‘[Team meetings] were every month and so and, you know, every step of the way it was explained what was going on and if you needed help, you know, there’s always a call afterwards’ (PPIE 1).

Patient and public involvement and engagement members held their relationship to the academics in high regard and described how ‘patient’ and ‘accommodating’ some members of the team were with them to foster a sense of partnership:

She’s very good teaching, training skills. […] Even if it was not of any real like, you know, oh isn’t the answer obvious, kind of thing, never showed that. I didn’t even consider that. I don’t know why I didn’t think of that, let’s just hold on to that, she’d says oh really, just ‘really good’, it really just always made me think. ‘Oh, right, okay, I said something that was … that took me ages to say and I said it’ and she’s, you know, she’s heard it. […] she’s really good.

PPIE 3

Other PPIE members also provided examples of times when they had felt supported. This was particularly important given the emotive topic of death and dying:

I knew I had support there that if I got upset by anything that was said, I think it came out [in the study data] that one young carer said that she had a fear around meeting death for the first time with a resident. And I found that very moving, you know, because it was how I felt when I was meeting death for the first time with my grandmother. So, I think I had lots of empathy.

PPIE 2

It is noteworthy that this was the only example of a PPIE member reporting an emotional response regarding the study’s focus.

Academics made a concerted effort to build and maintain relationships:

The team at Stirling responds really quickly to emails from them, and I think that’s again a kind of support thing, [….] So to be really responsive and quick, we don’t let emails from [PPIE members] sit in our inboxes, whereas we might with other emails. And I think again that’s relationship building and respect, and just making it clear that they’re valued.

Academic 6

The PPIE debrief sessions held after monthly team meetings were also mentioned as a core supportive mechanism. Academic interviewees described how the debrief sessions provided a space where PPIE members could ‘express their opinion’ or ‘unpick’ anything they did not understand (academic 7 and 6, respectively). PPIE 3 described this in the interviews:

We developed these hourly meetings or half-hour meetings after the main meeting, just to decipher in plain English that we knew what was going on. And that was a really good idea. So, I think we’re always well looked after.

PPIE 3

The debrief sessions were credited with strengthening the relationships in the team, making them feel valued and cared for. Academic 1 said ‘It’s also just having a chat, getting to know each other, and I think that’s where the kind of informal building up relationships happens’, while a PPIE member noted:

It’s just nice to have a chat if somebody was there and sort of find out a little bit about them as well, because I think it’s the human face, isn’t it? And, you know, sort of being able to relate to somebody.

PPIE 1

The ‘informal’ tone of the monthly research meeting was also important in building reciprocal relationships and enabled the team to get to know each other. All participants reflected on the ‘good ethos’ and ‘really good dynamic’ within the team (academic 9 and 1, respectively). Patient and public involvement and engagement members described feeling at ease within the team:

It’s quite a nice group, quite a relaxed group of people. And there always was there’s all humour in the group, […] There’s always a bit of a joke. And that does help, so the chairing – it’s quite a relaxed atmosphere. It was good. You’re encouraged to speak.

PPIE 3

Likewise, several academics felt that the team provided a ‘safe space’, engendering the ‘feeling [of being] able to talk’ (academic 3 and 1, respectively). This dynamic furthermore appeared to be intertwined with the lack of ‘a hard and fast distinction between academics and PPIE members’ (academic 6), as well as the notion that the PPIE members and academics involved in the study were on equal footing: ‘My feeling is that in this project, having people treated as being more equal members and making less of a distinction, was probably more, yeah, contributed to its success’ (academic 2).

Porous positions

The lack of a distinction was mentioned by five of the eight academics interviewed for this evaluation. Academics, like PPIE members, could also draw on their own lived experiences of CHs and palliative care:

There’s kind of weird insider–outsider sort of business, and that’s what I’ve mentioned before about these kind of blurred boundaries with–within this project. Like, you know, [academic colleague]’s parents who are in the care home just now and she supports them and that’s lived experience. So having these kinds of boxes of what is your … You have lived experience; you have academic experience. […] It’s blurred. It’s not neatly cut.

Academic 1

The quote illustrates how both academics and PPIE members can straddle multiple positions and speak from various standpoints. With team members occupying multiple identities and knowledges, there is a porous membrane between having academic knowledge and lived experience.

There was an acknowledgement that PPIE members’ perspectives, and their corresponding role, could change over the course of the project. Academic 6 emphasised that ‘personal experiences might shift as time goes on’. A PPIE member reflected this idea regarding how her input changed when someone close to her, who lived in a CH, died: ‘I was giving [the PI] quite a lot of detail in that because it was the exact sort of resident who should be taken to a Needs Rounds’ (PPIE 2).

PPIE 2 draws attention to her multivocality, as she speaks from a variety of different perspectives. She, furthermore, described her familiarity of the research processes in the study and her ability ‘to spot things on the research side of things’, ‘I’ve been doing [PPIE] for 15 years, so it’s very gradual. You do pick up the research terminology and you kind of know the research processes’ (PPIE 2).

Although a ‘lack of distinction’ and the ‘notion of equality’ were consistent across most of the interviews, some academic interviewees perceived the role and input from PPIE members to be focused on personal experience, whereas academics would focus on other elements:

There’d be some members of the PPIE team that might be more likely to be able to draw on past experience of being in PPIE work or in practice, or […] they might be talking about their own personal experiences of loss and death and draw on those side of things. Whereas you might have a senior academic who might talk more about or be able to contribute more in terms of theories and knowledge and expertise. […] So, it kind of creates a more rounded circle.

Academic 4

PPIE 1 emphasised that her lack of familiarity with research was something that was an asset and markedly distinctive to the academics: ‘Sometimes it’s good to not have some idea of what’s going on, but, you know, if you know too much, it’s just like when you’re a researcher, aren’t you, really?’ (PPIE 1). PPIE’s 1 positionality as ‘newcomer’ was also discussed by PPIE 2, who felt that her perspective was helpful and ‘novel’.

This focus on complementary inputs showcased that while team members could draw on and speak from multiple standpoints, they still held primarily different roles. The subsequent theme will discuss these roles further, illustrating how academic interviewees understood and negotiated their respective remit when working alongside PPIE members.

Negotiations

Negotiations refers to how academic team members held and negotiated their remit alongside PPIE members. Several academics drew attention to the clarity of remit extended to them in the context of the monthly meetings:

[The PI] knows what she’s wanting so she’s very clear and not wishy-washy about what it is that she needs and the roles that everybody has and what they need to deliver within that. So [the PI] very much makes a good chair of a meeting because she was in there, like this is the agenda with a lighter one, okay whose responsibility was it to do x, y and z […] I think that helps, you know, just structure the meetings that people were clear and knowing, I guess, sort of … that there wasn’t any sort of faff […] It was an hour meeting it meant that we got … the agenda was always got through.

Academic 7

The above interview extract foregrounds the discussion of the tasks involved in the research projects. It, moreover, exemplifies the responsibilities assumed by the PI, which include managing the meeting’s schedule, structure, time management as well as the contributions made by PPIE members. Some interviewees described how, on occasion, the contributions of PPIE made in the monthly meetings could be tangential requiring management by the chair. Academic 6, for example, considered the time spent in meetings explaining aspects of the study to PPIE:

I think sometimes [PPIE] comment on things which are more within I think the academic methodological clinical sphere, and they’ll ask questions which demonstrate that they’re kind of maybe moving beyond where their role sits. So they’ll say, ‘Oh, I don’t understand this term,’ and it would be [referring to] a methodology paper. And they’ll have been included for information and then they’ll say, ‘Oh this doesn’t make sense,’ and it’s actually for an academic journal, and PPIE people aren’t the audience for this output […] The purpose of this communication isn’t that everybody understands, this is that it fits with this journal. So I think there’s a bit of blurring sometimes about what’s meant to be for everybody to understand, and what are certain outputs for certain people. […] you don’t need to understand all of everything. And that’s I guess part of how I think about teams, is that we all bring different strengths, so, as long as we understand how we’re contributing, then that should be okay.

Academic 6

Two academics suggested that the role of the PPIE members would differ if the project was exclusively quantitative:

I think it’s a more challenging relationship for the quantitative elements because so much of what we do does have to be more rigid, and particularly in a sort of health study where you have to pre-register your hypotheses and how you’re going to analyse it. And a lot of that is kind of determined, there’s not as much choice. So, the value comes later, I guess, when you’re then trying to understand the findings or think through the implications, and that doesn’t work so well with the idea of being involved throughout, I guess. Now, I suppose one encounters that by having people involved really early on at the start when you’re conceptualizing the project, and actually writing the funding application, because there, I suppose, you’re deciding what the hypotheses are, and what the primary outcome should be. But there’s that kind of bit in the middle where there’s not a lot of scope for input. And I guess, that doesn’t sit as well, with the idea of kind of being involved, I suppose, throughout […] If the analysis you’re doing is quite statistically complex, then it’s quite a big ask for people to be up to speed with that, and we don’t necessarily expect the other academic members of the research team to understand all the ins and outs of the models that we’re using or whatever.

Academic 2

The remit of the PPIE members shifted throughout the study. The study had a lot of ‘moving parts, with various tasks to be involved with’ (PPIE 3) and that, as a result, their ‘input changed a lot’. Patient and public involvement and engagement members were involved in coding transcripts, the coproduction of interview schedules, co-authoring of reports and facilitation of workshops. Consequently, the negotiation of tasks, time frames and skills was ongoing throughout the study.

PPIE 3 explained that the time invested in explaining various aspects of the study made her feel like her contributions were ‘valuable’. Thus, the work done by academics could be considered an important part of PPIE’s involvement in the study. It, nonetheless, was something that academics had to attend to and negotiate in the monthly team meetings.

Academics and PPIE members were considered on an equal footing, despite also holding specific roles, but which required explicit work to signpost appropriate elements for PPIE members to contribute meaningfully to the study.

Professional development and learning

All interviewees were asked if the study had generated any transferable learning for them regarding the PPIE. Several academics answered by referring to the central and integral part PPIE members played and how this would inform other research projects: ‘The PPIE element had been tokenistic [in past research projects] and now it’s made me think we need to have this much more in the realms of people actually on the project with us’ (academic 5).

Academics reflected on the underpinning values of the study, for example:

I think that that whole thing about paying attention to the kind of values base to begin with. Paying attention to, you know, the integration. Paying attention to language, really actively working on that the whole way through. That whole business of checking out and checking in. And just not taking … so make not taking it for granted. Make it really explicit without being patronizing. You know, that actually work with our colleagues. So really building that kind of relationship-based work.

Academic 9

The quote underlines the need for active engagement in values, and ‘continual checking in and checking out’ throughout the study. Another academic described how the project used terms of reference (see Report Supplementary Material 15) to explicitly outline how the team would position themselves ‘in relation to the power and authority’ (academic 6).

Academics thus reflected on how this experience of PPIE would be used by them in future studies, and the underpinning ethics and processes, PPIE members named specific skills they developed over the course of the study. For example, sitting on a recruitment interview panel:

That very first [interview] was with [the PI] and it was a valuable opportunity because it led to other [research projects in which PPIE was required]. Once I knew how to do it and then develop the skill, then other people started asking me would I sit on the interview panel. And because I’d already done it with [the PI], it made the whole thing easier.

PPIE 2

PPIE 2 also had the opportunity to attend the NHS ethics committee meeting, though was disappointed by their lack of involvement at the meeting:

I was so looking forward to it. I helped with the ethics application, obviously, and on the day, we attended the ethics committee virtually, but it wasn’t a positive experience for me. It was positive in that we did get ethics approval, but I wasn’t included in that meeting at all (by the REC). I wasn’t acknowledged. And there was another PPI member on that Ethics committee, and you could easily identify who he was, and I thought he might have asked me a question about the PPI, but he didn’t speak to me at all […] I know that the tone of it was very business-like, but it was quite unfriendly.

PPIE 2

High levels of support in advance of the REC from the team were contrasted with poor engagement experienced at the ethics committee meeting. The PPIE member noted that ‘We had a pre-meeting, you know, [researcher and PI], and I saw that I was very well prepared and then it was just as if nothing, just that was it’. Despite being involved in drafting the ethics application and attending the meeting, there was no acknowledgement of the PPIE member during this, nor were any questions directed at her. So, although the intention was to enhance the PPIE member’s development in relation to ethical governance, processes and structures external to the team hindered this being fully realised.

Another PPIE participant focused on the ‘insights’ and ‘little glimpses’ she gained throughout her involvement of the Needs Round study:

I have learned quite a lot about what goes on and, you know, it could, you know, go towards other projects because I think things are changing all the time with technology and virtual wards and things […] I’m always open to new things, but I think I hadn’t realized how complex it was. And I have got a better understanding now.

PPIE 1

What she learnt related both to the study’s content as well as to the research processes that academics drew on and could be applied to other studies.

Withdrawals from the study

Forty-five CHs were recruited over the entire study period, 16 of which withdrew (11 prior to implementation commencing and 5 withdrew during implementation). All six hospices had at least one home withdraw from the study (Table 17). Of the CHs that withdrew, the majority (94%) were for-profit/privately owned and provided nursing care (81%). Half were in Scotland and half in England, mostly located in central Scotland (38%) or East England (38%). The average number of beds across the 16 CHs was 53, the smallest providing care for up to 26 residents and the largest up to 123 residents. When divided into categories, nine were classified as medium-sized homes (11–49 beds) and seven large (50 + beds). All homes provided care to older adults over the age of 65 years, but 50% of the withdrawals also provided care to adults younger than 65, two of which accepted residents from the age of 18 years.

Descriptive analysis

Examining the survey data, the highest rate of participation was observed for the CAPA surveys, and 23 CHs (79.3%) returned one or more CAPA survey. Eighteen CHs (62.1%) returned one or more QODDI survey. In comparison to the survey data (CAPA and QODDI), lower levels of participation were observed for the return of baseline (n = 14, 48.3%) and follow-up data. Baseline data were collected on people residing in study CHs in March 2021, and included data such as the status of residents, sociodemographic characteristics, staffing costs and interactions with healthcare professionals. The follow-up data were collected from March to May 2022 and included a repeat of the sociodemographic questions about CH residents. Data on staffing costs and interactions with healthcare professionals were not requested at follow-up, due to feedback from CH staff and the steering group. Only 12 of the CHs (41.4%) provided follow-up data, of which 9 (75.0%) had provided baseline data.

Table 18 shows the descriptive characteristics of the CHs participating in different aspects of the study. Percentages shown in the table represent the response rate for each factor, for example 50% of large CHs in the sample provided baseline data. Participation in the family perceptions survey (CANHELP Lite) was not included when calculating participation, since so few were returned to the team, and reporting on these data would risk maintaining site anonymity. A descriptive overview is included in Families’ perceptions: CANHELP Lite.

Overall participation scores (as described in Care home participation) ranged from 3 to 48 whereby a higher score indicated greater engagement and participation in different aspects of the study. Mean participation scores (Table 19) were higher in large CHs compared to medium and in voluntary compared to privately owned homes. Smaller differences were observed between the location variables, whereby CHs in Scotland had higher average participation scores than in England and CHs in urban areas scored higher than rural.

Regression analyses

None of the included predictor variables were statistically significant in predicting participation at baseline (Table 20). However, the provision of residential care was significantly negatively associated with overall participation score (Table 21), suggesting that those describing themselves as nursing-only homes had higher levels of engagement with the study overall, compared to residential-only or combined nursing and residential homes. The lack of statistically significant results is likely partially attributable to the small sample size (n = 29); thus, the patterns and directions of non-significant coefficients are still important to discuss.

When examining the results predicting participation at baseline and overall participation score, there are several discrepancies. For example, the results suggested that Scottish CHs are less likely to participate at baseline compared to English homes, but the coefficient for overall participation is positive, which would imply Scottish CHs have a higher level of engagement when participating in different aspects of the study. A total of 18 Scottish CHs were recruited into the study, 44.4% (n = 8) of which withdrew. However, in England, 27 CHs were recruited into the study, 29.6% (n = 8) of which withdrew. As such, there were fewer Scottish CHs participating in the study (n = 10) compared to English (n = 19). Additionally, some CHs were recruited onto the project later than others. Since baseline data were collected retrospectively (i.e. data on residents were collected in relation to March 2021), many CHs completed this at a much later date, potentially impacting their ability to provide accurate baseline data. Thus, CHs joining later may have been less likely to provide baseline data if they did not have this readily available. Descriptively, although a higher number of English CHs provided baseline data, the response rate when examined by country (see Table 18) was higher in Scotland (50.0%) than in England (47.4%) which does not align with the trend suggested in the regression analysis.

Large CHs were also found to be less likely to participate at baseline but had higher participation scores. Descriptive analysis revealed that participation scores for staff in large CHs were, on average, 8.2 points higher than in medium CHs (see Table 19). One interpretation of this might be that the number of staff is likely to be higher in large compared to medium homes; therefore, for research activities like the CAPA which were completed by different members of staff, larger CHs have the potential for higher participation scores simply because of a larger staff pool. Future analysis should look to equivalise the overall participation scores based on the CH size to avoid potential skews in the data.

In contrast, the direction of the results for both rural CHs and residential CHs aligned for baseline and overall participation. The CH being in a rural location compared to urban was positively associated with baseline and overall participation and residential care provision was negatively associated with both participation outcomes.

The lack of statistically significant results makes it difficult to draw any substantial conclusions about the factors that predict participation in palliative care research among CHs.

What works for whom, under what circumstances and why?

What works? Monthly Palliative Care Needs Rounds can provide CH staff and SPC staff with protected time and the opportunity to collaborate to plan for residents’ last months and weeks of life. Fidelity to the Needs Rounds checklist24 was low for case conferences and providing clinical interventions. However, 158 clinical interventions from SPC clinicians were triggered. Implementation showed that additional family meetings/multidisciplinary meetings did not work as the third of the three-component model. Both direct clinical work and family/MDT meetings were hampered by COVID-19 restrictions. Specialist palliative care was welcomed, but some input was generalist palliative care delivered by the SPC clinicians.

For whom? Qualitative data showed that Palliative Care Needs Rounds work for CH staff (greater understandings of palliative care, improving collaborations and trust), and SPC staff (improving collaborations and strengthening relationships and trust). Palliative Care Needs Rounds benefit CH residents (better informed staff, improved quality of dying). Palliative Care Needs Rounds also benefit relatives (giving greater confidence in the quality of care). There was insufficient evidence to demonstrate impact on healthcare systems, notably health economic impact.

Under what circumstances? The study was conducted during COVID-19 restrictions, and hence Palliative Care Needs Rounds were implemented and assessed during extreme workforce challenges, including high turnover and recruitment difficulties, exacerbated by poor pay and conditions. COVID-19 also meant that core parts of the model were not tested, such as family meetings.

Beyond COVID-19, CH staff have variable skills and confidence and sometimes have insufficient support within the wider health system. Flexibly arranging Needs Rounds to suit CH needs, having manager buy-in and building collaborative relationships can help to encourage time to be protected for the monthly meetings. Facilitation by SPC staff which is responsive to the complexities of CH context helps build collaborative relationships.

Why? Palliative Care Needs Rounds can provide timely expert input to CH staff and residents. The learning/educational component of Needs Rounds can strengthen understandings of ACP, the dying process, symptom and pain management, and communication mechanisms between CH staff, families, SPC staff and primary care. Needs Rounds can offer protected time and collaborations. Needs Rounds can shift power dynamics between CH and hospice services by harnessing complementary expertise. For relatives, Needs Rounds can work well because they increase confidence in quality care. Needs Rounds can augment existing practice by building CH staff confidence for those who attend.

Facilitation and facilitators are important mechanisms for successful implementation. Facilitation styles that worked included developing positive relationships, demonstrated through jokes, compliments and referring to what they had learnt from each other. Facilitative approaches were collaborative and involved gentle work in determining people’s current knowledge and areas for extending insights as well as which residents were appropriate to discuss. Managers needed to be key facilitators to enable staff to attend Needs Rounds, through rotas or overtime.

The programme theory is as follows:

While CHs experience workforce challenges, like high turnover and recruitment difficulties, exacerbated by poor pay and conditions, staff have variable skills and confidence and sometimes having insufficient support within the wider healthcare system. Needs Rounds can provide CH staff and SPC staff with the opportunity to collaborate during a protected time, to plan for residents’ last months and weeks of life. Flexibly arranging Needs Rounds to suit CH needs, having manager buy-in and building collaborative relationships can help to encourage time to be protected. Needs Rounds augment existing practice by building CH staff confidence for those who attend and prompting anticipatory planning and care. A facilitation approach that is responsive to the difficult context in which CHs operate can improve collaboration. In addition, Needs Rounds can strengthen relationships and trust and shift power dynamics between CH and hospice services by harnessing complementary expertise. Integrating learning into Needs Rounds strengthens understandings of the dying process, symptom and pain management, ACP and communication mechanisms between CH staff, families, SPC staff and primary care. This leads to higher-quality resident-centred care, enabling residents to be cared for and die in their preferred place. Needs Rounds also benefit relatives by increasing their confidence in care quality and CH staff by providing timely expert input.

The implementation of Needs Rounds was grounded in flexible facilitation to respond to CH contexts. Integrated Promoting Action on Research Implementation in Health Services underpinned the development of the model and approach to implementation. The innovation construct within i-PARIHS128 recognises that evidence is often adapted in line with recipients’ underlying knowledge sources, including their contexts, motivations and perceptions regarding the innovation. This innovation construct was explicitly harnessed within the study in the coproduction approach used to elucidate a range of stakeholder views, understand how Needs Rounds fit with existing CH organisational processes and explicate how these intertwine with external systems. Consequently, the intervention was designed to enable Needs Rounds to be responsive to wider contexts to improve the chance of successful uptake into practice.

Specialist palliative care staff were aware of the difficult context that CHs work within and tailored their facilitation approach around this. Building and engaging in reciprocal relationships, where collective expertise is acknowledged and harnessed, helped to build trust and improved interagency working. Needs Rounds enhanced communication mechanisms with primary care through learning about how best to communicate with GPs and using the expert status of the SPC clinicians, resulting in more responsiveness to requests to meet residents’ wishes and preferences, so that they could receive end-of-life care in their preferred place.

Over the 12-month intervention period, we found statistically significant increase in CAPA scores, indicating that CH staff grew more confident in their ability to adopt a palliative approach. This was a clinically important difference, as reflected in the qualitative interview data.

Deaths occurring later into the intervention were rated more positively than those at the beginning of the implementation period. Thus, Needs Rounds appear to improve the quality of death/dying for CH residents. Family members are mostly satisfied with their relatives’ care. However, since limited data were available for analysis, it is not possible to conclude the scale of this or impact of Needs Rounds on family perceptions of care.

Direct costs attributed to running Needs Rounds include staff time, from both CHs and SPC, alongside expenses such as equipment, medications and travel. The sample size of health service use data was too small to be able to conduct a meaningful inferential test for statistical significance of the treatment effect. The limited data available indicated a slight rise in hospitalisation rates.

The PPIE evaluation concluded that involvement and engagement were woven throughout the study, with three PPIE people included as full team members engaging in a wide range of activities. The impact of their engagement was varied and included adjustments to study measures, recruitment of research staff and being invited to participate in other research studies.

Analysis of patterns of data returns indicated that CH staff struggled to engage with some of the research processes, notably returning quantitative data. The conduct of the study during COVID-19 national lockdowns and ongoing impact on workforce resulted in more difficulties than typical in this already challenging environment. Fidelity to the coproduced UK model of Needs Rounds was variable, and in assessing a random selection, many sites achieved low scores despite qualitative interviews providing evidence of managing to achieve important outcomes.

Sitting these findings alongside outcomes from the prior Australian work enables a holistic assessment of ‘what works for whom under what circumstances and why’. The Australian evidence base is drawn out in subsequent subsections around palliative care in CHs, building interagency relationships and hospital transfers.

Palliative care in care homes

The study identified that Needs Rounds can be implemented in CHs, improving access to and provision of palliative care for residents, benefiting them, the staff and relatives. The Needs Rounds approach includes components that are recognised as necessary for integrating palliative care into CHs, namely facilitation, education/learning, alongside internal and external engagement. 130

The Needs Rounds study helps to address micro and meso needs identified in the European Association of Palliative Care’s (EAPC) white paper on palliative care in long-term care facilities,131 while also indicating the intervention is congruent with the macro context. The EAPC’s proposed futures include, at meso level, a focus on identifying facilitators and raising awareness, while developing a shared palliative care vision. Greater team working and support for managers introducing change are also noted at meso level. At micro level, EAPC identifies the creation of a culture of learning or development, delivering palliative care education/training and building confidence as requirements. Embedding palliative care into delivery, encouraging reflexivity among staff and planning for sustainability are also core elements of the micro level context advocated for by EAPC and addressed by Needs Rounds.

Provision of responsive education, tailored to staff needs, is woven throughout each monthly Needs Rounds meeting. A recent review of literature on palliative and end-of-life care education programmes demonstrates the need for such flexibility in addressing CHs’ needs and organisational context,35 alongside implementation frameworks. Thus, Needs Rounds help address this recommendation, and the strengthened relationships between CHs and SPC means CHs are better placed to draw on wider hospice education programmes, beyond the learning provided in Needs Rounds meetings. Education, while clearly a priority,30 remains a challenging area, given the precarity in workforce retention. This study augments evidence on Needs Rounds from Australia where it was demonstrated to provide education and support to CHs leading to more confident/competent staff25 and better-quality deaths for residents. 85

Strengthened relationships and interagency working

At a meso level, Needs Rounds showed improvements in interagency working between hospices and CHs. Collaborative facilitation and partnership working flattened the hierarchies that frequently permeate multidisciplinary team working with CHs. 132 Although models of improving healthcare for older people in CHs often point to primary care as a core partner, including EHCH in England,133 realist reviews challenge this perspective. General practitioners are rarely perceived as being central for delivering improvement in CHs,134 and there is insufficient evidence to support the belief that linking GP practices to CHs improves resident outcomes. 75 Being present and responsive is more important than having a named primary care provider, particularly when staff training was focused on areas of their interest and need. 75

Nevertheless, interagency working with primary care may continue to have an important place in policy initiatives and practice developments for aligning CHs and primary care. In England, the EHCH programme was introduced to standardise some aspects of healthcare for CH residents, and as an extension to the core GP contract, it promotes the provision of a clinical lead for each CH, weekly home rounds and identification of residents for MDT review, holistic assessment, development of care plans and prioritisation of residents for structured medication reviews. EHCH thus looks well positioned to dovetail with Needs Rounds. However, EHCH implementation started in 2020 and is due for completion by 2024 with patchy uptake to date. None of the English sites in this study had operational EHCH. In Scotland, the approach to palliative care in CHs has prioritised holistic assessment, MDTs, ACP and prompt access to generalist and specialist input. 83 The key implication lies in the integration with current practice in both countries. This would involve primary care integration with CHs being complemented by the SPC input from the Needs Rounds model.

Needs Rounds demonstrate relative advantage over other innovations designed to enhance interagency working between CHs and SPC. Palliative Care Needs Rounds provide a structure to be present and responsive, with case-based education focused on emergent and pressing topics. Thus, Needs Rounds create a mechanism where staff felt they had a common purpose for change to occur. Needs Rounds also improved communications with primary care, which resulted in increased responsiveness to CH staff requests for support.

The outcome of improved interagency working and associated mechanisms stand counter to studies which report ongoing hierarchical structures, and dichotomies between health and social care, reported in Canada. 135

The Needs Rounds facilitation approach was based on fostering mutuality and reciprocity to enable CH staff and SPC clinicians to learn together, and thereby disrupts and reformulates negative attitudes towards care work. 136 These negative stereotypes are exacerbated by unequal hierarchies that exist between MDT working in the social care sector, with care staff often being viewed as inferior. 132 Needs Rounds counter such stereotypes by adopting an approach that acknowledges CH expertise and champions the vital role of CH staff. The important role of communication in establishing and enhancing interagency and interprofessional relationships meshes with evidence from Australia on core mechanisms that enable Needs Rounds to work. 137

Hospital transfers

Although in this study Needs Rounds did not reduce hospitalisation rates, there is evidence that they have improved processes. That is, CH residents were able to be cared for and die in their preferred place, which is predominantly in the CH rather than hospital. The Needs Rounds evidence base for reducing hospitalisations and length of stay was established in two Australian studies. 23,138

Decision-making around hospital transfers is complex. 139 Hospital transfers, which are not clinically indicated, should be reduced to improve patient care and increase opportunities for people to die in their preferred place. Practices around hospitalisation have been significantly impacted by COVID-19. A realist review into ACP through COVID-19 found that residents and relatives opted against hospitalisations due to higher risks of infection and fears of having a lonely death. 60 However, ACPs risk not being sufficiently robust to enable preferences is to be fully understood and honoured. 140 A relative advantage of Needs Rounds is that learning is tailored to take account of these contextual factors and support decision-making and actions around hospital transfers. This tailored learning also increases the confidence and competence of CH staff to facilitate better-quality lives and deaths and have timely conversations to increase family confidence in the ability of CHs to meet their relatives’ need.

COVID-19

Needs Rounds took place in the context of COVID-19, with the study commencing 6 months after the initial UK lockdown. When the intervention started, CH staff had accrued considerably more experience in caring for dying residents and improved the use of ACP reinforcing the need to engage in such conversations rather than solely produce a document reflecting resuscitation preferences. 141 Needs Rounds augmented this improvement through discussions about residents’ symptoms and needs.

COVID-19 also led to depleted support from primary care, while increasing workload on the CH staff. COVID-19’s impact on the CH workforce significantly reduced available resources including personnel to attend from both within the CH and to host the visiting SPC clinician. Use of online video platforms, to accommodate local lockdowns during the implementation phase, enabled Needs Rounds to continue, which reflects Needs Rounds use in Australia. 27,28

Engaging care homes in research

Structural challenges such as limited time to complete research tasks were reflected in our data, despite offers to reimburse CHs for their time collating data. Difficulties of conducting research in CHs have been documented including the limitations on staff time to complete data returns. 130,142 Other core challenges and solutions include management engagement, ensuring research processes work for CH staff, managing ethical hurdles and structural challenges to collecting research data in CHs. 143 While the research team’s relationship with each site was very positive, feedback from some CH staff indicated variable understanding about what data were requested and the rationale.

Analysis returned limited significant results regarding the prediction of CH participation across the study. Comparison with the wider literature is difficult; previous studies focus mainly on the recruitment of CHs, rather than participation levels. 144 Overall, most CH managers were found to have little experience of research,145 which is also likely to impact participation, not just recruitment into research. Largely, the factors examined in relation to recruitment were administrative, such as staff time and turnover146 or linked to processes within the CH such as lack of understanding, compliance and gate keeping. 147 Moreover, in some cases, it was believed that certain research areas, for example, death and dying, could result in unfavourable publicity for the home,145 highlighting a lack of education or awareness on the importance of such research and potential stigmas which researchers may not be aware of. A substantial gap in previous literature was identified for studies that examined CH specific factors, such as size or type of care provided, in relation to the home’s participation in research activities.

The SPC clinicians were asked to audio-record each Needs Round meeting, yet only 33% of meetings were recorded. Hesitancy was explained in part due to the impact it might have on relationship building between CH and SPC. This points to learning regarding the kind of data which might be reasonably expected to be returned from such sites, where clinical practice/knowledge and skills are being displayed and where the need for trust development is a greater priority for the clinicians than producing research data. This may be particularly pertinent in studies which involve new relationships, rather than interventions delivered within existing and established professional networks. The research team encouraged recordings to take place, but also did not wish to jeopardise relationships, and hence the intervention and determining the primary outcome. Consequently, where one SPC clinician who was adamant that recording would be detrimental to the clinical relationship, the research team prioritised the intervention over data collection.

The data demands from CHs to conduct a full cost–benefit analysis were quite intense, and this may have impacted participation in monthly data returns. While an economic evaluation is important in contributing to the UK evidence for Needs Rounds, the burden of data collection on CHs does need to be considered.

Methodology

Conceptualising and reporting fidelity has been identified as an area for further development in palliative care interventions in CHs. 130 This study offers an outline of conceptualising fidelity in critical realist domains and how fidelity articulates with the development of programme theories. Critical realism challenges the ‘epistemic fallacy,’148 which occurs when the real and the known are conflated. It is based on ontological pluralism and seeks to avoid reductionism. 149 Acknowledging the inherent limitations of the knowledge production process, this study aimed to move beyond a partial account based on participant recollections or experiences. Thus, fidelity was assessed using both semistructured interviews and transcripts of the Needs Rounds recordings. The analysis was based on how participants reported the facilitation happening, as well as observations of the interactions between CH staff and SPC clinicians within the Needs Rounds. Attention was also paid to social, cultural and temporal dimensions which are often marginalised within fidelity analysis. 150

Alternative economic evaluation approaches (such as Social Return on Investment) could be considered that would balance the robustness of economic evidence against the impact on CHs providing data. But the need to accurately account for costs will remain with any approach, and so strategies and support to mitigate the difficulties experienced by CHs will be required. These could include streamlined data collection instruments, additional training and support for data collection by CHs or where possible researchers onsite to help gather the data.

Implementation objectives

Across Phase 1 and 2, we co-designed and built a programme theory for a UK version of Needs Rounds, which is responsive to the different (macro, meso and micro) contextual characteristics of the UK CH sector.

In Phase 2, we implemented the UK model of care. We determined that Needs Rounds were feasible to run in the UK and acceptable to both CHs and SPC staff. We were unable to determine all the anticipated outcomes, due to the paucity of data impacting economic analysis and analysis of family perceptions of care. However, the qualitative interview data indicated that Needs Rounds can improve the provision of palliative and end-of-life care in CHs.

Intervention objectives

We determined the transferability of the core elements of the Needs Rounds intervention in a UK context in Phase 1 and 2. For example, the model was adapted to adjust to primary care relationships with UK CHs and attended to the UK CH workforce configuration.

The programme theory delineated the mechanisms that enabled Needs Rounds to be used, such as providing the time and space to engage in proactive ACP, augmenting existing good practice and skills and developing communication, collaboration and reciprocity, alongside adjusted power dynamics and interagency trust.

The programme theory depicts the mechanisms common across sites. Individual differences were observed in CHs pertaining to regularity and accessibility of primary care staff, yet sites reported benefits of Needs Rounds despite this important variable. The outcomes of interest to stakeholders and observed in the data were stronger connections between CHs and hospices, improved trust and better connections between CHs and other services. Needs Rounds can improve communications between CH staff and GPs and CH staff and families.

Process evaluation objectives

Due to insufficient data, we were unable to determine whether UK Needs Rounds impacted hospitalisations including any cost implications. Quality of death and dying improved, with analysis indicating that deaths occurring later into the Needs Rounds intervention were rated more positively than those towards the start. We determined that the size of the CH also impacted quality of death and dying, with large CHs scoring lower than medium CHs. Not-for-profit homes scored on average lower than private CHs. We determined that CH staff grew more confident in their ability to adopt a palliative approach.

While there were insufficient cost data to conduct the planned cost–benefit analysis, the qualitative data highlighted the wider staff costs within the CH required to support Needs Rounds beyond the direct intervention costs. A fuller economic evaluation would be required to balance both these direct and indirect staff costs against the benefits of the intervention.

The potential for improving care

Needs Rounds can improve the provision of palliative and end-of-life care in CHs. The relative advantage of Needs Rounds has three dimensions: (1) has no direct cost to CHs, (2) provides direct clinical work and (3) focuses on specialist (rather than primary) palliative care. These features distinguish it from other approaches, such as the GSF. 151

Needs Rounds can help to develop stronger connections between CHs and hospices, improve trust and increase better connections between CHs and other services. Needs Rounds improve communications between CH staff and GPs, and CH staff and families.

Care home staff were widely praised by SPC clinicians for their expertise, insights and ability to provide high-quality care. With this strong foundation, Needs Rounds still had clear additional value and CH staff readily identified the usefulness of input from SPC clinicians. However, improved communication mechanisms are required to ensure that SPC can identify their added value. Despite precarious workforce arrangements,152 CHs enabled staff to have time away from their routine duties to attend Needs Rounds, though the numbers attending were often limited by staffing capacity.

Implications for commissioning and service delivery

Needs Rounds can provide a protected time where there can be increased collaboration between CH and SPC teams, focused on improving CH staff confidence, resident quality of living and dying and augment approaches to ACP. It is incumbent upon commissioners and service providers to ensure that CH residents with complex needs receive equitable care, when compared to older people living in their own homes. Needs Rounds can be one such approach and could contribute as one of the range of ways that CH residents gain the support they need to live and die well.

Future work could therefore focus on the sustainability of Needs Rounds, engaging with commissioners and policy-makers to secure funding to integrate and embed Needs Rounds into CH and hospice practice across the UK. This is needed over and above links with primary care to ensure access to specialist support and input, especially while primary care is so overwhelmed with work. Further establishing how Needs Rounds could work alongside primary care would be an important development for commissioning and delivery.

Interagency working between CHs and SPC should be routine elements of service level agreements. Using the Needs Rounds checklist to identify residents to discuss and provide biopsychosocial support, provide learning and develop clinical action plans will improve resident care, staff competence and relative satisfaction. Needs Rounds may not be required every month (e.g. depending on acuity of residents and size of CH) and could be linked with primary care rounds where they exist.

Provision of learning is beneficial, particularly given the increase in CAPA scores observed and qualitative feedback. CH managers should seek to ensure care staff can attend to benefit from the Needs Rounds meetings, and managers of larger homes should ensure rotation of staff attending to address the lower CAPA scores observed in CHs with more residents.

Needs Rounds can also be used to meet regulation requirements. While the CQC in England153 and Care Inspectorate in Scotland154 do not provide an inspection rating on end-of-life care, CHs should have a strategy in place to show how personalised support is provided to CH residents who are dying. Needs Rounds provide a formal mechanism to achieve this and align with several of the key CQC/Care Inspectorate aims, including enabling the person to express their preferences, reducing preventable hospital admissions, maximising quality of life, including through symptom and pain management, and supporting carers. Needs Rounds address 32 recommendations of the Scottish 'Healthcare framework for adults living in care homes'. 155

Scalability and sustainability will be dependent upon resources available for staffing Needs Rounds. This may require redeployment of existing resource to CHs, rather than an expectation of additional funding, particularly given the context of austerity in the UK. Both scalability and sustainability will also be related to how Needs Rounds articulate with primary care services. For scale, since primary care extends across the UK and since all UK residents should have access to general practice, there are key opportunities to wrap Needs Rounds into GP contacts. This approach would extend the building and maintenance of multidisciplinary working and trust, rather than SPC.

Implications for care home research

Despite the prevalence of CHs as a site of death, palliative care remains a marginalised focus of work, often receiving only a passing mention. 156 Future research needs to increase the evidence regarding living well and dying well in CHs across the UK in order to address issues of equity in access to specialist care and support.

Lower CAPA scores for CH staff than managers indicate that this is an urgent area for development. While Needs Rounds can support care staff learning and development, longer-term follow-up will be needed to determine how much knowledge and learning are retained and applied by those attending Needs Rounds.

Given the importance of relationships and facilitation to successful implementation of interventions, a discourse analysis of the Needs Rounds recordings data set could be conducted to further understand relationships dynamics and how the facilitation approach shapes delivery.

The project highlighted several common factors in working with CHs. The minimal engagement with study materials may reflect the lack of capacity staff have to read and retain information given the other pressures on their role. While this may be explained with reference to the pandemic, it is also possible that different strategies (beyond e-mails, phone calls and videos) are needed to communicate with CHs. Future studies conducting research in CHs may consider using fewer data collection tools or varying the methods of data collection, that is, some on-site researchers, to reduce the burden on staff completing surveys. Validating shorter versions of the questionnaires, if possible, would also mean that surveys with fewer items are easier to administer and complete by CH staff and should yield less missing and incomplete data.

Care home manager buy-in was essential to recruitment and retention of sites. While national CH research networks (such as ENRICH) have potential to support managers understand the importance of research, this remains suboptimal. Strategies to resource engagement in research may go some way to providing incentives for managers to prioritise this as a core function. Providing finance to backfill managers’ time may not be sufficient, given the lack of quantitative data returned to the team despite such resource.

Systemic workforce issues also hamper research, particularly studies such as this which run over several years, including 12 months of intervention time. The high turnover of CH staff means that many were only present for some of the study, and many others will have moved in the post-study dissemination period. As noted above, minimal engagement in study information is further complicated by staff moving in and out of CHs during the study. The task-oriented and busy shifts of care staff do not allow time for reading and reflecting on research materials. Evidence of this was seen across the sites, whereby different staff members were responsible for the completion of the data elements. For example, many sites’ data were completed by managers. Many CH managers may find they do not have the capacity to engage fully in the research activities due to other responsibilities within their role; however, the research tasks may fall to them by default. Reimbursing CH staff for their time attending Needs Rounds interventions (as a parallel to the reimbursement of SPC clinicians’ employers) may also be necessary in future research to promote engagement.

COVID-19

The study was conducted when UK health and social care systems were impacted by COVID-19. The project commenced between the first and second national lockdowns. The impact of COVID-19 was more long-lasting and consequential within the social care sector than almost any other part of UK society.

Multiple adaptations to the study’s data collection and implementation were made to accommodate COVID-19. This included moving all data collection online, developing new strategies to develop and strengthen relationships between the research team and sites by use of Teams, bespoke videos, e-mails and phone calls. All workshops (including training in delivering Needs Rounds and co-design of the programme theory) were redesigned to allow for delivery via Zoom, and the steering group only met online.

While restrictions lifted in different domains at different times as the project progressed, the perseverance of impact in CHs was marked. Care homes often remained closed to external visitors, including any clinician who was not routine care staff or attending in an emergency capacity. The consequence of this was that the SPC nurses were unable to attend Needs Rounds in person for some of the study timeline. Academic researchers were unable to attend sites to build relationships or gather qualitative data. Families and friends of residents were unable to visit and support residents, meaning that the care tasks they would routinely conduct were added to the responsibilities of paid care staff.

With the use of barrier nursing, increased staff illness due to infection with COVID-19 and increased resident acuity due both directly to COVID-19 and indirectly as hospital services sought to reduce admissions, the impact on CHs was considerable.

The COVID-19 context of the study meant that CH staff had considerable pressures on their time, which impacted their ability to fully engage in the study. This included reduced ability to complete CH data returns (with consequent impact on the amount and quality of data available for analysis), reduced desire to distribute and follow-up questionnaires for families, lower numbers of staff attending monthly Needs Rounds meetings and shorter/fewer interviews with CH staff during Phase 2.

Doubtless, COVID-19 also impacted both recruitment and retention of CHs to the study. Recruitment was all conducted online or via phone, eradicating the ability to build relationships more fluidly through personal contact. The extreme levels of CH staff sickness, staff distress and increased workload meant that engaging in a research study was unable to be a priority.

Data returns

Throughout the quantitative analysis, the small sample sizes are identified as a limitation, and in many cases, these restricted the possible analyses with the data. Particularly, the CANHELP Lite survey asking about family members’ perceptions of care was only completed by 13 family members, and this limited sample size meant that only descriptive analysis could be conducted and, to avoid identification, responses could not be compared across sites. Future research might consider supporting CHs to recruit family members, providing promotional materials and information about the benefits of participating in research or offering incentives to participate.

The CAPA survey had the highest response rate across the survey data, with 171 individual responses, allowing for working samples of 155–158 across the different regression analyses. The main limitation with the CAPA survey was that only 28 staff members completed the survey on multiple occasions; thus, pre–post comparison before and after intervention was limited. Researchers might consider administering the CAPA at the start of the project and then again to the same staff members at the end, rather than leaving it open throughout.

Eighty-one responses were received for the QODDI survey; however, only 5 completed all 16 items. As discussed in Quality of death and dying, this required the analysis to be conducted with a subset of the items, limiting the validity of the findings. Due to the missing demographic data at follow-up, as many of the sites did not return data on the status or characteristics of service users, it is not possible to calculate how many service users died over the period of Needs Rounds to determine a response rate for the QODDI. During the monthly check-ins further prompts to CH staff to complete QODDI surveys and would give more complete data on the service users within the homes across the whole intervention period. As noted elsewhere, the limited demographic and health service use data meant that it has not been possible to conduct a cost–benefit or other form of health economic analysis. Further, CHs reported demographic data that indicated a bias towards White British residents, meaning that the theories derived from the data are not based on a representative sample of the UK.

It was not possible to analyse fully what happened within each Needs Round. All SPC clinicians were asked to record Needs Rounds and provide that recording to the research team. Specialist palliative cares reported that this was fraught, with several barriers to providing these recordings.

Specialist palliative cares were concerned that recording would disrupt relationship development and trust with CH staff. For one site, this meant that no recordings were undertaken. Other sites were variable in their ability to record and then share the audio file. Specialist palliative care staff reported that their priority was on building trust in order to provide a clinical service.

Further, many of the recordings included use of residents’ names. Due to the ethical approvals requiring no identifiable details to be shared with the research team, this meant that some recordings were unable to be shared with the researchers. One site’s SPC clinician would routinely record organic Needs Rounds discussions and then redact any names used. Other sites asked CH staff not to use resident names at the start of the Needs Rounds but found it hard to adhere to this atypical manner of speaking.

The study was also limited by the lack of involvement of CH residents and relatives. Throughout Phase 1 and 2 interviews and workshops, attempts to recruit people residing in CHs were not successful. Consequently, the theory was developed without direct insight from service users. Recruitment of relatives (both to interview in Phase 1 and to complete questionnaires in Phase 2) was also suboptimal.

Three lay members were part of the team, despite extensive work to recruit more people with direct experience of CHs, including people residing in CHs. As described in COVID-19, the PPIE input was therefore focused on lay members or people with a relative living in a CH for older people. Input from people with lived experience of CHs was similarly restricted in Phase 1 (interviews with key stakeholders). The study was therefore impaired by the paucity of first-hand experiences of residing in residential care as a context, their preferences for outcomes or gaining their insights into mechanism which might assist in changing care practices. Difficulties in recruiting a wider sample of PPIE members are likely related to COVID-19 and gatekeeper access to residents in CHs that would be willing and able to consent to such involvement.

Realist methodology

Focusing on ‘what works for whom under what circumstances’ allows a thorough analysis and representation of effective components of the intervention. However, this risks underreporting of negative findings, including when elements did not work or where outcomes were counterproductive. Reporting this study, we manage this tension in producing an ethically sound programme theory by way of inverting elements which did not work. Thus, if data showed communication was poor because managers lacked time to engage with them, then this was reframed as time being required for managers to engage with the materials. Therefore, examples of what did not work were formulated as what could work. Key connectors and examples of what did not work describes what did not work and how this shaped the development of the programme theory.

Patient and public involvement and engagement members

Three PPIE coinvestigators participated in core activity throughout, commencing with proposal development. For example, PPIE members contributed to ethics paperwork, attended monthly team meetings and co-facilitated the co-design workshops. They drew on their personal experiences of CHs and social care.

Patient and public involvement and engagement and the wider literature

The PPIE in this study exceeded the expectations laid out in the National Standards. 161 The evaluation of PPIE indicated that integrating people with lived experience and lay members into the study team was valued by all parties and was an ‘invited space’ curated by academics and inhabited by people with personal experience as well as clinical/academic expertise to create ‘new solidarities’ (p. 452). 162 The independent evaluation of PPIE offered prospective insights into the CMOs of involvement.

Commencing involvement and engagement during the development of the protocol is a common feature of PPIE,163 and in this study, early involvement allowed time to develop relationships, which became a core mechanism for ongoing engagement. 164 Recognising the porous nature of people’s identities and experiences meant that academics/clinicians were also able to speak about their lived experiences, without invalidating PPIE members’ contributions. 127 The explicit values165 of the research team ensured that PPIE members’ views were respected, alongside expertise brought by academic/clinical members of the team. Thus, PPIE involved both the hardware (methods, approaches, guidelines) and software (values, relationships)166 of involvement expected in NIHR studies.

The epistemological traditions of research projects can interlink with, and bear implications for, the role of PPIE members. Questions of how technicalities or methodological expertise sit alongside PPIE involvement have been foregrounded by others. 167 Evaluation interviews regarding whether PPIE would be as engaged in the quantitative data as qualitative rest on assumptions about the ease of articulating underpinning epistemology and methodology. As other authors have noted,168 simplifying some of the language may provide enough of a bridge to allow people without technical knowledge enough insights to inform the process and focus of a range of methodologies within palliative care research. Simplifying language was embedded throughout the current study, including research team meetings and all workshops with diverse stakeholders.

Interviewees in the PPIE evaluation offered only one reflection on the emotional content of a study focused on CHs, dying and death. This absence may reflect the level of support offered to PPIE members (see Patient and public involvement and engagement training and support) or indeed that talking about death/dying/care was manageable without being burdensome. Consequently, there were no barriers to full and extensive engagement of PPIE members in this study focused on palliative care, which has historically been constructed as a sensitive topic. Providing support and debriefs are likely to have mitigated any potential difficulties, which are reported in other PPIE palliative care studies. 169,170 In describing the supports provided to PPIE, this report seeks to render visible the otherwise imperceptible activities undertaken by academic research team members to enable PPIE to flourish. 171

The study achieved success criteria of increased knowledge, agency and confidence of PPIE and academic/clinician team members developing inclusive practices, extended reach of PPIE in research and changes to how the project proceeded. 172