A critical realist evaluation of advance care planning in care homes

Aims: To evaluate care planning in advance of end- of- life care in care homes. Design: A qualitative study. Methods: Qualitative data were collected from January 2018– July 2019 (using focus groups and semi- structured interviews) from three care homes in the South West of England. The data were analysed using thematic analysis followed by Critical Realist Evaluation. Results: Participants comprised of registered nurses ( N = 4), care assistants ( N = 8), bereaved relatives ( N = 7), and domiciliary staff ( N = 3). Although the importance of advance care planning was well recognized, the emotional labour of frequently engaging in discussions about death and dying was highlighted as a problem by some care home staff. It was evident that in some cases care home staff's unmet emotional needs led them to rushing and


| INTRODUC TI ON
End of Life Care (EoLC) in care homes is becoming the focus of growing attention due to the rapidly ageing population across developed countries (United Nations [UN], 2017). In Europe, care homes are increasingly becoming the most common places of death (Andreasen et al., 2019), and in the UK it is predicted that care homes will be the most common place of death by 2040 (Bone et al., 2018). In the UK, the generic term 'care home' describes both residential and nursing homes. Both of these types provide food and board, 24-h care cover and assistance where required with activities of daily living. In addition, nursing homes provide care by registered nurses (Laing & Buisson, 2016). Both provide EoLC.
Collectively, the care home sector in the UK provides care for approximately 418,000 people which represents 4% of the population aged over 65 and 15% aged over 85 (Laing & Buisson, 2016).
Given age is synonymous with progressive chronic illnesses (World Health Organisation [WHO], 2017), more people are now living in care homes with multiple chronic conditions (multiple comorbidities), such as Dementia and frailty (Kingston et al. 2018).
It is therefore important that care homes staff are able to provide quality, person-centred care for residents approaching the end of their life often with multiple chronic conditions (Aasmul et al., 2018). Although the definition of end of life (EoL) differs between countries, is generally considered that people are probably to die in 12 months (National Health Service [NHS], 2018). Care provided during this period is termed EoLC which comforts and supports people who are dying from a progressive life-limiting condition (Fisher et al., 2000). Furthermore, EoLC should be person-centred which involves considering the persons wishes, values, beliefs and family circumstances; and incorporating these factors into their care (Fisher et al., 2000;Kitson et al.,2013). Advance care planning is the most common tool used by care home staff to support person centred EoLC (Spacey et al., 2020).
Although advance care planning practices can vary depending on the cultural and legal contexts of different countries (Sharp et al., 2013), efforts to standardize the definition have been made (Sudore et al., 2016). An international multidisciplinary panel described advance care planning as a process that supports individuals to share their personal values, life goals, and preferences about future care (Rietjens et al. 2017;Sudore et al., 2016). The advance planning process usually consists of a discussion between care home staff, the resident and/or their family, which is documented in an advance care plan (National Institute for Health & Care Excellence [NICE], 2019).
Regardless of country, it has been well established that advance care planning can improve care quality, reduce unnecessary admission to hospitals, and decrease stress and anxiety for both residents and their family members (Aasmul et al., 2018).
Thus, the care home staff delivering EoLC need to plan, gather and record information about a person's wishes and preferences and take them into account to provide the best care possible for that person (Cloninger, 2011). Despite the importance of advance care planning in care homes, research has highlighted a lack of insight from the care home staff engaging in advance care planning, as well as the perceptions of family members (Spacey et al., 2018).

| Background
Given the increasing numbers of residents dying in care homes, and the complexities associated with advance care planning a plethora of interventions have been designed to support care home staff to engage in advance care planning discussions with residents and their family members (Braun & Zir, 2005;Cronfalk et al., 2015;Mayrhofer et al., 2016;McGlade et al., 2017;O'Sullivan et al., 2016;Temkin-Greener et al., 2017). The main focus of these interventions tends to be on educating staff on the importance of advance care planning, how to approach such conversations with residents and the legal and cultural factors associated with care planning (Sharp et al., 2013). For example, some residents may not want to discuss death and dying due to cultural or religious beliefs. Education also informs care home staff of The Mental Capacity Act (2005), and in cases where a resident lacks mental capacity, a legally appointed decision maker can engage in the care planning process on their behalf.
Both registered and non-registered care home staff are involved in advance care planning and are responsible for delivering patientcentred care for people at the end of their life (Leaman 2011).
Despite this, existing studies have not explored to perceptions of both registered and non-registered care home staff engaging in care planning (Spacey et al., 2020). The UK care home workforce is made up of non-registered care assistants (76%) and registered nursing staff (12%); the remaining workforce is made up of ancillary staff such as housekeeping and administrative staff (Griffiths et al., 2019).
Despite these complexities, current research tends to be outcome rather than process focused with the number of advance care plans in place for residents receiving EoLC being the most commonly reported marker of the quality care (Garden et al., 2016;Kinley et al., 2014). A recent systematic literature review evaluating advanced care planning interventions found that these outcomes were often short lived and that the numbers of advance care plans dropped post-intervention suggesting limited sustainability (Spacey et al., 2020). The reasons for this are currently unclear due to a sparsity of research exploring the underlying processes behind advance care planning in care homes. Specifically, current evidence about the rationale behind intervention design, the effect of context, and the response of stakeholders is sparse. This means that the impact of these factors on outcomes has thus far not been adequately established.

| Aims
This study aims to evaluate care planning in advance of end-of-life care in care homes. The following research questions were set to achieve this aim: • What mechanisms currently exist to support care home staff to effectively engage in advance care planning? (intervention mechanisms) • Can the context of care homes influence advance care planning? (contextual mechanisms) • How do stakeholders tend to respond to advance care planning? (human agency)

| Design
This study reports on the results developed over the course of a two-phased study that adopted a Critical Realist design (Porter, 2015), which assumes that the outcomes of interventions result from the interaction of a plurality of causal mechanisms. Mechanisms are embedded in the intervention itself (intervention mechanisms); in the social and organizational context (contextual mechanisms); all of which are filtered through individuals' interpretations and responses (Human agency;Porter, 2015). Thus, each of the above causal categories will be used to explore advance care planning in care homes. Qualitative data was collected from January 2018-July 2019 using serial semi-structured interviews and focus groups to explore advance care planning in three care homes in the South West of England. Participants included bereaved relatives (N = 7), registered nurses (including managers; N = 4), care assistants (N = 8) and domiciliary/administrative staff (N = 3) who were all interviewed twice.
All participants had been involved in advance care planning. The qualitative data were transcribed verbatim, analysed and results were reported in accordance with the Consolidated criteria for reporting Qualitative Research guidelines (Tong et al., 2007).

| Setting
The Care Quality Commission's (CQC) electronic database was used to search for care homes in one county in the South West of England to include both residential and nursing homes. The CQC is an independent regulator of health and social care services in England.
Purposive sampling was also used for the selection of data collection sites (Tongco, 2007) to ensure variation in the care homes settings.
Moreover, the local Care Commissioning Group (which are government run organizations responsible for commissioning most hospital and community services in England) provided advice throughout the process which was used to develop the following purposive sampling variables and eligibility criteria (  Table 3).

| Sample/Participants
The selection of participants was informed by two systematic literature reviews carried out by the authors (Spacey et al., 2018(Spacey et al., , 2020, which revealed that few studies reported on experiences of advance care planning from the perspectives of bereaved relatives, particularly those from a residential setting. Additionally, whilst current literature explored the experiences of registered care home staff, the perspectives on non-registered care home staff, especially those with non-formal caring roles were often overlooked (Spacey et al., 2020). After discussion with the participating care homes and the ethics committee, it was deemed inappropriate to include residents who were receiving EoLC in this study. Specifically, the researcher would not be aware of residents' level of knowledge as to whether they knew they were receiving EoLC, so it was deemed insensitive and unethical to ask them about EoLC. However, the inclusion of bereaved relatives added a broader service user's perspective. All included participants and their allocated pseudonyms are detailed in Table 4.
To aid staff recruitment, researchers attended staff meetings to inform the staff about the study and hand out invitation letters to potential participants. This enabled those who expressed an interest to contact the researcher independently of the care home manager. Bereaved relatives were contacted through managers using an invitation letter which included the researcher's contact details. When the potential participants expressed an interest in taking part in the study, they were provided with participant Terminology Definition

Intervention mechanisms
The mechanisms contained in interventions which are designed to change the behaviour of those at whom they are targeted

Contextual mechanisms
The resources and restrictions embedded in the social and organizational context may inhibit or promote the effectiveness of intervention mechanisms Human agency People's experiences, interpretations and responses to the intervention and contextual mechanisms

Outcome
The changes in behaviour that result from how people respond to the intervention and contextual mechanisms TA B L E 1 Critical Realist terminology and definitions information sheets, and a consent form. No participants refused to participate.

| Data collection
There were two phases of data collection meaning participants were interviewed twice. The questions used in phase one for the focus groups and semi-structured interviews were developed from two systematic reviews of the literature (Spacey et al., 2018(Spacey et al., , 2020. Phase two involved presenting the initial findings to the participants (from phase one) using focus groups and semi-structured interviews. Murray et al. (2009) found that the use of repeating or serial interviews helped to uncover the complexities of individual situations by allowing narratives to develop.
Both sets of focus groups with registered nurses and care assistants lasted 1 h and were conducted by two researchers, one acting

| Data analysis
Analysis was carried out at the same time as data collection for each phase, and was discussed during fortnightly meetings of the research team; this helped determine sample size and data saturation. Throughout the process, analysis was triangulated by the four members of the team. A different mode of analysis was used for each of the two phases of focus groups and semi-structured interviews. The primary purpose of the first phase was to attain descriptive data about the use of advance care plans in care homes. The data gathered were thematically analysed (Braun & Clarke, 2006) by coding to identify recurring patterns. These patterns were arranged into themes and subthemes that augmented, refuted or refined the theoretical insights about intervention mechanisms, contextual mechanisms, human agency, and outcomes that had been developed previously from our critical realist review of the literature (Spacey et al., 2020) to establish what works, for whom and in what circumstances. The primary purpose of the second round was to present these theories to participants and discuss their validity and implications. Credibility was enhanced by ongoing interaction with the data during collection as well as during analysis.

| Rigour
A variety of strategies were used to ensure rigour (Lincoln & Guba, 1985), for example, the initial analysis by the lead researcher (AS) was triangulated by the other members of the research team (JS, MB, SP), to confirm key themes and sub-themes. Furthermore, a reflexive record during data collection and analysis was kept, and fortnightly meetings to discuss and record the decisions made during analysis were held. These strategies helped to confirm the credibility of the emerging findings. Lastly, the second round of interviewing that involved a discussion with participants about the adequacy of the researchers' theories helped improve credibility and validity.

| FINDING S
Findings are organized according to the critical realist evaluation categories of intervention mechanisms, contextual mechanisms and human agency.

| Intervention mechanisms
This section examines the generative mechanisms embedded in interventions that were hypothesized to promote behaviour conducive to supporting more effective advance care planning. These consisted predominantly of educational initiatives.

| Contextual mechanisms
This section will explore the resources and restrictions embedded in the social and organizational context which were hypothesized as either promoting or inhibiting the effectiveness of advance care planning. Two significant contextual mechanisms were identified: financial variability between care homes and the diversity of residential populations.

| Financial variability in the care home sector
Analysis of the initial interview transcripts identified financial and organizational barriers to the implementation of education and training to support staff to engage in advance care planning. These findings imply that some care home staff's perceptions prevented residents with dementia from being involved and contributing to their own care. Rather than attempt to engage or assume capacity, these staff assumed a lack of capacity and went straight to the family, excluding the resident living with dementia. Similar findings in regard to discussing death and dying as part of advance care planning were evident when analysing the bereaved relatives' experiences. One bereaved relative explained that the care home staff went directly to her as the next of kin to ask about her husband's preferences who was living with dementia.
BR2 C1: They wanted to know what my wishes were you know for him I didn't mention any of this to him and I don't think he knew I let them know.

| Human agency
Human agency represents how stakeholders interpret and respond to the identified intervention and contextual mechanisms.
The most significant factor identified was the emotional difficulty that staff experienced in initiating discussions about death and dying.
3.3.1 | Emotion and reluctance to discuss death and dying It was evident that some care home staff found it emotionally difficult delivering EoLC to residents, especially towards the end stages of their life. The emotional labour of delivering EoLC was expressed by a range of care home staff from housekeeping, to care assistants to registered nurses and managers with their years of experience ranging from 2-29 years.
Care assistant 2 C1: In some ways you become their family because you're spending more time with them than their family. Often families stay away as they don't want to see their loved one dying so often you take their place. You try being their family for them. So, losing a resident like that can be very emotional yes.
Housekeeping C3: We get very emotional because we get attached to them, you can't help that you're human.
It was apparent staff's emotional attachment to residents impacted on their ability to discuss death and dying as part of advanced care planning. Analysis implies that staff's close attachments with residents led to them feeling uncomfortable discussing sensitive and potentially upsetting topics, such as death and dying.
Despite the reluctance to discuss death and dying it was well understood by all care home staff that advance care planning was part of their 'professional job', thus it was not avoided. However, it was evident throughout the transcripts that discussions about death and dying as part of advance care planning were sometimes rushed and overlooked. Whilst the numbers of advance care plans in place are important (Finucane et al., 2013;Kinley et al., 2014;O'Sullivan et al., 2016;Temkin-Greener et al., 2017), findings presented in this study suggest that more needs to be done to support staff to sensitively engage in discussions about death and dying to improve the quality of advance care planning discussions.
Although current education mechanisms such as shadow- Although, the findings from this present study support Hall and colleagues it was found that reluctance to talk about death and dying was experienced by many care home staff.
There was evidence to suggest that some care home staff's avoidance of conversations about death and dying was also about protecting their own emotional wellbeing. Specifically, analysis suggested that some care home staff avoided the subject of death and dying with residents as a defense mechanism. Similar behaviour has been found in professionals who are regularly exposed to emotional and traumatic experiences (Drury et al., 2014). It has been found that being continually exposed to traumatic experiences without support can lead to compassion fatigue, which is described as a state of mental exhaustion triggered by a decreased ability to cope with everyday settings (Baranowsky & Gentry, 1999

| Limitations
The study did not include residents receiving EoLC. This is recognized as a limitation given that service users' perspectives and experiences are often overlooked (Spacey et al., 2020). Nonetheless, bereaved relatives were included in the study which provides valuable serve user insight whilst also acting as a proxy for residents themselves. The number of care homes involved in the study was also a limitation, three care home were selected, limiting the transferability and generalizability of the findings. However, these homes were carefully selected to represent the main care home types in England. This study did not explore the implementation of any specific intervention to support advance care planning, therefore inferences about causality cannot be made. Nonetheless, our findings have highlighted some of the processes which may influence the effectiveness of care planning in advance of EoLC in care homes.

| CON CLUS ION
Findings provide care home staff, policy makers, regulators and researchers with a deeper insight into what needs to be done to support advance are planning in care homes. For example, findings from our study suggest that more research and education initiatives need to go into supporting care home staff emotionally to improve their ability to engage in discussions about death and dying as part of advance care planning.
In the longer term, the findings generated in this study can be used to inform the development of future interventions designed to support care home staff engage in advance care planning. However, further research is recommended to replicate the study in different contexts such as rural care home settings to assess the generalizability of the findings.

CO N FLI C T O F I NTE R E S T
No conflict of interest has been declared by the author(s).

AUTH O R CO NTR I B UTI O N S
All authors have agreed on the final version and meet at least one of the following criteria (recommended by the ICMJE*): (1) substantial contributions to conception and design, acquisition of data, or analysis and interpretation of data; (2) drafting the article or revising it critically for important intellectual content.

PE E R R E V I E W
The peer review history for this article is available at https://publo ns.com/publo n/10.1111/jan.14822.