A qualitative study exploring the difficulties influencing decision making at the end of life for people with dementia

Abstract Background Dementia is a progressive neurodegenerative condition characterized by declining functional and cognitive abilities. The quality of end of life care for people with dementia in the UK can be poor. Several difficult decisions may arise at the end of life, relating to the care of the person with dementia, for example management of comorbidities. Objective To explore difficulties in decision making for practitioners and family carers at the end of life for people with dementia. Design Qualitative methodology using focus groups and semi‐structured interviews and thematic analysis methods. Settings and participants Former (n=4) and current (n=6) family carers of people with experience of end of life care for a person with dementia were recruited from an English dementia voluntary group in 2015. A further 24 health and care professionals were purposively sampled to include a broad range of expertise and experience in dementia end of life care. Results Four key themes were identified as follows: challenges of delivering coherent care in dynamic systems; uncertainty amongst decision makers; internal and external conflict amongst decision makers; and a lack of preparedness for the end of life. Overarching difficulties such as poor communication, uncertainty and conflict about the needs of the person with dementia as well as the decision maker's own role can characterize decision making at the end of life. Conclusions This study suggests that decision making at the end of life for people with dementia has the potential to be improved. More planning earlier in the course of dementia with an on‐going approach to conversation may increase preparedness and family carers’ expectations of end of life.


| BACKGROUND
Health and social care professionals report they lack confidence in decision making with respect to end of life care amongst people with dementia, and this is compounded by a lack of professional guidance in this area, with much of the guidance on end of life care focussing primarily on people with cancer. 1 Dementia presents a global health challenge. Although incidence is declining, 2 the prevalence of dementia is still rising across the world because of the ageing population. [3][4][5] It is a progressive neurodegenerative condition characterized by a gradual decline in cognitive abilities and an accompanying increase in health and social care needs. Median life expectancy from the time of diagnosis is estimated to be around three and a half years. 6 However, it is often difficult to estimate the prognosis for someone with dementia. Unlike diseases such as cancer, where the prognosis and course of illness are relatively well-understood, the trajectory of dementia is much more uncertain and heterogeneous and can be punctuated with declines related to acute illness. 7,8 End of life in the UK is considered to be the final 12 months of life, as defined by the National End of Life Care Strategy. 9 The end of life phase in dementia is often recognized late 10 and therefore may not be managed optimally. In this study, we define end of life care as a period of time during which the individual or family and professionals recognize that the person is dying, which could potentially be a period of up to 1-2 years. 11 End of life symptoms experienced by people with dementia may not be different to those of other terminal conditions such as cancer, but may be experienced for a much longer period of time. 12 Common clinical complications include recurrent infections (such as pneumonia), difficulties with swallowing, incontinence and pressure sores. 7 As the person with dementia may lack capacity, difficult decisions the decision-making process. 14,15 The care process has also been framed as a triad involving the person with dementia, any family carers and professionals from the health and care systems. 16 Few clinical guidelines focus specifically on end of life for people with dementia and those that exist have been developed by organisations such as the National Council for Palliative Care and the Alzheimer's Society. [17][18][19][20][21] In this article, we identify and explore influences on decision making for practitioners and family carers at the end of life for people with dementia.

| Design
A qualitative methodology was chosen for this study using semistructured interviews and focus groups to explore personal and professional accounts in a sensitive manner. Both interviews and focus groups were analysed using thematic analysis methods.

| Participants and setting
Former and current family carers of a person with dementia were recruited from a national dementia voluntary sector group, purposively inviting those with experience of providing care at home or experiencing it in hospital, and those who had experience of providing end of life. Invitations were sent either in the post or by email from the voluntary sector organisation to potential participants. Interested participants returned a postal expression of interest to the research team or replied in email to the voluntary sector, and they were then followed up with a telephone call from the research team to explain the study in more detail and confirm eligibility.
We purposively sampled a broad range of professionals working with people with dementia at the end of life in England including; general practitioners, palliative care nurses and physicians, geriatricians, speech and language therapists, hospital nurses, health-care assistants, community nurses and pharmacists. Professionals were recruited through the Dementias and Neurodegeneration (DeNDRoN) coordinating centre and the Comprehensive Local Research Network (CLRN). Finally, the research team invited experts within the field of dementia and end of life care to take part in the study, using snowballing methods to complete recruitment. 22 Invitations for professionals were sent via email from the research team or DeNDRoN and asked to reply via email or telephone to the research team.

| Procedure
One focus group (n=3) and three individual semi-structured interviews were conducted with current carers; one focus group (n=4) was conducted with former carers. Individual semi-structured interviews were offered to family carers who preferred not to participate in a focus group. Four focus groups were conducted with professionals with experience in end of life care in dementia, who were purposively sampled to include a broad range of expertise in dementia end of life care. The practitioner focus groups consisted of between four to eight participants (focus group 1, n=4; focus group 2, n=5; focus group 3, n=8; focus group 4, n=7).
Interviews were conducted by an experienced qualitative researcher (ND), and focus groups were facilitated by a researcher experienced in conducting focus groups (ND, SI, JW), observed by an additional researcher who took field notes (JW, RM).
A topic guide was developed based on findings of preliminary work with family carers of people with dementia 23,24 and a review of the literature. 25 The topic guide was used to facilitate the focus group discussions and was modified for use in individual interviews. It was presented to the research development group consisting of practitioners working with people with dementia as well as family carers for feedback and developed further based on these discussions. The same topic guide was used with family carers and professionals, but also explored the family carers' experiences of caring for a person with dementia within the individual interviews. The guide explored decision making related to the following specific areas: Groups were presented with the words of each topic visually using a PowerPoint presentation with a brief explanation from the facilitator and asked to discuss difficult decisions which needed to be made around this topic. This study adopted a 'think-aloud' strategy for the focus groups, which encourages participants to vocalize their thought processes while being presented with a problem or situation. 26 This method has previously been used to understand difficult decision-making processes in health care. 27 Traditional semi-structured interviews allow people to rationalize their thought processes before speaking and therefore may not provide an understanding of how an individual arrived at their decision. 28 The results from these interviews highlighted the views and experiences of both family carers and practitioners. The findings formed the basis for the development of a series of heuristics to aid decision making for practitioners caring for people with dementia at the end of life. 29 Ethics approval for the study was obtained from King's Cross and Camden REC committee (Ref 15/LO/0156). Informed written consent was obtained from all participants prior to conducting the focus groups and interviews. The focus groups lasted between 60 and 80 minutes and individual interviews lasted around one hour.

| Analysis
Focus group and individual interviews were audio-recorded and transcribed verbatim by an external transcriber and checked by (ND). An inductive approach was taken using thematic analysis methods. Each Themes were revised iteratively, searching for negative and deviant cases to ensure the themes were supported by the data, increasing the rigour of the findings. 30 Family carer and professional data were analysed together as following initial reading of the transcripts and discussions within the research team, and it was clear there was a high level of overlap between professional and carer views.

| FINDINGS
Four key themes were identified which contributed to making decisions particularly difficult or complicated at the end of life. These include the following: challenges of delivering coherent care in dynamic systems, uncertainty amongst decision makers, internal and external conflict amongst decision makers and a lack of preparedness for the end of life.

| Challenges of delivering coherent care in dynamic systems
As dementia progresses, the person with dementia can find themselves moving through the health and social care system. Movement through the care pathway is not linear and is often punctuated by a series of sudden moves between different care environments including hospital admission.

| Uncertainty amongst decision makers
Amongst the possible triad of decision makers, the views of the person with dementia are needed, but are most often inaccessible. This means family members or advocates and practitioners must make decisions.
Often decisions were based on the family member's insight about/ or knowledge of the values or preferences of the person with dementia. However, they expressed feelings of uncertainty in how to best meet the needs of their relative. Further complications resulted if formal discussion had not taken place or if legal arrangements were not in place. As one carer explained:

Manager)
A strong driver which appeared to increase internal conflict amongst clinicians was their expectation of themselves to treat the patient and make them better, keeping to a stereotype that the role of the clinician is to prolong life. This was reported to be most common amongst doctors.

Some practitioners described the implications of such a situation:
I think it's very easy to make the decision to feed, to treat so they're seen to be doing something and it's not always the right decision for that patient. (P8, Nurse Ward

Manager)
In an acute hospital [people] expect interventions and treatment and they don't necessarily expect there to be a lot of multi-professional discussion before that happens, which is often the case, because it needs to be right for the individual, rather than a reflex, they're not eating, we must whatever happens, get some nutrition in to them.

(P4, Palliative care clinical nurse specialist)
In some settings what were described as rigid routines/standard practices, guidelines (rules and regulations) and performance targets limited practitioners' autonomy and complicated decision making. There was a sense of conflict between working outside guidelines and thereby potentially exposing oneself to potential litigation and complaints vs acting in the best interests of the person with dementia. For example, as one nurse specialist explained, guidelines on how often a person should be turned to prevent pressure sores tended to be followed even it was judged to be causing undue distress for the person with dementia.
Ascribing this to fear of the potential legal implications, practitioners said other practitioners (not themselves) were sometimes making decisions to safeguard themselves, rather than what was in the best interests of the person with dementia: .. If

| Lack of preparedness for the end of life
Preparing early for a progressive decline in health and the inevita- Practitioners seemed to suggest that most professional dialogue around dementia was centred on the expected cognitive deterioration over time with less recognition that dementia affects physical capabilities, and in particular, that it can be a condition that people die from.
Such emphases seemed to lead to some families feeling that they had been 'caught off guard' when end of life seemed to be on the horizon that that they were placed in the position of having to make decisions which they felt they were not adequately prepared to make. Timing of discussions is critical but if a person with dementia is not recognized as nearing end of life, then these conversations may not take place.
Practitioners acknowledged that there tended to be poor recognition of the dying process even amongst clinicians, which meant that conversations with some families were delayed or absent altogether: Many distressing situations arose because people weren't identified as dying, so that's something that, as professionals, we need to get better at, is to recognise…not just recognise but to communicating that to family, making sure..

It's not an ambiguous message. (P10, GP liaison)
It's an uncomfortable subject for us; it's also uncomfort-

| Communication
At the end of life, people with dementia are likely to move across care settings. This movement limits the possibility of developing relationships between the triad of potential decision makers (person with dementia, family carer and practitioner), and this is reflected in the findings of the current study. Caron et al. (2005) suggest that quality of the relationship between carers and practitioners is established from the first interaction but that a trusting and supportive relationship between these two parties can take months to years to develop. 14 This supports our finding that there is likely to be increased difficulty in sharing decision making between staff and family carers who are unfamiliar with each other.
Moves between different settings can limit the knowledge and insight into a person's underlying social and medical history. Having little or no knowledge of the person and not being able to communicate with them can lead to 'disease labelling' 31  Poor interpersonal communication between practitioners in the same setting as well as across settings, and between practitioners and family carers, not surprisingly appeared to increase the difficulty of decision making. Good communication has been consistently highlighted as an important aspect of shared decision making in various health conditions [36][37][38] as well as in end of life care. 39 Our study suggests that good communication needs to be considered at system level (eg information sharing) as well as interpersonal levels.

| Uncertainty and conflict amongst decision makers
Uncertainty was often a core component of much of the discussion that took place with family carers and practitioners. Not only was uncertainty a potential cause of sudden movement between settings with a direct impact on the person with dementia but it was also likely to create conflict within individuals as well as between the decision makers, indirectly impacting upon the person with dementia.
The unpredictable nature of the disease course and at times a lack of recognition of the dying phase appeared to be the main source of uncertainty for practitioners. Other studies have also found that limited understanding, recognition and preparedness of the end of life amongst both practitioners and family members are common. 40,41 People with dementia can have a decline in health with little expectation of recovery; however, on occasion, there may be a sudden and unexpected improvement. Our findings indicate that the unpredictable nature of the dementia trajectory may lead to practitioners fearing that families will not trust their opinions and decisions. This fear appeared to make communication between carers and practitioners difficult. The theory of cognitive dissonance suggests that individuals strive to ensure that their beliefs and behaviours are consistent. Inconsistency (dissonance) or disharmony results in the decision maker becoming psychologically uncomfortable, motivating them to reduce this dissonance by actively avoiding situations that are likely to increase it. 42 It may be that this is reflected in our findings amongst some practitioners. The uncertainty might result in avoidance of difficult conversations. This fear also extended to overarching concern about the risk of wider repercussions such as litigation or complaints.
The complexities relating to decision making at end of life amongst those with dementia seemed to leave some practitioners feeling both disarmed and conflicted. Practitioners may feel disarmed because they do not have all the necessary information to make the decision for the person with dementia. In this situation, they must rely on secondary sources for information regarding the preferences of the person with dementia. However, some practitioners appeared to assume that the views and expectations of relatives were shaped by their own previous experience of dementia amongst people known to them. Other A lack of preparedness can set up a cascade of events, whereby the person with dementia moves between care environments and is exposed to investigations and treatments that are not in line with the goals of care. Once the person with dementia is admitted to hospital, there may be an impetus to do something, in the way of treatment.
This may spiral and lead to a situation in which it is more difficult to take a step back and initiate end of life care.

| Strengths and limitations
The focus groups carried out in this study included a mix of practitioners from different disciplines. The interdisciplinary mix of the focus groups was valuable to obtain a range of views and perspectives. 44 Mixed groups have the potential to silence individuals; however, this study found that a mixture of participants encourage conversation, debate and friendly questioning.
There are possible limitations with the sample in particular those who agreed to participate may encompass stronger opinions regarding end of life dementia care and therefore present a potential sample bias. Regional variations in end of life care have been recognized 45 and participants in our study were all practitioners experienced and working in either dementia or palliative care in London, Greater London, or Essex; therefore, their views may therefore not necessarily be representative of practitioners across England. However, teams did vary in the level of services provided.
Only two focus groups and three interviews were conducted with current and former family carers; it would have been useful to conduct more focus groups and interviews with family carers to allow for an in-depth comparison of family carer and practitioner views. Family carers were recruited from a national organisation and therefore may provide a biased range of opinions, for example, relating their views to extreme negative or extreme positive experiences they had of caring.
Previous research has shown that family carers are more likely to take part in research if they have positive views to report, 46 however other studies have demonstrated a bias towards those with more negative experiences. 40

| Implications and further work
This study echoes other research conducted on needs and decision making relating to end of life care in dementia. 15 It suggests that the physical impacts of dementia, beyond cognitive decline, may need to be better recognized by practitioners and that there should be more efforts to engage families in such discussions if they wish. In terms of changes in care settings, decision makers need to consider the impact of moving as weighed against the potential gains. It is likely that some conversations with relatives need to be revisited multiple-times, as appropriate. Although increased importance is being given to advance care planning, it is evident that the uncertainty around decision making continues; therefore, important conversations between the triad of decision makers need to take place at an early stage. Movement through care settings is likely to complicate decision making and make it unclear as to whether end of life conversations have taken place.
The role of GPs may extend to forestalling unnecessary movement through different care settings, facilitating a more seamless journey of care when necessary, and ensuring better transfer of information about the person with dementia. Additionally, there appears to be a pressing need for improvements in informational sharing practices and policy. Practitioners should reflect on their own values and whether the expectations they place on themselves are in line with good decision making for their patients.
Research often concludes within this field that more training is needed for professionals, or that more information is needed for family carers about dementia and end of life in particular. However, we suggest that training is not always enough and guidelines can only guide to some extent. We suggest what is needed is more practical assistance, a tool such as a decision aid that encourages more engagement between professionals and carers, to have difficult conversations and carefully consider difficult decisions which need to be made. A tool such as this may enhance the engagement with advance care planning, and encourage both more professionals and people with dementia and their families to forward plan. Similarly, such a tool may be useful when planning has not taken place and decisions need to be made later on in the course of dementia when the person no longer has capacity.
Finally, such a tool could be used as a means of engaging those practitioners and or family in difficult conversation which many so often actively avoid.