Narrative inquiry on case studies of crisis in dementia.

30 Purpose: This study explored the coping styles that can be inferred from the discourse of dyads with 31 dementia, and how these appear to impact on care management. 32 Design: This was a case study approach. Participants were recruited from two teams managing crisis 33 in dementia in the UK. We employed multiple qualitative interviews with people with dementia and 34 their family carers over the course of one month. The analysis was first performed through thematic 35 analysis. Data were further analysed through narrative inquiry to create a story line, or play in our 36 case, for our findings. 37 Findings: Five dyads were interviewed and a total of 16 interviews were conducted. Three dyads 38 were husband-wife and two were daughter-mother relationships. The mean age was 67.4 for carers 39 and 79.8 for people with dementia. In these cases, the carer assumed responsibility for managing the 40 episode and was more likely to seek formal help if a pre-existing plan was in place. Otherwise, when 41 a crisis arose, dyads preferred to avoid involving professionals. 42 Originality: This is the first study using qualitative interviews of dyads to inquire into their 43 experience of mental health crisis. 44 Practical implications: Psychosocial interventions should aim to identify and replace unhelpful 45 strategies used by dyads to manage crisis episodes. 49 50 51

wandering, dysphoria, resistance to care (Backhouse, 2018). Crises can further lead to increased 68 hospital admissions and decreased quality of life (Ledgerd et al., 2016). A crisis is 'a process where 69 there is a stressor(s) that causes an imbalance requiring an immediate decision which leads to a 70 desired outcome, and therefore crisis resolution. If the crisis is not resolved, the cycle continues' 71 (MacNeil et al., 2013). 72 In the UK, professional support for dementia crisis is provided by a range of secondary care teams 73 managing crisis in dementia (TMCDs). TMCDs provide multidisciplinary assessment and a range of 74 short-term care services (telephone consultation, at home or care home visits), that are usually 75 provided for up to 6 weeks. They may then refer the person with dementia to appropriate longer-term 76 mental health teams or other care providers if necessary. Despite this potential to effectively manage 77 critical episodes in people with dementia, few such teams currently exist (Streater et al., 2017). 78 Crisis episodes altering the equilibrium of the family. As the primary providers of care for people 79 with dementia, family carers may experience a sense of burden and decreased mental health. Crises, 80 in turn, may destabilise the support systems of the dyad (Toot et al., 2013). When the demands of 81 care made by the person with dementia do not adequately match with the resources available to the 82 carers (personal resources such as self-efficacy or external support from friends or peers), the carers 83 may experience a series of deleterious consequences for their wellbeing (Au et al., 2010). 84 Fingerman and Bermann (2000) frame this in terms of disruption to the familial system; individuals 85 expect the family system to maintain a certain tone, which is based on a series of expected behaviours. 86 This can be seen as systemic homeostasis within the family. Research on dementia caring reveals that 87 family members tend to favour homeostasis when facing critical changes engendered by dementia 88 rather than challenging each respective role with respect to the situation (Vroomen et al., 2013). 89 Therefore disruption to the family system has ill effects on carers' mental health. 90 Dyadic coping with dementia. Bosco et al. (2018) described how relational processes influence the 91 coping strategies that each member of the care dyad enacts when dealing with adverse events. For 92 example, when the dyad is confronted with a situation (be it challenging or of everyday concerns) the 93 members of the dyad engage in an inter-dependant process, depending on their respective role in the 94 relationship, to appraise and manage the situation. This process of joint coping requires a certain 95 degree of shared responsibility towards the situation (Lyons et al., 1998). It follows that when both 96 the carer and the person with dementia assume co-ownership and responsibility for managing it 97 (dyadic coping style), there is more equal distribution of power and the event is less likely to 98 negatively affect the person with dementia or the carer. Conversely, when only one member or neither 99 engages in coping (individual coping style), and both the person with dementia and the carer maintain 100 differing views on how to manage the situation, it is more likely that the stress posed by the event will 101 have a major impact on them. Research on dyadic coping has pointed to a positive association 102 between perceived quality of the relationship and wellbeing in the person with dementia and the carer; 103 poor relationship quality being associated with increased carer strain and feelings of being unable to 104 cope (Litwin et al., 2014). 105 Narrative studies on dementia caring. Narrative approaches have been used in dementia caring to 106 explore how individuals make sense of their life and the way they negotiate their roles when 107 managing complex situations. These approaches facilitate the structuring of individuals' experiences 108 according to past events and in view of future events (Clark, 2001). In participants with dementia, 109 narrative methodology has been associated with positive experience in terms of reminiscing about 110 past and future events. This approach may also aid the quality of communication within dyads and 111 their sense of identity as a couple (Scherrer et al., 2014). 112 Despite the research cited here into familial roles and identities, little is known about how people with 113 dementia and their carers cope with mental health crises. This project therefore aimed to explore the 114 ways in which the family carer and the person with dementia coped with episodes of mental health 115 crisis across time. We also set out to identify barriers that could influence the likelihood of seeking 116 professional help from TMCD. 117

Study design 119
This study employed a case study design to undertake interviews with the person with dementia and 120 their carer at the same time (dyadic interviews) over the course of one month. We complied with the 121 COREQ checklist for qualitative studies to comprehensively report on study methods for data collection 122 and analysis used, context of the study, and findings (Tong et al., 2007). The study was granted ethical 123 approval from the East Midlands -Derby Research Committee (reference 18/EM/0023).

125
Participatory research as meta-method. 126 The study utilised a community-based participatory approach informed by the tenets of co-operative 127 inquiry into health care practice (Borg et al., 2012). It employed two co-researchers with lived 128 experience of dementia in the delivery of interviews. 129

Study participants 130
The person with dementia was eligible to join the study if s/he: 131 • Was on a caseload of TMCD. 132 • Was 50 years old or above.

133
• Received support from an informal carer.

134
• Had a formal diagnosis of dementia.

135
• Was identified by staff as being cognitively able to participate in the interview.

136
• Was aware of their diagnosis of dementia and willing to talk about dementia.

138
The carer was eligible if s/he: 139 • Was a family carer of a person with dementia (either spouse or child) who met the above 140 inclusion criteria 141 • Was willing to discuss dementia and their experiences of mental health crisis. The co-researchers with dementia were identified during PPI meetings and asked if they were willing 152 to participate in the research study. They were recruited if they were 50 years old or above, had a formal 153 diagnosis of dementia, and were cognitively able to conduct the interview in English. The carer co-154 researcher was a family carer of a person with a formal diagnosis of dementia. No restrictions on age 155 or sex were applied. A total of six co-researchers were identified (two carers and four people with 156 dementia) as eligible to participate. Four agreed to take part in the initial briefing about the study (i.e. 157 to explain the study and the implications of acting as co-researcher) but did not continue because they 158 were unable to commit to it regularly. Hence, only two co-researchers received training in delivering 159 qualitative interviews and helped to analyse data (CB helped with the organisation of themes for the 160 thematic analysis, and one carer DC helped deliver the interviews by prompting participants with further 161 questions around managing crisis care). DC could help deliver interviews for only one dyad as the other 162 dyads preferred not to have more than one interviewer in the room to avoid confusion in the person with 163 dementia. Thus, AB delivered the remaining interviews alone.

165
Training of co-researchers was informed by the literature (Di Lorito et al., 2017) and consisted of two 166 pre-interview and one post-interview sessions. The pre-interview sessions were held on consecutive 167 days, lasted for two hours each, and consisted of role play scenarios of dyadic interviews. In addition, 168 a 20-minute briefing was held on the day before the interview to discuss the process in detail. The follow 169 up session was held post-interview and consisted of a 10-minute de-briefing around what was the 170 purpose of the study and participants' right to withdraw to take part in follow-up interviews.

172
Data collection 173 174 The recruitment of participants took place with the support of TMCDs. The qualitative interviews 175 with dyads were conducted by AB and DC. As the therapeutic input of TMCD lasts up to 6 weeks, 176 up to four interviews were conducted with each dyad. This was to ensure that each dyad was receiving 177 support for crisis from the team at the time the interviews took place. In concert with Legard et al.

178
(2003), we used an individual-oriented perspective for saturation and stopped conducting new 179 interviews with each dyad when no new information was being generated during the interviews. We 180 consulted the TMCD about recruiting new participants. Interviews lasted around 90 minutes for each 181 of the four interview sessions.

183
Findings from a previously published systematic review (Bosco et al., 2018), guided the topics for 184 the interviews. For example, questions around dyadic coping were used during the interviews (e.g.

185
'how do you manage the crisis as a couple'?). We used multiple, semi-structured interviews, as they 186 allowed us to explore how each member of the dyad negotiated their respective role and the role of 187 teams in supporting them over time (e.g. 'compared to last time we met, has any new difficult 188 situation or crisis arisen?', and if so 'how did you manage it?' and 'how did you both feel about that 189 situation?).

191
This line of inquiry enabled a first exploration of the type of coping mechanisms used by each dyad.

192
For example, the response was deemed a dyadic coping style when both members of the dyad showed 193 a shared understanding of the crisis and agreed on how to manage the event. between cases. Hence, the first coding of the interview transcripts was performed by paragraph, and 210 memos (or short comments) were recorded for each code, to clarify the relationship between the code 211 and the relevant piece of text. Memos proved key to the analytic process, enabling more abstract 212 categorisation of the data. The thematic analysis lent the interview data definite structure around 213 themes. Themes were then organised in a diagram to explore how they linked to one another to form 214 a coherent flow for the argument around managing crisis in dementia. This flow helped create a 215 storyline or play through narrative inquiry.

217
Explanatory narrative methods by Kleinman (1980) and Williams (1984) were used to describe how 218 dementia dyads retrospectively constructed crisis events. These were recent memories of crisis 219 occurring during the month the dyad was receiving support from TMCDs. Three features of narrative 220 inquiry: temporality, social aspect and place, constituted the analysis of complex relational dynamics 221 of individuals' stories.

223
We adopted the two-stage narrative analysis advocated by Burnell et al. (2009) by looking at both 224 the content and the form of the narrative. We explored the individual and dyadic coping mechanisms 225 identified by Bodenmann (1995).

227
Results 228 Five dyads were interviewed over one month, a total of 16 interviews. The mean age for carers was 229 67.4 and for people with dementia was 79.8 (Table 1). The interviews lasted on average 46.3 minutes. 230 All the married dyads were living together (married for more than 20 years) but in a different 231 household from their adult children. The daughter-mother dyads lived separately but in the same 232 neighbourhood. 233 Table 1 here. 234 From the analysis of power relations, we found that participants fell into two categories: a dyadic 235 coping style (N=2) or an individual coping style (N=3). The narrative approach led us to structure 236 findings into three 'acts': the setup, confrontation and resolution of crises ( Table 2). As the crisis 237 unfolded, narratives differed according to whether couples used dyadic or individual coping styles. 238 Table 2 here. 239

First act -Setup and antecedent of crises (Past narrative pre-crises) 240
The first act concerns the life-story or 'play' of each dyad, their characteristic roles and relationships. 241 The first act also reports how dyads typically negotiate everyday life in the absence of a crisis. The 242 events that led to the crisis under examination represent the end of the first act. 243 Different stimuli acted as triggers for a mental health crisis in people with dementia. Such stimuli may 244 be caused by the symptoms of dementia or be external to the individual. What seems to be a common 245 factor is the disequilibrium in the person with dementia associated with such stimuli. People with 246 dementia in three dyads reported triggers for crises that were associated with hallucinations. These 247 were described as 'alterations of reality' by one participant. The same person spoke of 'seeing' his 248 wife cheating on him with other men from his past when in fact she was sleeping in the bed next to 249 him. This couple exemplified a dyadic-focused coping style, illustrated by the wife's account of how 250 she felt before the situation escalated to a crisis: 251 'He was very threatening but in my heart I knew he was not going to do any harm. He thinks that I 252 am cheating on him. It is just an hallucination…' (D3S1). 253 When hallucinations are coupled with memory problems, the person may feel disoriented and unable 254 to experience the environment clearly. This was the case for a woman with dementia who reported 255 'seeing' her husband in the room at night when in fact he had died some time ago. As her daughter 256 explained: 257

'My mum would call you at night if she thinks she has seen my dad next to her, she would get scared 258 because he was not an easy man but she is not able to remember that my dad is now gone' (D2S1). 259
An individual-specific coping style characterised this (mother-daughter) dyad, with the mother 260 maintaining her firm belief that that situation, as described by the daughter, was not related to crisis: 261 'That was not a crisis for me, and was not going to be one, you say it was, but again for me it was 262 normal' (D2S1). 263 Triggers may also come from the social environment, such as when the person feels s/he is not being 264 acknowledged. A wife explained that the feeling of 'being ignored by her' upset her husband so much 265 that he would cry in his chair (D1S2). 266 Constant anxiety was a feature of the situations studied. This affected even distant family members 267 who were apprehensive about receiving phone calls concerning a crisis. Anxiety could prove 268 overwhelming for those providing more direct care. This feeling was common to dyads with either 269 coping style. different coping styles, the analysis next developed accounts of how the crises unfold. 280

Second act -Confrontation -Being in crisis (Present narrative at the time of appraisal and 281 response) 282
The second act featured the 'rising action' of the play, a narrative around the way dyads made sense 283 of the crisis and how they attempted to cope with the difficult situation. Sometimes they failed, 284 sometimes they learned new skills to manage the situation. Through confronting the crisis, some 285 dyads developed a sense of awareness of the critical situation and reached a point that may be seen as 286 transformational. This phase of development seemed to be promoted by external support, be it close 287 friends or TMCD. 288 Appraisal. Dyads displayed either a shared appraisal or an individual appraisal of the crisis. Open 289 communication was usually fostered for both appraisals of the event. Dyads with shared 290 understanding of the situation tended to first acknowledge the existence of the crisis and to evaluate 291 together whether or not they had enough resources to cope with the situation through a perspective-292 taking process. In one case, warning signs included clenching teeth and prolonged eye contact 293 (D3S1), or irritability (D1S2). While these signs were not always acknowledged by the person with 294 dementia, he knew that something was not right with his behaviour because the carer would sit next to 295 him to calm him down: 296 'Yes the armchair is the safe place we sit together and then I calm down…I now know when 297 something is wrong because I usually sit with her on here' (D3S1). 298 This illustrates shared appraisal and dyadic coping. Conversely, dyads with individual approaches 299 sometimes differed in their appraisal of whether a given episode was a crisis. In one case, the memory 300 of the participant with dementia was not at issue, rather, that person firmly held to their own 301 experience of the event, 'It was OK for me, full stop. No crisis' (D2S1). This was in direct contrast to 302 the carer's appraisal. 303 The 'second appraisal' of the event, the phase in which dyads evaluate whether or not they have 304 enough resources to manage the situation, is when they are most likely to engage in help seeking 305 behaviour. At this point, dyads evaluated their response options to the event, weighing the advantages 306 and disadvantages. Data from the interviews suggested that the advantages were seen in terms of 307 assistance with problems, while disadvantages consisted of having professionals involved. 308 Calling for help to manage crisis. Although the intensity of the crisis may lead to a more rapid call for 309 help, dyads with either coping style seemed to decide based on when the decision needed to be made. 310 During an episode of crisis, if the call for help needed to be made, dyads balanced the presence of 'too 311 many problems to manage' against the consequences of making such call for help for example, having 312 many professionals in the house, changes to routine and relationships. A carer reported: 313 'If I see the crisis is so intense of course I call the ambulance…but usually…the first thing I do is to 314 think "Okay, he is in crisis now." I definitely plan to call the team tomorrow or the day after. There is 315 no way I call for help the same moment he is in crisis, that would make him even more confused' 316 (D2S2). 317 During an episode of crisis, if the dyads were presented with the scenario of calling for help 'right 318 now' the members of the dyad (usually the carer) had to weigh up the undesirable consequences 319 against the utility factor (help anticipated). The dyad or the carer were more likely to delay calling for 320 help when episodes were more intense and unmanageable: 321 'When I feel agitated, I prefer my wife to help me and not having all 'strangers' (team) here, they 322 would get me more confused than I already am, so if she asks I tell her not to call for help' (D3S1). 323 Dyads seemed to reach a threshold for help-seeking behaviour, and this was dependent upon the level 324 of awareness that either member of the dyad (or both members) had of the crisis. It was usually the 325 responsibility of the carer to call for help even when both members reached full awareness. As 326 recounted by one carer, this was mainly the case when the person with dementia had decreased 327 capacity to call for help (in case of dyad-focused coping style). Similarly, this could happen in dyads 328 with individual-specific coping style if, for instance, the person with dementia felt that the crisis was a 329 family matter and not something to share with strangers, yet the carer because unable to manage the 330 situation and decided to contact the crisis team. 331 Level of support. When the professional support was offered or provided before an episode of crisis 332 arose and information was shared within the couple from the time of diagnosis, dyads seemed to seek 333 help more swiftly when needed. One couple (D1S1) commented on how 'receiving information 334 [support] from a professional right from the beginning takes away so much responsibility' for action 335 and reduced the risk of disappointment because an outsider was involved. This viewpoint was shared 336 by all dyads irrespective of their individual-or dyad-specific appraisal and responsibility. 337 Where the dyad's individual coping style was led by the carer, that person tended to delay the search 338 for support because the person with dementia did not necessarily recognise or acknowledge the seeking professional support in a crisis. 394 In our study, when a crisis was imminent (e.g. in the next few days), carers engaged in help-seeking 395 behaviour. However, when in the midst of a crisis, they chose to postpone professional support until 396 the crisis was even more intense and severe. As reported in the narratives of our participants, this 397 occurred because the detractors (e.g. having strangers in the home, fear of changes in relationship) 398 'won' over utility (i.e. receiving help for crisis). 399

Conclusions 400
This is the first qualitative inquiry into the experience of mental health crisis for people with dementia 401 and their carers. We used two qualitative methods for data analysis based on thematic analysis and 402 narrative inquiry and reliability of coding was high. A limitation is the lack of diversity in the 403 participants who were all middle-class white British in heterosexual marriage, or in daughter-mother 404 relationship. Also not included were other family members providing care, or paid carers. These 405 groups might have added new knowledge and insights into coping with crises and accessing 406 professional help. 407 Our findings suggest that it is timely to develop psychosocial interventions to identify and replace 408 unhelpful coping strategies used by dyads for crisis management. In addition, family carers of people 409 with dementia and health professionals operate in a complex care environment requiring key care 410 decisions to be made in short periods of time, which is immediate for the carer and usually around 6 411 weeks for the health team.