Death and the Oldest Old: Attitudes and Preferences for End-of-Life Care - Qualitative Research within a Population-Based Cohort Study

Introduction Increasing longevity means more people will be dying in very old age, but little is known about the preferences of the ‘oldest old’ regarding their care at the end of life. Aims To understand very old people’s preferences regarding care towards the end of life and attitudes towards dying, to inform policy and practice. Methods Qualitative data collection for n = 42 population-based cohort study participants aged 95–101 (88% women, 42% in long-term-care): topic-guided interviews with n = 33 participants and n = 39 proxy informants, most with both (n = 30: 4 jointly + separate interviews for 26 dyads). Results Death was a part of life: these very old people mainly live day-to-day. Most were ready to die, reflecting their concerns regarding quality of life, being a nuisance, having nothing to live for and having lived long enough. Contrasting views were rare exceptions but voiced firmly. Most were not worried about death itself, but concerned more about the dying process and impacts on those left behind; a peaceful and pain-free death was a common ideal. Attitudes ranged from not wanting to think about death, through accepting its inevitable approach to longing for its release. Preferring to be made comfortable rather than have life-saving treatment if seriously ill, and wishing to avoid hospital, were commonly expressed views. There was little or no future planning, some consciously choosing not to. Uncertainty hampered end-of-life planning even when death was expected soon. Some stressed circumstances, such as severe dependency and others’ likely decision-making roles, would influence choices. Carers found these issues harder to raise but felt they would know their older relatives’ preferences, usually palliative care, although we found two discrepant views. Conclusions This study’s rare data show ≥95-year-olds are willing to discuss dying and end-of-life care but seldom do. Formal documentation of wishes is extremely rare and may not be welcome. Although being “ready to die” and preferring a palliative approach predominated, these preferences cannot be assumed.


Introduction
Living longer means dying older. Demographic changes mean more people are living their last years in very old age. [1] How does this affect their attitudes towards death, dying and care towards the end of life? The experiences and preferences of 'older old' people, or their carers, are rarely heard, but their voices are crucial to shaping end-of-life care services to reflect the priorities of increasingly older people approaching death.
The existing literature on end of life preferences and experiences of older people is limited, and particularly so in relation to very old people. [2;3] Research to date has typically sought bereaved relatives' views, [4][5][6][7][8] though some pioneering studies with older people have been largely in long-term care settings, [9][10][11][12][13] occasionally with just community-dwelling older people, [14][15][16][17][18] but to our knowledge none so very old and none except our cohort [19] taking a population-based approach.
The present study informs policy, planning and practice by exploring thoughts on death and dying and preferences for end-of-life care with a representative sample of very old people-over-95-year-olds, their relatives and formal carers.

Methods
A qualitative interview was added to the regular quantitative survey data collection at the 7 th wave follow-up (Year 21) of a longitudinal study of ageing, the Cambridge City over-75s Cohort (CC75C) study. Full study methods have are described elsewhere [20] (http://www. cc75c.group.cam.ac.uk/). The original population-based sample (n = 2166, 95% response rate) enrolled in 1985/87 using general practice lists have been re-interviewed every few years following Cambridge Research Ethics Committee approval and renewed consent. To avoid underrepresentation of the most frail, proxy informants were also interviewed. Interviews were conducted face-to-face in the interviewee's usual home or occasionally, with proxies, by phone. Mortality was the main source of attrition.
In this Survey 7, following completion of the CC75C interviewer-administered structured questionnaire (http://www.cc75c.group.cam.ac.uk/pages/questionnaires), participants were invited to a further interview "more like a conversation than the usual survey" with the aim of understanding "what it is like to be so old, your experiences of care and your views on some of these issues that are more complex than a survey can explore." A topic guide of open-ended questions (http://www.cc75c.group.cam.ac.uk/pages/additionaldata) explored issues Funding: The authors thank all the past CC75C sponsors for financial support spanning two decades (see http://www.cc75c.group.cam.ac.uk/pages/grant/ default.htm for full list of project grants), particularly the Abbeyfield Society for current funding and the BUPA Foundation for support under their Health and Care of Older People grant. CC75C was a member study of the National Institute for Health Research funded Collaboration for Leadership in Applied Health Research & Care (CLAHRC) for Cambridgeshire and Peterborough. All researchers were independent from funders. No funder sponsor played any role in study design, data collection and analysis, decision to publish or preparation of the manuscript.
participants had raised in past surveys, where quantitative methods had been inadequate to fully explore the depth of responses. When proxy informants were interviewed, they were asked what they believed would be the response from the older person as well as their own views and experiences. Multiple strategies maximised inclusion of these very old and often frail participants in qualitative data collection [see Box 1].
To maximise the value of these data sources, the entirety of this relatively large qualitative archive was included in the analysis. A descriptive thematic analysis was conducted using a framework approach, [21] which facilitates working with large data sets, particularly where data collection is more structured, and where co-researchers share analysis. NVivo software (http:// www.qsrinternational.com/products_nvivo.aspx) was used to manage and index the data prior to charting, mapping and interpretation. The analysis team, trained in framework analysis, consisted of two to three analysts at any one time (four individuals trained in qualitative methods) working independently, in parallel and together at various steps of the analytic process.

Results
Of 54 study participants alive when Survey 7 started, 6 died before they could be interviewed. Of the 48 remaining, 44 (92%) took part in the usual survey. Qualitative data were collected for all but two people. For the majority (n = 30, 71%) both the participant and a proxy informant were interviewed, usually separately (n = 26) or for a few jointly (n = 4). For a small minority (n = 3, 7%) only the participant was interviewed and for one in five (n = 9, 21%) only proxy interviews were possible [see Fig 1]. All but six proxy informants were relatives, most commonly daughters (n = 20) or sons (n = 7); non-family included care-home staff, a live-in careassistant and a friend. Participants were aged 95-101, mainly women (n = 37, 88%); 57% were community-dwelling (n = 20 in private dwellings, half of these living alone, n = 4 alone in sheltered accommodation) [see Table 1 for further demographics]. The full archive for qualitative Box 1. Multiple strategies used to source qualitative data.
Data sources: 1. Transcribed audio-recordings of topic-guided interviews with study participants and proxies: a data usually collected at a second visit following the standard CC75C survey interview b data collected over several short visits for those with sensory or cognitive impairment c mixed method interview for those too frail to manage more than a single visit.  * SD = standard deviation † Missing social class data for 1 respondent (2%) reflect missing data from baseline interview ‡ MMSE = Mini Mental State Examination (score range 0-30), from which cognitive impairment was categorised as severe (0-17), moderate (18)(19)(20)(21) and mild (22)(23)(24)(25) or cognition was rated intact (26)(27)(28)(29)(30). § ADL = Activities of Daily Living, IADL = Instrumental Activities of Daily Living, PADL = Personal Activities of Daily Living.
Disability was defined as needing assistance in any 1 task. **Self-reported Health is in comparison to others of same age.
analysis comprised 112 source documents (1269 pages) including anonymised transcripts from 95.5 hours of audio-recordings from 91 interviews relating to 42 older people. Analyses yielded three main topics: 1) the context, beliefs and outlooks framing the lives of our very old study participants; 2) their attitudes towards dying and death; 3) their preferences concerning end-of-life care. Findings are presented below in sub-themes of each topic with sample comments (all names and identifying details have been changed) [see Box 2 for thematic overview].

1) Context, beliefs & outlooks
Participants vividly described the context in which they were living their later years, a context where contemporaries' deaths heightened awareness of dying, though interviews revealed considerable variation in views on readiness to die or not, medicalisation of dying, euthanasia and life outlook.
1a) Everyone dying. The age of these older people was so great that most of their contemporaries had died. Their circles of friends and family had diminished and death was a regular feature of life, leading some to question why they were still alive. One proxy measured this by the size of her mother's Christmas card list-"down to five now"-and quoted her frequently asking "for at least ten years 'Why am I still here? All my friends have gone.'"(1077-proxy) However, there were positive outlooks. One proxy described her mother as a survivor, acknowledging deaths of others but able to carry on; others celebrated their survival: ". . .they became great buddies. And I do think she was sad when she died, but [. . .] she didn't miss a beat [. . .], she carried on with her life. So I think there's obviously the survival thing there." (3103-proxy) Thus, for many, death and dying framed their outlook on their remaining lives. Some spoke, more negatively, of being on borrowed time: "Maybe she feels she's on borrowed time.[. . .] She's had underlying heart problems over time, so she must have had a sort of taste of her own mortality." (142-proxy) Most knew, and accepted, that they were going to die soon. Their lives had been long and were coming to an end. For some this was only a recent realisation, sometimes triggered by illness. A few indicated that, although towards the end of their life, they did not see the end as imminent: Some were more desperate in their desire to reach the end, sometimes linking this to their longevity, or suggesting they had simply lived too long: "'Please don't let me live 'til I'm a hundred' she said." (148-proxy) Others described having nothing to live for and having had enough: Some had cried about it: "I wake up in the morning and think what the hell am I here for and I'll cry." For others these feelings varied-they wished for death on bad days but on better days the will to live remained.
"Some days she gets a bit fed up or cross. She'll say, 'Oh I wish I wasn't here!' Or something like that [. . .] And another day she's completely the opposite." (3504-proxy) 1c) Medicalisation postponing death. Well-intended medical interventions were viewed negatively as prolonging life by a handful of interviewees, an issue usually, but not exclusively, raised by proxy informants. One care home manager offered the explanation that "in hospitals they do feel [. . .] death is not something they want, so they will over-medicalise because it's better than doing nothing"(142-proxy). She exemplified this with the ethical issues posed by enteral tube feeding decisions: "It's harder to take the thing out than to put one in."(142-proxy) However, relatives questioned the logic of some medical practices: (3403-proxy) One 98-year-old couldn't "see any point in keeping people alive"(2999), confirmed in a separate interview with her niece: "She thinks it's absolute nonsense being kept alive. And that everybody here [care home] is being kept alive. [. . .] she will come out every now and then with [. . .] 'you look after us too well'" (2999-proxy). This was reiterated more strongly by a daughter whose mother thought it "bloody silly keeping people [alive]"(1077-proxy).
1d) Euthanasia. A step on from this, another daughter struggled with her perception that "it seems to be society's mentality that they keep you alive anyhow [. . .] regardless"(2961-proxy). A son with similar doubts described a vivid memory of accompanying his mother to visit one of her friends with dementia following several strokes: "When she had her marbles, she said 'Gordon, [. . .] if I ever get like that, for goodness sake put a. . .', it was her words, not mine, 'put a pillow over my head, will you?'. . . But of course you can't. If she was compos mentis enough to realise [. . .] the state she was in, she would be very upset." (2916-proxy).
Another daughter said her mother would ask 'why don't they put me out of my misery? It's pointless me lying here"(3211-proxy). One relative took a stern line in response to such comments, voicing a key dilemma: Two daughters whose mothers had repeatedly expressed similar wishes considered potential legal repercussions and expressed doubts: "Someone would be assisting, wouldn't they, if they gave her something to end her life?" (1162-proxy) "We've sort of been in sympathy with euthanasia [. . .] [but] when it actually comes to that moment, to one's nearest and dearest, it is actually another factor. I don't know." (3124-proxy).
Another daughter raised assisted dying: "If you offered her to go to Switzerland or Holland or wherever, she'd go. She's been ready [. . .] for a long time."(1077-proxy), but recognised her mother's ability to act on her wishes had diminished as her desire to die increased.
1e) Not ready to die. However, for some the will to live remained strong: 1f) Outlook on remaining life. Many of the older people referred to "taking each day as it comes", expressing thankfulness for where they were in life and content, at this stage, to take life one day at a time, not worrying too much about tomorrow. There was a sense of life ticking along until something drastic happened.
"It is only day-from-day when you get to ninety-seven." (3103) Faith, or a lack of faith, provided a framework for end-of-life care preferences for some.
Interviewer: How do you feel about the future, how do you think things will work out for you?
Participant: Well, being a Christian, I leave it to the Lord, yeah, what he wills.
"I'm an atheist, or an agnostic, so I've no belief in a second. . . in a further life. I think it will be just as it was in the thousands of years before I was born." For some, however, their preferences for how they would like to die outweighed their beliefs: "She used to say 'if I ever got cancer now, forget the chemotherapy and I wish there was something that I could be given to end my life peacefully' [. . .] And yet she was quite a religious woman." (1162-proxy)

2) Attitudes towards dying and death
Thinking about death, talking about death and what might follow death emerged as separate themes in the interviews. 2a) Thoughts about dying. As well as framing their thoughts on their remaining life, the deaths of others and faith perspectives provided a context for thoughts about death and dying. Most were not afraid of dying, either reporting it did not worry them or proxies saying they had not expressed any worries or fears about it: But where the experience of others' deaths had been negative, there was worry-not about death itself, but the process of dying: "[My aunt] found my mother's end quite distressing, because she [pause] was not herself for quite a long time before she died. My mum was in hospital for a long time, and then in the home. [. . .] I think she worries that she's going to be like that when the end comes." (2882-proxy) Worries about death also related to the impact death might have on others. This usually related to the emotional impact on the family or resulting loneliness for a remaining partner, but also to financial concerns.
"The only thing I'm worried about is my sister. I hope that she'll be not sad and be able to come to terms with it." (1502) "[. . .] he got awfully worried about money" (3185-proxy) A few reported thinking about death and dying, suggesting it was a worry, but for most it was something occasionally crossing their minds and not a great concern.
"I suppose it comes across everybody. . . occasionally, but I wouldn't make a mountain out of it" (1502) "Never given it much thought." There were some who stated explicitly that they didn't think about, or preferred not to think about, or discuss, death. A care manager proxy suggested this was born out of fear. Death was welcomed where it was seen as a release. One older person said he was not afraid of death as it meant there was probably something wrong. Others felt death was both inevitable and approaching.
The older people and their proxies seemed to accept this inevitability quite philosophically as beyond their control: "Well, it's no good making a fuss. If you can't alter it, abide by it."(1079) Despite the inevitability of death, its trigger, process and timing remained uncertain. 2b) Talking about death. Proxies reported that death was rarely talked about: "That generation, they didn't actually discuss death much, I don't think, my parents, at all." (1162-proxy) Some described conversations that alluded to death but didn't go as far as explicitly discussing it: Others had no preferences or were content to leave arrangements to their families: "I can quite safely leave it with them to make arrangements." Some had simply not discussed it: Interviewer: Has she talked to you about the funeral? 2d) The manner of dying. The manner of death was of more concern than its imminence. Although some said they had not really thought about dying, many explicitly expressed the wish to die peacefully, pain free and preferably while asleep-to "just slip away quietly."(3194) Indeed death coming suddenly was seen as positive: "Make me comfortable"(2952) was a far more typical response and proxy informants tended to echo the older people's dominant preference for comfort rather than life-saving treatment: "I would say she would just like to be made comfortable." (1540-proxy) The term "life-threatening" in our questioning signified to some that quality of life would be threatened, prompting qualifying responses: "I wouldn't want a thing like these terrible things where people go on living and deteriorating"(2930). Some felt their preference would very much depend on specific circumstances. ". . .whether it could be treatment that I could get about or I had to be bedridden" "I haven't thought about it. It's a decision you can't make unless you're in that position." Relatives were similarly aware of the complexity of the uncertainties. "If you had a bad stroke, mother, and I knew you were gonna be paralysed, and you couldn't think for yourself, and they said to me 'Shall we resuscitate your mother?', I would say 'No.'" (645-proxy) Older people were fatalistic about decisions being made by others.
"Well, whatever's thought necessary." (1433) Some stated that they would leave such decisions to others-relatives, health professionals or even God.
"Her son's been taking the decisions for the last ten, fifteen years perhaps. [ It was relatives who reported never having had any such discussion: Asking about life-saving treatment often elicited views, mostly negative, regarding hospital admission. Only one woman said she had "not given it any thought"(148) and one daughter thought her mother would positively welcome it: "That's where she thinks you get proper care."(1077-proxy). The minority happy to go to hospital qualified this with particular circumstances: "If I had an accident or something like that."(1110). Most were quite adamant about not wanting this-"I should hate it [. . .] I just wouldn't like it."(2916)-but rarely gave reasons, though one care home manager explained: "If she went to [hospital] quite poorly, I think she would be full of anxiety and I think it would exacerbate any illness she had. I think she would find it alarming actually." (3103-proxy) Many older people recognised hospital admission might arise despite preferences against this: "Well, I'd have to go I suppose"(1516). This was sometimes voiced in tones of acceptance, albeit reluctant: Another used the phrase "I wouldn't put her in hospital anymore"(148-proxy), often heard in relation to moving older relatives into care.
3c) Family members' wishes for their relatives. There were clear parallels between what the proxy informants tended to say they thought their older relatives wanted, as described above, and what they wished for on behalf of their older relatives. Very similar themes emerged as their most common hopes-avoiding any suffering, hospital admission or prolonging of life, and achieving a peaceful death at home or in a familiar care home. Relatives echoed the older participants' awareness that, regardless of preferences, decisionmaking may be taken over by others: "I mean, in the end it may ultimately be our decision anyhow, rather than hers" (2961-proxy) Only one proxy informant, not a relative but a care home manager, expressed more general views about the manner and place of death: "To die in the home she regards as home is definitely more dignified and gentler, because she is surrounded by people she knows." (3103-proxy) 3d) Family members' understanding of their older relatives' preferences. Occasionally a relative explicitly conveyed the difficulty of understanding how their elderly parent felt: "I wonder sometimes if she wants to live to 100. [. . .] to say that she's done it, you know." (1079-proxy) A few mentioned views expressed in the past, for example before dementia had limited the ability to discuss preferences: "Knowing Mum's opinion from the past, I said I thought it would be kinder just let her go without this extra effort of resuscitating the heart and so on." (3149-proxy) Far more commonly family and paid carers commented that the older person they cared for had never specifically discussed care preferences with them.
"She knows she's going to die one day. I mean, we all do. But she never really mentions saying that she wants to go here or any of that." (3504-proxy) However, by far the majority felt they knew their relative well enough to predict their views: Most tended to voice what they understood the older person's preferences to be as definitive statements, rarely phrased with any uncertainty: "She doesn't believe in a lot of highfaluting interventions.[. . .] Nature will take its course. And if it is on the cards that any particular infection makes her shuffle off her mortal coil then that is the way it is supposed to be." (3103-proxy) Again these conveyed a clear preference for palliative care, rather than life-saving interventions, and a particular abhorrence of severe dependency: "She wouldn't. . . wish the life support machine to be switched on[. . .] to be resuscitated, because she's been saying now for so long that she wants to go. I mean what is there for her? She's got nothing, has she?[. . .] Her quality of life is zilch really." (1077-proxy) In a couple of noteworthy instances there was marked contrast between what older people said and what their relatives thought would be their views. One daughter said "I don't think she'd really and truly want to be resuscitated"(1110-proxy), whereas her mother was one of the minority in favour of active treatment (see 3a) above). Another daughter thought both her parents would "like treatment to make them comfortable rather than life-saving"(1516+-1523-proxy), though they hadn't discussed it, but her mother in fact wanted to "have treatment as long as [she] could."(1523).
3e) Discussing end-of-life care preferences.
A few examples were mentioned of how older people had discussed, or would like to discuss, their end of life preferences. One woman and her niece wanted the care agency to be "aware of [her] policy decision" that she "would rather be nursed at home"(3154-proxy), and a wife described how her husband raised his concerns with their GP: "When Leonard first came back and heard he had this other cancer Dr Warner came to see him and Leonard said to him 'Two things I want, please, Andrew, is stay at home and if I'm in pain, would you try [. . .] if you could keep me out of pain. And I want you to promise me.' And Andrew Warner was very sweet and he said 'Leonard, I can't promise, but I will do my very best.'" (3185-proxy) These discussions were usually with care professionals, but a care home manager commented that it was not necessarily with older people themselves that professionals broached the topic: "Historically what happens in hospitals is they tend to. . . if it's not asked beforehand they go straight to the relatives. They won't broach it with the individual, which is actually [. . .] against the Data Protection Act, and it's nobody else's business. But it is how the culture of hospitals works unfortunately." (142-proxy) However this care home manager pointed out not all her residents wished to have such discussions: "Each individual person deals with it differently. Some people do feel that they've lived a decent age and they're not too worried about it, and other people have a great fear of death and even in their late nineties would be absolutely terrified. So I think that's an individual thing." (142-proxy) She had found that a variety of approaches could be useful: This experience provided considerable insights into potential pitfalls. The timing of these conversations was felt to be particularly crucial and often problematic, especially given that views may change: "It's nice to elicit a conversation, but it can be quite scary. I could imagine if I said 'If I have a dense stroke tomorrow, leave me', but then if I had the dense stroke tomorrow I may decide. . . you know, it is a worry." (142-proxy) Within families it appeared there was very rarely any specific discussion: "We haven't discussed it in so many words, but by the way it's come into conversation here and there. One relative commented that the study interview was making her think it was something she needed to bring up. She felt unsure about how to raise it and who in the family might best do this: (1150-proxy). The only example of when specific wishes had been documented was in a care home where: " However, these recorded wishes had then been overlooked by agency nursing staff who called an ambulance when concerned about the resident's condition one night.

Discussion, Implications and Conclusions
We sought thoughts on death and dying and preferences for end-of-life care with a representative sample of very old people, their relatives and formal carers. This rare dataset of qualitative interviews described 95-year-olds' lives as contextualised by death, with the majority wondering why they were still there and a minority celebrating their survival. Death was now a part of life for these very old people who were mainly living day-to-day. Most were ready to die, reflecting their concerns for quality of life, not being a nuisance, having nothing to live for and feelings of having lived long enough. Contrasting views were rare exceptions but voiced firmly. Unwanted medical prolongation of life was an issue for some and mention of euthanasia related to loss of quality of life, usually raised by proxy informants for whom legal implications were a concern. Most were not worried about death itself. Concerns, coloured by experience of others' dying, were more about impact on those left behind and the dying process; a peaceful and pain-free death, preferably during sleep, was a common ideal. The wide variation in attitudes ranged from not wanting to think about death, through accepting it as inevitable and approaching, to welcoming or even longing for its release. Preferring to be made comfortable rather than have life-saving treatment if seriously ill, and wishing to avoid going into hospital, were both commonly expressed views, again with a few firm exceptions. Some stressed choices would be affected by circumstances, such as likelihood of severe dependency. They fatalistically accepted decisions might well be taken by others, family and healthcare professionals, and trust in decision-makers was important. Some felt their destiny lay with God or fate. Uncertainty hampered end-of-life planning even when death was expected soon. There was little or no planning for the future, some consciously choosing not to. Written documentation of wishes was virtually unheard of, except sometimes for funerals or wills; the only mention of living wills was in relation to a reluctance to prepare one, and the only example we came across of an advance care plan, in nursing home notes, had been ignored. Despite a notable willingness to discuss end-of-life preferences in most interviews, and death being so ubiquitous, previously having talked openly about death was uncommon, often only alluded to or couched in humour. Not all the older people wanted to have such discussions; for the minority who had, these explicit discussions had usually been with care professionals. Families and care professionals found these issues hard to raise, generally saying preferences had never been discussed. They felt they knew their older relatives well enough to predict their preferences, usually palliative care, mirroring what most older participants wanted, although we found two discrepant views.
Rarely are the views of the very old heard. Our sizeable qualitative dataset uniquely evidences the perspectives of a representative sample of over-95-year-olds. In contrast to much qualitative research that aims to capture diversity through purposive sampling, these were population-based cohort participants and moreover were well-characterised by the mixed methods CC75C study's quantitative epidemiology. Although a single-centre study, sample characteristics are in line with national data available for this age-group and other developed countries. Non-response potentially limits research validity, but 79% of the surviving cohort were interviewed in person and proxy interviews maximised participation to 92%, ensuring people across the cognitive and health spectrum were included. Proxy informants frequently quoted participants themselves, but there are inevitable uncertainties regarding the validity of perspectives heard second-hand. Reliance on informants to represent older relatives' preferences, as we found most could, is not only a research issue; the occasional discrepant views we found have implications for practice, highlighted in other studies too. [22][23][24] However, including these reported views, despite the limitations of proxy reporting, ensured we captured some of the voices of those least able to express themselves. A further limitation was the study's inability to explore how very old peoples' views might change over time, particularly as they neared death, a question of importance given their unpredictable circumstances that merits further research.
Our study adds to the few studies to date that have sought the views of older people themselves, [9-19;24] rather than just through relatives. [4][5][6][7][8] Although previous research generally interviewed younger old people, we found overlapping themes, also highlighted in the few relevant reviews.[2;3;10] Heterogeneity of views is unsurprising,[2;3;25;26] but some common themes are striking: even in their late 90s or beyond our participants generally conveyed a 'living for today' approach previously reported [2;12-14;27] and comfortable acceptance of death. [2;3;26] Others too have found a readiness to talk about dying and end-of-life care, [2;3;11;16] though this rarely happened, [2;12] and willingness to make plans was less clear-cut [3;16;28]: one UK study found people living alone in their 80s "wanting to prepare for and to have a choice with regard to where and when they might die", [15] but a US study found housebound older people were reluctant to plan for serious illness or dying, especially resisting decisionmaking on hypothetical future dilemmas, preferring to wait till they "cross that bridge". [14] Other studies of people with dementia and their family carers have also reported the perceived difficulty of making formal plans for unknown futures. [24;27] Other researchers found, as we did, a minority who preferred not to discuss dying at all [3;14] and an expectation by some that others might make decisions when necessary. [3;11;12] As in our study, generally people were more likely to plan for death-funerals or wills-than dying, [2;11;14] although only a minority worried about death itself, most were more likely worried about the manner of dying. [15;19] Our study found that an over-whelming majority of our very old sample had strong preferences for palliative care-"being made comfortable"-versus "life-saving treatment". Others have also reported a majority preferring not to have interventions [11] and found preferences for life-sustaining treatment declined with decreasing physical or psychological functioning. [29] Some questioned the value of intervention: "Why are we being kept alive?" [18] Although only a minority wanted life-saving treatment, these were strong preferences but had not been communicated. Although current policy encourages early discussion and documentation of endof-life care preferences, [30][31][32] our study and others illustrate how multiple barriers to advance care planning [3;8;24;33;34] apply also to other relevant options rarely discussed, such as moving into long-term care, [35] not least how unpredictability and changing contexts may change choices. [9;11;14;23;29;36-40] Others have given thoughtful consideration to the need to recognise that more dying people are following longer frailty trajectories and the implications for care services. [6;16;26;41;42] Previous research has highlighted challenges of improving communication about preferences [7;25;26;34;40;41;43-46]; we noted with interest other researchers' comments that participants welcomed interviews as an opportunity to express thoughts, [11] as did many of ours. We found questions on specific topics could open up dialogue with people hesitant to discuss dying in general. Interviewees themselves suggested that "being surrounded by death" provided opportunities to start conversations, and both family and professional carers admitted being hesitant to broach the matter, at times leaving an older person with concerns and information needs, for instance pain relief, that no-one addressed.
We and others have described the complex transitions often experienced by very old people approaching the end of life, [38;47-52] and there is widespread recognition that transfers between care settings may be detrimental. [42;53-55] Most of our participants viewed the prospect of hospital negatively, but supporting older people to achieve current preferences in changing circumstances is a major challenge for research and practice. Prognostication uncertainties in frailty trajectories make determining when hospital admission is 'inappropriate' challenging, and too much 'admissions avoidance' risks denying very old people diagnosis, treatment and care that may be entirely appropriate. [38;42;56] Half of UK deaths aged 85 are in hospitals and a third in care homes [53] so improving end-of-life care for very old people in all settings is a priority. [7;30;41;42;54;57-60] Prospective population-based research with older old people, and their informal and formal carers, examining determinants of staying-in-place or transitions is needed to understand what would enable older people to die in the place and with the support they would choose.