Do the elderly have a voice? Advance care planning discussions with frail and older individuals: a systematic literature review and narrative synthesis

Background Recent years have seen marked improvements in end-of-life care, however concerns have been expressed that services are focused on the needs of patients with cancer. This review focuses on conversations about end-of-life care with frail and older people who have no main overriding diagnosis who are estimated to account for around 40% of deaths. Aim To investigate the attitudes of the public and healthcare professionals to advance care planning discussions with frail and older people. Design and setting Systematic literature review and narrative synthesis. Method Articles that related to frail or older individuals and either advance care plans or discussions on end-of-life care were included. Studies of specific conditions or that focused on prognosis, capacity, or resuscitation decisions were excluded. Results While a significant minority of frail older individuals would find them unwelcome, the majority would appreciate the chance to discuss end-of-life care, yet most do not have this opportunity. Attitudes to the timing of these discussions were variable, but most perceived the risk of leaving them too late. Most doctors believed it was their professional responsibility to initiate discussions, but felt limited by time pressures and the absence of a precipitating event. A wide range of barriers were identified including the reluctance of family members to discuss end-of-life care, the passive expectation that someone else would decide on an individual’s behalf, and significant uncertainty concerning future illness and decline. Conclusion The marked disparity between the majority of older individuals who would like the opportunity to discuss their end-of-life care and the minority that currently have this opportunity raises important questions if the wishes of this large group in society are to be respected. The challenge is to find effective ways of encouraging dialogue and choice within the constraints of the current healthcare systems and personal circumstances.


Aim
To investigate the attitudes of the public and healthcare professionals to advance care planning discussions with frail and older people.

design and setting
Systematic literature review and narrative synthesis.

Method
Articles that related to frail or older individuals and either advance care plans or discussions on end-of-life care were included. Studies of specific conditions or that focused on prognosis, capacity, or resuscitation decisions were excluded.

results
While a significant minority of frail older individuals would find them unwelcome, the majority would appreciate the chance to discuss end-of-life care, yet most do not have this opportunity. Attitudes to the timing of these discussions were variable, but most perceived the risk of leaving them too late. Most doctors believed it was their professional responsibility to initiate discussions, but felt limited by time pressures and the absence of a precipitating event. A wide range of barriers were identified including the reluctance of family members to discuss end-of-life care, the passive expectation that someone else would decide on an individual's behalf, and significant uncertainty concerning future illness and decline.

conclusion
The marked disparity between the majority of older individuals who would like the opportunity to discuss their end-of-life care and the minority that currently have this opportunity raises important questions if the wishes of this large group in society are to be respected. The challenge is to find effective ways of encouraging dialogue and choice within the constraints of the current healthcare systems and personal circumstances. 2. What are individuals' attitudes to discussions?
3. What are individuals' preferences to timing of discussions?
4. What are healthcare professionals' attitudes to discussions? 5. What are healthcare professionals' attitudes to timing of discussions? 6. What are the barriers to and facilitators of discussions?
An electronic literature search of Medline, CINAHL, PsychINFO, and ASSIA databases from January 1991 to September 2012 was undertaken to cover published research in health and social science. The challenge of developing appropriately sensitive and specific search terms for 'frail elderly' with no overriding medical condition was supported by an information officer. The second stage of the search sought articles that either included terms for 'advance care planning' or that mentioned words synonymous with both 'end of life' and 'conversations' or 'discussions'. Box 1 outlines this search strategy used for the Medline database. Appendix 1 details all search terms used for each database.
Exclusion criteria included studies of participants with specific conditions such as cancer, heart failure, chronic obstructive pulmonary disease (COPD), and dementia. Although many of the frail and older individuals will have such conditions, studies of those with a single overriding diagnosis have been excluded. Also excluded were studies that focused on prognosis, capacity or resuscitation decisions and studies not originally published in English. Appendix 2 details the full list of exclusion criteria. The electronic database search generated 12 694 titles which were screened twice to identify potentially relevant papers. One hundred and eighty-six abstracts were reviewed independently. The review protocol was applied and agreement reached on 30 papers to be read in full, seven of which were excluded. A citation search of the 23 included papers identified a further three for inclusion: the final 26 articles were from 20 publications, five from Archives of Internal Medicine: no other journal had more than two included papers. The flow chart at Figure 1 shows the distillation to 26 articles. Although there were no geographical criteria, all of the included papers are from US or UK studies.
Data relevant to the review questions was then extracted from these 26 included papers into a study-specific data extraction sheet. Initially double data extraction was undertaken independently on six papers, and standardisation of analysis procedures was ensured. The remaining data extraction was undertaken by the lead author, who led work on the narrative synthesis of the data from each paper in discussion with the review team.

how this fits in
This is the first known systematic literature review to look at the attitudes of the public and healthcare professionals to advance care planning discussions with frail and older people towards the end of their life. It found that although a significant minority would find end-of-life care conversations unwelcome, a majority of this growing population would appreciate the chance for such a discussion with healthcare professionals but only a minority have the opportunity. This is despite doctors seeing these conversations as part of their professional responsibilities. The review identified barriers to end-of-life care conversations with frail older patients that were not found in studies of other populations including the reluctance of family members to discuss end-oflife care, the passive expectation that others would decide on their behalf, and the significant uncertainty concerning future illness and decline were particular barriers in frail older individuals. The paper discusses the issues connected with healthcare systems, individual autonomy and personal circumstances that will need to be addressed if the care wishes of this important group in society are to be respected. OR AND discuss or discussions or conversation* or exp decision making/ or exp treatment refusal/ Each included paper was weighted using Gough's 'Weight of evidence criteria'. 9 This includes an attempt to assess the risk of bias within individual studies. The weightings of each paper are shown in the final column of Appendix 3, with the overall weighting given for each study in bold.

reSultS
The results of the systematic review are analysed by research question with the number of papers addressing each question given in brackets (n). Appendix 3 presents a full list of the papers included in this review, including citation, sample, key findings, and the weighting given to each paper.
Are end-of-life care discussions being held? (n = 16) Seven papers found between 2% and 29% of frail older people had discussed some form of end-of-life care plans with a healthcare professional. [10][11][12][13][14][15][16] The highest proportion was from a study of 600 people admitted to US nursing homes between 1990 and 1994: before the introduction of the Patient Self Determination Act (PSDA) 20% had a discussion of future treatment wishes documented in their notes: after the PSDA this rose to 38%. 10 The disparity between conversations being held with family members and with healthcare professionals is marked. 11,15 In one study 28% (24/86) of community dwelling senior citizens had discussed the terminal care they wanted with a family member but only 2% (2/86) had discussed it with their physician. 11 This was the lowest reported rate of discussions with healthcare professionals and in each case was patient initiated.
A further five papers found between 15% and 66% 17-21 of older people had end-of-life care plans documented in their records. Meanwhile between 40% and 79% of doctors report having discussed end-of-life care preferences with older patients. 22,23,24,25 What are older individuals' attitudes to end-of-life care discussions? (n = 15) The majority of papers reported between 61% 19 and 91% 24 of older individuals wanted to discuss their endof-life care. 11,12,14,16,19,23,24,26,[27][28][29][30] Some expressed comfort even enthusiasm 14 for such conversations. In contrast, three studies 18,21,31 found a reluctance to have such discussions: including housebound individuals who preferred to live 'one day at a time' 21 or people over 50 years who preferred to postpone making plans until they were older or in worse health. 31 Older people saw the benefits of discussions to include assurance that their wishes would be respected, 30 the opportunity to address important issues of care and treatment before becoming cognitively impaired 26 or physically seriously unwell, 24 and to assist loved ones in making decisions. 30 They saw the responsibility of initiating discussions to lie with doctors 12,28 who they want to talk in an honest and straightforward manner. 29 What are older individual's preferences for timing of end-of-life care discussions? (n = 9) Most older individuals wanted discussions sooner rather than later: 12,14,26,28,29,32 perceiving the risk of 'leaving it too late' 26  though time and information was needed to make decisions. 32 Some felt that discussions should happen routinely: 'Advance care planning discussions should be routine questions such as screenings like mammograms and colonoscopies. When somebody is X years old, discussions should begin. ' 14 In contrast, three papers reported older individuals would rather defer discussions, preferring to 'cross that bridge' only when they had to, 21 when the onset of a debilitating or terminal illness precipitated the need to make plans. 31,33 What are healthcare professionals' attitudes to end-of-life care discussions? (n = 4) Most doctors felt these discussions to be an important part of their professional responsibility, 23,29 enabling preferences to be well informed and decisions made in the patient's best interest. 20,23 Many view discussions to be important with patients who have severe chronic illness (91%) or terminal illness (97%), but fewer (64%) felt these conversations were important with older patients regardless of their health status. 23 While some doctors do not find these conversations stressful, 23 others comment on the difficulty with frail older people with multiple comorbidities rather than a clear terminal diagnosis, given the uncertainty over future decline and prognosis. 25 What are healthcare professionals' attitudes to the timing of end-of-life care discussions? (n = 4) There was considerable diversity of opinion. Some thought that discussions should start early, before the onset of serious problems. 28,20 Others describe the lack of a clear threshold event, such as a diagnosis, to prompt discussions leaving them to rely on physical or social cues. 25 While acknowledging their responsibility to initiate discussions, many feared that early discussions may damage the hope that older people bring to the patient-physician relationship. 29

What are the barriers to and facilitators of end-of-life care discussions?
A number of themes emerged from the literature: Families (n = 10). The most frequently identified barrier to discussions are the families of older frail people. It was felt they were sometimes unwilling to have discussions, to accept that their relative is near the end of their life or wish to protect their loved one from upsetting conversations. 14,16,20,26,27,34,35 Breakdown in family relationships and lack of close family were further obstacles identified. 17,31,33 Professional and time limitations (n = 9). Concerns over healthcare professionals' proficiency and willingness for end-oflife discussions 20,27,29,35 and perceived lack of continuity of care and support 23,31 are identified as barriers. Some physicians describe being uncomfortable with the 'paradox of promoting health and discussing its inevitable failure'. 29 Health professionals also reported the pressure to see a large number of patients and difficulty of scheduling timely follow-up visits conflicts with the time needed for these conversations and so greatly reduced their ability to hold them. 14,22,23,25,27 Patient reluctance to discuss (n = 8), feeling 'others' would decide (n = 4). Older frail individuals were found to sometimes be unwilling to discuss their end-of-life care 17,20,21,24,25,27,31,33 not wanting to talk about such 'upsetting' 21 and 'negative' 17 issues, not feeling 'ready to do it', 21 or wanting to put off discussions to a time 'if I ever have a terminal illness'. 33 They sometimes saw end-of-life care discussions as the responsibility of others, commonly family members. 26,33 Some reported feeling content to leave such matters 'in God's hands', 18 or that 'my doctor will decide for me'. 18 Difficulty planning for uncertain future (n = 5). Dementia/lack of capacity (n = 4). The problems of unforeseen medical scenarios and the difficulty of making well-informed decisions before illness occurs were felt to inhibit end-of-life care planning. 16,20,21,26,33 While cognitive impairment and a lack of decision making capacity were felt to be important barriers to planning. 20,27,31,35 The onset of dementia was identified as a prompt for early planning. 31 Administrative barriers (n = 4). A lack of information, inadequate time to consider decisions and the legalistic paperwork involved in completing advance care plans were all felt to be off-putting. 16,17,29,32 dIScuSSIon Summary Important key themes emerge from this review. A minority of frail and older individuals had end-of-life care conversations with a healthcare professional. Most would welcome the opportunity for such discussions, although a significant minority would find them unwelcome. The preferences for timing are highly variable. The few studies that have investigated healthcare professionals' attitudes report that doctors see these conversations as part of their professional responsibility, although workload pressures and uncertainty over prognosis inhibited healthcare professionals initiating these discussions.
This review identified important barriers to end-of-life care conversations with the frail and older individuals that were not found in previous reviews. The reluctance of family members to discuss end-of-life care was the most common barrier while the passive expectation that family members, God, or others would decide on their behalf and the significant uncertainly concerning future illness and decline were particular barriers in frail older people.

Strengths and limitations
Frail and older individuals, unlike groups with a single main diagnosis, are a diffuse group that are challenging to target in a literature search. Low specificity generated a large number of titles to screen: good sensitivity is evidenced by the few additional papers identified by citation searching. The explicit exclusion of older people with a single main diagnosis such as cancer or heart failure, and the focus on those with the multi-comorbidity frailty of old age represents a subset of the older population approaching the end of life: a subset of increasing prevalence as the population ages. Study quality was variable: most were small scale qualitative studies whose primary focus was other than that of this review. The 'weight of evidence' scoring 9 of all included papers was thus either low or medium: no paper achieved a high rating. There is a risk of publication and selective reporting bias by parties with a policy agenda, there wasn't the capacity to undertake a grey literature search to help eliminate this risk.

Implications for research and practice
The frail and elderly are an important and growing population with challenging care needs. That most would like the opportunity to discuss their end-of-life care but few currently have this opportunity is a marked disparity that raises important questions if the wishes of this large group in society are to be respected.
The issues raised by this review would benefit from further larger-scale studies and in settings other than the US and UK. It would seem reasonable to consider how more can be done to understand the advance care wishes of the frail and older individuals.
The challenge for policy makers and healthcare practitioners is to find effective ways of encouraging dialogue and choice within the constraints of the current healthcare systems and individuals' circumstances.
Healthcare system issues. Healthcare systems across the world have different approaches to how to respond to the needs of the frail and older individuals but the significant pressure on professionals' time is common. It is therefore not difficult to understand how advance care planning conversations are overlooked, especially when the future is uncertain, there's no precipitating event and some individuals may find these discussions unwelcome. Healthcare professionals are likely to need support and encouragement to find appropriate opportunities to initiate these discussions. Moreover, increasing expectations of choice among the frail and older individuals, although not necessarily increasing cost, could well have implications for healthcare resource allocations.
• Should professionals have more support and encouragement to initiate these conversations?
• Is there an optimum time to hold these discussions? At a particular age? Or health event (for example, hospital admission or new diagnosis)?
• What are the potential cost and resource implications of more end-of-life discussions?
Individual autonomy issues. As well as professionals initiating them, advance care conversations could be promoted as a right, which individuals are encouraged to seek for themselves when they are in good health. Where advanced care planning issues are raised individuals need information and support to help them make informed decisions and these need to be documented and communicated so an individual's wishes are respected. However, it is important that individuals feel able to refuse such a conversation and don't feel obliged to make a particular decision or infer that they will receive lower levels of care.
• How can it be ensured that individuals have all the information and time they

ethical approval
Ethical approval was not required for this study.

Provenance
Freely submitted; externally peer reviewed.

competing interests
The funding for this research has been stated above. This research received no other specific grants from any funding agency in the public, commercial or not-for-profit sectors. • To what extent should an individual's family be involved in these discussions?
• How can discussions reflect medical uncertainty or a changing prognosis?
• Should a diagnosis of dementia be a prompt for discussions?
Given the findings of this review it would be difficult to argue that more shouldn't be done to understand and promote the wishes of the frail and older people towards the end of their life. In the context of increased pressure on time and resources, policy makers and healthcare practitioners will need to consider the healthcare and individual circumstance issues raised by this review if the personal autonomy of this important group in society is to be promoted. To study the prevalence of specific barriers that prevent home-bound older adults from obtaining advance directives.

Interviews
• 66.2% had an advance directive • Of the 530 who didn't almost 60% of the barriers consisted of the answer "never thought of it", "chose not to execute" or "do not want to think about a negative topic".

A) Medium B) Medium C) Low
Overall: Medium Gordon 13 Arch Overall assessment of study contribution to answering review question (low, medium or high) Criteria 1 involved an attempt to assess the risk of bias within individual studies. The weightings of each paper are shown in the final column of Table 1 with the weighting given for overall assessment of study contribution (criterion 4) in bold.