Eight systematic reviews were included. Of this total, three systematic reviews predominantly focused on palliative care of older adults in community, acute or aged care settings. [2,3] Each of the three reviews focused on end-of-life (EoL) communication for older adults either without a definitive diagnosis or prognosis [1,2] or where a diagnosis, such as chronic heart failure, makes a clear prognosis hard to predict. 
A further 4 reviews [4-8] were included which primarily focused on palliative care and communication in adults, not specifically older adults or in an aged care setting. These provide additional context to the synthesis.
Overall the quality of the evidence was mixed mainly due to lack of methodology or quality reporting in the papers. In general, the quality was adequate with two reviews found to be of high quality. [1,5]
Two systematic reviews [1,2] discussed whether end-of-life discussions are being held. A review by Sharp et al.  found that <29% of frail older adults were engaging in discussion around end-of-life planning with a care professional. Similarly the review by Barclay et al.  suggests that the majority of patients do not perceive they have discussed end-of-life care with their healthcare professionals.
Despite these low numbers, the review found that 61 – 91% older adults wanted to discuss their end-of-life care not wanting to ‘leave it too late’.  Adults 50yrs and over were less likely to wish to engage in discussions believing they could wait until they were older or in poorer health.  This is at conflict with findings from de Graaff et al.  who found the younger adults were more inclined to seek out information than their older peers.
The systematic review by Garland et al.  also found that older adults were less likely to ask questions and often deferred decision-making while also waiting for health care professionals to raise the end-of-life discussion. Whereas many patients in the studies wished to have information regarding their condition, not all patients wanted full disclosure and in particular wished to avoid prognostic information. 
In the review by Sharp et al.,  four papers noted that 91-97% health professionals identified the importance of end-of-life discussion and care planning but <64% were as confident of the need for these conversations with older adults, particularly those without a certain prognosis. It is proposed that discussion is more likely to occur with older adults diagnosed with a malignancy, where the dying trajectory is clearer than for those with no overriding diagnosis.  Another concern raised by doctors was that initiating discussion with older adults may damage the hope that this cohort bring to the patient-doctor relationship. [1,3]
The review by Walczak et al.  focusses its attention on interventions aimed at improving end-of-life communication and education. Findings in this review indicate research with stronger methodology is needed for it to be clear what works best and for whom. This review provides evidence that end-of-life communication, for older people, is linked to more comfort with end-of-life planning and greater action in decision-making and documentation of decisions.  This review also highlights the challenge of communicating with family and carers as their need for and readiness for information may be different to that of the person receiving palliative care. 
Barriers to initiating end-of-life discussions are proposed by Sharp et al.  Family are often reluctant to accept that their relative is near the end-of-life or feel protective and do not wanting to engage in upsetting conversations. Similarly, the patient may be reluctant to engage in discussions, preferring to leave decisions to their family or medical team. [1,3] In the case of dementia these discussions do not often occur before a person has reduced decision-making capacity.
Barriers to timely communication
Inadequate understanding of a condition can mean that people are ignorant of likely complications or trajectory of the disease. Educating patients and families, and linking them in with resources for self-education can initiate discussion of end-of-life management and generate questions.  It is important to recognise that not all doctors are confident in their communication skills when discussing these topics and often resort to euphemisms or excessive use of medical jargon, which can mean the nature of what is discussed is not explicit or well understood. 
Health professionals are concerned they don’t have adequate time to have such conversations [1,3] and may even assume other health professionals have had necessary conversations re- EOL.  Health professionals also cited the uncertainty of diagnosis and prognosis as a barrier to timely conversations.  This was particularly highlighted in reviews by Stephen et al.  and Garland et al.  who recognised that the fluctuating trajectories of non-malignant diseases made it challenging to recognise when EoL discussions were merited.
Fear of causing distress to the patient and their family is also cited as a barrier to initiating discussion around end-of-life, with concern that hope is removed from the patient-health professional relationship. The systematic review by Stephen et al.  notes a randomised controlled trial (RCT) whereby an active end-of-life intervention resulted in improved functional status, preparation for end-of-life and reduced anxiety and depression. This is further dispelled as a misplaced concern, as several studies cite themes of most patients preferring to have the discussions rather than information being withheld. [4,5]
Conflicting information exists around when the best time for end-of-life discussions should take place. Waiting for the patient to initiate conversations is often preferred by clinicians but in practice patients rarely do so. [2,3,8]
Timing of end-of-life discussions
In the paper by Garland et al.,  studies found that patient understanding of their condition or disease influenced how early EoL conversations were initiated. People with an understanding of the life-limiting nature of their condition were more likely to want to put affairs in order early in the disease process.  Limited studies have been conducted on when the best time to discuss end-of-life options is. The literature suggests it is more likely to occur when a person is unwell and in hospital which is the least likely time when information can be processed. [1,2,8] Under these circumstances, the favourable option is that a health professional initiates the conversation and continues at follow up appointments. [2,4]
In the systematic review by Stephen et al.,  the style and content of end-of-life discussions was deemed one of three key themes determined from 14 qualitative and quantitative studies. Successful discussions were those where patients felt listened to and where questions were encouraged. Noted was a split in consensus as to the detail of prognosis, [3,5] and whether discussion should include the withdrawal of aggressive life-sustaining treatments  and information on the changes the patient might expect in daily living. 
Content of end-of-life discussions
Consideration of multicultural contexts of palliative care is an important facet of end-of-life care.  In the systematic review by de Graaff et al.,  37 of the 57 studies looked at communication in end-of-life care in Turkish and Moroccan immigrants in the Netherlands. Five themes emerged: communication of diagnosis and prognosis; communication about pain, sorrow and mental health problems; language barriers; and communication patterns within the family. Prognosis or diagnosis was not discussed with patients up to 63% of the time, mainly due to the doctor or family not wanting to distress the patient, sentiments similar to those expressed in reviews by Murray et al.  and Stephen et al.  A higher socio-economic status and education level, or the type of illness i.e. terminal malignancy, increased the likelihood of these discussions.  These older adults were less likely to want information on their disease whereas younger adults were more likely to ask questions and encourage discussions. In addition, the relatives did not like the direct nature of delivery of these discussions from the Dutch care providers.
Overcoming cultural or language barriers
Cultural values and beliefs are to be included in end-of-life decisions. De Graaff et al.  found that Turkish patients were unlikely to complain about pain and did not want to discuss psychological symptoms such as depression or dementia. Interpreters may not always be available. Relatives acting as interpreters may cloud decision-making, particularly if young children take this role. [6,7] A recent systematic review by Silva et al.  discussed a significant improvement in quality of care when a professional translator was used for end-of-life patients with malignancy. Less errors and improved understanding were noted for both parties. This study also raises the need for professional translators to receive some support and education around delivery of end-of-life information and emotionally-charged discussion as many felt uncomfortable relaying sensitive information.
Health professionals may lack confidence in initiating end-of-life discussions particularly where prognosis is not known, for fear of distressing the person and their relatives. The majority of older adults are open to discussion around end-of-life and want to know what to expect as their health deteriorates. They may need to have these discussions with a person they trust and feel comfortable with and perhaps seek these discussions over more than one occasion. People at a social disadvantage may be less likely to want to engage in end-of-life discussions. If needed, a neutral interpreter is recommended where possible. Conversations in all scenarios, should take place with ample time to invite questions, and the opportunity to summarise what was understood. Regular follow-up meetings are recommended as the health status changes, as terminal phase begins and if treatment is to be discontinued.
Page updated 24 May 2017