Communication in Palliative Care and About End of Life: A State-of-the-Art Literature Review of Conversation-Analytic Research in Healthcare

ABSTRACT

I review conversation analytic research on healthcare interactions in palliative care, and end-of-life preferences and plans conversations in other settings. The review process drew on established systematic review methods. Twenty-two publications were included. Key themes were initiating and managing the topics of dying, of prognosis, and of advance care planning, and interactionally managing emotions. There is substantial, cumulative evidence about patterns and practices for initiating sensitive conversations, managing emotions, and indirectly referring to death. Two of 22 studies examined interactional consequences of companions accompanying patients, and two examined pain assessment. Current evidence is restricted: 21 of 22 studies were in secondary/tertiary care and all were in high-income countries, and most involved specialist physicians. Nevertheless, findings contribute to conversation analytic scholarship on delicacy, emotion in interaction, and indirect reference. Healthcare applications include contributions to training in communicating about sensitive topics and to policy on talking about dying. Data presented are in British English, U.S. English, Canadian English, Swiss-French, and Japanese.

View correction statement:

Correction

Overview and background

This state-of-the-art review examines published findings from conversation analytic research on palliative and end of life (EoL) healthcare interactions. The studies reviewed examine recurrent difficulties entailed in these high-stakes, important conversations—conversations that are often daunting for all parties involved. Most of this work aims to describe and understand interactional means by which people limit, circumvent, or overcome these difficulties, and thereby achieve palliative care and EoL conversations of substance and sensitivity. A small number of the studies examine ways in which healthcare professionals (HCPs) can exacerbate the difficulties entailed.

My focus is on interactions in which people approach and address current and future states of affairs and events that are arguably the most painful, distressing, frightening, and—in a sense—unknowable matters that humans deal with in our lives. Besides being emotionally charged, other interactional challenges arise. These include uncertainties about at least three matters: how long someone will live, the precise nature and course of future symptoms, and the availability and effectiveness of EoL care for the patient (for instance, whether sufficient support will be available to allow someone to die at home if that is his or her wish). No matter how expert the HCP, these cannot be predicted with a high degree of accuracy. Another challenge is that, in many contexts, people tend to avoid, or rapidly close off, negative and distressing topics including dying and death, and instead turn to or stick to contrasting orientations by focusing on positive aspects of the situation with “bright side sequences” (Holt, 1993), or turning to concrete actions and efforts (see Anderson et al., 2020)

Current healthcare policies and guidance recommend that HCPs engage patients and their informal carers in conversations about EoL and associated care plans (e.g., Australian Government Department for Health and Aged Care, 2020; National Palliative and End of Life Care Partnership, 2021). The kinds of matters covered in conversations about care plans and decisions include a wide range of matters, including cardiopulmonary resuscitation; what kinds of treatment one would want toward the end of life; preferred place of dying; who one would want there; and condition-specific decisions, such as whether someone with a neurodegenerative condition would want to have ventilation and when one would want it to stop. Unfortunately, absent or very poor communication on these matters is frequently reported (e.g., National Palliative and End of Life Care Partnership, 2021), with both professionals and patients reporting a lack of skills for these conversations (Almack et al., 2012). Fortunately, the CA approach can produce detailed description and insights into the difficulties of, and effective practices for, engaging people in these conversations, with huge potential to change and improve practice and policy (Pilnick, 2022).

Three previous literature reviews have examined CA research adjacent to that which I review here. Maynard (2017), in a nonsystematic, purposive review (Cook, 2019), considered delivery and reception of bad news in an article written for an emergency medicine practitioner audience. Maynard focused on the strategy of “forecasting”—“giving an advance indication of the bad news to come” (Maynard, 2017, p. 3)—and explained how and why forecasting tends to result in less friction, disruption, and misunderstandings. The focus of my review is not bad news delivery, although it is evident in some of the studies, especially those that examined interactions in which the patient’s death is likely to be within minutes, hours, or days (Anderson et al., 2020; Kawashima, 2017). These studies evidence HCPs using forecasting practices.

The other two reviews (Ekberg et al., 2021; Parry et al., 2014) also overlap with this state-of-the-art review. Both of these aimed to review CA research in order to distill specific and concrete implications for practice and, for both, practitioner and clinical education audiences were the primary audience. The earlier of the two (Parry et al., 2014) systematically reviewed conversation and discourse analytic research on practices entailed in effective, sensitive communication with patients about illness progression and EoL. The later, rapid review (Ekberg et al., 2021) aimed to update Parry and colleagues’ review and narrowed the focus, excluding discourse analytic studies and only including studies of interactions in which the patient had a life-limiting illness. Rather than focusing on distilling clinical implications, my review is focused on overviewing the current state of knowledge generated by CA research in palliative and EoL care related interactions, and to consider its chronological development over time, its implications for CA scholarship more broadly, and its applications, and to suggest directions for future research (Barry et al., 2022).

Review question, aims, rationale, and definition of key terms

The review asks: What knowledge and insights have been established by CA research in palliative care and in other healthcare settings where there are interactions about the end of the patient’s life?

I aimed to systematically collate, examine, and summarize the state of knowledge to date in this field by examining published CA studies. The main purpose of doing so was to pave the way for future CA research questions, proposals, and analyses, by summarizing what has already been well or partially investigated and established, and also exposing weaknesses and gaps in knowledge.

Some readers may be unfamiliar with the key terms and topics central to this review. Furthermore, there are some differences between different healthcare systems and countries in their definition and scope. For this reason, I provide the definitions that I used for these terms.

• End of life is “the time in a person’s life when they are impaired by an eventually fatal condition and care is shifted to focus on optimising quality of life and the relief of suffering when curative treatment is no longer wanted or appropriate” (Heufel et al., 2022, p. 1). My review also included analyses of EoL interactions in which patients involved did not have a diagnosis of a life-limiting condition, but EoL preferences and plans were nevertheless discussed (e.g., cardiopulmonary resuscitation discussions).

• Life-limiting conditions, also known as terminal conditions, include “advanced, progressive, incurable conditions; … general frailty and co-existing conditions; … existing conditions [in which the person is] at risk of dying from a sudden acute crisis in their condition; … [and] life-threatening acute conditions caused by sudden catastrophic events” (National Palliative and End of Life Care Partnership, 2021, p. 38).

• Palliative care has been described by the the World Health Organization (WHO) as “an approach that improves the quality of life of patients and their families facing the problems associated with life-threatening illness” (World Health Organization, 2020, p. 13); these problems include the “physical, psychological, social or spiritual” (WHO, 2020, p. 43). Palliative care “can be provided by a range of health and social care staff” (National Palliative and End of Life Care Partnership, 2021, p. 38), and can be relevant whatever the patient’s condition (i.e., it is not confined to cancer care). WHO (2020) has asserted that palliative care contributes to the quality of people’s lives and deaths and affects service-use: Early delivery of palliative care reduces inappropriate healthcare provision and hospital admissions. However, only 12% of the global palliative care need is currently being met (WHO, 2020).

• Advance care planning (ACP), of several terms in use, I use Advance Care Planning and this consensus definition: “a process that supports adults at any age or stage of health in understanding and sharing their personal values, life goals, and preferences regarding future medical care,” with the goal of helping to “ensure that people receive medical care that is consistent with their values, goals, and preferences during serious and chronic illness.” (Sudore et al., 2010, p. 2). Accumulating evidence indicates that, provided they happen before the last few weeks of life, EoL and ACP discussions lower costs of healthcare in the last weeks of life; are associated with less aggressive care (Gonella et al., 2019; McDermott & Selman, 2018); and improve patient and family wellbeing and the match between preferences for care and actual care (McDermott & Selman, 2018).

Two final points contextualize this review: Even in higher-income countries, despite the evidence of positive effects of ACP, it is rare in practice (Rietjens et al., 2021), and specialist palliative care is an expensive and limited resource; with few people with life-limiting conditions receiving it (Quill & Abernethy, 2013). Therefore, developing nonspecialist staff’s skills in palliative and EoL care is a means to provide more equitable care. This review provides powerful evidence that could be put to good use in developing such nonspecialist staff’s skills.

Methods

I conducted my review using procedures for systematic reviews (Parry & Land, 2013; Petticrew & Roberts, 2006) but without a second reviewer. Inclusion and exclusion criteria are given in Additional File: Part A. I wanted to include technically rigorous CA research, and therefore imposed specific criteria to assess whether studies included established steps, processes, and components of CA. I used Parry and Land’s (2013) checklist and its revision by Stortenbeker et al. (2020). In addition, I only included empirical studies published in peer-reviewed journals. I excluded studies in which the only participants were practitioners and those that examined interactions where prognosis and EoL could have been raised but were not (e.g., Cortez et al., 2019). I excluded studies involving pediatric patients for reasons of feasibility and scope, because pediatric healthcare studies so often involve parent/practitioner interaction, rather than interactions with the patient (see Jenkins et al. 2024).

In October 2022, I searched Google Scholar, the EMCA Bibliography Database,¹ and my own and colleagues’ databases. No publication date limits were imposed. Search terms are provided in Additional File: Part A. To make inclusion decisions, I read title and abstract and, where necessary, the full article. For the set of studies thus identified, I searched for further studies via their reference lists and citations.

I adapted a previously designed data extraction form (Parry & Land, 2013: Additional File: Part B), completing one for each study. Data collected included details of setting and participants, the contexts and topics of the interactions analyzed, the practices and actions documented, and their functioning. I summarized findings via tabulation then aggregated findings into themes (Barnett-Page & Thomas, 2009).

Current state of the art

In total, 22 studies were identified. The table in Additional File: Part C details their characteristics, including settings, participants, and number of cases/episodes analyzed. This table lists the studies in chronological order, and in doing so shows that the earliest study identified was Peräkylä’s (1993) examination of how issues such as illness and death are introduced and built as sensitive in counselor interactions with HIV-positive people (at a time when HIV constituted a terminal diagnosis). Only three of the included studies were published before 2016, with a subsequent rapid growth in numbers. Review analysis did not find any contradictory findings and found that, in the most thoroughly studied areas, findings were cumulative and complementary to one another.

The studies identified focused on European (including the United Kingdom), Australian, Canadian, United States, and Japanese secondary and tertiary healthcare (except for one in U.S. primary care (Maynard, 2006). The languages studied were British, American, and Canadian English; Dutch; Japanese; and Swiss French.

Introducing the themes

Table 1 shows the six themes into which I sorted findings, and which studies contributed to which theme(s). Three themes focus on patterns and practices entailed in initiating and managing the topics of dying, prognosis, and ACP. These topics are relatively specific to palliative and EoL interactions both within and outside palliative care. The remaining three themes are the interactional management of emotion, the interactional consequences of companions accompanying patients, and pain assessment. Although derived from investigations of palliative care and EoL interactions, these latter themes consider topics, clinical tasks, patterns, and circumstances that feature across a wide range of healthcare provision.

TABLE 1 Publications Included in Chronological Order, and Mapped to Review Themes, Showing Which Publications Contribute to Which Theme/s

Publication	Themes
	Initiation and management of the topic of dying	Initiation and management of the topic of prognosis	Initiation and management of the topic of advance care planning	Interactional management of emotion	Interactional consequences of companions accompanying patients	Pain assessment
Peräkylä (1993)	●		●	●
Lutfey and Maynard (1998)	●	●		●
Maynard (2006)				●
Pino et al. (2016)	●	●		●
Singh et al. (2017)		●
Kawashima (2017)	●	●		●
Shaw et al. (2017)	●		●	●
Land et al. (2019)	●		●
Pino and Parry (2019a)		●		●
Ford et al. (2019)				●
Shaw et al. (2019)			●	●
Anderson et al. (2020)		●		●
Pino et al. (2020)					●
Anderson et al. (2021)		●		●
Briedé et al. (2021)	●		●	●
Jenkins et al. (2021)						●
Sterie et al. (2021)	●		●
Beach (2022)				●
Jenkins et al. (2022)						●
Pino and Land (2022)	●				●
Tate (2022)	●			●
Van Meurs et al. (2021)				●

Below, I examine each theme individually. It will become evident that there are three interactional features that recur across themes. These are (1) people recurrently move into delicate topics via cautious, gradual, step-by-step sequences; (2) both HCPs and patients recurrently refer to death and dying indirectly—that is, without using the most direct terms available; and (3) HCPs recurrently communicate in ways that show a preference for patients or companions to volunteer topics, plans, and decisions—that is, they work to encourage patients or companions to “go first” in explicitly introducing or proposing them. As these three features have been examined in considerable detail in studies of initiating and managing talk about dying and about prognosis, I consider these two themes first.

Initiation and management of the topic of dying

This theme describes findings about interactions in which people move toward, engage in, and move out of talk about the patient’s death and related matters, including feelings about it. Frequently, interactions in which dying is raised happen within interactional episodes in which people also talk about prognosis and ACP—thus, my division of these into separate themes is artificial. However, I have done so in order to map out similar and different features in how these matters are interactionally handled.

The studies included in this review are not about death and dying in the abstract. They analyzed interactions about dying in very specific, sensitive contexts: when an HCP, a patient, and (sometimes) a companion refer—or attempt to refer—to the fact that the patient is going to die. This theme included 10 articles. In one (Tate, 2022), a single case was examined; the others analyzed multiple cases. Most studies have focused on the initial stages of these conversations, in which people attempt to raise matters relating to the patient’s dying to the interactional surface. Although I have used the term “dying” in naming this theme, it is unusual for people to actually use the words “dying” or “death” (or the equivalent to these in other languages); people use indirect terms.

When engaging in talk about a patient’s future death, people sometimes approach it with such caution that it is impossible to claim definitively that they are, in fact, definitely alluding to or hinting at their dying. Rather, people refer to things that could imply or might merely prompt reference to dying and death. The research reviewed found that when interactional participants attempt to move toward these matters, they regularly and recurrently do so in ways that convey and build the topic(s) as extremely sensitive. This is evident in features including a gradual, step-by-step trajectory, beginning with moves that subtly—often ambiguously—hint at or probe for dying/death-implicative responses (Briedé et al., 2021; Kawashima, 2017; Lutfey & Maynard, 1998; Peräkylä, 1993; Pino & Land, 2022; Pino et al., 2016; Shaw et al., 2017; Sterie et al., 2021) this building of sensitivity is also evident in the interactional caution, hesitancy, and effort that feature in these sequences (Briedé et al., 2021; Lutfey & Maynard, 1998; Peräkylä, 1993; Pino et al., 2016; Tate, 2022). Although these interactions involve a gradual trajectory, this does not necessarily mean they are always lengthy (Kawashima, 2017, is an example). The step-by-step trajectory usually reaches its destination, although two studies (Lutfey & Maynard, 1998; Pino et al., 2016) reported occasions when HCPs’ moves are met by patients/companions consistently avoiding any move toward dying- and death-implicative responses, and when HCPs end the conversation without the patient’s dying reaching the interactional surface. However, in most interactions studied, the topic is arrived at sooner or later; furthermore, if patients/companions consistently avoid opportunities and resist encouragements to talk about it, the HCP “pushes harder” by introducing it more explicitly (Pino et al., 2016).

When the fact that the patient is going to die reaches the interactional surface, multiple studies, in different languages, have found that HCPs, patients, and companions rarely use the most direct and explicit reference forms—in English these being “die,” “dying,” and death” (Briedé et al., 2021; Kawashima, 2017; Lutfey & Maynard, 1998; Pino and Parry 2019a; Pino et al., 2016; Shaw et al., 2017; Tate, 2022). Instead, somewhat indirect reference is made, on at least some occasions, in ways that are evidently taken by the parties involved as unambiguously referring to the patient’s dying and death (for more on this, see S. Ekberg et al., 2019; Pino and Parry 2019b). Examples of practitioners’ indirect references include talk referencing care when the patient is or might be dying, such as, “making sure you are comfortable” (Lutfey & Maynard 1998, p. 332, Excerpt 2d). Or, in the context of ACP conversations, “what if there is an emergency, let’s say, something happens to your heart” (Briedé et al., 2021, p. 2048), raising “the possibility you might not get better”. Or, in the context of Emergency Department conversations with a relative, “there’s very little chance that her heart will start again” (Kawashima, 2017, p. 318, simplified). Examples of patients’ indirect references include, “Sometimes I feel like, ‘Oh this is it now’” (Pino et al., 2016, supplementary file; Extract 1, Lines 52–53, simplified); “I feel kind of scared, what’s gonna happen” (Shaw et al., 2017, Extract 1, p. 216); and, in a context in which a doctor asks a very frail patient to expand on her report of feeling fear, the patient replies: “Just the reality of knowing what’s happening and not being able to do anything about it” (Pino et al., 2016, Extract 2, lines 18–21, simplified).

The features and practices that make up the cautious, step-by-step sequences have been analyzed in depth. Recurrent features and practices include hesitancy, perturbations, quiet tone of voice, and self-repairs (Briedé et al., 2021; Peräkylä, 1993; Pino et al., 2016; Sterie et al., 2021; Tate, 2022). Initial actions in these sequences include raising and describing (chronologically) the patient’s overall trajectory, deterioration, and current state (Anderson et al., 2020; Kawashima, 2017; Lutfey & Maynard, 1998). Extract 1 provides an example. It is from a study of Japanese emergency department interactions, in which a doctor takes a patient’s daughter through recent events and toward the patient’s current state.

Extract 1.

From Kawashima (2017, Excerpt 1, p. 310–311).(DOC = emergency medicine doctor, FAM = daughter)

Other studies documenting this initial action (describing what has been going on for the patient) examine interactions in which the patient’s death was nowhere near as imminent as in Example 1. For instance, Tate (2022), examining U.S. oncology consultations in late-stage cancer documents the oncologist describing various treatments that have already been used but that are no longer working as an initial move in a trajectory that moves toward introducing palliative care. This resembles “exhausted current treatment statements” documented in U.S. oncology settings (Cortez et al., 2019). As I discuss in the section, “Interactional management of emotions,” raising a patient’s past and current state, deterioration, and non-response to prior/current treatments can sensitively bring all parties to shared understandings of the situation and facilitate acceptance, thereby reducing the chances of prompting acute distress, shock, or other strong emotional expression with resultant interactional breakdown (Kawashima, 2017; Lutfey & Maynard, 1998; Tate, 2022).

Step-by-step sequences in interactions that attempt or achieve talk relating to dying and related topics (albeit through indirect terms) recurrently entail a series of HCP enquiries, descriptions, and “my-side tellings” (Pomerantz, 1980) that encourage dying/death-implicative responses; sometimes, these are quite lengthy series. Patients’ and companions’ responses do not always support progression toward this topic, and sometimes they actively move away from it. A clear example of the latter is seen in Line 15 in Extract 2. In the extract, an oncologist is speaking with a patient who is nearing the end of his life and is recovering from an unsuccessful attempt at treatment. The oncologist has, prior to this extract, already made several explicit enquiries that encourage responses relating to future dying; for example, “What do you see as (0.8) as the- happening in the future?” (Lutfey & Maynard, 1998, Excerpt 1C)

Extract 2.

From Lutfey and Maynard (1998, Excerpt 1D, p. 328), abbreviated and simplified.DOC = oncologist, PAT = patient

The oncologist describes aspects of the patient’s recent circumstances (Lines 3–4), tells of his own concerns (Line 1) and their upshot, and does so in ways that leave multiple slots in which the patient could respond (Lines 1–14). The patient has not taken these opportunities, and at Line 15 onward responds in a way that turns sharply away from any dying-relevant matters. After this extract, the oncologist and patient collaborate in closing the conversation, the oncologist does not attempt to resurrect the dying-relevant enquiries.

Across studies, HCP enquiries recurrently focus on the patient’s/companion’s perspective (e.g., Lutfey & Maynard, 1998; Peräkylä, 1993), and thus initiate perspective display sequences (Lutfey & Maynard, 1998). In this context, if and when the patient’s/companion’s response has made it possible, the HCP draws on what they have said in order to progressively “unpack” dying-/death-implicative matters. Pino et al. (2016) observed that HCPs’ perspective enquiries recurrently pick up on patients’ mention of physical symptoms in such a way as to lead from talk about those symptoms toward talk about dying/death and related matters. Furthermore, if on-topic talk has not been achieved via their earlier broad enquiries, HCPs progressively narrow the field of projected responses toward a focus on difficult feelings, worries, and inner experiences (Lutfey & Maynard, 1998; Peräkylä, 1993; Pino et al., 2016; Shaw et al., 2017). One way they do so is to pick up on patient/companion references to low mood, anxiety, and panic amid shortness of breath. HCPs use these as opportunities to shift the focus toward matters that are more clearly related to dying (Pino & Land, 2022; see also Shaw et al., 2017). Extract 3 gives a brief example of a perspective enquiry in which a doctor asks a question that works to move the focus from the patient’s reported symptoms to that of his inner experience, his perspective.

Extract 3.

From Pino et al. (2016, Supplementary File 1., Extract 1, p. 1).

In this example, the doctor explicitly picks up on something the patient has already said. This is one of those practices through which practitioners can develop an interactional environment wherein if and when the patient’s dying and related matters finally reaches the interactional surface, it comes off as the patient’s own topic and concern—as (ostensibly) raised “first” by the patient, as volunteered by the patient, rather than imposed by the HCP. Encouraging patients/companions to expand on something they have already raised in such a way as to connect those to dying/death-implicative matters frames subsequent talk as stemming from concerns they themselves have raised. As Pino and colleagues noted, “experienced doctors withhold introducing the topic of EoL in places where they could do so, and instead, they use forms of talk that leave it to the patients to take the initiative to introduce EoL considerations” (Pino et al., 2016, p. 13).

Raising future-oriented hypothetical scenarios and then making enquiries that push toward dying-implicative responses is another practice used in moves toward on-topic talk (Briedé et al., 2021; Lutfey & Maynard, 1998; Peräkylä, 1993; Pino et al., 2016; Sterie et al., 2021). The hypothetical mode works to put distance between, on the one hand, the patient and this current conversation and, on the other, the reality of the patient’s dying and death. A similar practice involves HCPs framing a dying-related topic or question as a universal matter that anyone in this situation could face—again distancing it from the reality of this patient’s death (Peräkylä, 1993; Sterie et al., 2021).

Pino et al. (2016) documented that, when other practices have not resulted in the topic surfacing, HCPs may go further toward raising it themselves. In Extract 4, the HCP suggests the patient might already have dying-related thoughts. The patient has advanced cancer and has been admitted to the hospice with marked worsening in his condition. Clinically, it is assumed that his cancer is advancing rapidly, and that he is nearing the end of his life.

Extract 4.

From Pino et al. (2016, Extract 5, Supplementary File 1, p. 6), abbreviated.

Here, the doctor’s introduction of what—in the context of the patient’s deterioration—are clearly dying-implicative matters puts the patient in a position of having to either respond on-topic or overtly refuse to do so (which he does). This practice is different to those described earlier, which all provide patients/companions with opportunities to engage in dying-related talk, but which neither oblige them to do so nor require them to go on record in declining to. The fact that HCPs (including the doctor in Extract 3, data not shown) try other approaches first is in keeping with the recurrent pattern of HCPs encouraging patients to go first, to volunteer dying-related matters, rather than the HCP raising them first.

Two other studies (Shaw et al., 2017; Tate, 2022) examined how talk about dying-related matters is brought to a close. Both found that this can be accomplished by shifting toward talk about concrete practicalities, processes, and care.

Initiation and management of the topic of prognosis

In this review, prognosis concerns expected life expectancy (or, more bluntly, expected time until death), predicted by HCPs by applying their professional knowledge and experience to the individual patient’s signs, symptoms, and trajectory. This is different to the kind of prognosticating that is done by oncologists (and presumably other specialists, such as those working with degenerative neurological conditions, such as motor neurone disease) that is based on survival statistics for large groups of patients in combination with the individual’s scan results (in oncology) and professional judgment (Hagerty et al., 2005).

Seven articles (see Table 1) included analysis of interaction about prognosis. Interactions studied ranged from interactions in which the patient was likely to die within minutes (Kawashima, 2017), to hours or days (Anderson et al., 2020, 2021), to those in which death was anticipated in weeks or months (e.g., Lutfey & Maynard, 1998; Pino et al., 2020).

Within this theme, I outline findings about when and how prognosis is raised, how prognosis is delivered and received, and how the topic is closed. The studies led by Anderson et al. (2020, 2021) and Pino et al. (2020) provide more in-depth and specific examination of the initiation and management of prognosis talk than do the other publications shown in Table 1 to be relevant to this theme. Although the extracts and analysis of these other publications touch on prognosis, their main focus is on practices, sequences, actions, and challenges in raising the topic of dying. For instance, in Kawashima’s (2017) study, the patient is imminently dying (and the clinical team is attempting cardiopulmonary resuscitation), and the practices studied move the relative toward understanding that the patient is imminently dying. Conveying imminent dying could be framed as “conveying the prognosis”; however, Kawashima’s analysis is framed as examining resources practitioners use to allow family members to understand the situation and to engage in treatment decisions. I have opted to include Kawashima’s detailed findings, especially those about the practitioners’ forecasting practices (including telling the story of what has happened so far), in the theme on initiating and interactionally managing the topic of dying. In Pino and colleagues’ work, the focus is on a set of practices experienced doctors use to “navigate the dilemma of promoting discussions about dying” (Pino et al., 2016, p. 1); it is included in this theme because in one extract, a patient’s (indirect) reference to prognosis is explored because it is salient to the authors’ primary focus on the practices for promoting dying talk.

Anderson et al.’s (2020, 2021) work concerns contexts in which hospice inpatients are in their last hours or days of life, and interactions about prognosis take place between experienced HCPs and patients’ family members. Reference to prognosis was recurrently prefaced by talk about the patient’s deterioration over (recent) times; the reader will note that similar “forecasting” practices were described in the theme initiation and management of the topic of dying. In Anderson and colleagues’ context, this reference to deterioration was usually raised in response to a relative’s request for an estimate.

In hospice consultations in which patients were generally months from dying, Pino and Parry (2019a) found prognostic talk commonly arose following a doctor-provided opportunity to the patient/companion to raise agenda items. Rather than asking questions along the lines of, “How long have I got?” patients made statements or assertions that cautiously expressed an interest in receiving an estimate. Patients and companions also put interactional effort into conveying they know it is not possible to provide a precise time estimate—for instance, by asking for an opinion rather than a statement of fact (Anderson et al., 2020), and/or by directly referring to uncertainty: “I know you can’t say how much time I’ve got” (Pino & Parry, 2019a, p. 228, Table 2). Besides showing they anticipate that the doctor may be unable to give an estimate (or a precise one), Pino and Parry found patients recurrently referred to prognostic estimates they had received before, to reasons why they wanted an estimate now, and to their readiness to hear about prognosis—for example, “I wanna know” and, “I don’t mind because I know I’m going to die” (Pino et al., 2016, p. 228, Table 2). In these ways, patients establish prognosis as “requiring special, interactionally cautious treatment, which should not be raised abruptly but, rather, worked up to.” (Pino & Parry 2016, p. 230). This could equally be said of the topic of dying, which is initiated and managed in interactionally cautious ways.

TABLE 2 Recurrent Interactional Features Documented in 18 of the 22 Included Publications

	The interaction progresses via gradual, step-by-step sequence.	Patients and HCPs recurrently refer to death and dying via indirect reference forms.	Rather than raise or impose a topic or propose a response, HCPs encourage patients/companions to be volunteer the matter, and to bring it to the interactional surface.
Peräkylä (1993)	●		●
Lutfey and Maynard (1998)	●	●	●
Pino et al. (2016)	●	●	●
Singh (2016)		●
Kawashima (2017)	●	●	●
Shaw et al. (2017)	●	●
Land et al. (2019)	●		●
Pino (2018)	●	●	●
Shaw et al. (2019)	●
Anderson et al. (2020)	●	●	●
Pino et al. (2020)	●
Anderson et al. (2021)	●
Briedé et al. (2021)	●
Jenkins et al. (2021)			●
Sterie et al. (2021)	●		●
Jenkins et al. (2022)			●
Pino and Land (2022)	●
Tate (2022)	●	●

Pino and Parry (2019a) also found that when patients/companions had not first conveyed what they already know, how they feel about it, their readiness to know more now, and a recognition of uncertainty, doctors encouraged patients to provide this information before giving their estimate. They argued, “The doctors, patients and companions in our recordings cooperate in preparing conversational environments for the delivery of estimates” (p. 234), and that the doctors “treat life-expectancy as information that patients can legitimately request, they nevertheless treat estimates as information that should not be delivered point blank but, rather, communicated in a prepared conversational environment—an environment where patients’ prior knowledge, emotional state, and their readiness to receive an estimate have been ascertained” (p. 235).

Pino and Parry (2019a) found that once the environment had been prepared, doctors recurrently formulated their own estimate as agreeing with what patients report they already know—for example, “I think what he said is probably right” (Pino & parry, 2019a, p. 230). Anderson et al. (2020) found that in hospice conversations with relatives of dying patients, HCPs gave absolute (as opposed to relative) time references, most commonly “days” or “hours”—but not a specific number of those. Both Anderson and colleagues and Pino and Parry found that neither HCPs nor patients/companions used the direct terms “death” and “dying” in interactions about prognosis, although they clearly recognized these were being talked of. Anderson and colleagues argued that avoiding the direct words “death,” “die,” or “dying,” and giving “hours” or “days” estimates—without specifying how many—provided a sensitive approach to prognostic awareness.

Anderson et al. (2021) found that after the HCP had given a prognostic estimate, relatives recurrently responded with acceptance and acknowledgment; the HCP then summarized and acknowledged the prognosis, and finally shifted to certainties and “brightsides” (Holt, 1993), such as assurances of care and comfort and proposed future concrete actions. The parties thereby moved from uncertain and uncontrollable prognostic estimation to action-oriented talk about more certain and controllable topics.

A final point in this theme is that in the studies I have focused on, it is the relatives or patients who, albeit cautiously, raise prognosis. In the interactions studied, cooperation and collaboration between the parties is evident. This contrasts with the resistance that is often seen in interactions in which attempts are being made by the practitioner to raise the topic of EoL, and to get it raised in ways that mean the patient is ostensibly the first to actually raise it all the way to the surface of the interaction. My review procedures did not identify studies in which raising the topic of prognosis involved resistance; however, I would argue that there are likely to be times when resistance to the topic of prognosis occurs in real-life EoL and palliative care interactions.

Interactions about advance care plans

Before launching this theme, I should clarify that it is restricted to talk about future matters: future care and treatment, and preferences relating to these. I did not include studies in which decisions were being made in the moment, although two studies examined this: Kawashima (2017) studied interactions involving decisions by relatives in the here and now about whether the team continued to attempt cardiopulmonary resuscitation on the patient, and Tate (2022) examined interactions involving oncologists and people with advanced (Stage 4) cancer in which there was decision making regarding next steps in the treatment. The rest of this section considers what has been studied regarding making preferences, plans, and decisions about future care and treatment.

ACP conversations are uncommon in clinical practice. Six publications provided evidence on interactions that include ACP talk; two of them (Shaw et al., 2017, 2019) did not involve directly making plans or decisions but, rather, talk between counselors and women with advanced cancer about their perspectives and thinking about future treatment choices. I report below on the other four studies. Compared with studies of the interactional management of the topics of dying and prognosis, the contexts in which ACP has been studied are more diverse, and a diverse set of different practices has been documented in separate studies. Unfortunately, there is insufficient material to synthesize findings together; therefore, I describe practices and studies one by one.

In the context of making decisions about future treatment attempts to extend life, that is “code status (patient preferences regarding resuscitation, mechanical ventilation and Intensive Care admission),” HCPs were found recurrently to give accounts and justifications for raising the topic (Briedé et al., 2021, p. 2045). The setting was outpatient consultations in a hospital with a well-publicized policy that advance care decisions were (supposed to be) discussed and documented, regardless of whether the patient was terminally ill. The HCPs made extensive efforts to justify raising EoL plans, despite their lack of immediate and current relevance to the patient; the researchers pointed to a “relevance paradox” faced by the HCPs. In terms of the justifications, they described two types. External justifications referenced organizational matters (e.g., hospital policy and/or electronic care record requirements). Internal justifications invoked patients’ health-related matters (e.g., a patient’s heart condition as justification for talking about cardiopulmonary resuscitation plans).

Doctors’ explicit justifications for raising ACP have also been examined (although briefly) in a very different context. Pino et al. (2016) examined interactions between doctors and patients who had actively deteriorating life-limiting conditions—so, the relevance of considering future treatment/care was clear. Rather than resuscitation preferences, ACP here entailed matters such as preferred place of dying. They found doctors providing explicit justifications for engaging in ACP in contexts in which patient or companion were evidently reluctant to talk about it. They found doctors talking of engaging in ACP in terms of “just in case” or “rainy day” matters (Pino et al., 2016, p. 6).

Sterie et al.’s (2021) study took place in an older people’s rehabilitation hospital—that is, a non-palliative setting in which patients were unlikely to have been given a life-limiting illness diagnosis. Like Briedé and colleagues’ study, the focus was on talk about preferences for/against resuscitation and other attempts at life-extending treatment, such as dialysis. These episodes of talk occurred within “intake assessments,” of which resuscitation preferences formed only a small part. Sterie and colleagues focused on how the junior doctors studied used more or less presumptive framing of ACP decisions. Most commonly, doctors introduced CPR presumptively—that is, in such a way that it assumed what the patient’s decision would be: for example, “So uh we- if there’s a problem with uh the heart or something, generally we resuscitate you” (Sterie et al., 2021, Excerpt 1, p. 3, simplified), and they found patients tended to respond with agreement. Less frequently, doctors used less presumptive formats: for example, “If something very very serious happened to you, what would you wish that the physicians do (.) or don’t do?” (Sterie et al., 2021, Excerpt 3, p. 4., simplified). Occasionally, patients preempted ACP discussion by stating their preference in the midst of a doctor’s introduction of the topic.

Finally, ACP-related practices were documented by Land et al. (2019). The setting was hospice consultations involving specialist doctors and patients who already had some form of ACP in place. They found a practice HCPs used to address possible problems with the patient’s plans or expectations: hypothetical future scenarios. For instance, with a patient who preferred and planned to be an inpatient for EoL care, the doctor posed a hypothetical scenario of no bed availability at the crucial time. The number of cases was small, but those identified and examined suggested that, having raised the hypothetical scenario, doctors tended to encourage the patient to volunteer and propose the solution to the scenario. Once again, as within other themes, this study evidenced ways that practitioners in palliative care settings encourage patients to volunteer important matters and avoid proposing them unilaterally.

The CA evidence about the challenges and practices of ACP is sparse and rather disparate, given that it examines two different elements of ACP: resuscitation preferences in settings outside palliative care (specifically, medical outpatients and older people’s rehabilitation) and more broad plans and preferences for dying (including preferred place of death and preferences about who will be with the patient) in palliative care settings. However, there is enough to suggest that the challenges and practices of ACP include managing the sensitivity and the relevance of ACP; interactionally managing which party is making the decision(s); and, more broadly, matters of accountability.

Interactional management of emotions—in particular, distress, shock, and anger

Distressing emotions form a key backdrop to many healthcare settings and interactions. The emotional stakes are particularly high in palliative care and EoL conversations, and so distressing emotions are particularly likely to impact interactions. In this theme, I examine actions and practices that, researchers argue, reduce the likelihood of provoking emotional expression to a degree that interrupts or derails the ongoing conversation and therapeutic agenda. I will also overview the smaller body of evidence that examines interactions in which a patient expresses considerable distress within a healthcare interaction, and in which progressivity of the clinical agenda is halted. One study examined HCP practices that can provoke such distress.

Many of the strategies, sequences, actions, and practices described above contribute to managing distressing emotions; I briefly revisit them now with a focus on how they manage emotions. The gradual, step-by-step sequences through which delicate matters are worked toward function to bring sensitive, delicate topics to the surface in a gentle cautious manner. They enable gradual movement into the sensitive matter, so that when it reaches the interactional surface, it has been forecast, rather than unilaterally and bluntly announced or imposed on the patient/companion.

Within these step-by-step sequences, practices used include quiet, soft intonation and hesitancy. Through their practices, HCPs convey they are treating and recognizing the patient/companion as likely to be experiencing difficult emotions, and they are constructing the topic to be addressed as a sensitive one (Pino et al., 2016; Tate, 2022). Cautious, stepwise entry into the sensitive topics via telling the story of what has happened to the patient up to this point, and the patient’s state now—often collaboratively produced by both HCP and patient/companion (Anderson et al., 2020, 2021; Kawashima, 2017)—forecasts what is to come and builds gradual realization. In addition, perspective display enquiries and other questioning practices allow HCPs to gather information they can use to assess the likelihood of sudden acute distress if the topic is fully brought to the surface (Lutfey & Maynard, 1998; Pino et al., 2020).

Practices that frame dying and related matters and choices as hypothetical, or as something that anyone would face, also function as part of this protection against provoking severe emotional responses. As Peräkylä (1993) pointed out, hypothetical scenarios “create a discursive space marked as hypothetical to accommodate talk about the patients’ future. Within the space so marked, assertions can be made about the clients’ future without the participants openly committing themselves to them as accurate descriptions of what will happen” (Peräkylä, 1993, p. 313).

Practices that invite patients/companions to engage with the sensitive delicate topic but that do not impel them to do so also contribute to avoiding the provocation of severe distress. Peräkylä (1993) has succinctly and insightfully argued that patients’ and HCPs’ use of indirect terms for dying and death also contribute: “Words have the power to invoke realities. Perhaps that power makes us sometimes very reluctant to put things in words: we may want to avoid bringing about realities that are difficult for us to cope with” (Peräkylä, 1993, p. 292).

Surprisingly, given the importance placed on it by clinicians, patients, and educators, empathy was examined in only one CA study fitting the criteria for this review. Ford and colleagues (2019) focused on the role of doctors’ empathic talk in achieving therapeutic tasks. They found that when there is some disparity between a patient’s experience and the doctor’s expertise, empathic talk can work to bridge this gap. This analysis of one function of empathy within a specific context is one of a sizable and cumulative body of CA evidence about practices for, and the functioning of, expressing empathy (key sources include Heritage, 2011; Kupetz, 2014; Ruusuvuori, 2007.

Three studies examined situations in which a patient’s powerful emotional expression hampers progressivity of the consultation; only one focused on how an HCP’s practice contributes to this. Maynard (2006) examined a case in which a relatively inexperienced primary care doctor announces a diagnosis of cancer to a young man. He noted that the doctor does not use practices commonly seen in such interactions: He does not build gradually, step by step, toward the announcement; he does not convey the matter as highly sensitive and difficult; and he does not follow the sensitive news with concrete information about what will be and should be done next. The patient suffers a severe emotional response, and extreme disruption of the consultation occurs.

Beach (2022) took a different tack, in a very different context. He examined ways in which one senior oncologist responds to expression of great distress. In this single-case study, a young mother is attending an oncology clinic; she is already aware that her cancer is now advanced. As the consultation agenda moves from practical matters to the doctor conveying that he recognizes how difficult the patient’s and her family’s situation is, the patient begins to sob and cry at length. Beach documented how the doctor stands close to the patient, offers reassurance with soothing and concerned prosodic features, softly rejects the patient’s apologies for crying, offers her tissues, and avoids verbal or embodied moves that might be understood as attempts to move on to next business.

Finally, Pino and Land (2022), documented a practice for managing a patient’s distress: In consultations in which the patient is accompanied by a companion, and the patient’s distress may be hampering his or her ability to respond to an HCP’s questions, the HCP sometimes interactionally manages this by moving to seek the response from the companion.

Companions’ co-presence is salient

Two studies examined episodes within palliative care interactions in which the fact that the patient is accompanied by a companion gives rise to interactional challenges. Although their analytic focus was not primarily on the palliative care nature of these interactions, the palliative care context is one in which the companions’ role is particularly crucial to effective assessment, care, planning, and decision making.

Pino et al. (2020) examined dilemmas HCPs face when companion and patient take different positions—one supporting, one opposing—on healthcare measures (such as instituting or changing home care services or using walking aids). The study’s focus is on how HCPs manage this situation in ways that avoid escalating conflict, and how they try to avoid overtly taking sides: Given the importance of companions’ collaboration in making and implementing care decisions, it is important for HCPs to maintain relationships with both patients and companions.

The other study (Pino & Land, 2022), mentioned earlier under the “managing distressing emotions” theme, examined episodes in which an HCP has addressed a question or recommendation to the patient, but there is some problem in the patient’s responding. Pino and Land considered when and why companions respond, concluding that, “companions’ interventions on patients’ behalf are collaboratively constructed as warranted departures from an otherwise prevailing normative orientation to patients’ right to speak for themselves. The practices we identified support and maintain this right and can thus be considered one fundamental way in which the principle of patient autonomy is put into practice within health-care interactions” (Pino & Land, 2022, p. 410).

These two studies form part of a growing body of CA work on companions’ roles in health and social care interactions (Chinn, 2022; Ekberg et al., 2015).

Assessing physical pain

The final theme covers two studies of doctors assessing patients’ physical pain. This is arguably an underrepresented area of study because, like empathy, managing pain is a major and abiding concern for practitioners and those who train them. Both pain studies were of hospice-based palliative care consultations. One (Jenkins et al., 2022) examined how the “0 to 10” numerical pain intensity scale, originally designed to elicit a single number response, is introduced and responded to in real-life healthcare. The study found patients rarely provided single number responses; instead they gave more than one number: for example, “At the moment, now, maybe three four” (Jenkins et al., 2022, p. 2). Patients thereby provided more than just pain intensity; for instance, they conveyed their pain’s variability and that position and movement affected it. The doctors supported patients to add this information by letting them expand on their responses, rather than insisting on a single figure. Jenkins et al. (2021) examined one way in which doctors sometimes repeat part of a patient’s response to a question about pain. They found that this provided opportunities for patients to elaborate on what they had said, and also made it easy for patients to pass on the opportunity without justifying or accounting for doing so.

Recurrent interactional features across the themes

Multiple studies have documented in considerable detail three interactional features that recur across several of the themes. I have attempted to highlight these within individual themes. To further emphasize these findings, Table 2 depicts which studies provide evidence on these features: the cautious, gradual, step-wise movement toward sensitive topics; the fact that patients recurrently refer to dying and death via indirect reference terms and phrases; and that, rather than unilaterally impose a topic or a particular response on a patient, HCPs encourage patients and companions to be the ones to volunteer the topic or response, and thereby to be the ones who bring it to the interactional surface. Table 2 shows these themes and the 18 out of 22 studies that contributed to them.

Applications

Some activities, practices, and features entailed in end-of-life and palliative care interactions have been studied sufficiently to derive applications in the form of practical teaching and learning insights that could inform guidance and practice. Multiple studies have shown how patients and HCPs can move stepwise toward sensitive topics and increase patients’ readiness to engage in them while avoiding the distress this might cause.

In this area, findings from Peräkylä’s (1993) study onward have been cumulative, not contradictory. These studies show that experienced HCPs use step-by-step, cautious strategies; enquiries that focus on patients’ perspectives; and other practices that encourage the patient to be the one who eventually raises the sensitive matter to the surface. They show that hypothetical scenarios and questions offer effective and, indeed, kindly means of engaging people in sensitive matters and decisions. Furthermore, within the reviewed studies, there are insights into why people use these practices; some of these insights are clearly evidenced and some are logical extrapolations made by researchers on the basis of studies’ contexts and analytic findings. These include insights that step-by-step movement toward topics, hypothetical scenarios, and perspective display questions encourage patient engagement and allow practitioners to gauge the degree to which patients and companions are ready and prepared to engage in sensitive topics, But they don’t only gauge readiness; in fact, they can build patients’ readiness and preparedness to engage. Other functions of these practices include reducing the interactional risk of provoking distressed, shocked, or angry emotional expression sufficient to disrupt the interaction itself. Another practice that reduces the risk of provoking strong emotional expression entails referring to dying and death indirectly, at least when it is first being mentioned; after first mentioned, such indirect terms may be “unpacked” (Lutfey & Maynard, 1998). I will return to this shortly.

In other areas, the evidence is not so strong. On interaction about prognosis evidence is patchy, and there are no studies of interactions in which patients/companions resist engagement in the topic. Also understudied is ACP, which is widely advocated, as research has shown it reduces the chances patients will receive hospital interventions they do not want and increases the chances a person will die where he or she wishes to die. The evidence on how to move toward sensitive topics is relevant to the build up toward ACP, but there are only limited pointers toward practical implications in regard to ways to elicit, express, and negotiate preferences, plans, and decisions about future care and interventions.

I turn now to concrete applications of the findings: using them to enhance HCP training in sensitive and effective healthcare communication and an example of using them to reshape assumptions and conventional thinking in palliative and EoL interaction.

Some of the work included in this review (including Ford et al., 2019; Jenkins et al., 2022; Pino et al., 2020, 2016) has been used to develop training resources for EoL and palliative care communication trainers in a program known as RealTalk² (Parry et al., 2022; Parry et al., 2024). RealTalk’s primary audience comprises professionals who teach “communication skills” to HCPs and to healthcare practitioner students. The resources include modules (“Talking about dying,” “How long have I got?” “Empathy,” and so on). Each module contains cases, and each case focuses on a single video-recorded healthcare consultation. Clips from a consultation are provided alongside verbatim transcripts and CA-based learning points. The aim is to enable trainers to incorporate CA evidence into their work; use the clips to increase the authenticity and realism of the materials they use in training; and support a more analytic approach to communication.

When using clips, the trainers are asked to encourage trainees to answer the questions, “What did you see?” and, “What did you hear?” as a way of discouraging them from focusing on what they would do differently, and whether the video showed “good” or “bad” practice. Extensive detail on RealTalk resources can be found in Parry et al.’s (2022) report. Preliminary qualitative evaluations (Parry et al., 2018) have found that RealTalk materials enhance active engagement in learning, and that “train-the-trainer” workshops improved professional trainers’ self-rated competence in training and supervising others in good communication, as well as their confidence in giving detailed feedback to trainees (Parry et al., 2024). A closely related program entails starting to use CA insights and evidence to shape and build materials on EoL conversations, including scripts for video materials for the National Healthcare Communication Programme³ of the Republic of Ireland’s healthcare system (the HSE) Module 3 Challenging Conversation.⁴

In our editorial (Parry & Barnes, 2024), we discussed CA evidence’s role in informing policy and guidance. A brief example illustrates how some of the practices and actions identified via direct observation contradict (well-meaning) guidance based on assumptions and indirect evidence. This review showed that HCPs, patients, and companions across several languages rarely refer to death and dying by using the most direct terms available in the particular language. This has been found in studies of U.S., British, and Canadian English speakers, and in Japanese speakers (see also Ekberg et al., 2019). However, HCPs are strongly advised to use those direct terms, “death” and “dying”, at all times (e.g., Sutherland, 2019), particularly in palliative care. The recommendation to use direct language, and to use the most direct terms rather than alternatives (Sutherland, 2019), is grounded in a moral position—that conversations about EoL should be honest—and on a priori assumptions that direct reference provides for effective communication and ensures that patients and companions understand what is being referred to.

In contrast, this review found that people recurrently use indirect terms and phrases. In the context of palliative care, there is interactional evidence (Anderson et al., 2020; S. Ekberg et al., 2019; Pino and Parry, 2019a) that indirect terms can be utterly unambiguous for the parties involved. For instance, in certain contexts, reference to “not here anymore” and “passed away” cannot reasonably be heard as meaning anything other than “dead”; and similarly, “coming to the end” and “won’t survive” can only mean dying. Good reasons for these alternative and implicit references to death and dying include avoiding prompting expression of acute and strong emotion. As Peräkylä wisely put it, “Words have the power to invoke realities. Perhaps that power makes us sometimes very reluctant to put things in words: we may want to avoid bringing about realities that are difficult for us to cope with” (Peräkylä, 1993, p. 292). As a caveat, there are times in palliative care when patients and practitioners do use direct terms (for example, “I know I’m gonna die” (Pino and Parry, 2019a, Extract 2, lines 37–8)

Issues for reflection

Currently, state-of-the-art in CA research on palliative care and EoL conversations examines only a small portion of the settings, HCPs, patients, and locations in which these conversations occur. The reviewed studies were entirely based in healthcare systems in high-income countries (rather than middle- and low-income countries), in secondary care in large university cities, and the majority examined cancer care and medical doctors. This is unfortunate, given that it is in community/primary care settings where most interactions about the end of life occur. On the other hand, there is good reason to assume that, wherever these sensitive conversations happen, similar interactional features, practices, and the functioning of those are likely to occur. If we accept the latter point, then applying the findings is likely to be beneficial.

The findings of this review contribute to several areas of foundational CA scholarship, in particular, emotions in interaction, the initiation and management of sensitive topics, the form and functioning of indirect reference, and talk about the future and the hypothetical mode. The findings can also contribute to CA in healthcare; in particular, they overlap with and build on work examining highly sensitive interactions—for instance, Weijt’s and colleague’s insightful work on talking about sexuality and the “expressive caution” that professionals and patients accomplish collaboratively during gynecological examinations (Weijts et al., 1993).

Limitations of the review process itself may limit its applicability. I worked alone, and this is not optimal for systematic reviewing procedures: I may have either excluded or included relevant articles and data, including within the quality assessment phase. Having second or, optimally, third review team members would have helped reduce the chances of this.

The future of CA in interactions with adults in palliative care and/or about EoL

From the point of view of healthcare practice, policy, and training, future research in this area should prioritize locations, settings, healthcare systems, and participant groups that have been overlooked to date. I have already noted the lack of research in lower-income countries and healthcare systems.

It is also worth noting that, in actual practice, a large proportion of palliative care and EoL conversations take place in the community (e.g., people’s homes, residential care) and primary care settings; that the kind of care involved is primarily delivered by nurses, health and social care assistants, and HCPs who are nonspecialists in palliative and EoL care; and that it is delivered to older people, many of whom are living with frailty and dementia.

Also, and particularly for people in underserved and minoritized communities, there is a steady increase in people (and their families) presenting at end of life to emergency and acute services. Future research should also prioritize these common but as yet understudied or unstudied components of care. It should also include studies of ACP conversations, and the growing use of telehealth via video and other remote modes to deliver palliative care and EoL care (see Seuren et al., 2024).

In addition, one EoL matter that has not been studied at all—but that is a matter of considerable concern and controversy—is assisted dying. Assisted dying is available in a growing number of countries. We can expect interaction to be central to the processes involved, for instance, in how patients and practitioners construct the patient as eligible or not, of independent mind or not.

Finally, healthcare interactions about end of life, dying, and death could contribute to deepening sociological and social psychological understandings of fundamental matters, including how we construct the meaning of living and dying for ourselves and others, and how healthcare differs when not aimed at cure, and when spiritual and existential matters form part of the agenda. Studies to date have not generally delved into these fundamentals.

Acknowledgments

I would like to thank Dr Rebecca Barnes, Dr Aija Logren, and Dr Rebecca Anderson-Kittow for their wise and thoughtful comments on earlier versions of this paper.

Disclosure statement

No potential conflict of interest was reported by the author.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/08351813.2024.2305048

Correction Statement

This article was originally published with errors, which have now been corrected in the online version. Please see Correction (http://dx.doi.org/[10.1080/08351813.2024.2305048)

Notes

¹ https://www.emcawiki.net/EMCA_bibliography_database.

² https://www.realtalktraining.co.uk/.

³ https://www.hse.ie/eng/about/our-health-service/healthcare-communication/.

⁴ https://www.hse.ie/eng/about/our-health-service/healthcare-communication/module-3/.