Inviting the Patient to Talk About a Conversation They Had with Another Healthcare Practitioner: A Way of Promoting Discussion About Disease Progression and End of Life in Palliative Care Interactions

ABSTRACT

Discussing disease progression is a core task in palliative care. This is especially important when there are indications that a patient considers their death as less imminent than the clinical team does. This article examines a communicative action that palliative medicine doctors use to address such discrepancies in knowledge and understanding of the patient’s prognosis: inviting the patient to talk about the contents of a conversation they had with another healthcare practitioner. The study used conversation analysis to examine five consultations in which this action was identified. These were part of a larger data set of 37 consultations recorded in a large UK hospice and involving patients with palliative care needs, sometimes accompanied by family or friends, and palliative medicine doctors. Findings are that the action of inviting the patient to talk about a previous conversation creates an opportunity for patients to articulate what they know and understand about their disease progression – but without requiring them to do so. Discussing such sensitive matters is thus made a matter of ‘opting in’ (rather than ‘opting out’). Doctors thereby avoid being interactionally accountable for directly initiating a potentially distressing topic. The article shows how the task of discussing disease progression and end of life is intertwined with the delicate management of patients’ displayed states of awareness regarding their disease progression. The study thus has practical implications by documenting ways in which clinicians can help patients realign their expectations about such delicate matters.

Introduction

Healthcare interactions with patients who have life-limiting illnesses can involve discussions about disease progression and end of life (EoL). Studies in Western settings have documented how healthcare practitioners (HCPs) inform patients and/or their family that curative treatment is no longer effective (Cortez et al., 2019), give prognostic information (Anderson et al., 2020), assist in decisions about EoL care (Shaw et al., 2016), and support the processing of feelings and thoughts about mortality (Shaw et al., 2019). There is consensus that these discussions benefit patients and families (Ghandourh, 2016; Jimenez et al., 2018), but HCPs experience them as difficult (Hancock et al., 2007). Documented reasons include challenges in providing accurate prognoses, lack of training, and fear of taking away hope (Almack et al., 2012; Brighton & Bristowe, 2016; Ghandourh, 2016).

Studies using conversation analysis (CA) use direct observation of recorded interactions to document ways in which patients, companions, and HCPs navigate practical problems in initiating, sustaining, and closing discussions about disease progression and EoL (for a review see Ekberg et al., 2021). In the Western settings where this research has been done, some HCPs display reluctance to discuss these matters with patients. For example, when delivering news about scan results, oncologists sometimes invite appreciation of how cancer treatment has prolonged the patient’s life (Maynard et al., 2016), which can result in foregoing discussion about prognosis and implications for future care (Singh et al., 2017). Additionally, when oncologists inform patients that current treatment is not working, they sometimes bypass discussions of the prognostic implications of such news by quickly transitioning to discussion about alternative treatment options (Cortez et al., 2019). This may sustain hope (on hope work, see Peräkylä, 1991), but it also limits opportunities to promote awareness about likely disease progression and to discuss EoL care.

By contrast, some HCPs actively promote discussions about disease progression and EoL. Western settings where this has been documented include adult palliative care (also the context of our present study), pediatric palliative care (Ekberg et al., 2017), and some types of counseling and psychotherapy (Peräkylä, 1995; Shaw et al., 2017). In these settings, discussing the future is part of practitioners’ remit and works as a preliminary to institutionally relevant activities, including planning for EoL care. As we discuss in detail later, practitioners nevertheless treat these activities as delicate.

Discussion about disease progression and EoL shares properties with talk about personal difficulties (or troubles-telling), which has been investigated in everyday, informal interactions (Jefferson, 1988). People treat personal troubles as a special topic; they do not introduce it abruptly and rather work to promote conversational contexts that are favorable to its articulation and reception (Jefferson, 1980). Similarly, patients and family members may allude to EoL matters rather than introducing them directly (Pino & Land, 2022). When patients articulate matters more explicitly, such as when requesting an estimate of their life expectancy, they can still exhibit caution, for example by channeling their requests through declarative rather than interrogative formats (an example of declarative is: “No one’s ever talked or said anything about the actual cancer […] how rapidly it grows and that sort of thing,” Pino & Parry, 2019; but see Anderson, 2020, pp. 82–88 for documented cases where relatives use interrogatives to ask about the life expectancy of a dying patient).

HCPs also treat discussion about the future as delicate. For example, it has been shown that HCPs ask about matters that bear a potential relationship to disease progression and EoL rather than introducing them directly (e.g., “They’re hoping to open up this sort of respite house called Humming Bird […] Have you heard about that at all?,” Ekberg et al., 2017, pp. 6–7; “Did she talk to you about the program called hospice?,” Lutfey & Maynard, 1998, p. 325). This gives patients and companions opportunities to discuss disease progression and EoL without making it a requirement (Ekberg et al., 2017). When patients and companions raise these matters, this emerges as something that they did by their own initiative (rather than the HCP’s initiative). Some practices that HCPs use to maximize this outcome have been documented, and one that presents important similarities to those we examine in this article is the use of a follow-up question.

HCPs use follow-up questions to invite patients to elaborate on something that they have raised in the current interaction (they work as “elaboration solicitations;” Pino et al., 2016). This practice, documented in several healthcare settings (Peräkylä, 1995; Pino et al., 2016; Shaw et al., 2017), enables HCPs to show attentiveness to patients’ talk whilst shaping the direction of the conversation. Follow-up questions are selective (Peräkylä, 1995) and can thus be designed to pick up on aspects of patients’ talk that can be related to disease progression and EoL based on context and shared knowledge (e.g., “And when the pain’s bad and you start to feel a bit panicky, can you remember what’s going through your mind at that time?” asked to follow up on a description of panic during episodes of breathlessness; Pino et al., 2016, p. 6). These follow-up questions are relatively open or unrestricted (Shaw et al., 2017) because they do not nominate specific issues on behalf of the patient (thus deferring to their primary right to articulate their own experience; see Pino & Land, 2022); importantly, they do not specify disease progression and EoL as foci. They nevertheless make their discussion possible because of their placement in contexts where patients or companions have raised matters that can be related to them. Furthermore, they make relevant talk that is specific to the patient’s own disease progression and EoL because they invite elaboration in terms of the patient’s subjective experience – their feelings or perspective (Pino et al., 2016; Shaw et al., 2017). Follow-up questions thus create opportunities to discuss disease progression and EoL without making it a requirement; patients and companions can relevantly respond by raising matters that are not future-related. Discussion about disease progression and EoL is made a matter of ‘opting-in’ rather than ‘opting-out’—privileging the patient’s (or companion’s) initiative. This practice is therefore also similar to perspective display solicitations (Maynard, 1992), by which HCPs invite a patient to articulate their understanding of a delicate matter before sharing their own.

In this article, we examine a communicative action that resembles follow-up questions: inviting the patient to talk about the contents of a conversation they previously had with another HCP, who is not present in the current interaction. Use of this action is mentioned in studies focusing on discussions about disease progression and EoL (Lutfey & Maynard, 1998; Pino & Parry, 2019), but its design and functions have never been made the primary focus of an investigation. We demonstrate that, like follow-up questions, invitations to talk about a previous conversation provide opportunities for patients to volunteer talk about disease progression. Because of the incurable, progressive nature of the patients’ conditions, discussion about disease progression concurrently makes EoL considerations relevant. We argue that this action reduces HCPs’ accountability, that is, the extent to which HCPs can be seen as responsible for unilaterally initiating discussions that might be undesirable or even emotionally harmful for patients (on accountability, see Robinson, 2016, especially pp. 12–13). We also show that HCPs invite patients to talk about a previous conversation to address discrepancies in knowledge and understanding of the patient’s prognosis – a matter to which we turn in the next section.

Discrepancies in knowledge and understanding in palliative care interactions

Patients commonly access palliative care after being diagnosed with an incurable disease. In principle, palliative care interactions thus happen in a context of “open death awareness” (Glaser & Strauss, 1964; Seale et al., 1997). In practice, though, the ways in which patients describe their situation can suggest they are in a different “awareness context” (Glaser & Strauss, 1964). For example, although a patient might understand their disease as incurable, they might treat their death as less imminent than an HCP does. Such inconsistencies in anticipated life expectancy can have ramifications for the patient’s preparedness for their deterioration, and the extent to which they see EoL discussions and decision-making as relevant. Addressing such discrepancies in knowledge and understanding is an important task in palliative care (Brighton & Bristowe, 2016). We will refer to these as epistemic discrepancies (on epistemics, Heritage, 2013).

Research on everyday social interactions outside health care provides a useful backdrop for analyzing epistemic discrepancies. This research has shown that people hold one another responsible for maintaining shared understandings of matters that are relevant to their social interactions (Heritage, 1984, p. 54). For example, people hold others accountable for failing to update them about matters that are consequential for their activities together (Smith, 2013). One type of epistemic discrepancy that is relevant for our investigation involves participants displaying different knowledge or understanding of a particular matter. One circumstance in which this happens is when a participant claims or implies that something is true, but this diverges from what a recipient otherwise knows (Deppermann & Reineke, 2020; Smith, 2013). Some ways in which recipients address this have been documented. For example, recipients can repeat parts of another’s turn that has evidenced an epistemic discrepancy, thus promoting talk that reconciles it (Robinson, 2009). Recipients can articulate the state of affairs that they hold to be true and that another participant’s talk has contradicted (e.g., “I thought you were going;” Smith, 2013, p. 319; see also Deppermann & Reineke, 2020). These practices make the other responsible for resolving the discrepancy on the basis that they have first-hand knowledge of the matter being discussed – they are epistemically responsible for it. These practices are comparatively direct in the ways in which they point out a discrepancy. By contrast, the HCPs in our study address discrepancies in subtler ways. They do not point out discrepancies directly; they rather promote talk that can lead to reconciling them through a series of steps – without making the HCP’s understanding of the patient’s prognosis explicit at first. This is consistent with the cautiousness generally exhibited by HCPs in raising EoL matters (see previous section).

In this article, we examine how palliative medicine doctors in a UK hospice setting invite patients to talk about a conversation they had with another HCP as a way of addressing epistemic discrepancies regarding the patient’s prognosis – specifically, in cases where the patient appears to treat their prognosis as being longer than the doctor does. Our study contributes to research on ways in which HCPs promote discussion about disease progression and EoL. Additionally, it is the first to document ways in which healthcare practitioners address epistemic discrepancies regarding the patient’s prognosis.

Materials and methods

Data for this study include 37 consultations, 34 video-recorded and three audio-recorded in 2014 at a large UK hospice, with ethical approval from NRES Committee West Midlands – Coventry & Warwickshire, UK (ref 14/WM/0128). The recordings involved 37 adult patients previously diagnosed with an incurable disease and having an outpatient or inpatient consultation with a doctor for review or management of symptoms (physical and/or emotional) and/or help with planning future care. The consultations involved 17 companions and five palliative medicine doctors. Patients and companions were invited to participate if they spoke English and if the patient was not in acute distress and had capacity to consent (as judged by the clinical team). All participants provided written informed consent to have the consultation recorded and for pseudonymised transcripts to be published. The hospice is an independent charitable organization providing services to the UK National Health Service; patients and their companions are not charged for those services.

Our method, conversation analysis (CA; Sidnell & Stivers, 2013), relies on recorded naturally occurring interactions to examine social actions that people use to accomplish activities and navigate practical problems. Analysis focuses on resources (vocal and bodily) that participants use to implement actions. The analyses we report are part of a broader study of how palliative care practitioners promote discussion about disease progression and EoL. Whilst mapping out ways in which they do so, we identified instances where they invite patients to talk about a conversation they had with another HCP. Three of the five participating doctors used this action in five consultations. Two were women, one was a man. The patients in these consultations were diagnosed with cancer (2), a rare lung disease (1), lymphoma (1), and motor neurone disease (1). One was a woman, four were men. They were aged 56–71. Two attended with a companion, three were unaccompanied. Four met the doctor as outpatients, one as an inpatient. We transcribed relevant parts using conventional procedures in CA (see Appendix), detailing aspects of speech delivery including timing and prosody (Jefferson, 2004), and visible conduct (Mondada, 2018); in this article, we limit representation of the latter to aspects that are consequential for our analyses to maximize readability.

In our analyses, we observed that the doctors not only invite patients to talk about a previous conversation to promote discussion about disease progression and EoL; they also use it to address a discrepancy between their own and the patient’s understanding of the prognosis. To support this claim, we used conventional procedures in CA, including qualitative analysis of the design of the doctors’ actions, their relationship with the immediate context of the interaction, and the patients’ responses. Attention to responses is a central validation procedure in CA (known as next-turn proof procedure; Sidnell, 2013). It enables researchers to support the claim that a participant’s turn implements a certain social action by showing that other participants, within the same interaction, observably treat that turn as indeed implementing that action. Using this procedure, we show that patients treat the doctors’ invitations to talk about a previous conversation as creating opportunities to raise disease progression and EoL in their responses.

It was nevertheless difficult to rely solely on the next-turn proof procedure to validate our claim that the doctors invite patients to talk about a previous conversation to address epistemic discrepancies. For example, they do not use this action immediately after an epistemic discrepancy has arguably emerged (but rather later), and they do not immediately make explicit that their understanding diverges from the patient’s. To support our claim, we thus drew on additional types of evidence. This includes the relationship between the doctor’s action and the broader context of the interaction; its relationship with aspects of knowledge that the participants arguably share; and its recurrence in extended sequences of talk where the doctor addresses epistemic discrepancies through several actions. For this reason, after using a concise example to demonstrate basic properties of the action in focus (Figure 1), we present several extracts from two additional consultations. These provide evidence that inviting the patient to talk about a previous conversation is the first in a series of attempts through which the doctors address an epistemic discrepancy in increasingly explicit ways. As a final note, CA is not an introspective method. We thus focus on knowledge and understandings that the participants display to one another through their actions; we do not make claims about what they might know or understand in an introspective sense. This approach enables us to ground our analysis of epistemic discrepancies in the ways in which the participants observably handle them.

Figure 1. VERDIS23 23,18 VT539 EL23.2 MP

Pat = patient (Seth). Doc = doctor (Michelle). Seth’s wife (Violet) and a healthcare assistant are also present. Audio only. 19 minutes into the consultation.

Analysis

We report our findings by examining three cases, selected because they exemplify recurring features in our collection. The first shows how inviting the patient to talk about a previous conversation creates an opportunity for discussion that focuses on their prognostic understanding and EoL awareness. The second and third cases show how the doctors use this action to address an epistemic discrepancy regarding the patient’s prognosis. In the extracts that follow, the doctors’ turns inviting the patient to talk about a previous conversation are highlighted in gray.

Inviting the patient to talk about a conversation with another HCP

Inviting the patient to talk about a previous conversation creates an opportunity for them to display their prognostic understanding and to engage in discussion about disease progression. Because of the incurable, progressive nature of the patients’ conditions, discussion about disease progression concurrently makes EoL considerations relevant. Figure 1 is from a consultation with Seth, a patient in his mid-sixties diagnosed with a rare, progressive lung disease. He is accompanied by his wife, and they are attending a first appointment with a doctor at the hospice. Before the extract in Figure 1, they have been discussing Seth’s cough (data not shown). The extract in Figure 1 begins with the doctor inviting Seth to talk about a conversation he had with another HCP (Dr Fairclough; lines 1–3).

The doctor’s turn in lines 1–3 of Figure 1 uses interrogative format (a grammatical design used for questions and recognizable by the presence of question words and subject-auxiliary inversion). It displays second-hand knowledge of a conversation Seth had with Dr Fairclough; it treats Seth as having first-hand knowledge of it and thus as having the ability to share its contents (Pomerantz, 1980). It does not specify a focus on disease progression. However, Seth responds by reporting that Dr Fairclough told him no curative treatments are possible (lines 4–6), and that his condition is progressive (lines 8 and 10–11). He thus treats the hospice doctor’s action as making discussion about prognosis a relevant possibility. Seth’s answer further links his prognostic understanding to EoL awareness, albeit allusively through the litotes “nothing can be done” at lines 5–6 (suggesting a state of affairs by negating the contrary; Lutfey & Maynard, 1998) and the figurative expression “it just will go […] downhill” (lines 8 and 10). This exemplifies how disease progression is linked to EoL considerations in our data.

The design of the doctor’s turn (lines 1–3) contributes to making discussion about prognosis possible. It contains two questions, one after the other (Raymond & Heritage, 2013). The first asks about what Dr Fairclough told Seth about the “problem” with his lungs. Given shared knowledge that this is an advanced progressive disease, the question can be heard as promoting discussion about its progression. The second question narrows the focus to Seth’s understanding (rather than a factual report) of those matters. It further raises the possibility that what Dr Fairclough said differs from what Seth understands. Also, the consultation occurs in a hospice setting, whose remit includes helping patients prepare for the EoL. This further contributes to hearing the doctor’s action as creating an opportunity to focus on Seth’s prognostic awareness. Importantly, though, the doctor’s action displays a degree of open-endedness. Although a focus on prognosis is made available as a response option through the aspects of turn design, shared knowledge, and context we have described, the doctor stops short of naming disease progression. Her action thus creates an opportunity for talk about prognosis and EoL rather than restricting Seth’s response to it.

Addressing discrepancies in knowledge and understanding

We have shown that inviting a patient to talk about a previous conversation creates an opportunity for talk focusing on the patient’s prognostic understanding and EoL awareness, which is an important task in palliative care. In this section, we examine how this also helps the doctors address an epistemic discrepancy regarding the patient’s prognosis.

Our second case involves Eashan, a patient in his fifties diagnosed with neck cancer and lung metastases, attending a hospice appointment with a doctor whom he has met before. He is accompanied by his brother Rajesh. From the start, Eashan updates the doctor about his deteriorating physical condition but also shares his hope to obtain curative treatment through a clinical trial. Throughout the consultation, there is evidence of a mismatch between Eashan’s and the doctor’s focus. The doctor asks questions that create opportunities to talk about disease progression, apparently to promote discussion about Eashan’s short prognosis and to facilitate reassessing the decision to join the clinical trial on this basis – something that the doctor later makes explicit. By contrast, Eashan maintains a focus on curative treatment, and this suggests that he might consider his prognosis as being longer than the doctor does. In this case, inviting the patient to talk about a previous conversation does not immediately lead to addressing this epistemic discrepancy, but it is the first in a series of actions through which the doctor attempts to do so (and eventually succeeds). We will show three extracts whilst summarizing other parts of the consultation. Figure 2 shows evidence of the epistemic discrepancy. Figure 3 shows the context leading to the doctor inviting Eashan to talk about a previous conversation. Figure 4 shows Eashan’s response.

Figure 2. VERDIS31 EL31.1 04.32 VT75 MP

Pat = patient (Eashan); Com = companion (Rajesh); Doc = doctor (Mick). Royal = hospital where the clinical trial takes place. 4 minutes into the consultation.

Figure 3. Approximately 21 minutes after Figure 2.

Figure 4. This extract continues from Figure 3. Lines 19-21 are repeated and renumbered from 01.

Before the extract in Figure 2, Eashan shared frustration over delays in being enrolled in the clinical trial. He added that his condition has recently deteriorated, and that he now has to use a wheelchair (data not shown), thus conveying the urgent need to join the trial (before his condition deteriorates further). In Figure 2, the doctor asks whether Eashan would feel well enough to go to the clinical trial site, which is in a different city.

The doctor’s hypothetical question in Figure 2 (lines 1 and 3–4) shows that his focus diverges from Eashan’s (on uses of hypothetical questions to raise delicate matters, see Parry et al., 2014). Rather than supporting Eashan’s project to join the trial, the question creates an opportunity to problematize it. Coming after discussion of Eashan’s deteriorating condition (before the extract in Figure 2, in data not shown), the question gives him an opportunity to conclude that he might not be fit to travel (on ways in which HCPs give patients opportunities to draw similar inferences, see Gill & Maynard, 1995; Lutfey & Maynard, 1998). Eashan nevertheless declares that he is fit enough (lines 12–13). With the “oh” preface (line 12) he frames the matter as a given, thus treating the doctor’s question as inapposite (Heritage, 1998). Rajesh further voices Eashan’s position (line 14) but notably, he focuses on Eashan’s determination to join the trial, thus eschewing the matter of his fitness.

Eashan’s assertion in Figure 2 (lines 12–13) can be heard as overestimating his ability to receive curative treatment in the context of severe deterioration and thus as underestimating the severity of his prognosis (later in the consultation, the doctor will estimate Eashan’s life expectancy in terms of months). This raises the possibility of an epistemic discrepancy, which presents the doctor with a practical problem. As a specialist in palliative care, his role includes helping patients weigh the potentially slim benefits of aggressive treatment against risks associated with it, particularly in terms of reduced quality of life. However, raising this could be treated as destroying hope, especially in the context of Eashan’s focus on survival. We will show that the doctor handles this without making explicit his understanding of Eashan’s prognosis at first (he will do so later). Rather, he creates opportunities for Eashan to take the initiative in showing prognostic awareness. Inviting him to talk about a previous conversation is the first step in this process. Discussion following the extract in Figure 2 focuses on problems Eashan is experiencing. Just before the extract in Figure 3, the doctor asks about his loss of appetite, and Eashan responds that this started after the onset of a persistent cough (data not shown). As the extract in Figure 3 begins, Eashan is sharing that the cough scared him (line 1).

Eashan alludes to the nature of his fear by associating it with something he experienced in the past, when he started coughing and was later diagnosed with lung metastases (lines 4–5 and 9–10). He thus implies that his present cough has led him to fear that his cancer has spread further. The doctor acknowledges this whilst preserving the allusiveness (“you made that connection,” line 13). Eashan further reiterates the present relevance of these concerns by reporting that he is coughing (line 16). His further report of being “too scared” likely circles back to his reference to loss of appetite (data not shown), thus referring to being too scared to eat (line 18).

The doctor subsequently invites Eashan to talk about a conversation he had with an oncologist, Dr Walton (lines 19 and 21), who had been mentioned earlier (data not shown). The doctor’s turn uses an interrogative format. He treats Eashan as someone who (unlike him) has first-hand knowledge of the conversation and can thus report its contents. The question is not marked as initiating a new topic and can therefore be understood as related to Eashan’s immediately preceding talk (line 18). The question (lines 19 and 21) treats Eashan’s reference to being scared as alluding to concerns about disease progression. Indeed, in the context of those concerns, asking about the oncologist’s report on the X-ray creates an opportunity to focus on any prognostic matters discussed in that meeting.¹ Because of the incurable, progressive nature of Eashan’s condition, this could lead to discussion about EoL matters. The question also creates the possibility of reducing the epistemic discrepancy which emerged in the extract in Figure 2 because it gives Eashan a chance to display prognostic awareness. The doctor nevertheless does not specify prognosis as a focus, thus leaving this to Eashan’s initiative. This concurrently reduces the doctor’s accountability for introducing such delicate matters. Figure 4 shows Eashan’s response.

In Figure 4, Eashan starts to respond but displays difficulty through a word search (“he said there’s this,” lines 4–5; Lerner, 2013). Rajesh comes in to assist (line 6) by reporting that Dr Walton diagnosed Eashan with a “rough cough,” which Eashan confirms (line 7). This collaboratively constructed answer thus focuses on symptom management whilst foregoing discussion of disease progression. Consistent with this, Eashan goes on to report on Dr Walton’s prescription of radiotherapy as a way of easing his cough (lines 8–20). He even concludes on a positive note, saying that he has been able to tolerate the radiotherapy (lines 30–32). This contrasts with earlier discussion about deterioration, but it lends support to his earlier claim that he is well enough to join the trial (see Figure 2).

Although in this case the patient’s response does not address the epistemic discrepancy, there is other evidence that the doctor’s question was designed to create an opportunity for this. He pursues the matter through an additional question (lines 41–42). This makes it possible to retrospectively see the doctor’s earlier invitation to talk about a previous conversation as the first in a series of increasingly explicit attempts to focus on Eashan’s disease progression. The doctor first summarizes Eashan’s talk thus far in terms of his deteriorating health (lines 37–38) and his wait for the clinical trial (lines 38–39). He then invites Eashan’s considerations about what “might be coming next” (lines 41–42). Placed after the reference to deterioration, this question can be heard as alluding to disease progression. It thus promotes talk about the future more directly than the question about the conversation with Dr Walton. It still stops short of suggesting specific concerns, thus leaving discussion about disease progression to Eashan’s initiative. However, in subsequent interaction not shown here, Eashan’s answer does not focus on disease progression. Rather, he shares frustration with not having a scheduled appointment with Dr Walton (thus conveying that for him this is what should “happen next”). This answer once again embodies a focus on curative treatment and thus creates an unfavorable environment for discussing disease progression (Maynard, 1996). The doctor nevertheless pursues this with an additional question, which further supports our claim that his invitation to talk about a previous conversation was the first in a series of attempts to promote reconciliation of the epistemic discrepancy. Due to space constraints, we only summarize these developments.

After the extract in Figure 4, in data not shown, the doctor asks, “So that’s one thing […] hoping that the Royal will have treatment that’ll work […] Do you ever wonder what will happen if they don’t have treatment that works?.” This question contrasts with the previous ones by suggesting a possible concern with disease progression. Eashan responds with a statement of awareness that he cannot be offered any more treatment (except for the trial) and shares that sometimes he worries that he might die. This further exemplifies how talk about disease progression is linked to EoL considerations in our data. There is evidence that the doctor treats this as a display of prognostic awareness and as reducing the epistemic discrepancy because he next promotes decision-making about EoL care. This later leads to offering Eashan the option to come to the hospice as an in-patient when his condition deteriorates further, to which Eashan expresses some openness. Eashan later asks the doctor whether he thinks his cancer can still be treated or whether he is “just counting days.” The doctor responds that there is only a small chance of the clinical trial helping, and that the decision to join it should be weighed against the effort required to travel there. He then estimates Eashan’s life expectancy in terms of months. This confirms (albeit retrospectively) that the doctor’s understanding of Eashan’s prognosis diverges from the one embodied in Eashan’s focus on treatment. It is notable that the doctor voices this understanding after the epistemic discrepancy is reduced and Eashan’s question about life expectancy further licenses him to provide such sensitive information.

To summarize, Eashan’s focus on curative treatment suggested a discrepancy in the understanding of his prognosis. His focus on survival further created an unfavorable context for discussing disease progression. The doctor’s solution was to create opportunities for Eashan to take the initiative in raising this. Inviting Eashan to talk about a conversation with an oncologist gave him an opportunity to share his understanding of his condition in a way that could facilitate reconciliation of the epistemic discrepancy. Whilst it did not achieve this initially, it was the first in a series of actions that increasingly narrowed the focus to disease progression, eventually leading to EoL discussion and decision-making.

Addressing an epistemic discrepancy emerging from outside the current interaction

We end with a case that differs from the others in our collection. First, evidence of an epistemic discrepancy does not initially emerge in the consultation but from information the doctor had gleaned before it. Second, the doctor uses a declarative rather than an interrogative format to invite talk about a previous conversation. Third, he invites talk about what the patient told the other HCP (rather than what the HCP told the patient). Fourth, the doctor specifies EoL as a focus. Despite these differences, this case supports our claim that inviting a patient to talk about a previous conversation is a way of addressing an epistemic discrepancy regarding disease progression.

The patient, Lynn, is in her early seventies and lives with frailty and multiple co-morbidities secondary to radiotherapy for lymphoma, which first occurred fifty years before. At the start of the consultation, she reports that her condition has deteriorated. This includes breathing difficulties leading to a hospitalization, and reduced mobility (data not shown). We present four extracts from this consultation whilst summarizing other parts. Figures 5 and 6 show the doctor’s invitation to talk about a previous conversation and Lynn’s response. Figures 7 and 8 show later parts where the doctor goes back to matters relating to Lynn’s prognosis; this provides evidence that inviting Lynn to talk about a previous conversation is the first in a series of actions designed to address an epistemic discrepancy.

Figure 5. VERDIS16 00,01 VT1 EL16.1 MP

Pat = patient (Lynn). Doc = doctor (Mick). 5 minutes into the consultation.

Figure 6. Approximately 2 minutes extract after Figure 5. 7.50 minutes into the consultation.

Figure 7. 14 minutes into the same consultation as Figures 5 and 6.

Figure 8. 15 minutes into the same consultation as Figures 5, 6 and 7.

Just before the extract in Figure 5, Lynn has reported that her mood has recently dropped (data not shown). In Figure 5, the doctor invites Lynn to elaborate on her report of a mood drop (lines 1–2). Whilst proposing that this could be related to her breathing difficulties, the question also gives her an opportunity to link it to other matters (“something else,” line 2). In response, Lynn links her mood drop to “fear” (lines 3–4). After the doctor’s initiation of repair seeking confirmation (line 6; Schegloff et al., 1977), Lynn confirms (“Just fear,” line 8) and, after a silence (line 9), further elaborates on the nature of that fear (“the reality of knowing what’s happening and […] not being able to do anything about it,” lines 10–11 and 13).

The doctor then invites Lynn to talk about a conversation with another doctor (lines 15–19). He does not use interrogative but rather declarative (i.e., statement) format. He mentions a comment he read in a letter from Lynn’s heart doctor, reporting that Lynn had not felt ready to discuss EoL issues (lines 15–19). By raising this now, the doctor implicitly treats Lynn’s prior reference to “knowing what’s happening” (lines 10–11) as alluding to disease progression and EoL and as possibly evidencing prognostic awareness – and thus (similar to the extract in Figure 3) as an opportunity to engage in more discussion about these matters. By contrast, the heart doctor’s report that Lynn had not felt “ready” to discuss EoL issues raises other possibilities, for example, that she might consider her prognosis as being longer than the doctors do (the hospice doctor will make explicit his understanding of Lynn’s prognosis later). By raising this, the hospice doctor creates an opportunity to address this possible epistemic discrepancy. The doctor’s declarative (lines 15–19) displays that his knowledge of the previous conversation is second-hand and limited (Pomerantz, 1980). It implicitly invites Lynn to elaborate on the basis that she has first-hand knowledge of it (Pomerantz, 1980). Additionally, it frames the heart doctor’s report as puzzling by mentioning that Lynn had previously discussed EoL issues at the hospice (lines 17–18) and thus makes relevant for Lynn to provide an account (Smith, 2013).

After a silence (line 20), the doctor pursues a response (Pomerantz, 1984) by seeking confirmation that Lynn recognizes the events whilst allowing for the possibility that the heart doctor might be wrong (lines 21–22). Following another silence (line 23), the doctor further invites an account by again mentioning his contrasting experience of Lynn discussing EoL matters with him previously (Smith, 2013), although he frames this cautiously as something that was discussed “a little bit” and “a long time ago” (lines 24–26). Lynn responds by reporting that she does not remember (lines 27–28). After the extract in Figure 5, she asks the doctor to check the letter. He opens the letter on his computer and reads it out (data not shown). Following that, in Figure 6, Lynn displays that she now remembers (“Oh yes,” line 1) and proceeds to report what had happened with the heart doctor.

In Figure 6, Lynn shares what she told the heart doctor (lines 1 and 3–5). This demonstrates awareness that she is “terminally ill” (line 4) and that her condition cannot improve (line 5). With this, she treats the heart doctor’s letter (and possibly the hospice doctor’s mention of it) as reflecting the concern that she might not be fully aware of her prognosis. Her report thus appears designed to reassure. However, she proceeds to report that she did not feel “in imminent danger of dying” (lines 8 and 9). She contrasts her condition to cancer (line 12), describing herself as less at risk of dying comparatively (lines 14–16). Lynn provides this as an account for her alleged lack of engagement in discussion about EoL with the heart doctor. She further displays awareness that she could die (of an infection or heart failure, lines 19–23) but again articulates the understanding that she could survive for an indefinite amount of time (alluded to with the idiomatic “open ended situation,” lines 26–27 and 30–31).

Inviting Lynn to talk about the previous conversation has brought her prognostic understanding into view. However, her response in Figure 6 provides further evidence of an epistemic discrepancy (rather than resolving it). It provides a complex picture, in which Lynn sees herself as terminally ill, but not in danger of dying imminently. We now examine how the hospice doctor manages this in later parts of the consultation. These additional analyses further evidence that inviting Lynn to talk about a previous conversation is the first in a series of attempts at addressing an epistemic discrepancy. After the extract in Figure 6, in data not shown, Lynn talks about how she has learned to live with her conditions and not worry too much about dying. The doctor later asks about Lynn’s earlier reference to fear (raised in the extract in Figure 5), and Lynn shares that recently she has been more frightened of dying. After more discussion about this, the doctor asks about the mood drop she mentioned previously (back in the extract in Figure 5). Just before the extract in Figure 7, Lynn states that she does not need anti-depressants, although she again alludes to having been less able to cope lately (data not shown). In Figure 7, the doctor directly addresses the epistemic discrepancy evidenced earlier in the extract in Figure 6.

In Figure 7, the doctor brings the focus back to Lynn’s prognosis (lines 1–6) by mentioning other HCPs’ perspectives on it. He builds a contrast between the other HCPs’ view that Lynn could die soon (lines 4–6) and Lynn’s own contrary view (lines 8–9), which she voiced in the extract in Figure 6. After Lynn confirms this view (line 10), the doctor contrasts it with Lynn’s own displayed awareness (also in Figure 6) that she could die of an infection or heart failure (lines 11–12, 13, and 15). By calling this “realistic” (line 11), the doctor implicitly supports this understanding. The doctor thus selectively formulates (Heritage & Watson, 1979) and reinforces aspects of Lynn’s earlier talk that embody awareness of the risk of dying. This both shows that he supports a different understanding to Lynn’s, and that he actively works to promote reconciliation between the two. He simultaneously exhibits caution by attributing that understanding to other HCPs, and by reinforcing aspects of Lynn’s own earlier talk that also support it, rather than directly supporting it himself.

Lynn confirms (line 14) and starts a story (16 and 18–19) about being admitted to hospital to treat a chest infection (not shown in full here). Because of its placement and framing (see the use of “Because,” line 16) it can be heard as further confirming Lynn’s awareness that she could die from an infection. This is further embodied in her mention of being told by the hospital doctors that she is “a woman standing on the edge of the cliff and just fell over.” However, the story also shows Lynn’s focus on survival. Indeed, she emphasizes that she now keeps antibiotics at home, should another infection happen (data not shown). Once again, Lynn’s talk embodies a complex understanding in which awareness of the risk of dying is accompanied by a focus on survival. We re-join the interaction in Figure 8 The start of the extract shows the end of the story about the hospital admission.

In Figure 8, the doctor makes explicit his understanding of Lynn’s prognosis for the first time in the consultation (lines 6–7). He acknowledges Lynn’s focus on coping (lines 10–11), which has been a major theme in her talk. He then contrasts this with what he sees as a risk that Lynn could die (lines 14–15). He proposes that given her frailty, a complication (like the chest infections they talked about) could kill her. He alludes to this by using the metaphor of the cliff, which the hospital doctors reportedly used with Lynn (data not shown). Lynn promptly confirms (line 16). This works to bring the doctor’s and Lynn’s understanding and focus into alignment, thus reducing the epistemic discrepancy. Subsequent discussion, not shown here, further focuses on Lynn’s awareness of dying and plans for her EoL care.

To summarize, the doctor did not make explicit his understanding of Lynn’s prognosis straight away. Rather, he gave multiple opportunities for Lynn to articulate her own understanding (a pattern similar to perspective display solicitations; Maynard, 1992). Lynn responded by expressing a complex understanding, showing awareness of a terminal diagnosis accompanied by an expectation that she is unlikely to die soon. The doctor selectively built on and reinforced parts of Lynn’s talk that embodied awareness of the risk of dying, and eventually made explicit his understanding of Lynn’s prognosis in Figure 8. Inviting Lynn to talk about a previous conversation was the first in a series of actions to address the epistemic discrepancy regarding Lynn’s prognosis.

Discussion and conclusion

This article has examined a communicative action whereby palliative medicine doctors in a UK hospice invite patients to talk about a conversation they had with another HCP. This action creates an opportunity for discussion about disease progression and EoL. In this section, we compare the action we examined with others that promote these types of discussion. We then explore implications for understandings of HCP accountability, epistemic discrepancies in healthcare interactions, and palliative care practice. We also discuss some methodological implications of our study.

Inviting the patient to talk about a conversation they had with another HCP shares properties with follow-up questions that promote elaboration of something a patient (or companion) has raised in the present interaction (Pino et al., 2016). The difference is that invitations to talk about a previous conversation promote elaboration of matters the patient had discussed with a different HCP outside the current interaction. Despite this difference, one feature that all these actions share, is that they favor the patient’s initiative in raising disease progression. Our findings thus resonate with Drew’s (2022) proposal that in many social interactions doing something voluntarily is valued over the alternative of being directed to do it; in our cases, the doctors’ actions afford patients opportunities to volunteer discussion about disease progression without inviting this directly. This is because the doctors’ invitations to talk about a previous conversation usually do not specify disease progression as a focus (see Figure 5 for a partial exception to this pattern). This is nevertheless made a relevant possibility through the design of this action (mentioning matters that bear a potential relationship to disease progression such as the X-ray in Figure 3), its use in contexts where the patient has alluded to concerns about disease progression (Figures 3 and 5), and because the conversation the patient had with another HCP was one where such matters are likely to have been discussed. Additionally, the invitation is to talk about a conversation of which the patient (unlike the hospice doctor) has first-hand knowledge (Bergmann, 1992; Pino, 2016; Pomerantz, 1980). These features give the patient latitude to respond in different ways, including the option of not raising disease progression (as happens in Figure 4). Inviting the patient to talk about a previous conversation thus resembles practices of “topic shading” documented in pediatric palliative care (Ekberg et al., 2017); these promote discussion about matters that can be related to deterioration but without specifying it as a topic. Similarly, in a study of oncology consultations, Lutfey and Maynard (1998) showed an instance of an oncologist asking a patient about a previous conversation they had with another HCP (“Did she talk to you about the program called Hospice?,” p. 325), and noted that with this, “the physician introduces a topic associated with death and dying” (p. 325; emphasis in the original). Inviting the patient to talk about a previous conversation thus has a “fishing” quality (Pomerantz, 1980)—it creates an opportunity for the patient to raise disease progression without inviting this specifically. When disease progression is raised, discussion about EoL issues is also made relevant because of the incurable, progressive nature of the patients’ conditions.

Inviting the patient to talk about a previous conversation further enables HCPs to manage their accountability in promoting discussion about disease progression and EoL. Because this action makes such discussion a matter of ‘opting in’ (rather than ‘opting out’) HCPs cannot be seen as responsible for unilaterally introducing a topic that some patients may find distressing. Compared to other practices that facilitate a similar outcome (e.g., follow-up questions), inviting the patient to talk about a previous conversation presents special affordances: if difficult matters (such as a short prognosis) were discussed with the other HCP, and the patient reports them in the present interaction, then the hospice doctor does not need to introduce them. The doctor can design their next contributions as further elaborations of what had already been discussed with the other HCP. The action we examined thus shares properties with perspective display solicitations (Maynard, 1992), with which HCPs invite a patient to articulate their understanding of a delicate matter before sharing their own. Inviting the patient to talk about a previous conversation is therefore one of several actions that promote favorable interactional environments for discussions about disease progression and EoL (Maynard, 1996). These findings resonate with Pino and Parry’s (2019) study, which identified invitations to talk about a previous conversation in a different context of activity from the one examined here: following patient requests for a life expectancy estimate. In that context, the invitation to talk about a previous conversation promotes articulation of what the patient already knows as a preliminary step to delivering prognostic information. This strengthens our finding that inviting the patient to talk about a previous conversation affords patients opportunities to articulate their prognostic understandings.

Another contribution of our study is to show how HCPs can address epistemic discrepancies regarding the patient’s disease progression. Previous studies examined how speakers work to resolve epistemic discrepancies in everyday, informal interactions. To our knowledge, ours is the first to document how HCPs do so in healthcare interactions. There are important differences between the action we have examined and those documented in other studies, such as the practice of articulating the state of affairs that a participant held to be true and that another participant’s talk has just contradicted (e.g., “I thought you were going,” Smith, 2013, p. 319). The latter is explicit in pointing out a discrepancy and attributes responsibility for it to the recipient. The action we have examined does not directly point out the discrepancy, and it does not treat the patient as responsible for it. This makes sense in the palliative care context, where HCPs’ actions commonly embody cautiousness and sensitivity in raising delicate matters such as disease progression. Invitations to talk about a previous conversation create opportunities to resolve an epistemic discrepancy by giving patients a chance to show prognostic awareness. As such, our findings have parallels with Shaw et al.’s (2019) finding that therapists use questions to afford cancer patients opportunities to explore alternative understandings of their disease progression (rather than challenging their perspectives directly).

The indirectness embodied in the doctors’ actions in our study has raised methodological complexities in demonstrating that they create opportunities to address epistemic discrepancies. This is because the doctors do not invite patients to talk about a previous conversation immediately after an epistemic discrepancy has been evidenced (e.g., Figure 2) but rather later. Additionally, the doctors do not immediately make explicit that their understanding of the patient’s prognosis diverges from the patient’s. Our position is that these complexities do not suggest that the action we have examined is a fuzzy phenomenon. Rather, we argue that the indirectness of HCPs’ actions is intrinsic to palliative care (and discussions about disease progression and EoL more broadly); neglecting to examine them would fail to capture essential aspects of these settings. We supported our analyses by drawing on several types of evidence including the design of the action; its relationship with the broader context of the consultation and with knowledge that the participants arguably share; and its recurrence in extended sequences of talk where doctors address epistemic discrepancies through several actions. Our study thus joins a tradition of studies examining indirectness in discussions about disease progression and EoL (Ekberg et al., 2019; Lutfey & Maynard, 1998).

Our findings have practical implications for palliative care. Earlier research suggests that patients’ lack of understanding of a prognosis limits the extent to which they can make informed decisions about treatment (Ghandourh, 2016). It can also lead patients with short prognoses to still seek aggressive treatments, which can reduce quality of life and cause carer burden and financial costs (Ghandourh, 2016). Addressing displayed discrepancies in knowledge and understanding of a prognosis is thus important in palliative care. The cases we examined reinforce this conclusion, because we see doctors proceeding to promote planning of EoL care only after the epistemic discrepancy has been reduced (see our second and third cases). Our analyses nevertheless suggest that detecting and addressing epistemic discrepancies is not a straightforward job. The patients in our data sometimes conveyed complex understandings in which awareness of dying was accompanied by a focus on survival. This meant that their divergent understanding was evidenced in subtle ways, such as when they maintained a focus on curative treatment. Addressing these understandings is a complex task, which arguably touches a paradox at the core of palliative care: helping patients affirm life whilst preparing for death (Bye, 1998). Within this framework, HCPs should help patients gain realistic expectations whilst avoiding destroying hope (an orientation to hope was strongly embodied in both Eashan’s and Lynn’s talk in the cases examined in this article). As noted in our introduction, these are some of the challenges that HCPs reportedly experience in addressing disease progression and EoL. The contribution of our study is to document a communicative action that HCPs can use to navigate these complexities. By affording patients opportunities to go first in engaging in discussion about disease progression and EoL, HCPs can show sensitivity to the emotional consequences of establishing that the patient’s prognosis is shorter than they apparently make out.

Some limitations are worth mentioning. The action of inviting the patient to talk about a previous conversation is arguably not always available as a resource. Its use is contingent on the HCP knowing that the patient has had a conversation with another HCP where disease progression is likely to have been discussed. The action we have documented should be seen as part of a larger toolkit (for a review see Ekberg et al., 2021). Additionally, our analyses show that the action is not always successful in promoting discussion that addresses the epistemic discrepancy (Figure 3 is instructive in this sense). This makes sense because the action is designed to create an opportunity for discussion about disease progression rather than making it a requirement. Because the action does not specify disease progression as a focus, it might also present some ambiguity for patients. Our analyses have nevertheless shown how HCPs can create additional opportunities through subsequent questions that increasingly narrow the focus to disease progression and EoL.

Some limitations to the transferability of our findings include the Western context of the study and a relatively small set of recordings from one large hospice. We only identified six instances of the action we examined, raising the possibility that we were unable to exhaustively identify its properties. Additional research is needed to examine the distinctive affordances of the two linguistic formats (interrogative and declarative) that we identified. Finally, it would be important to establish whether and how this action is used in interactions with patients at earlier stages of their healthcare journey. We have nevertheless advanced understandings of a central area of palliative care practice (addressing epistemic discrepancies regarding the patient’s prognosis) and presented analyses that further illuminate key tensions in the promotion of discussion about disease progression and EoL. Ultimately, we have shown that expert communication practice entails sensitive management of patients’ displayed states of awareness regarding their disease progression.

Acknowledgment

We would like to thank the staff, patients, and their companions who consented to being recorded. We are grateful to Ruth Parry for her overarching work in establishing and leading the VERDIS project and building and supporting the team that made this paper possible. We are also indebted to Ruth for her invaluable input on the analyses leading up to this paper. We thank two anonymous reviewers for their detailed and helpful comments on an earlier version of this article. Our gratitude also goes to Elizabeth Jenkins, whose insightful clinical perspectives greatly improved the clarity of this paper.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

Part of this work was funded by The Health Foundation [Grant IDs: AIMS 1273316, AIMS 1273743/GIFTS 7210], an independent charity committed to bringing about better health and social care for people in the UK, and by the National Institute for Health Research Academy Career Development Fellowship to project lead Ruth Parry: ‘Enhancing staff-patient communication in palliative and end of life care’ (Grant ID CDF-2014-07-046). The views expressed are those of the authors and not necessarily those of the NIHR or the Department of Health and Social Care.

Notes

1. This may also explain why the doctor did not do so right after the epistemic discrepancy emerged in Figure 2; the interactional environment in Figure 3, after Eashan has shown some orientation to concerns about disease progression, is more favorable.

Appendix

Transcription conventions adapted from Jefferson (2004):

	=	Slightly upward intonation
¿	=	Upward intonation (more marked than a comma but less than a question mark)
?	=	Upward intonation
	=	Falling intonation
_	=	Level intonation
[	=	Overlapping talk begins
]	=	Overlapping talk ends
(0.8)	=	Silence in tenths of a second
(.)	=	Silence less than two-tenths of a second
wo:::rd	=	Lengthening of the sound just preceding
wo-	=	Abrupt cut-off or self-interruption of the sound in progress
word	=	Stress or emphasis (usually conveyed through slightly rising intonation)
↑ ↓	=	Marked pitch rise or fall
=	=	Latching
( )	=	Talk too obscure to transcribe
(word)	=	Best estimate of what is being said
hhh	=	Hearable out-breath
.hhh	=	Hearable in-breath
h	=	An italicised h at the end of a word indicates an aspiration sound
((words))	=	Transcriber comments
WORD	=	Louder talk
°word°	=	Quieter or softer talk
>word<	=	Faster or rushed talk
<word>	=	Slower talk
#word#	=	Talk with a creaky voice quality

Conventions for the transcription of visible actions adapted from Mondada (Mondada, 2018):

% %	=	Descriptions of visible action are delimited between
+ +	=	two identical symbols (one symbol per participant’s line of action) and are synchronized with corresponding stretches of talk/lapses of time
*—>	=	The action described continues across subsequent lines
—>*	=	until the same symbol is reached.
>>	=	The action described begins before the extract’s beginning
–>>	=	The action described continues after the extract’s end
……	=	Action preparation
——	=	Full extension of the action is reached and maintained
pat	=	Participant doing the embodied action is indicated in lower case when they are not the speaker