Creating space to discuss end-of-life issues in cancer care

Abstract

Objectives:

Analyze entire oncology clinical visits and examine instances in which oncologists have to break the bad news that patients’ treatments are no longer effective.

Methods:

Using conversation analysis we examine 128 audio recorded conversations between terminal cancer patients, their caregivers, and oncologists.

Results:

When oncologists break the bad news that a patient’s treatment is no longer effective, they often use a conversational device we call an “exhausted current treatment” (ECT) statement, which avoids discussing prognosis in favor of further discussing treatment options. Analysis suggests that improving and prioritizing patient-centered care and shared decision making is possible if we first understand the social organization of clinical visits.

Conclusions:

ECT statements and their movement towards discussing treatment options means that opportunities are bypassed for patients and caregivers to process or discuss scan results, and their prognostic implications.

Practice Implications:

When oncologists and patients, by fixating on treatment options, bypass opportunities to discuss the meaning of scan results, they fail to realize other goals associated with prognostic awareness. Talking about what scans mean may add minutes to that part of the clinic visit, but can create efficiencies that conserve overall time. We recommend that oncologists, after delivering scan news, ask, “Would you like discuss what this means?”.

1. Introduction

As patient-centered care¹ becomes routine in the provision of modern, quality healthcare [1–5], shared decision making (SDM) and patient-centered communication (PCC) are promoted as key vehicles for meeting the standard. [6–8]. As a consequence, attention has increasingly been paid to the patient-physician interactions, to ensure that a patient’s voice, agency, and rights are considered. Despite the concerns, this ideological shift has met major impediments [9] and is unrealized in routine practice. Some researchers argue that patients may not want their physician to burden them with complete transparency or the onus of making medical decisions [10,11 ]. Not all patients value autonomy and self-determination [12,13]–the cornerstone of patient-centered care. Researchers and practitioners call for a more adaptive and nuanced approach to patient-doctor interactions [14–17], including respect for different styles of decision-making [16,18]. All this points to a clear need for further research on patient-centered care.

Conversation Analysis (CA) has contributed to this research by its focus on the structural components of patient-doctor interactions [19]. Researchers specifically concentrate on the interactional devices participants use to generate and react to treatment recommendations “across a range” of clinical settings [20]. For example, Bergen and Stivers [21] demonstrate the importance that disclosure plays when patients attempt to bring up specific topics and negotiate particular treatment options. Gill [22] shows how patients are capable of exerting pressure on clinicians for diagnostic tests that would make a particular treatment relevant. Lindstrom and Weatherall [23] show that physicians lexically shape their treatment proposals to favor their own knowledge and preferences, while limiting patients’ responsive expressions. Koenig [24] establishes how a patient’s silence can introduce resistance and agency into a medical encounter. Stivers et al. put forth a basic taxonomy of the treatment recommendations in primary care: pronouncements, suggestions, proposals, offers and assertions all of which “represent different social actions” [25]. This research is informative along a number of dimensions, but none has addressed the topics of this paper, which include (a) a specific disease (lung cancer), (b) a particular period in the treatment of the disease, namely when the treatment so far has stopped working, and (c) how oncologists immediately address this matter as bad news and as a way of formulating a different treatment recommendation.

Our main phenomenon is what we call an exhausted current treatment (ECT) statement, which embodies two interrelated social actions: 1) breaking the bad “scan” news that a treatment protocol is no longer effective, and 2) discussing different options through bypassing or shrouding explicit discussion of this bad news. We show how clinicians, by truncating the explanation of diagnostic images and quickly transitioning to treatment options, miss opportunities to elicit and address patients’ (and caregivers’) understanding of the current state of the disease. Patients and caregivers, through silence and other displays of passive resistance, contribute to such missed opportunities. While passive resistance may lead to more information about the scan results, the implicit nature of the action (i.e., silence) is left for the clinicians to decipher. Even though more information of scans might be transmitted, the bad news aspect of that information is likely to produce another ECT or a quick exit [26], which diminishes the possibility of discussing what the patient understands about the new prognostic information. These missed opportunities are co-constructed–i.e., products of how scan results are delivered and received, and the collaborative way that ECT statements manage a quick pivot away from further discussion of scan results. These findings have implications for patient-centered care, which we address in the discussion section.

2. Data and methods

Our data were selected from 128 audio recorded conversations between cancer patients, their caregivers, and oncologists, which are drawn from a larger study that evaluated the effects of an internet-based support system for patients and caregivers with various forms of cancers. The Comprehensive Health Enhancement Support System (CHESS) study [27] was a multi-site, non-blinded randomized trial of lung cancer patients and caregiver dyads recruited from four cancer center hospitals in the East, Midwest, and Southwest US between September 2004 and May 2009. Our main eligibility criteria were patients over the age of 18 with locally advanced stage IIIA, stage IIIB, and metastatic stage IV non-small cell lung cancer (NSCLC), which we refer to as lung cancer hereafter. For more details about the data, see Refs. [28] and [29].

Consistent with the CA approach [30–32], including its use in medical settings [33,34], our study utilized the audio recordings to capture detailed interactional features and identify recurring patterns [28]. After listening and coding 128 conversations, we excluded any conversations that were devoid of scan results. Due to the intensive nature of CA transcription², we further reduced this number of cases to 50%, which left us with 64 conversations. After using conversation analytic conventions to transcribe and analyze these 64 conversations, we had 22 recordings in which scan results showed tumor growth or metastasis. ECT statements were present in approximately one-third (7) of those “bad” news cases. This was the most consistent way of exiting bad news. Other forms of closing out bad news were varied and had very little uniformity. Although this is a small number by traditional research standards, CA works with collections rather than samples, and the procedures of analysis ensure the possibility of generalizing as researchers focus on practices of talk that are both context-free and context- sensitive [35]. We do not mean that our findings generalize to all oncology visits, but that, when a specialist is presenting the bad news of treatment ineffectiveness using an ECT statement, it can forestall what the statement means in terms of end-of-life discussion in favor of raising the possibility of different treatment modalities. What we identify below as so-prefacing or upshot statements are mechanisms that can generalize to similar contexts in which physicians (with at least partial encouragement from patients) bridge to the good news of other therapies.

3. Findings

Past research shows that medical interviews in primary care have an ordered structure of component activities [19,33,36,37]. Keeping with that tradition, our analysis suggests that medical visits in lung cancer clinics tend to follow a similar order [29]. After the opening, an initial phase involves the reporting of (or the inquiry about) the patient’s current or recent symptoms. During this “symptom-talk,” patients and caregivers report the physical and psychological difficulties the patient is exhibiting, whether from the cancer or side effects from its treatment. Next is the clinician’s reporting and explanation of the most recent diagnostic images–magnetic resonance imaging (MRI), positron emission tomography (PET), and/or computed tomography (CT) scans–during a phase that we call “scan-talk.” Then, the conversation shifts to “treatment-talk,” when clinicians recommend relevant options (which may only suggest staying with the current plan). Another phase is “logistic-talk,” when arrangements are made for further visits or scans or chemotherapy, and the like³. Finally, there is closure to the visit. This phase structure appears often in, and across, numerous oncological visits regardless of the types of scan news being delivered or treatment recommendations made.

Exhausted current treatment (ECT) statements are utterances in which clinicians tell patients and the caregiver(s) that the chemotherapy or other treatment is no longer affecting the cancer. Examples of ECT statements include: “I don’t think this treatment is going to cut it,” “we have gotten the most we are going to out of this drug,” “not seeing any further improvement using this drug,” and “I think we’ve gotten as much as we are going to get from this treatment.” As we will show, there are three features to ECT statements and their aftermath: 1) they are formed as bad news deliveries; 2) they elicit little or no vocalized response from patients and caregivers; and, 3) they truncate and/or bypass explicit discussion of scan results, regularly using so-prefaced upshots, which are utterances that start with the particle “so” and provide the main point that the speaker is suggesting the recipient (s) should understand from the previous talk. ECT statements regularly occur just after scan-talk and lead into treatment-talk. In the next section we discuss two cases, which exhibit the patterns and are representative of other cases in which ECT statements appear.

3.1. Relationship between ECT statements and treatment-talk

Our first case involves a 34-year old female patient with stage IV lung cancer and her husband (caregiver). They are in the clinic speaking with the oncologist after a new round of scans. Along with cancer, the patient also has hypothyroidism, back pain, and numbness on the side of her body. Prior to the transcript in extract 1, there is a review of the patient’s symptoms, and the nurse performs a physical exam. In this extract, we see that the first ECT statement, but now as an upshot to the news (lines 22–24). This strategy may be similar to instances in which physicians in primary care medicine “recycle” treatment recommendations in an “attempt to gain acceptance” by patients who have not noticeably responded to the aforementioned recommendation [25]. The ECT statements also remind the patient that the treatment was, at some point, effective.

Then, in lines 25–26, we see a silence, an “uhm,” and still another lengthy silence, following which the oncologist pivots statement (bolded) occurs in response to a caregiver’s question about the scans.

An indication that scan results reveal bad news is the way the oncologist does not directly answer the caregiver’s question about the scan results. Instead, referring only initially to the scans (line 13), the doctor tacitly suggests adverse results by using an ECT statement (lines 13–15). This ECT statement is epistemically downgraded with “I think” and “unfortunately,” contains several pauses, and has a marker of dubiousness (“probably,” line 14) that the oncologist stretches. The central message (“done as much as it’s gonna do”) is pushed to the very end of the physician’s utterance. The way the scan results are initially presented “shrouds” [44] the bad news; the features are markedly different than when good news is delivered in an “exposed” way during cancer care or other types of medical visits. For an example of the way good news is presented in an oncology visit, as opposed to bad news, see [29].

In packaging the answer to the caregiver’s question about the scans in terms of efficacy of the treatment, the oncologist’s statement at lines 14–15 obscures the more detailed news that is eventually reported, which is that the cancer has progressed. Following a silence at line 16, which can indicate the patient and/or caregiver withholding any vocal responsiveness, the oncologist (line 17) initially formulates such news with negative polarity: the clinician’s (“we”) are “not seeing any further improvement” with regard to the original lesion in the lung. After another silence (line 18), he formulates the matter with positive polarity in that they are “seeing” metastases in the liver (line 19), whose recent appearance is spoken at a near whisper (line 20). After the oncologist progresses from what they are not seeing to what they are seeing, another silence (line 21)–the longest one yet–develops. Then, the oncologist, with initial hesitations at line 22, reproduces the ECT away from scan-talk and introduces treatment-talk by formulating it as “good news.” This fits a regular pattern whereby clinicians may exit bad or difficult news by producing a “bright side” account [38,39]. Bright side accounts might also work towards presenting a more complex way of understanding the overall news as a mix of both good and bad news components [40].

3.2. Bypassing discussion of scan results

A central matter is the way that the doctor starts his utterance at line 22 (extract 2), first hesitating and then producing a so-prefaced upshot or formulation that eventually embeds the ECT. These upshots, as Heritage and Watson [41] suggest, work to summarize what the speaker wants the listener to take away as the essence or gist of the previous sentences. Leading up to that upshot, the news that there is no “further improvement in the chest,” and that there are “new areas of disease in the liver,” could provide justification for the doctor to introduce what literature defines as end-of-life discussions [12,42]. Clinically, that is, given that the cancer progressing, and assuming the patient and/or caregiver are open to such discussions, it would be an appropriate moment for prognostication, planning ahead, discussing whether more aggressive treatment or clinical trials are available, or whether it might be time to stop treatment and spend time with family and loved ones. Later we suggest one way this can happen. Here, however, rather than saying that a discussion of end-of-life issues is missing, we analyze what is present in the conversation. We take an “interaction order” [43] approach by analyzing the organization of everyday talk as it takes place in the setting of oncology care [29], rather than making a normative argument about why discussing end-of-life issues is important and could happen here.

The essential device that pivots away from further discussing the scan results involves the use of upshot utterances, “so”-prefaced statements that formulate a relationship between prior talk and what is next relevant to discuss. That is, when previous talk contains different potential trajectories, speakers can use these “so”-statements to mark or resuscitate which of these trajectories is to be pursued [44–46]. A type (1) so-statement may connect to an immediately preceding utterance, while a type (2) so-statement ties to an utterance made before the immediately preceding utterance. In extract 1, the mechanism involves both (1) and (2), and we reproduce a portion of extract 1 to illustrate.

The target of analysis is the so-prefaced turn-of-talk at line 22. Initially, by saying “… that tells us … ” (our emphasis), the oncologist rather minimally refers to the scan news at lines 17 and 19, 20–type (1). However, the entirety of the so-prefaced turn at lines 22–24, rather than exploring implications of “not seeing any further improvement in the chest” (line 17) and “new areas of disease … ” (lines 19–20), instead is produced with aversion of–and thereby links back to–the turn-of-talk at lines 13–15, formulating the ECT statement as the important aspect of the scan news–type (2). In technical terms, when the oncologist positions the ECT at the end of his current turn (22–24), he makes it “sequentially implicative” [31 ] or relevant for any next turn of talk to address–type (2). After the patient and caregiver stay silent at line 25, the oncologist draws on the implicativeness of the ECT to fully pivot to “good news” treatment options. Accordingly, by making relevant the “good news” of “other options,” the so-prefaced upshot turn from 22 to 24 is like the proverbial “bridge over troubled waters”–the news that the cancer is progressing.

Even if a patient asks about what a set of scan results “means,” and thereby indicates the possibility of addressing wider issues than treatment, the physician may use conversational devices (such as an upshot) to reconnect to one of their previously stated ECT statements. Such a maneuver may be prompted by a patient whose turn to talk, despite the question about meaning, may not encourage unpacking the general, and possibly incomplete, reports of the adverse scan news [47]. In other words, a patient may ask their physician to further explain scan results, but the general formulation of that question may prompt the physician to avoid elaborating on their previously general answer. The physician can do this avoidance by recycling an ECT statement. The next excerpt, which involves a 75-year-old man with stage IV lung cancer, is an example. Prior to the transcript in extract 3, there was symptomtalk regarding the “really bad” fingernails and toenails that the patient had (due to previous chemotherapy).

The oncologist transitions (lines 10 and 11 below) to announce “stable” x-ray results, appending a “laudable event proposal” [29] to the effect that the patient had a “very good response” (line 12). Then, he produces an ECT statement at lines 12 and 13 and follows with an upshot (“So at this point,” line 14; arrowed) encompassing a recommendation to stop treatment and “sit tight” to give his “body a rest” (lines 15–16). This use of “so” is not the type (2) bridging-over kind; it is a type (1) upshot that goes directly from the ECT to an immediately subsequent recommendation to “stop the chemo” but “resuming” at a time when the cancer starts “to grow again” (lines 17–23).

Although this recommendation may hint at end-of-life concerns, the oncologist once again refers to what the scan showed (“stable”) and reverts to another ECT statement (lines 25–27, 29).⁴ At this point, the patient intersects the statement (line 30) to ask about the meaning of “stable.”⁵ The patient’s question is a “yes-no interrogative” [48], which strongly elicits confirmation in two ways. As an action, it is proposing a candidate understanding [49], and making an inference that develops the context in which the answer is to be delivered. The context it supplies presupposes an affirmative answer. And, in terms of grammar, it is a question that has what linguists call “positive polarity,” which is structured to prompt an affirmative, yes-type answer.

Thus, when the oncologist confirms the patient’s interpretation (lines 32–33), it is a regular response, not resisting end-of-life talks as such, except as it aligns to a trajectory that makes a so-prefaced upshot (lines 33–35, 37) relevant. This upshot is another ECT statement, during which the doctor increases the pace of his talk—performing a kind of “rush through” [50]. Bridging back to the turns at both 25–27 and lines 12–13, this turn embodies a type (2) upshot that re-situates the ECT as the “news” of the sequence rather than the lack of improvement in the chest and new areas of disease in the liver, or (per lines 18–23) the inevitability of the cancer going from being dormant to growing again. Thereby also, this type (1) upshot implicates a next turn, which is a treatment recommendation (to “take a break,” line 39). Together, the upshot-marked turns (lines 33–35 and 37–38) tie that recommendation to the initial ECT statement while undercutting or sequentially deleting the intermediate turns in which the disease itself is discussed.

This is not necessarily a unilateral matter, as if it were something the oncologist were “doing” to the patient. If we understand the news of disease progression or even stability as glosses whose “unpackaging” could mean engaging in end-of-life talk, Jefferson’s [47 ] work has shown that such explication takes encouragement from a listener. That is, unpackaging only occurs when someone listening to the speaker tacitly or otherwise provides an auspicious environment for it. Here, the way in which the patient asked what stable “means”— by asking for confirmation of a negative statement (“it didn’t shrink”)—suggests an orientation to bad news. The utterance “it didn’t shrink” at the end of the patient’s question provides an opportunity for the oncologist to ignore discussing the implications of the scan results (i.e., what “stable” might mean), and simply confirm that the tumor did not shrink. Conversation analytic research shows that when bad news is reported, the person who is revealing the bad news tends to add some token of hope [38,39]. In other words, a bad news orientation can encourage a “good news exit [38,39]”–treatment options still available–rather than exploration of implications regarding the scan news as such. Accordingly, in the context of ECT statements, upshot turns may help to bridge from these statements across the troubled waters of unfavorable scan results to optimistic treatment recommendations–other therapy or even “taking a break” before resuming other therapy. However, this bridging may be described as something that is co-produced rather than as the physician’s imposition. The silences we saw in excerpt 2 also can be seen as discouraging the unpackaging or explication of the scan news delivered there.

While not all treatment recommendations are presented with the same explicit degree of optimism across transcripts, they do consistently act as good news exits. We consider these treatment recommendations to be optimistic because they follow bad news [39] and also implicitly provide a token of hope. They are presented in a way that counters the bad news that the current treatment is no longer working, and in some cases, the notion that the patient’s disease is spreading, by presenting the potential for other treatments. Therefore, even when doctors recommend halting treatment for a time, the ECT statement has a positive tone as it works to transition to treatment-talk and the implicit “good news” that there may still be time to try something else. Taken together, an ECT statement and treatment recommendation provide tokens of good news, even when they are modest or present things as having both good and bad news components [40]. Additionally, and equally as important, this also closes the door on discussing what the scan results mean.

4. Discussion

4.1. Discussion

To talk about dying is a difficult, yet daily task for oncologists treating patients with incurable NSCLC. For these patients, the best they can expect is temporary cancer control. Given the seriousness of the situation, it is not surprising that interactions between doctors and terminal lung cancer patients are challenging and confusing. Studies suggest that patients with advanced and incurable cancer generally overestimate their probability for long-term survival and misunderstand the goals of anti-cancer therapy [51,52]. Additionally, Weeks et al. [53] showed that 69% of patients with metastatic lung cancer and 81% of patients with metastatic colorectal cancer believed their chemotherapy was curative in intent. According to Singh [28], patients with poor illness understanding and “prognostic awareness” may develop unrealistic expectations, which, in turn, can mean pursuing inappropriately aggressive treatment towards the end of life.

4.2. Conclusion

Our research aims to fill a gap in knowledge about how such interactions between oncologists and their patients and the patients’ caregivers actually work. As a qualitative study employing CA, we came upon the regular use of what we call “Exhausted Current Treatment” statements. Following news regarding the patient’s latest scan, oncologists use such statements to break the bad news that a patient’s treatment is no longer effective, and to almost immediately occasion the relevance of other treatments or temporary cessation until it is time to try something else. This allows for a kind of “good news exit” from the bad news [39], and as patients and caregivers often remain silent at places where they could respond to the bad news, they may discourage unpackaging or explication of what the scan news means in terms of the disease’s progression, and other considerations, such as working on closure with loved ones, desires for resuscitation, and the like. That is, such development of the talk results in foregoing prognostic awareness. Even when patients raise tumor related questions (as it happens in extract 3), it can occasion the recycling of an ECT and making treatment-talk the relevant option for further discussion. More generally, such patterns contradict the main principles and moral imperatives at the heart of patient-centered care and shared decision making.

4.3. Implications

Patient-centered care means we must factor in individual hopes, values, and goals. Many of these hopes relate to goals that center around families and celebrations of a life lived and legacy preparation in advance of death. When patients have not had an opportunity to discuss and unpack serious news they may remain fixated on treatment, missing opportunities to realize other moral goals. Spending time talking about what scan results mean may require several minutes during a single encounter. Yet, when considering the time spent building trust and the patient-doctor relationship across clinic visits, there may be other, incalculable efficiencies—considerations about time with family, desires for mental awareness, avoidance of encumbering others, minimizing one’s own pain and suffering, to name a few. Furthermore, research shows that, despite the difficulty associated with discussing death and end-of-life, patients with advanced cancer generally want these issues on the table [54], prefer alleviating pain and discomfort over curative-based treatment [55–59], and are open to the topic of emotions associated with their experience of the disease. While dealing with emotions might be difficult for the physician, training on these topics is increasingly available, and the associated skills are consistent with the principles promoted by patient-centered care.

Within bad news visit, we have found that treatment-talk occupies a disproportionate amount of time (50%), especially compared with the portion of the visit spent on discussing scan results (less than 10%) [28]. By decreasing the preoccupation with treatment, especially at the latter stages of a patient’s cancer, more time in the situation can be available for talk about meaning and prognosis. Openly discussing the meaning of scan results and prognosis may also create a working consensus regarding illness management and generate occasions to talk realistically about resources and goals. These types of discussions provide room for patients to reflect on their illness experience, make sense of the current state of their illness, and contribute to improving the doctor-patient relationship.