Conversations on Cancer Chemotherapy Cessation in Advanced Cancer Patients: Qualitative Findings from a Multi-Institutional Study

Abstract

Purpose:

As many as 20% of oncology patients receive chemotherapy in the last 14 days of their lives. This study characterized conversations between patients and cancer clinicians on chemotherapy cessation in the setting of advanced cancer.

Methods:

This 3-site study captured real-time, audio-recorded interviews between oncology clinicians and cancer patients during actual clinic visits. Audio-recordings were reviewed for discussion of chemotherapy cessation and were analyzed qualitatively.

Results:

Among 525 recordings, 14 focused on stopping chemotherapy; 14 patients participated with 11 different clinicians. Two types of non-mutually-exclusive conversations elements emerged: direct and specific elements that described an absence of effective therapeutic options and indirect elements. An example of a direct element is as follows: “… You know this is… always really tough…. But I – I think that you may need more help…. I think we’re close to stopping chemotherapy….. And hospice is really helpful to have in place….” In contrast, the second conversation element was more convoluted: “…transplant is not an option and surgery is not an option….. The options…are taking a pill…. It doesn’t shrink the tumor…. It may help you live a little longer. But I’m worried if [you] had the pill, it’s still a therapy and it still has side effects. I [am] worried if I give it to you now, that you’re so weak, it will make you worse.” No relationship seemed apparent between conversation elements and chemotherapy cessation.

Conclusions:

Conversations on chemotherapy cessation are complex; multiple factors appear to drive the decision of whether or not to stop.

INTRODUCTION

As many as 20% of oncology patients receive chemotherapy in the last 14 days of their lives, and >50% receive it within the last 30 days [1-3]. Although some of these seemingly unexpected deaths after chemotherapy administration might be a direct consequence of a therapy-induced complication, the vast majority of these patients appears to be receiving cancer treatment that only exacerbates suffering [1-3]. Indeed, the frequency of this practice has made the administration of chemotherapy within 14 days of death a quality of care metric [4]. Furthermore, trends suggest that the percentage of patients prescribed such ineffective cancer therapy is rising [5,6].

Several factors explain these overly-aggressive prescribing practices. These include the emergence of better tolerated antineoplastic agents; patients’ and family members’ unrealistic expectations and hopes; clinicians’ time constraints and emotional discomfort in broaching the topic of chemotherapy cessation; and perhaps even financial incentives on the part of clinicians [6]. Despite the above, most thought leaders agree that cancer clinicians serve a pivotal role in discussing the cessation of chemotherapy with patients and in assuring patients that “stopping chemotherapy” does not equate with “giving up” [6]. Earle and others have commented, “By shepherding many patients through the journey towards death, oncologists have a broader perspective and experience than their patients can possibly have…. oncologists must be prepared to tell patients when they would be better off without the next line of possible chemotherapy” [1].

The current study was undertaken to characterize these conversations between patients and cancer clinicians and to gain a sense of the types of conversations that occur. Relying on a unique collection of multi-institutional, real-time, transcribed clinic visit conversations between cancer patients and cancer clinicians, we undertook this qualitative study with the goal of documenting the content of these discussions on stopping chemotherapy near the end-of-life.

METHODS

Overview.

The current analysis was part of a larger parent study that the Mayo Clinic and University of Southern California Institutional Review Boards (IRB’s) had approved. The parent study has been described in detail and entailed real-time, cross sectional, audio-recorded interviews between oncology clinicians and cancer patients during actual clinic visits at Mayo Clinic in Rochester, Minnesota, the University of Southern California, and the Los Angeles County Hospital [7,8].

Audio-Recordings, Demographics, and Outcomes.

From 2012 through 2014, the parent study accumulated audio-recordings that represented a cross-section of medical oncology outpatient visits. A single study team member listened to the audio-recordings, reviewing each recording for end of life discussions that alluded to such topics as hospice, stopping chemotherapy, goals of care, and end of life. When in question, a larger group of investigators adjudicated the decision of whether to include a specific audio-recording. Transcripts from identified recordings were analyzed with a focus on the portions of the transcript that dealt with cessation of chemotherapy in the context of the broader, overall conversation.

Demographics from patients and clinicians were recorded. An experienced oncologist (AJ) reviewed each transcript for chemotherapy outcome: chemotherapy to be prescribed in the near future, not prescribed, or cannot tell.

Qualitative Analyses.

The study team relied on an inductive thematic analytic approach that entailed an iterative process, copious note-taking, and hand-coding of transcripts [9]. Two investigators (JS and AJ) read each retrieved interview transcript multiple times. Then by means of conversations that took place during multiple team meetings, these two investigators developed a code book to capture the relevant data elements; this code book was iteratively modified until both investigators agreed on its content. Both investigators then independently coded transcripts and met regularly to review coding and reach a consensus on coding. Data elements with the same code were merged and re-reviewed. Both investigators then reviewed all the data to look for relevant themes and relationships between themes, which were later shared more broadly with other team members either in person or in writing for interpretation and confirmation of conclusions. In view of the relatively small sample size and qualitative methodology, a deliberate decision was made to avoid inferential testing of relationships between clinician/patient demographics and qualitative themes as well as between chemotherapy outcomes and qualitative themes.

RESULTS

Audio-Recordings.

Among the 525 recordings from the parent study, 14 were relevant to stopping chemotherapy at the end of life. The mean duration was 33 minutes and 8 seconds minutes (range 8 minutes and 52 seconds and 56 minutes and 22 seconds). Features relevant to each recording appear in Table 1.

Table 1:

Patient and Clinician Demographics and Interview Outcomes

TRANSCRIPT NUMBER	PATIENT SEX (m/f)/AGE (years)	CLINICIAN SEX/AGE/ PROFESSIONAL ROLE**	INSTITUTION	INTERVIEW DURATION (minutes:seconds)	OUTCOME AT END OF CONVERSATION
USC-GI-008	f/49	f/43/attending physician	University of Southern California	24:31	chemotherapy prescribed
LAC063	m/49	f/29/attending physician	Los Angeles County Hospital	13:00	not prescribed
LAC057	m/51	m/37/attending physician	Los Angeles County Hospital	10:45	chemotherapy prescribed
LAC012	m/53	m/33/trainee	Los Angeles County Hospital	8:52	not prescribed
MCR498	f/52	f/39/nurse practitioner	Mayo Clinic	39:08	chemotherapy prescribed
MCR471	m/67	f/49/nurse practitioner	Mayo Clinic	33:18	cannot tell
MCR470	f/62	f/49/nurse practitioner	Mayo Clinic	40:45	not prescribed
MCR651	m/59	m/36/trainee	Mayo Clinic	54:06	not prescribed
MCR248	m/56	m/50/attending physician	Mayo Clinic	27:28	cannot tell
MCR280	f/54	m/50/attending physician	Mayo Clinic	20:47	chemotherapy prescribed
MCR332	f/63	m/37/attending physician	Mayo Clinic	29:23	cannot tell
MCR319*	f/58	m/37/attending physician	Mayo Clinic	56:09	chemotherapy prescribed
MCR162	f/55	m/40/attending physician	Mayo Clinic	49:11	chemotherapy prescribed
MCR002	m/58	m/43/attending physician	Mayo Clinic	56:22	chemotherapy prescribed

^*Did not appear to be an end of life conversation.

^**One clinician conducted interviews 6 and 7; another conducted interviews 9 and 10; and a third conducted interviews 11 and 12.

Eleven clinicians participated in 14 audio-recordings, and the average clinician age was 40 years (range: 29, 50). Seven were male, 7 were attending physicians, 2 were nurse practitioners, and 2 were fellows (Table 1).

Examination of the outcome of these conversations revealed that 7 patients went on to receive chemotherapy, and 4 did not. In 3 conversations, the outcome was difficult to know (Table 1). Below we summarize types of conversations based on qualitative interpretations.

Direct and Indirect Conversations.

Two types of conversations emerged: those that were direct and specific in describing an absence of effective therapeutic options and those that were indirect in describing this absence. These two categories were not mutually exclusive; elements of both at times appeared in a single conversation. Additionally, no relationship between elements of these conversations and final outcomes with respect to chemotherapy administration seemed to emerge.

Direct Conversations.

Direct conversations were characterized by statements of the gravity of the patient’s situation with, at times, recognition of the patient’s long cancer journey.

In one interview, the clinician again acknowledges a long trajectory but describes this trajectory from the vantage point of all the prior cancer therapy:

“…you have tried everything so you’re not going to suffer from the what if I would have tried it. You have given everything a great try and you are a miracle, though. I mean, you know that… for what you have…you wouldn’t have been here if we hadn’t done the things we’ve done.”(transcript MCR470)

This patient goes on to acknowledge the conferred therapeutic benefits but also acknowledges reluctance to receive further chemotherapy: “And I know chemo has worked and stuff, but it…makes me so sick, and I don’t want to do it anymore.”

In a different interview, the clinician was direct in describing stopping chemotherapy, but included the patient in decision-making:

“Same chemo, different chemo, no chemo, I mean to summarize. I think it depends a little bit on the guy sittin in the chair right there -- your philosophy and your feelings about that because I could have the very same discussion with 6 different people and I might get three different answers, you know, and some spread along the way -- depending on, you’re the one who keeps getting chemo and you know what that’s like for you. You’re the one who has this disease…. You know, sometimes if you have a really bad headache, it doesn’t do any good to beat your head against the wall, even if it feels like you are doing something.” (transcript MCR002)

This discussion appeared to make it easier for the patient to declare, “Well, I’m going to keep fighting it.”

This same clinician/patient collaborative style was clearly described in yet another direct conversation where the clinician explained,

“… sometimes when people have cancer, sometimes, you know, you’ve been on chemotherapy a long time – and sometimes you feel like, ok, I want a break, I want to maybe not do so much. I want to enjoy my life….. And I want to make sure you’re telling me what you want.” (transcript MCR471)

This patient appeared empowered by the foregoing and responded, “No, I’m going to be the one making the decisions.”

Such direct conversations often revealed the clinician’s concern for the patient:

“… You know this is -- this is always really tough for me when I see you because… it’s hard for me to really gauge how you’re doing.” [clinician]

“Well, I don’t know either….”[patient]

“… But I think – and you’re really strong. But I – I think that you may need more help than we’re all kind of admitting to.” [clinician]

“Okay. Like what?” [patient]

“Well, I think we’re close to stopping chemotherapy….. And hospice is really helpful to have in place….” [clinician] (transcript MCR498)

Despite the direct nature of this conversation, the patient responded, “Yeah? Well, I know you had said that before.” Later in the conversation, this patient says, “But I’d kinda like to try it [chemotherapy]. I know you think I’m crazy, but.”

During direct conversations, clinicians sometimes directly assured the patient that hospice would not result in abandonment:

“And they have a nice hospice department…..and I think it’s important to have them come by and visit you and get to know you….. And I’ll still be involved….. So you know, we’re not done taking care of you.” (transcript MCR470)

This patient responded, “Yeah. I know… I know…. I’ve been preparing myself with this. It’s just a little tough today. But, uh… yeah, okay, this is what we will do then.”

Yet another direct conversation provided this same assurance that stopping chemotherapy does not translate into abandonment, but this conversation appeared far more candid with respect to stopping chemotherapy and discussion hospice:

“…. I would still be your doctor. You can still call me. I would still be happy to speak with you anytime. I would still continue to meet with you as we’ve been doing this year….. I strongly, strongly recommend [hospice] because I won’t be here 24 hours a day 7 days a week….. My colleagues will be here to take your phone call. Somebody will be here to answer your phone call. But if you need someone to go to your house and help you with things on a Saturday, nobody’s gonna be able to do that except hospice.” (transcript MCR651)

Despite the frank nature of this conversation, this patient expressed continued reluctance to be stopping chemotherapy and to be enrolling in hospice, saying, “Mm, if I need somebody that bad, I gotta, I gotta phone. I can call.”[Clinician: “Who you gonna call?”] “A friend of mine.”

Finally, these direct conversations did not necessarily occur immediately before a recommendation for stopping chemotherapy and enrolling in hospice. Rather, they could occur much earlier, serving as a presaging of future conversations within the context of an otherwise routine clinical visit:

“You know, there generally comes a time when we say, boy, the treatment really isn’t worth it. It’s not doing much good. It’s making you sick, and we need to essentially focus on symptom control, usually in the context of a hospice-type approach… I don’t think we’re there.” (transcript MCR319)

The patient’s response was as follows, “Nope.”

Indirect Conversations.

In contrast, the second type of conversation was more convoluted, avoiding direct statements, describing options that did not appear to be in the best interest of the patient but that ultimately did not seem to be a determinant of whether or not the patient went on to receive further chemotherapy, as seen in Table 1:

“…unfortunately, given the amount of disease in the liver and the blockage of the blood vessel, transplant is not an option and surgery is not an option….. The options that we have are taking a pill every day. It doesn’t shrink the tumor…. It may help you live a little longer. But I’m worried if [you] had the pill, it’s still a therapy and it still has side effects. I [am] worried if I give it to you now, that you’re so weak, it will make you worse. Okay?” (transcript LAC012)

In this discussion, a woman who accompanied the patient seemed to be speaking for the patient and chimed in with the following, “And, um, the primary doctor mentioned something about a hospice. Um, now, I, I-my understanding of the hospice is it’s just a place where people go when they’re gonna die, and they just make them comfortable.” It appeared the decision of no chemotherapy was guided not by the clinician but rather by a bystander who had accompanied this patient to her clinic visit.

In another example of an indirect conversation, the clinician juxtaposed “good news” with “your cancer’s growing,” illustrating the multi-dimensional nature of the conversation:

“Then essentially we just focus on, um, you know, supportive care, making you feel good. Now the good news is, yeah, your cancer’s growing but you’re doing okay.” (transcript LAC057)

This patient with a gastrointestinal cancer responded with, “Yeah,” and then went on to receive chemotherapy.

These indirect conversations sometimes seemed to deflect the issue of stopping chemotherapy, instead placing the burden of next steps on the patient. In other words, if a patient was a poor candidate for chemotherapy because of progressive cancer-induced weakness, the administration of chemotherapy was cast as contingent upon the resolution of weakness – over which the patient, in actuality, had little to no control. For example, a clinician stated,

“…. I can give you some medication to try to improve the nausea and your appetite and you eat more and we see if you’re getting stronger. Okay? If you’re getting stronger, then maybe we can start some kind of therapy.” (transcript LAC012)

The patient responded, “Right.”

A similar shift of burden to the patient is seen in the following quote, where the patient had hepatocellular cancer and long-standing liver disease, which precluded chemotherapy administration:

“But if you’re able to improve your diet to the point that you are able to get your protein improved, you might be able to become a candidate for this pill. So the thing that you have to do is, first of all, continue to abstain -- no more drinking alcohol. Okay?” (transcript LAC063)

In this situation, it appeared to be the patient who provided clarity when he voiced, “So, um, since it’s incurable, then it just means I’m just basically just dying, right?” To this, the clinician responded, “You know, um, yeah.”

Yet another example of an indirect exchange is seen below where the clinician resorted to a discussion on clinical trials as a default to a discussion of stopping chemotherapy and enrolling in hospice:

“Um, and so at this point, we’re fairly limited in what our treatment options are. Um now we h – we do at times have clinical trials where there are new drugs being developed….” (transcript LAC057)

This patient responded with, “Right.” In this instance, the patient was not a candidate for a clinical trial and went on to receive other chemotherapy.

DISCUSSION

These recordings provide a unique glimpse into these intimate conversations between patients and their cancer clinician. Although multiple studies have examined how oncologists broach the cessation of chemotherapy at the end of life, few employed the methodology used here -- namely, analysis of in-person, real-time, audio-recorded clinic visits that captured the verbal interactions between clinicians and cancer patients and that relied upon qualitative methodology [10,11]. Interestingly, although we had originally assumed the more direct the conversation elements (the first type of conversation), the more likely a patient would accept stopping chemotherapy, this was not the case. Interestingly, a few times, it was the patient or the person accompanying the patient who seemed to direct the conversation to less aggressive care. In reality, these conversations are complex and seemingly halting in their trajectory, but the final decision about whether to stop chemotherapy appears to be influenced by patients’ and family members’ unrealistic – and, at times, highly realistic -- expectations.

Importantly, this group of conversations from oncology clinics resulted in the continuation of chemotherapy in several patients. The fact that many of these conversations ended with chemotherapy administration is in keeping with what occurs in many oncology clinics, where the aggressive prescribing of cancer treatment has become commonplace.

Of note, this work has limitations. One limitation is the limited number of relevant audio-recordings. It appears implausible that among 500+ interviews that took place in medical oncology clinics, only 14 discussed chemotherapy cessation. However, it should be noted that all participants (patients and clinicians) were asked to consent to the audio-recording; therefore, an imminent end of life discussion might have prompted a clinician to decline study enrollment on behalf of the patient. A second limitation is the fact that these conversations occurred at academic medical centers, where clinicians might have had more time to talk with patients. A private practice setting might provide greater time constraints, resulting in fewer such conversations. Third, with a cross-sectional study design, we are unable to confirm the final decision about chemotherapy administration, survival, and whether it was truly -- in retrospect – time to have stopped chemotherapy. Instead, we relied heavily on the content of each interview, doing the best we could to discern outcomes with the available data. This last limitation underscores the importance of remaining non-judgmental about whether chemotherapy was ultimately prescribed to each of these 14 patients or not.

Despite such limitations, these 14 conversations offer a thoughtfully unique glimpse into how clinicians talk with patients about chemotherapy cessation. We serve as an audience to intimate discussions that cancer patients often undertake only once in their lives and that family members remember for a long time thereafter [9,10]. These 14 conversations point out the variability in how clinicians broach the topic of stopping chemotherapy and invite further research on this topic.