Bereaved Caregivers Perspectives of Negative Communication Experiences Near the End of Life for Adolescents and Young Adults with Cancer

Abstract

Purpose:

High-quality communication is a standard of palliative care for adolescents and young adults (AYAs) with cancer. Yet, few studies have characterized the negative communication experiences of AYAs near the end of life (EOL).

Methods:

We performed a secondary analysis of 27 qualitative interviews with bereaved caregivers of AYAs with cancer who died between 2013 and 2016 at 1 of 3 sites. Interviews focused on barriers to optimal EOL care for AYAs. We used thematic analysis using iterative consensus coding to analyze transcripts.

Results:

Participants were predominantly white (85%), non-Hispanic (93%), and female (74%). Half of the participants were bereaved parents, and 37% were bereaved partners or spouses. Overall, 23/27 (85%) caregivers described at least one negative communication experience related to one of three themes: (1) Insensitivity to patients' needs, preferences, and values; (2) Insufficient discussions of prognosis and/or EOL; and (3) Loss of support from the clinical team near EOL. Both clinician- and patient-related factors contributed to limited EOL discussions. Lack of care continuity related to both clinician factors and systems of care that required new or changing clinical care teams near the EOL.

Conclusions:

Caregivers report a desire for clinician sensitivity to their needs and values, information about the future, and longitudinal connections with individual clinicians. Clinicians might improve caregivers' EOL experiences by eliciting patient preferences, engaging in EOL discussions, adapting to the AYA's developmental and emotional needs, and demonstrating a commitment to AYAs and caregivers as they approach the EOL.

Introduction

Cancer is the leading cause of disease-related death for adolescents and young adults (AYAs) in the United States.^1,2 High-quality communication is necessary for effective patient care at the end of life (EOL).^3–5 The adult palliative care literature demonstrates that many patients use implicit cues or explicit requests for prognostic information.⁶ Although many adult patients desire EOL conversations, these conversations often do not occur.^7–9 As a result, patients might experience unexpected negative medical outcomes¹⁰ and might not achieve their treatment goals near the EOL.¹¹ Furthermore, an individual's EOL communication needs can be influenced by the clinical context, personal readiness for conversations, and patients' emotional defense mechanisms.¹²

Most studies of EOL communication have focused on adults rather than the experiences of AYAs. The few studies focused on AYAs' experiences have shown that many EOL discussions occur shortly before death, often within the last month of life.¹³ AYAs also report lacking information, feeling ignored, receiving unclear prognostic information, experiencing poor clinician demeanor or attitudes, needing continuity of care, and receiving inconsistent communication.^14–16

AYAs with unmet information needs report higher levels of psychological distress.¹⁷ AYAs might have unique concerns, needs, and responsibilities compared to older patients, such as lack of financial independence, social pressures, and school responsibilities.¹⁸ AYAs also span a broad developmental spectrum, ranging from complete dependence of caregivers to full independence. Their level of social and cognitive development might affect communication preferences, especially related to the desired role of the caregiver. It is imperative to understand communication experiences of AYAs with cancer near the EOL rather than extrapolating findings from older adults.

In the parent study, Mack et al. performed semistructured interviews with caregivers of AYA patients who died of cancer, focusing on barriers to high-quality EOL care.¹⁹ The study did not specifically ask about communication, yet, participants identified several communication challenges.¹⁹ Given the prevalence of communication-related issues, we performed a secondary analysis to better understand these communication challenges and inform future communication interventions.

Patients and Methods

We report this study following Consolidated Criteria for Reporting Qualitative Research guidelines²⁰ (Supplementary Appendix SA1).

Design

We conducted a secondary analysis of semistructured interviews of bereaved family caregivers of AYAs with cancer. We accessed these transcripts through the Palliative Care Research Cooperative (PCRC) data repository. Below, we describe the methods from the original study and our current analytical methods. In our prior work, we found that parents and clinicians in pediatric oncology identified multilevel barriers to communication and negative consequences of these communication failures on families affected by pediatric cancer.^21,22 Based on these data, we analyzed transcripts from caregivers of AYA decedents for negative communication experiences near the EOL. See Supplementary Appendix SA2 for our reflexivity statement.

Participants and recruitment

Researchers initially identified 252 AYAs aged 15 to 39 years who died of cancer between 2013 and 2016 at three academic centers.¹⁹ They identified bereaved caregivers from medical records and contacted them with a letter inviting them to participate in a survey and in-depth phone interview. For each patient, one caregiver was eligible to participate if they spoke English or Spanish and they identified as the primary caregiver. To ensure a broad range of socioeconomic statuses, the researchers purposively sampled for caregivers of AYAs who were uninsured, insured by Medicaid, or lived in a zip code with median income ≤200% of the federal poverty level. If multiple caregivers were listed in the medical record, the researchers contacted each sequentially to ensure only one caregiver participant per patient.

The team identified 260 caregivers for these 252 patients. Of these, 80 caregivers had inaccurate contact information in the medical record, 66 declined participation, and 74 did not respond. Overall, 40 caregivers agreed to participate in the overarching study, of whom, 28 completed in-depth interviews. The research team obtained written informed consent. Institutional Review Boards (IRBs) at all involved sites approved this study. The Washington University IRB approved our secondary analysis. No authors had clinical relationships with participants. Participants knew that the research team consisted of clinicians interested in palliative care delivery and EOL experiences.

Data collection

The original study included 28 semistructured telephone interviews between June 2017 and July 2018 (See Supplementary Appendix SA3 for semistructured interview guide). Interviewers continued until reaching thematic saturation for the original analysis of barriers to optimal EOL care. Trained research assistants conducted interviews via telephone in either English or Spanish, following an interview guide. Questions in this interview guide focused on (1) decisions about utilizing palliative care, hospice, and location of death; (2) ways in which care did or did not meet the AYAs' needs; and (3) barriers to high-quality EOL care. Researchers audiotaped and transcribed interviews verbatim.

Twenty-seven de-identified transcripts were deposited in the PCRC Qualitative Data Repository and available for use. One interview performed in Spanish was analyzed but not transcribed in the original study, and was not deposited to the repository. This secondary analysis was conducted using these data in collaboration with the original study team.

Data analysis

This secondary analysis aimed to examine negative communication experiences of AYAs with cancer near the EOL from the perspective of caregivers. We used directed content analysis^23,24 to develop, refine, and apply the codebook to all transcripts. We focused our analysis on communication experiences between the patient and/or caregiver and the clinicians, inductively analyzing for categories of content related to negative communication experiences near the EOL. Three authors (B.A.S., G.L.S., and M.A.K.) first read the transcripts to familiarize themselves with the transcripts, then the authors reviewed three to five transcripts in iterative cycles and independently developed categories of content. The authors met to review codes and modify the codebook through discussions. These 3 authors reviewed all 27 interviews to finalize this codebook. This finalized codebook was reviewed with the Principal Investigator of the original study (J.W.M.). Then two authors (G.L.S. and M.A.K.) applied the finalized codebook to all transcripts using consensus coding. The authors were initially blinded to each other's codes, and they met to review each other's codes at each iteration and resolved disagreements through discussion. The coding was performed using Dedoose software.

Results

Participant characteristics

Participants were predominantly white, non-Hispanic, and female. Half of the participants were bereaved parent and 37% were bereaved partners or spouses (Table 1). The majority of AYAs were male, white, and non-Hispanic. Ages of AYAs at the time of death were distributed across all ages from 15 to 39 years, with slightly more AYAs in the 30–39 years old category. Half of AYAs lived in zip codes with median incomes ≤200% of the federal poverty limit (Table 2).

Table 1.

Adolescents and Young Adult Caregiver Characteristics (N = 27)

Characteristic	n (%)
Sex
Male	7 (26)
Female	20 (74)
Ethnicity
Hispanic	2 (7)
Non-Hispanic	25 (93)
Race
White	23 (85)
Black	2 (7)
Asian/Pacific Islander	2 (7)
Other	0 (0)
Highest level of school
<High school	2 (7)
High school graduate	9 (33)
Technical school or AA degree	3 (11)
College graduate	6 (22)
Advanced degree	7 (26)
Relationship to patient
Parent	14 (52)
Spouse	9 (33)
Other relative	2 (7)
Partner	1 (4)
Child	1 (4)
Religion
Catholic	5 (19)
Protestant/other Christian	17 (63)
Jewish	0 (0)
Other	1 (4)
No identified religion	4 (15)

Table 2.

Adolescents and Young Adult Patient Characteristics (N = 27)

Characteristic	n (%)
Sex
Male	15 (56)
Female	12 (44)
Ethnicity
Hispanic	2 (7)
Non-Hispanic	24 (89)
Unknown	1 (4)
Race
White	21 (78)
Black	3 (11)
Asian/Pacific Islander	1 (4)
Other	0 (0)
Unknown	2 (7)
Age at death
15–21	7 (26)
22–29	8 (30)
30–39	12 (44)
Insurancea
Private insurance	14 (50)
Medicaid/public health insurance	12 (43)
Uninsured	0 (0)
Other insurance	1 (4)
Unknown	1 (4)
Median income by ZIP code of residence
≤200% federal poverty level	14 (52)
>200% federal poverty level	13 (48)
Site of care
University of California San Diego	9 (33)
University of Iowa	10 (37)
University of Alabama at Birmingham	8 (30)

^a One patient was dually insured by Medicaid and private insurance; therefore, a total of 28.

Communication-related barriers to EOL care

Caregivers described negative communication experiences in 23/27 interviews, related to 3 aspects of clinicians' communication behaviors near the EOL: (1) insensitivity to patients' needs, preferences, and values; (2) insufficient discussions of prognosis and/or EOL; and (3) loss of support from the clinical team near EOL (Table 3).

Table 3.

Negative Communication Experiences Near the End of Life

Code	Definition	Excerpts
Insensitivity to family needs	Caregivers perceived clinicians as insensitive to the AYA's needs, preferences, or values by failing to show compassion, making recommendations that did not align with AYA or family priorities, and excluding the AYA's family from conversations.	“I think one doctor might've told us, you know, bluntly, you know you might wanna think about the future of her not being here” [4266-13]. “[The hospice nurse] mostly talked about himself, and I thought that was really weird, you know. I mean, he showed us pictures of his daughter, and I don't know if he was trying to bond with us. It didn't work, though” [4268-27]. “Well, I think they just made a recommendation. We never asked for hospice, ever. We were told basically that while he was in hospice care, it wouldn't cost us anything. And were not really worried about the cost of anything. We would have you know we would have paid for his care if we had to. You know we had money to take care of him” [4267-77]. “Going on hospice, they said she couldn't do the transfusions—go in for the blood transfusions—so for her, it would be just like giving up, saying okay I'm going to hook up to this and die because a day or 2 without the transfusions, she would have been gone, and that's not what she wanted” [4266-63]. “A lot of that information was shared more with him because he was considered an adult than it was with his family. So, by law, there were probably a lot of things that couldn't be shared with us because our son was, you know, no longer a minor… I was surprised that there weren't any family meetings that we were ever invited to” [4267-77].
Insufficient discussions of prognosis and/or EOL	Caregivers described insufficient communication about prognosis and/or EOL, resulting in partial discussions, poorly timed discussions, or “false hope.” Sometimes disclosure occurred too late, and other times disclosure occurred before the AYA or caregiver was prepared to receive challenging information.	“The doctors had come in and continued to talk about hope. And I hate that word because of those doctors. Cause there was no hope. There was only preparing for what was gonna happen, and I got no support from the nurses, or the doctors” [4266-14]. “[The doctor] came in and you know, not in—you know, just came in and sat down, and he was the first one to sit down and actually start talking about palliative care and dying. And you know, we—that was oh, several weeks before [patient] died, but she was still very, you know, mobile and you know, with it and everything, and we weren't quite ready for that conversation” [4267-22]. “I said to him, ‘[Doctor], I want to ask you a question. Do you know of any patient who has recurred with [this cancer] and survived?’ And I'm not talking survived for a couple of months and then relapsed. You know what I'm saying? I'm not saying survived 3 years and then died. I'm saying. And I never got an answer” [4268-27]. (When asked what stood in the way of EOL discussions and care) “Me. Me refusing to accept that she was dying… (crying) if anything stood in the way, it was us just not wanting to accept the reality of what was going on” [4267-63].
Losing support from the clinical team near EOL	Caregivers described how clinicians became more distant and difficult to contact as the AYA approached EOL. This disconnection was exacerbated when new clinicians were incorporated into other care teams near the EOL, including the outpatient hospice and inpatient medical teams.	“I felt like when we first got there, they were very excited about seeing him and talking to him. I just I really felt like after they did the tests and they found out, you know, how bad he was, their whole demeanor to us changed. And I just felt like we were sort of pushed out the door. It was extremely upsetting. And I don't know how to describe. I felt like we were just being rushed out the door once they found out that his condition was too far gone” [4267-40]. “Just a personal connection with people. It felt like in the end there, other than palliative care they were great, but with the hospital in general and the bone marrow transplant team, I just felt like we were kind of like just hung out to dry; I mean, they couldn't do any more and they weren't very supportive at that point” [4266-63].

AYA, adolescents and young adult; EOL, end of life.

Insensitivity to family needs

Eighteen caregivers reported experiences with clinicians failing to meet the family's needs near the EOL. Some clinicians failed to show compassion or concern: “Have respect for the parent. If you can't talk to them right then, you let that parent know that I can't talk to you right now ‘cause I do have some other patients that I need to go to right quick, but I will come back to you and talk to you… Don't brush me off” [4268-42]. Another caregiver described insensitive disclosure: “I don't think the doctors really know how to talk to people, how to support them, or how to bring devastating news to them” [4266-14]. Some professionals lacked warmth in their demeanor: “The social worker was the one that was, not very passionate about it, you know not very compassionate about anything” [4267-77].

Clinicians also demonstrated insensitivity by making recommendations that did not align with the AYA's/family's values, often related to hospice decisions. For example, some caregivers felt that clinicians pressured the patient to remain in the hospital near the EOL unless they transitioned to hospice care. Yet, the caregiver felt the patient belonged at home, and they preferred to make a hospice decision later: “We were arguing to get him out of the hospital. They were not going to let him go because of his physical situation…unless we put him in hospice care. But that's when we were able to, you know, let them know, but now he wants to go home to die. We'll put [him] into hospice care when he is ready. We are going home” [4267-16].

Caregivers also described conflicts with clinicians over blood product transfusions near the EOL: “They said we couldn't, he couldn't have a blood transfusion because that would just extend his life. Well, we were trying to extend his life. And that was why… [the] social worker kept saying to us, well, that can't be paid for, that can't be paid for. You know my husband [and] I both said we'll pay for it out of our own pocket if we have to” [4267-77].

Insufficient discussions of prognosis and/or EOL

Seventeen caregivers described insufficient communication about prognosis and/or EOL, resulting in partial discussions, poorly timed discussions, or “false hope.” When asked when she realized that the patient would not survive, one caregiver responded: “Probably like a month before he passed away” [4266-74]. When asked if any specific event triggered this realization, she described how use of a morphine drip seemed like a euphemism for redirecting care: “He went into the hospital, and the doctor asked me ‘should we start him on a morphine drip?’” [4266-74]. Another caregiver described not knowing that the patient was going to die until the last day of life: “Until the last day, our doctors were, you know, doing the best they could to stabilize her so we could continue on with the experimental [treatment]… so they didn't really talk about anything” [4266-13]. This delayed recognition affected some patients' decisions regarding hospice care: “I guess that was the reason he was not put on that type of care until a week before he passed away was because he had not understood that he needed that kind of care” [4266-74].

Caregivers described how this focus on treatment rather than prognosis led to false hope: “When they sent us to [city], to me, that was like…even though you just told us it was a chance, but when you said it's a chance, they gave us false hope” [4268-42]. One caregiver described how a doctor came to the bedside and explained that the treatment was working on the same day the patient died: “We had actually just met with her oncologist the day that she passed away, and he had kind of, you know, talked to her a little bit about how her treatment was going, and it was actually kind of going in the right direction. So, I think that discussion kind of delayed any thoughts about hospice or palliative care” [4267-14]. This late disclosure also affected how caregivers spent remaining time with the patient: “I just wish… those doctors had been clear with me because I needed a little more time to say goodbye than what I did” [4267-14].

Although late disclosure was an identified problem, caregivers also described receiving prognostic information before they felt prepared to hear such difficult information: “[The doctor] told us that day that [patient] would not even make it to his next birthday which would be in July. So, that's when we were told he was going to die. We just were stunned because we never thought we were going up there for that type of information” [4267-46].

Caregivers attributed some of these insufficient discussions to clinician hesitance: “I still feel the one that betrayed me the most was the lack of honesty and transparency… They avoided telling the truth so that they didn't have to deal with anyone possibly being disappointed, angry” [4266-14]. Some caregivers interpreted physicians' silence as an indication of bad news: “We realized that he was not probably gonna make it when his doctor wouldn't give us no kind of information” [4268-62].

Insufficient discussions did not always result from lack of clinician intent or effort. Some caregivers described how the lack of patient and family readiness to hear difficult news created a barrier: “I think [the patient's] mind was still a little bit in denial, even right there. When he started showing really, really bad symptoms, he was in denial” [4267-16]. Another caregiver described a mutual pretense between the patient and caregiver: “Dying was something that we never talked about. [Patient] was a very vibrant person, was a very bubbly personality, and she was always positive, and for her sake, we always maintained that positivity, but she was a very smart girl, a very tech-savvy girl, and so I know that she had Googled her disease and knew what the outcome almost always was” [4268-31].

Loss of support from the clinical team near EOL

Eight caregivers described how clinicians became more distant near the EOL. One caregiver described how clinicians seemed to lose interest when it became clear that the disease was incurable: “They came in, they checked on her, they did what they needed to do, but there really wasn't the connection and support going through this end time. I mean, it was like they were supportive in there up until all the care they could do, and when they couldn't do anymore, I don't feel like they were there for us anymore, or her, and I was very disillusioned by that because I thought the team followed you through whether you lived or died” [4266-63].

This disconnection at the EOL seemed to be exacerbated by incorporating new clinicians, whether the patient was in the hospital or at home. For example, one caregiver described frustrations with meeting a new clinical team when the patient was admitted to the hospital near the EOL: “When you get admitted, you have a doctor that does your care, you know, that's not your doctor. I think most hospitals go with that, which I am totally against [it]… I don't feel like the doctor really understands everything to do with a patient, especially a terminally ill patient… You get a relationship with your doctor” [4267-16].

Discussion

We present one of the first studies to examine caregivers of AYAs who died of cancer to understand their negative communication experiences. We identified three categories of negative experiences near the EOL—sensitivity, disclosure, and support. These experiences led to frustrations and worries, created false hopes, and possibly affected the therapeutic relationship. In addition, communication with AYAs and caregivers near the EOL seems to be influenced by multiple complex factors, including clinician comfort, AYA and caregiver readiness to hear difficult news, AYA's developmental level, and the family's unique needs and preferences.

These findings are similar to previous studies in pediatric oncology that found negative communication experiences could lead to decreased trust, poor understanding, emotional distress, and false or decreased hope.²¹ Also, 85% of caregivers in this current analysis described at least one negative communication experience near the EOL, suggesting that negative EOL communication experiences might be common, and that communication challenges need to be elucidated and effectively addressed at the EOL for AYAs with cancer.

Several clinician behaviors demonstrated insensitivity. Building relationships is a core function of communication, and the quality of relationships influences other aspects of communication.^25,26 Clinicians can support these relationships by demonstrating understanding, concern, and competence while ensuring the patient feels unique and valued.²⁵ Caregivers also perceived clinicians as insensitive when they misunderstood the family's values and priorities. Clinicians often misinterpret or misunderstand AYA's and caregivers' preferences for information or decision-making roles.^27–29 These preferences might also change over time as AYAs and caregivers experience illness longitudinally.²⁷ Clinicians might provide better care for AYAs and caregivers by assessing EOL communication and care preferences over time in repeated, consistent, and gentle ways.⁴ Goal elicitation tools might help clinicians to open and facilitate these difficult conversations.³⁰

Caregivers also identified the lack of EOL discussions as negative experiences. Prior studies have shown that most AYAs with cancer and caregivers prefer detailed prognostic information presented sensitively.^31–33 Our findings suggest that families' needs are more complex and nuanced than disclosure versus nondisclosure. Caregivers described how hesitance from both families and clinicians could influence the timing and quality of these conversations. Furthermore, timing of the disclosure matters. Some caregivers had negative experiences related to late disclosure, but others received difficult information before they felt ready to hear. As such, deficiencies in EOL discussions could relate to timing and clinician barriers, including limited communication skills or confidence, worries about causing harm with disclosure, and complex personal emotional responses of the clinicians.^4,34,35

The lack of EOL discussions could be influenced by the perceived characteristics or needs of the patient and family. A prior study found that clinicians perceive many AYAs as hesitant to communicate about sensitive EOL issues.³⁶ In adult oncology, clinicians have identified several patient- or caregiver-related barriers to opening EOL discussions, such as difficulty accepting the prognosis, difficulty understanding complications and limitations of intensive EOL interventions, and intrafamily disagreements about EOL goals.³⁷ Several interventions aim to facilitate EOL communication.^38–40 Still, this field requires further study to develop, implement, and disseminate tools to support these conversations.

We also found that some clinicians distanced themselves as AYAs neared EOL. Caregivers attributed this distancing to the clinicians' loss of interest when cure was not possible. Others commented on the number of clinicians involved in care and the constant disruption of incorporating new clinicians at EOL, including the outpatient hospice team and the inpatient clinical team. Continuity of care and fidelity to the patient are critical aspects of clinical relationships that may be affected by rotations of clinicians.²⁵ This distancing of clinicians has received little attention in the AYA literature, but it could relate to physician discomfort with death, especially for AYA patients.

Many clinicians feel that AYAs dying of cancer is “unfair,”³⁶ and offering hope of a cure is central to the professional identity of oncologists.⁴¹ Other clinicians might determine that the family “wants space,” and they might worry about intruding during an important and sensitive time. Furthermore, the medical systems of EOL care impede continuity of care. For example, transition to hospice necessitates new care teams, and inpatient teams cycle routinely. This field needs further research to guide transitions of care and to ensure that patients and caregivers receive high-quality communication and EOL care.

These results should be interpreted in light of study limitations. First, caregivers were predominantly white women. Future studies should purposively sample to ensure diverse representation. Furthermore, caregivers provided their perspectives on EOL care, and these perspectives might incompletely represent caregivers' experiences. In addition, this qualitative study only enrolled 10% of eligible caregivers and we used purposive sampling. While purposive sampling supports the representation of important viewpoints, these results are not necessarily generalizable to all AYAs and caregivers. Caregivers might have experienced recall or conformity bias. Recall bias might cause participants to remember negative events more vividly than positive experience. Conformity bias might have pressured participants to de-emphasize negative experiences.

Furthermore, we performed a secondary analysis of a dataset initially developed to focus on barriers to EOL care, not specifically communication. Although we identified rich themes, we might not have identified all negative communication experiences at EOL. These findings should be further explored and corroborated with future studies.

Conclusion

Caregivers of AYA patients want to receive information about the future in a sensitive manner from the clinicians who know them best, and they want to maintain these longitudinal connections as the EOL approaches. Clinicians might improve AYAs' and caregivers' EOL experiences by eliciting patient values and preferences, engaging in EOL discussions tailored to the AYA's developmental and emotional needs, and demonstrating commitment to AYAs and caregivers as they approach the EOL. Future research should evaluate clinicians' personal experiences in caring for AYA patients at the EOL and fostering advance care planning as a routine part of care. Research must also focus on timing and effective methods to elicit communication needs and preferences.

Disclaimer

The content is solely the authors' responsibility and does not necessarily represent the official views of the National Institutes of Health.

Authors' Contributions

All authors have contributed significantly to this article, and all authors have reviewed and agreed with the final submission.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

Research reported in this article was supported by the NIH under award number R21 R21NR016580 (J.W.M.) and the Palliative Care Research Cooperative (PCRC) Group, funded by the National Institute of Nursing Research (U2CNR014637).

Supplementary Material

Supplementary Appendix SA1

Supplementary Appendix SA2

Supplementary Appendix SA3