Congruence Gaps Between Adolescents With Cancer and Their Families Regarding Values, Goals, and Beliefs About End-of-Life Care

Key Points Question Do families of adolescents with cancer know what the adolescent would want if they were dying? Findings In this cross-sectional study of 80 adolescent-family dyads (160 participants), critical congruence gaps were found between adolescents with cancer and their families. Families had a poor understanding of their adolescents’ preferences for the best time to bring up end-of-life decisions, dying a natural death, and being off life support if they were dying, but families’ understanding of adolescents’ relational needs was excellent. Meaning These findings suggest that family-centered pediatric advance care planning interventions are needed to close the gaps in families’ knowledge of adolescents’ end-of-life treatment preferences.


Introduction
Cancer remains the leading cause of disease-related death for adolescents. 1,2 For adolescent patients with cancer, death frequently occurs in the context of withholding and withdrawing life-sustaining treatment. 3,4 If parents are being asked to make these difficult decisions, a prior understanding of their child's preferences may ease the burden of decision-making. Unfortunately, the timing of these discussions is often very close to death, preventing thoughtful reflection about how these decisions affect the patient and their family. 3,4 Preparation for the possibility of death includes open and honest communication between adolescents with cancer and their families. Pediatric advance care planning (pACP) is the reference standard in the care of patients with life-limiting illnesses 5 and is internationally recognized as a need by clinicians. [6][7][8] Small pilot and qualitative studies [9][10][11] show that parents desire to have pACP conversations and to keep all options open. Nevertheless, policy recommendations 12,13 to include adolescents in shared decision-making remain aspirational, despite findings from qualitative and pilot studies [14][15][16][17][18] showing that adolescents with cancer want to engage in pACP. Most adolescents aged 14 years and older do not differ from adults in their capacity to make informed treatment decisions, and their understanding of death is no less mature than that of adults. 15,19 Benefits of pACP, as identified in the pilot trial 20 (30 dyads) of the current larger scale study, include families' improved understanding of their adolescent's treatment preferences and adolescents receiving earlier palliative care. Trials of the same pACP intervention with adolescents with HIV (105 dyads) also demonstrated improved congruence on end-of-life (EOL) preferences 21,22 and decreased HIV-specific symptoms among adolescents longitudinally. 22 Among children with complex chronic conditions, pACP decreased suffering at the EOL and improved families' quality of life. 23 Lack of ACP has been associated with poor communication, increased hospitalization, poor EOL quality of life, poorer adherence to patient's EOL preferences, and legal actions. [24][25][26] Conversations about hoping for the best while planning for the worst are emotional and often are avoided or considered taboo. 17 Structured patient-centered and family-supported pACP conducted by trained or certified facilitators safely elicited strong emotions among HIV-positive adolescents and their families. 27 It remains unknown whether early pACP builds families' capacity to make EOL decisions for their child, which could allay clinicians' concerns about burdening parents 7,28,29 or their lack of readiness. 30 The discomfort reported by clinicians about knowing what to say, especially when it comes to discussing resuscitation status, 6,30,31 may be minimized, if this is not the first time the adolescent and/or family has had this conversation. Parents of children with cancer define being a good parent to include making informed, unselfish decisions in the child's best interest and teaching their child to make good decisions. 16 Parental perspectives on pACP focus on what a loving parent would do. 31 Yet, few systematic pACP programs exist, 32,33 with only 1 model tested in a fully powered randomized clinical trial. 21,22 Most studies on pACP presume that families accurately represent the adolescent patient's goals, values, and EOL treatment preferences. 5,7,31 Only our small pilot study 8 from a single site has empirically tested this presumption for adolescents with cancer (17 adolescent-family dyads) and found it to be false. The current multisite cross-sectional study of pACP enrolled the largest sample of adolescents with cancer and their families to date, adding geographical and economic diversity to the pilot. We identified key areas of misunderstanding during session 1 of a 3-session pACP intervention, so as to close gaps in understanding during pACP conversations during sessions 2 and 3.

Study Design and Participants
The Family Centered Pediatric Advance Care Planning for Teens With Cancer (FACE-TC) intervention is a 2-group randomized clinical trial designed to evaluate the efficacy of an adequately powered pACP intervention (Figure 1)

Procedure
After consulting with a patient's primary oncology physician, research assistants approached potentially eligible participants face-to-face during hospital outpatient visits and stays. The trial comprises 8 visits over 2 years. The first visit included enrollment, secondary screening, and, if eligible, enrollment and completion of baseline questionnaires. Dyads were then randomized to the FACE-TC intervention or treatment as usual control, using a computerized 2:1 ratio. In the FACE-TC group, session 1 included the Lyon FACE-TC ACP Survey, session 2 included the Next Steps: Respecting Choices Interview, 35

Statistical Analysis
Descriptive

Participant Characteristics
Data presented here report on the 80 adolescent-family dyads randomized to the intervention group of the FACE-TC trial who attended session 1 (Figure 1). Adolescent participants had a mean (SD) age of 16 Table 1). Three hundred sixty-six adolescent-family dyads were approached, of whom 336 dyads met initial eligibility criteria ( Figure 1). Of these, 203 declined, 3 were ineligible, and 130 dyads enrolled (39% participation rate), achieving the predetermined sample size. 34 Of those who declined, 23% (46 of 203 dyads) had at least 1 member of the dyad report who did not want to talk about pACP ( Figure 1). The major reason given for declining was lack of time to commit to a 2-year   Figure 2 illustrates that most adolescents thought the best time to bring up EOL decisions was early, with 86% reporting "before getting sick, while healthy," "when first diagnosed," "when first sick from a life-threatening illness," or "all of the above." Only 39% of families accurately reported their adolescents' preference for early timing with poor dyadic congruence (PABAK, 0.18). Table 3

JAMA Network Open | Pediatrics
Congruence Gaps Between Adolescents With Cancer and Their Families Regarding End-of-Life Care

Discussion
Survey results about congruence in goals, values, and preferences are important for pACP because explication of patient's goals, values, and preferences helps prepare patients and surrogates to participate with clinicians in making the best possible in-the-moment decisions, as discussed by Adolescents with cancer were willing and able to engage in pACP with their families, even though some found the process uncomfortable, and not all would want to be told when they would die. Specific findings should be viewed in light of the 39% participation rate (130 of 333 participants), as discussed later in the Strengths and Limitations section. Of those who declined to participate, 23% (46 of 203 participants) had at least 1 member of the dyad who did not want to discuss pACP, so that the opinions of this subpopulation are not represented (Figure 1).
Most adolescents reported a willingness to forgo life-prolonging medical treatments if dying and to die a natural death. Families' understanding of these preferences was poor, confirming pilot findings with 17 dyads 8 and the teams' HIV study with 48 adolescent-family dyads. 39 In a medical record review 43 of patients with cancer aged 15 to 39 years, investigators found that 56% preferred comfort care to life-prolonging care in the month before their death. Yet, 75% of those who preferred comfort measures received at least 1 form of intensive EOL care. 43 Ongoing communication with families and clinicians regarding adolescents' preferences is needed.
The optimal timing for initiating conversations about EOL decisions from the adolescents' perspective is clear and consistent with our pilot cancer study 8 and adolescent HIV studies. 39,44 Among adolescents willing to enroll in a trial about pACP, adolescents preferred to talk about EOL decisions when healthy, first diagnosed, or throughout the disease process, rather than when hospitalized or dying. Families consistently thought their child preferred to delay the conversations. Clinically, hospitalization and life-threatening medical crisis are the most frequent triggers for clinicians to initiate EOL conversations, even while recognizing this is often too late. 30,45 These gaps in understanding have serious clinical consequences because modern medicine has medicalized dying, resulting in most adolescents who die from cancer dying after withholding or withdrawing life-sustaining treatment. 5,6 These emotional decisions are made by their families, and there is the potential for decisional regret 46 and long-term complicated grief. 47,48 The high premium adolescents placed on understanding their treatment choices and saying everything they have to say to their families underlines the importance of timely pACP, which may ease the families' burden of making EOL decisions on behalf of their child.
The importance of being at peace spiritually at the EOL was almost universal (92.5% of participants), consistent with findings of the cancer pilot study (100% of participants) 8 and adolescent HIV study (94% of participants). 39 Spirituality influences the experience of illness, pain, and pediatric EOL decisions. [49][50][51] Spiritual assessments and appropriate referrals should be provided to ensure culturally sensitive care. 51 Consistent with study findings, many adolescents with life-limiting conditions perceive themselves as a burden, 52-55 which may not reflect how families feel. 52,53 Some patients base treatment decisions on the perceived burden the treatment creates. 52 Resolving adolescent feelings regarding being a burden is a need that could be addressed through pACP.

Strengths and Limitations
To our knowledge, this is the largest dyadic EOL survey of adolescents with cancer and their families.
The completion of the surveys in 4 tertiary hospital-based settings increases validity and the likelihood of findings influencing clinical practice. Validity is further enhanced by replication of earlier studies. The overall study achieved the enrollment goal of 130 adolescent-family dyads.
Generalizability is increased by the geographically and economically diverse sample.
Limitations include a low participation rate (39%), which affects generalizability. Participation was below the 50% benchmark achieved in our pilot 8 and HIV trial. 39 Among those who enrolled, only 28% wanted to know if someone could tell them when they would die, in contrast to 50% in the cancer pilot 9 and 46% in the HIV study. 39 Overwhelmingly, adolescents wanted honest answers from their physician, suggesting that many adolescents may only want to know their prognosis, if they ask. Nevertheless, all were willing to participate in a pACP trial, which enabled a discussion of the hypothetical scenarios with their family, regardless of prognosis. These findings highlight the complexity of the pACP process.
Social desirability bias could have occurred with face-to-face administration. We chose this approach to enable monitoring of emotional reactions and to control for issues of literacy, impaired vision, item comprehension, and survey completion. Male adolescents with cancer were statistically significantly more likely to decline participation than female adolescents, although 45.0% of enrolled adolescents were male. Among adults, male patients are less likely to participate in ACP than female patients. 56 The sample included all adolescents receiving oncology care, reflecting recommendations that ACP occur at all stages for anyone with a serious illness. 57

Conclusions
Families had a poor understanding of their adolescent's values regarding their own EOL care with respect to when to initiate EOL conversations and preference for being off machines that extend life, if dying. Pediatric ACP could minimize these misunderstandings, potentially affecting the broader domain of clinical practice guidelines for quality palliative care. 58 Access to pACP to increase congruence for interested and ready adolescent-family dyads may be more beneficial than simply asking adolescents about their EOL treatment preferences by helping families with the burdens of making EOL decisions, ensuring that adolescents' preferences are heard, and opening up conversations on topics that both the adolescent and family member may be thinking about, but