You are on page 1of 8
Journal of Adolescent Health 54 (2014) 710 e 717 Original article A Longitudinal, Randomized,

Journal of Adolescent Health 54 (2014) 710 e 717

Journal of Adolescent Health 54 (2014) 710 e 717 Original article A Longitudinal, Randomized,

Original article

A Longitudinal, Randomized, Controlled Trial of Advance Care Planning for Teens With Cancer: Anxiety, Depression, Quality of Life, Advance Directives, Spirituality

Maureen E. Lyon, Ph.D. a , * , Shana Jacobs, M.D. b , Linda Briggs, M.S.N., R.N. c , Yao Iris Cheng, M.S. d , and Jichuan Wang, Ph.D. d

a Children s National Health System, Division of Adolescent and Young Adult Medicine, Children s Research Institute, Center for Translational Science and George Washington University School of Medicine and Health Sciences, Washington, DC

b Children s National Health System, Division of Oncology/Children s Research Institute, Center for Cancer and Immunology Research and George Washington University School of Medicine and Health Sciences, Washington, DC

c Gundersen Medical Foundation, Inc., La Crosse, Wisconsin

d Children s National Health System, Children s Research Institute, Center for Translational Science, Division of Biostatistics and Study Methodology and Department of Epidemiology and Biostatistics, George Washington University School of Medicine and Health Sciences, Washington, DC

School of Medicine and Health Sciences, Washington, DC Article history: Received July 13, 2013; Accepted October

Article history: Received July 13, 2013; Accepted October 28, 2013 Keywords: Adolescent; Advance care planning; Advance directive; African-American; Cancer; Communication; Decision making; End of life; Family intervention; Pediatric palliative care


Purpose: To test the feasibility, acceptability and safety of a pediatric advance care planning intervention, Family-Centered Advance Care Planning for Teens With Cancer (FACE-TC). Methods: Adolescent (age 14 e 20years)/family dyads (N ¼ 30) with a cancer diagnosis participated in a two-armed, randomized, controlled trial. Exclusion criteria included severe depression and impaired mental status. Acceptability was measured by the Satisfaction Questionnaire. General Estimating Equations models assessed the impact of FACE-TC on 3-month post-intervention out- comes as measured by the Pediatric Quality of Life Inventory 4.0 Generic Core Scale, the Pediatric Quality of Life Inventory 4.0 Cancer-Speci c Module, the Beck Depression and Anxiety Inventories, the Spiritual Well-Being Scale of the Functional Assessment of Chronic Illness Therapy e IV, and advance directive completion. Results: Acceptability was demonstrated with enrollment of 72% of eligible families, 100% atten- dance at all three sessions, 93% retention at 3-month post-intervention, and 100% data completion. Intervention families rated FACE-TC worthwhile (100%), whereas adolescents ratings increased over time (65% e 82%). Adolescents anxiety decreased signi cantly from baseline to 3 months post- intervention in both groups ( b ¼ 5.6; p ¼ .0212). Low depressive symptom scores and high quality of life scores were maintained by adolescents in both groups. Advance directives were located easily in medical records (100% of FACE-TC adolescents vs. no controls). Oncologists received electronic copies. Total Spirituality scores ( b ¼ 8.1; p ¼ .0296) were signi cantly higher among FACE-TC adolescents versus controls. The FACE-TC adolescents endorsed the best time to



Family-Centered Advance Care Planning for Teens With Cancer is the rst randomized clinical trial of pediatric advance care planning involving adoles- cents with cancer. This structured model meets the World Health Organization, Institute of Medicine, and American Academy of Pe- diatrics recommendations for early palliative care that involves the patient, surro- gate decision makers, and physicians, moving the eld of pediatric advance care planning forward.

Disclaimer: The contents of this study are solely the responsibility of the authors and do not necessarily represent the of cial views of the National Center for Advancing Translational Sciences or the National Institutes of Health or the American Cancer Society. Con icts of Interest: The Lyon Advance Care Planning Survey is copyrighted by the rst author and is available free of charge by request at mlyon@ . Linda Briggs, MSN, RN, receives no royalties for training

facilitators to use the Respecting Choices tool. This training is provided by the non-for-pro t Gundersen Health System, which does charge fees for the on-line and face-to-face trainings. * Address correspondence to: Maureen E. Lyon, Ph.D., Center for Translational Science, Children s Research Institute, Children s National Health System, Room M7658, 111 Michigan Avenue, N.W., Washington, DC 20010-2970. E-mail address: (M.E. Lyon).

1054-139X/$ e see front matter 2014 Society for Adolescent Health and Medicine. All rights reserved.

M.E. Lyon et al. / Journal of Adolescent Health 54 (2014) 710 e717


bring up end-of-life decisions: 19% before being sick, 19% at diagnosis, none when rst ill or hospitalized, 25% when dying, and 38% for all of the above. Conclusions: Family-Centered Advance Care Planning for Teens With Cancer demonstrated feasibility and acceptability. Courageous adolescents willingly participated in highly structured, in- depth pediatric advance care planning conversations safely. 2014 Society for Adolescent Health and Medicine. All rights reserved.

Palliative care (PC) currently exists in the context of a changing health care delivery system in which death is medi- calized. Four of ve children with complex chronic conditions in the United States die in the hospital [1] . Among children with cancer 72% of those without PC died in the ICU versus 21% of those with PC [2] . Advance care planning (ACP) is one dimension of palliative care that provides support in decision making about limiting burdensome medical interventions and advance di- rectives. Adolescents want a voice in choices for their end-of-life (EOL) care, if they cannot speak for themselves [3 e 5] . Avoidance of EOL conversations can lead to miscommunication and dis- agreements, which may result in families charged with neglect, court battles, and even legislative intervention [6] . The Institute of Medicine s report, When Children Die, recommends that these conversations become a routine and structured part of standard of care [7] . The landmark SUPPORT study demonstrated that focusing on advance directives alone is insuf cient to ensure that patients treatment preferences are honored at EOL [8] . Building on these ndings, research identi ed the critical importance of including the surrogate decision maker in structured conversa- tions with the patient about the patients goals for EOL care [9] . Nevertheless, ACP conversations are frequently not initiated before a medical crisis because of the strong emotions evoked [10] or fear of taking away hope [11] , which contrasts with empirically based research that early palliative care can instill hope [12] , maintain psychological well-being [13] , increase congruence in treatment preferences between patients and their surrogate decision makers [14] , prolong survival [15] , and improve quality of life and mood among surviving relatives [16] . Spiritual quality of life is another important aspect of PC among dying children [7] , because being diagnosed with a life-limiting illness can inspire re ections on death and spirituality. Only three structured pediatric ACP (pACP) programs are re- ported for adolescents with complex chronic conditions. First is FOOTPRINTS, a model of pACP and coordination speci c to pa- tients with muscular dystrophy (age 8 e 21 years) [17] . Second is Family-Centered (FACE) Advance Care Planning, a model speci c to adolescent patients with human immunode ciency virus (HIV) (age 14 e 21 years) and their family. This was the rst pro- tocol to test pACP using a randomized clinical trial [13,18,19] . The third is FACE for Teens With Cancer (FACE-TC), which adapted the HIV protocol for teens with cancer [20] . (The development of FACE-TC is described in Appendix 1, which can be found in the on-line version of this article.) Immediate outcomes for FACE-TC replicated the previous ndings from FACE [13] of increased congruence in treatment preferences between adolescents and their families, decreased decisional con ict, and a willingness to limit treatments in some situations compared with controls [20] . Standard hospital practice is for hospital administrative personnel to inquire about the presence of an advance directive when the patient is hospitalized, only if they are 18 years of age. Under the age of 18 years, the pACP process is often triggered

during a medical crisis. Many physicians believe that this is too late to begin this conversation [21] . Challenges to pACP unique to adolescents include their emerging intellectual and emotional maturity [22] . Nonetheless, evidence supports the practice of allowing adolescents to make decisions about their health. Cognitive capacities required to make informed treatment decisions do not differ from those of adults [23] . Many young adolescents have as mature an under- standing of death as those aged 18 years [7] . Unique challenges also arise from the adolescent e family dynamic if there are dis- agreements in treatment preferences, particularly in the context of the adolescents lack of legal authority to make treatment decisions under the age of 18 years in most states [6] . Cancer remains the leading illness causing death for adoles- cents [24] . In a review of 95 medical records of children with cancer who died, 53% had a do not resuscitate order at the time of death [25] . Effective EOL care is a public health issue because of cancer s high symptom burden and the potential to prevent suffering [26] . An age-appropriate communication model that includes adolescents with cancer and their families in making decisions about medical treatments that frequently precede death was recently tested for ef cacy in a randomized controlled trial [21] . The 3-month post-intervention outcomes of this study are reported here. Our aim was to demonstrate the feasibility, acceptability, and safety of a highly structured pACP conversation with teens with cancer and their families, which incorporated communication of the results to their oncologist and placement of an advance directive in the medical record. We explored the in uence of pACP on spirituality. Predetermined benchmarks for accept- ability were enrollment > 50% of eligible families, attendance > 80% at all three sessions, retention > 80% at 3-month follow-up, completeness of data > 90%, and > 90% rating sessions as worthwhile. Legal guardian(s) or surrogate decision maker(s) are referred to as family/families from this point forward. Hypotheses at 3-month post intervention follow-up were:

(1) Adolescent/family dyads in the intervention will maintain or increase quality of life and psychological adjustment compared with controls, a marker for safety. (2) Adolescents in the intervention group will be signi cantly more likely to engage in a robust pACP session, followed by completion of an advance directive accessible in their med- ical chart, compared with controls. (3) Adolescents in the intervention group will have signi cantly increased spirituality compared with controls.


From January 2010 through July 2012, we conducted a small, randomized, controlled, clinical trial in which adolescent/family


M.E. Lyon et al. / Journal of Adolescent Health 54 (2014) 710 e717

dyads were computer-randomized to either study arm, using an intention-to-treat design.


Data were collected from 30 adolescents and 30 family members. Inclusion criteria for adolescents were that they knew the diagnosis, were age 14 e 21 years, and had an available legal guardian (if < 18 years of age) or family member who was at least 21 years of age. Exclusion criteria were severe depression, homicidality, suicidality, psychosis, being in foster care, severe developmental delays, or impaired mental status. Eligibility was not based on prognosis. Potential participants were either actively treated or in long-term care. Most patients were receiving cure-directed treatments. The institutional review board reviewed and approved the curriculum/protocol and measures. Signed assent and consent were obtained before eligibility screening.


Potential participants were identi ed with the help of oncology physicians and other providers. Providers did not discuss the study directly with patients, to avoid risk of coercion. Research assistants approached the adolescent or family only after clearance with the treating physician. Five study visits were conducted in the hospital in-patient unit, outpatient clinic, or patient s home: Visit 1: baseline assessment; Visit 2 (Session 1), the Lyon Family-Centered Advance Care Planning Survey [27,28] ; Visit 3 (Session 2), the Respecting Choices Disease-Speci c Advance Care Planning Interview [29] ; Visit 4 (Session 3), the Five Wishes [30] ; and Visit 5: 3-month post-intervention follow- up. Family-Centered Advance Care Planning for Teens With Cancer consists of three weekly, 60-minute sessions in a dyadic format with a trained or certi ed facilitator ( Table 1 ).

Treatment as usual control

All participants were given an ACP education brochure during baseline. Assessments were administered at the same ve time points as intervention dyads (baseline, Sessions 1 e3, and 3- month follow-up).


Demographic data and diagnosis were collected from patient report and con rmed by chart review. The second author then con rmed diagnosis and relapse status. Enrollment, attendance, retention, and completeness of data were gathered from the study database . Measures were administered separately to the adolescent and family by a research assistant who was not the facilitator. The Satisfaction Questionnaire was developed and pilot- tested for the FACE protocol with HIV-positive adolescents [18] . Thirteen items are answered on a 5-point Likert scale, ranging from strongly disagree to strongly agree, after each of the three intervention sessions. Higher scores indicate higher satisfaction. Five Wishes [30] is a legal document that facilitates expres- sion of treatment preferences and served as a tool to facilitate the participation of adolescents under age 18 years, but was signed by the family to be legal.

The Beck Anxiety Inventory [31] is a 21-item measure of anxiety rated on a 4-point Likert scale of symptoms over the past week. The Beck Anxiety Inventory has demonstrated reliability and validity to assess anxiety in individuals aged 17 e 80 years. We extended downward to age 14 years to allow for consistency of data collection. Clinical score interpretation of levels of anxiety is: 0 e 7, minimal; 8 e 15, mild; 16 e 25, moderate; and 26 e 63, severe. The Beck Depression Inventory e II [32] is a 21-item scale self- report assessing the presence of symptoms of depression over the past 2 weeks on a 4-point Likert scale for adolescents > 13 years of age. Clinical score interpretation of levels of depression is: 0 e13, minimal; 14 e19, mild; 20 e 28, moderate; and 29 e 63, severe. The Pediatric Quality of Life Inventory 4.0 Generic Core Scales [33] is a 23-item measure of health-related quality of life (physical, emotional, social, and school). Norms exist for up to age 18.9 years, which was extended upward to allow for one measure across participants. The integrated Pediatric Quality of Life Cancer-Speci c Module measured cancer symptoms. Higher scores indicate better quality of life. The Spiritual Well-Being Scale of the Functional Assessment of Chronic Illness Therapy e Version 4 [34] , a 23-item scale, fo- cuses primarily on the existential aspects of spirituality (e.g., I feel hopeful, ” “ I feel forgiven for any harm I may have ever caused, ” “ I feel a sense of thankfulness for my life ). There are two subscales: Meaning/Peace (Items 1 e8 [e.g., I feel peaceful, ” “ I have a reason for living ]) and Faith (Items 9 e12 [e.g., I nd comfort/strength in my faith spiritual beliefs, ” “ My illness has strengthened my faith or spiritual beliefs ]). The higher the score, the better the spiritual well-being.

Trained or certi ed facilitators

Facilitators were four part-time, ethnically diverse, female graduate students. Facilitators were trained or certi ed through the Respecting Choices competency-based Disease-Speci c ACP Facilitator course, which included an on-line ACP course, a 2-day face-to-face training with role plays, and validation of skills through the submission of a role play video to ensure compe- tency and delity in implementing the Respecting Choices Interview [30] . Training also was provided for highly structured Sessions 1 and 3. Facilitators debriefed with the rst author.


Analyses used SAS 9.2 (SAS Institute, Inc., Cary, NC). The impact of FACE-TC on psychological well-being, quality of life, and spirituality over a 3-month observation period was assessed using the Generalized Estimating Equation (GEE) models. The GEE model accommodates longitudinal data with attrition. In the GEE models, the main effect of time represents the outcome change from baseline to the 3-month follow-up, and the outcome change may depend on intervention status. When the interaction between time and intervention is statistically sig- ni cant, the main effect of time only represents the outcome change in the control group. Outcome change in the intervention group is the sum of the time main effect and the interaction ef- fect between time and intervention. Data were missing for two dyads for the 3-month post- intervention follow-up. One patient died and another was too ill to participate. We conducted the 3-month outcomes analysis

M.E. Lyon et al. / Journal of Adolescent Health 54 (2014) 710 e717


Table 1 Description of Family-Centered Advance Care Planning intervention [19]

Session 1 Foundation

Session 1 Goals

Session 1 Process

Lyon Family-Centered Advance Care Planning Survey e Adolescent and Surrogate Versions, which engage the participant in EOL questions

1. To assess adolescents and surrogates values, spiritual and other beliefs, and life experiences with illness and EOL care

2. To assess when to initiate EOL discussion and planning.

1. Research assistant orients family to study and issues

2. Adolescent is surveyed separately from the surrogate

3. Surrogates are surveyed privately with regard to what they believe their adolescent prefers


4. Research assistant highlights the similarities in their responses, building cohesion

Session 2 Foundation

Session 2 Goals

Session 2 Process

Respecting Choices Interview (Briggs and Hammes)

1. To facilitate conversations and shared decision making between the adolescent and family about palliative care, providing an opportunity to express fears, values, spiritual and other be- liefs and goals with regard to death and dying

Stage 1 assesses the adolescent s understanding of current medical condition, prognosis, complications Stage 2 explores adolescents philosophy regarding EOL decision making and their understanding of the facts Stage 3 reviews rationale for future medical decisions the adolescent would want the surrogate to understand or act upon Stage 4 uses the Statement of Treatment Preferences to describe clinical situations common to acquired immunode ciency syndrome and related treatment choices Stage 5 summarizes the discussion or need for future discussions as situations or preferences change. Gaps in information are identi ed and referrals are made.


To prepare the family to be able to fully repre- sent the adolescent s wishes

Session 3 Foundation

Session 3 Goals

Session 3 Process

The Five Wishes is a legal document that helps people express how they want to be treated if they are seriously ill and unable to speak for themselves. Unique among living will and health agent forms, it looks to all of a person s needs:

1. Which person the teen wants to make health care decisions for him or her

For adolescents < age 18 years, the Five Wishes must be signed by their legal guardian. Processes, such as labeling feelings and concerns, as well as nding solutions to any identi ed problem, are facilitated. Appropriate referrals are made to help resolve disagreements over decision making (e.g., a hospital ethicist or their doctor) or spiritual issues (e.g., chaplain or their clergy).This session may include other family members or loved ones

2. The kind of medical treatment the teen wants

3. How comfortable the teen wants to be

4. How the teen wants people to treat him or her

medical, personal, emotional, and spiritual.

5. What teen wants loved ones to know

6. Any spiritual or religious concerns teens may have

EOL ¼ end of life.

for 16 adolescent/family dyads randomized to FACE-TC and 12 dyads randomized to treatment as usual. One adolescent did not complete the Spiritual Well-Being Scale of the Functional Assessment of Chronic Illness Therapy e Version 4 Sp-EX.


Participant characteristics

We approached 126 potential subjects, 34 of whom declined, 2 of whom had no surrogate, 18 of whom were interested but not ready, and 10 of whom were ineligible. A total of 31 dyads were screened for eligibility; one dyad failed screening because the teen did not know she had cancer. We then randomized 30 dyads: 17 to FACE-TC and 13 to treatment as usual ( Table 2 ). There were no statistically signi cant differences in baseline characteristics between study arms.


We achieved benchmarks, including 72% of eligible families enrolled, 100% attendance at all three sessions, 93% retention at 3-month post-intervention follow-up, and 100% data

completeness for the 56 participants who completed the 3- month follow-up.


We evaluated and compared the Satisfaction Questionnaire [18] scores immediately after Sessions 1 e3 between FACE-TC intervention adolescents and their families. A similar analysis was not done in the control group, because they received no facilitated conversations. All intervention families rated each of the three sessions as worthwhile and universally found FACE-TC helpful (100%). Adolescents worthwhile ratings increased over time from 65% for Session 1 (survey) to 71% for Session 2 (interview), and to 82% for Session 3 (advance directive); 88% e 94% of adolescents found the intervention helpful, depending on the session. Participants were asked whether there was anything else they would like to add. Comments from adolescents included: I m not use to talking about the dying part. ” “ I think it just refocused . my decisions. ” “ It was a good session. I learned a lot about what to think about d good thought process. Com- ments from families included I think this is very, very good and this is a way to bring up dif cult topics to the table. ” “ This was a relief . I always want to know what my son thinks. ” “ The


M.E. Lyon et al. / Journal of Adolescent Health 54 (2014) 710 e717

Table 2 Baseline characteristics for adolescents with cancer (n ¼ 30) and their families (n ¼ 30)











Age, years Mean Mode Range Gender, N (%) Male Female Race Asian Black White Declined Cancer diagnosis Leukemia Lymphoma Solid tumor Brain tumor Education In high school/no high school diploma High school grad/General Equivalency Diploma equivalent Some college/no bachelors College or higher education Income At or below federal poverty level 200% federal poverty level 300% federal poverty level > 300% federal poverty line Unreported














14 e 21

14 e 20

14 e 21

36 e 62

22 e 56

22 e 62

12 (71)

6 (46)

18 (60)

1 (6)

1 (8)

2 (7)

5 (29)

7 (54)

12 (40)

16 (94)

12 (92)

28 (93)

1 (6)

1 (8)

2 (7)


1 (8)

1 (3)

7 (41)

6 (46)

13 (43)

7 (41)

6 (46)

13 (43)

9 (53)

6 (46)

15 (50)

9 (53)

6 (46)

15 (53)




1 (6)


1 (3)

6 (35)

8 (62)

14 (47)

2 (12)


2 (7)

4 (24)

2 (15)

6 (20)

5 (29)

3 (23)

8 (27)

14 (82)

10 (77)

24 (80)

1 (6)


1 (3)

1 (6)

1 (8)

2 (7)

1 (6)

4 (31)

5 (16)

2 (12)

2 (15)

4 (13)

10 (59)

3 (23)

13 (43)




5 (29)

6 (46)

11 (36)


2 (12)

2 (15)

4 (13)

1 (6)

1 (8)

2 (7)

6 (35)


6 (20)

5 (29)

9 (69)

14 (47)

3 (18)

1 (8)

4 (13)

FACE ¼ Family-Centered Advance Care planning; SOC: standard of care control.

questions asked were important questions [I] had not thought about and [I] can now discuss with [my] son.

Anxiety and depression

From baseline to 3-month follow-up, anxiety scores of ado- lescents dropped signi cantly ( b ¼ 5.6; p ¼ .0212) in both groups because time effect does not depend on intervention assignment (i.e., time e intervention interaction is not statistically signi cant [ b ¼ 1.4; p ¼ .649). Tables 3 and 4 show the means and standard deviations (SD) and Table 5 displays the GEE results. The anxiety score of surrogates at 3-month follow-up dropped signi cantly

Table 3 Descriptive statistics of psychological at the baseline and 3-month Post intervention



Standard of care

advance care planning control (N ¼ 13) for teens with cancer (N ¼ 17)


3 month


3 month

Mean (SD) Mean (SD) Mean (SD) Mean (SD)

Beck Anxiety Inventory (range, 0 e 126) Adolescent Surrogate s own mood Beck Depression Inventory (range, 0 e 63) Adolescent Surrogate s own mood

6.8 (8.2)

2.6 (2.2)

9.8 (10.0) 4.0 (3.2)

3.4 (3.4)

4.0 (5.1)

4.3 (8.6)

3.5 (8.7)

5.5 (4.8)

6.3 (5.3)

10.9 (8.1) 7.4 (4.3)

5.4 (6.6)

5.3 (7.7)

5.8 (5.8)

5.3 (8.0)

Higher scores represent higher symptom levels. SD ¼ standard deviation.

from baseline. With respect to change in the anxiety score of surrogates, the time effect depends on intervention assignment, because the time e intervention interaction is statistically signi - cant ( b ¼ 2.0; p ¼ .039). That is, the anxiety score dropped signi cantly ( b ¼ 1.2; p ¼ .0314) in the control group but increased in families in the intervention group ( 1.2 þ 2.0 ¼ .8). The FACE-TC adolescents had a signi cantly lower depression score ( b ¼ 5.4; p ¼ .0268) at baseline and 3-month follow-up compared with controls. Table 3 shows the means and SD and Table 5 lists the GEE results. One control had just learned of a relapse before completing the baseline assessment and scored in the severe range. The patient was receiving psychological sup- port and wanted to be in the study.

Quality of life

There were no signi cant differences for adolescents in quality of life by group, controlling for baseline levels by inter- vention or time. Table 4 reports the mean scores and SD and Table 5 reports GEE results. School score evaluated by surrogate in the FACE-TC group improved signi cantly ( b ¼ 15.9; p ¼ .0287) compared with the control group. Overall adolescents rated quality of communication with their providers very highly, with means ranging from 83.3 (SD, 15.9) to 94.4 (SD, 8.9). Adolescents did not report signi cant differences in symptoms by group, nor were there signi cant differences over time. Table 4 reports mean scores and SD, not shown in Table 5 .

Advance directive

The FACE-TC adolescents completed the Five Wishes [31] advance directive with their families at a higher rate than

M.E. Lyon et al. / Journal of Adolescent Health 54 (2014) 710 e717


Table 4 Descriptive statistics of Pediatric Quality of Life Inventory: General and Cancer Symptom; and Functional Assessment of Chronic Illness Therapy e Spiritual Well- Being Scale, Version 4 at baseline and 3-months post-intervention

Quality of life

Family-Centered advance care planning for teens with cancer (N ¼ 17)

Standard of care control (N ¼ 13)


3 month


3 month

Mean (SD) Mean (SD) Mean (SD) Mean (SD)

Adolescent Total Physical School Emotional Social Surrogate perception of adolescent quality of life Total Physical School Emotional Social Quality of life: cancer symptom of adolescents Pain and hurt Nausea Procedural anxiety Treatment anxiety Worry Cognitive Perceived physical appearance Communication Quality of life:

spirituality of

71.9 (17.4) 77.2 (13.4) 68.7 (17.4) 76.2 (10.4) 70.2 (23.7) 77.9 (19.8) 62.5 (25.4) 69.0 (23.0) 55.3 (25.2) 60.4 (17.6) 58.1 (21.6) 70.6 (15.1) 75.0 (19.9) 79.7 (17.7) 82.3 (17.9) 84.1 (14.3) 85.0 (11.3) 86.9 (11.5) 75.8 (17.3) 84.5 (10.8)

68.9 (18.9) 74.7 (15.8) 61.7 (16.3) 66.9 (11.1) 65.3 (29.4) 73.6 (21.5) 55.3 (27.8) 57.6 (20.7) 65.9 (22.2) 73.5 (18.9) 50.0 (17.8) 60.1 (22.8) 73.5 (19.7) 72.2 (21.0) 69.6 (14.6) 72.5 (16.0) 75.0 (17.3) 80.6 (14.6) 76.5 (17.5) 81.4 (14.7)

75.7 (25.2) 78.1 (28.7) 74.0 (30.0) 79.2 (22.8) 82.6 (17.5) 89.1 (12.3) 70.8 (19.9) 76.3 (20.9) 81.4 (22.7) 79.7 (20.9) 88.5 (19.7) 84.0 (17.9) 86.3 (16.4) 88.0 (20.4) 91.7 (12.8) 91.7 (13.8) 735. (26.9) 67.7 (29.3) 71.2 (14.3) 70.1 (17.6) 66.3 (25.2) 71.9 (20.4) 76.9 (16.7) 74.2 (17.4) 82.4 (23.4) 87.5 (16.9) 88.8 (13.8) 91.7 (15.5)

83.3 (15.9) 84.4 (15.2) 89.1 (18.4) 94.4 (8.9)



28.2 (3.8)

27.6 (3.6)

24.4 (5.5)

25.4 (4.0)


13.2 (4.0)

12.2 (4.4)

11.8 (3.7)

9.9 (4.9)

Total score

78.9 (13.1) 78.2 (8.1)

70.8 (7.8)

67.2 (14.3)

Higher scores indicate better quality of life. SD ¼ standard deviation.

standard of care (100% vs. 0%) and the advance directive docu- ments were easily found in the medical record at 3-month post- intervention. All treating oncologists received an electronic copy with a process note. When asked, When do you think is the best time to bring up end-of-life decisions? intervention adolescents responded, Before getting sick (19%; n ¼ 3), At diagnoses (19%; n ¼ 3), When rst hospitalized (0%); When dying (25%; n ¼ 4), or all of the above (38%; n ¼ 6). Only one adolescent reported ever talking to anyone about wishes for care at EOL before the study.


Total score of adolescents in the FACE-TC group was signi - cantly higher ( b ¼ 8.1; p ¼ .0296) at baseline and 3-month follow-up, compared with controls (see Table 3 for means and SD, and Table 5 for GEE results). Peace subscale scores of ado- lescents in the FACE-TC group were signi cantly higher ( b ¼ 3.9; p ¼ .0239) at baseline and at 3-month follow-up, compared with controls. Faith subscale scores of adolescents at 3-month follow- up dropped signi cantly from baseline in both groups, because the time effect is statistically signi cant ( b ¼ 2.2; p ¼ .0466) and

the effect does not depend on intervention assignment (the time e intervention interaction is not statistically signi cant [ b ¼ 1.20; p ¼ .3654]).


This is the rst randomized, controlled, clinical trial of pACP for adolescents with cancer. This innovative model provided support in decision making about limiting burdensome medical interventions and completion of advance directives, using highly structured conversations that adolescents with cancer and their families found feasible, acceptable, and safe, with 100% atten- dance at all three sessions, high retention at 3-month follow-up, and universal completion of advance directives. Although it had a small sample, this study lls the gap in our knowledge about actual receipt of pACP by adolescents with cancer, answering questions that have been persistent barriers to pACP: How? Evidence-based structured conversations with the adolescent and family. When? From the onset of diagnosis. Prognosis is not necessary. Cure-directed care can be combined with pACP. Where? Hospital based in-patient and outpatient settings. By whom? Trained or certi ed facilitators. Family-Centered Advance Care Planning for Teens With Can- cer demonstrated that participating in a pACP study alleviated adolescents anxiety signi cantly. In contrast, intervention fam- ilies were statistically more anxious than controls at 3-month post-intervention. However, these differences are clinically insigni cant, with scores indicating minimal levels of anxiety. Furthermore, 100% of intervention families rated each session as worthwhile. Families especially valued the increased capacity to have EOL conversations with their children and to know what their child wanted. These ndings con rm our HIV pilot results, which also found that pACP decreased adolescents anxiety, yet increased families anxiety, but to clinically insigni cant levels [13] . This rise in anxiety is not surprising because the death of one s child is one of life s most stressful events [7] . Low depressive symptom scores and high quality of life scores were maintained by adolescents in both groups. Intervention families perceived improved school quality of life at follow-up, whereas their teens did not. Perhaps study communication generalized to the school domain, suggesting to parents that their adolescents school quality of life had improved. General- ized increases in communication as the result of a pACP in-depth conversation merit further study. As hypothesized, adolescents experience of re ection on death and dying sustained high levels of peace and spirituality. Intervention exposure appears to offer a protective effect regarding the general pattern of declining faith, but to an insig- ni cant degree. Future research should explore this phenomenon. Family-Centered Advance Care Planning for Teens With Can- cer demonstrated acceptability to African-American adolescents with cancer and their families, an underserved population who composed almost half of this sample. Our ndings also con rm that seriously ill adolescents understand treatment deci- sions [3 e 5,25,35,36] , want to be involved in pACP decisions throughout the course of their illness, and have the cognitive capacities necessary for an active role in decision making about their care [3 e 5,19,35,36] . Contrary to the HIV pilot, adolescents with cancer were less satis ed with the FACE-TC intervention (ranging from 75% in Session 1% e 82% in Session 3) than adolescents with HIV (ranging


M.E. Lyon et al. / Journal of Adolescent Health 54 (2014) 710 e717

Table 5 Selected results of Generalized Estimating Equation model


Intercept ( p value)

Main effect of Time a ( p value)

Main effect of Intervention b ( p value)

Interaction between time and intervention ( p value)

Psychological score Beck Anxiety Inventory Adolescent Surrogate s own mood Beck Depression Inventory e II Adolescent Surrogate s own mood Quality of life Adolescent Total Physical School Emotional Social Surrogate perception of adolescent s quality of life Total Physical School Emotional Social Quality of life: spirituality of adolescents Peace subscale Faith subscale Total

9.8 (.0002)

5.6 (.0212) * 1.2 (.0341) *

3.1 (.3542) .9 (.6973)

1.4 (.6490) 2.0 (.0392) *

4.3 (.0609)

10.9 ( < .0001) 5.8 (.0001)

3.0 (.1007) .9 (.5375)

5.4 (.0268) * .4 (.8424)

3.8 (.0642)

1.2 (.5007)

98.7 ( < .0001) 62.5 ( < .0001) 58.1 ( < .0001) 82.3 ( < .0001) 75.8 ( < .0001)

5.9 (.1123)

3.1 (.6132) 7.7 (.3784) 2.7 (.7472) 7.3 (.2741) 9.2 (.0830)

1.5 (.7339) 3.1 (.6043) 9.3 (.3674) 3.6 (.5193) 6.2 (.1552)

3.7 (.4087)

13.6 (.1085)

.2 (.9731)

7.2 (.0630)

61.7 ( < .0001) 55.3 ( < .0001) 50.0 ( < .0001) 69.6 ( < .0001) 76.5 ( < .0001)

3.8 (.2435)

7.2 (.2475) 10.0 (.3249) 15.9 (.0287) * 3.9 (.5179) 1.5 (.8037)

.8 (.8741) 6.4 (.4889) 2.0 (.8445) 4.3 (.5586) 1.1 (.8464)

.3 (.9643)

9.8 (.1594)

2.3 (.6373)

4.0 (.3912)

24.4 ( < .0001) 11.8 ( < .0001) 70.8 ( < .0001)

1.0 (.6228) 2.2 (.0466) * 3.9 (.3614)

3.9 (.0239) * 1.3 (.3286) 8.1 (.0296) *

1.6 (.4380) 1.2 (.3654) 3.0 (.5389)

The Generalized Estimating Equation results of scores about quality of life for cancer symptoms of adolescents are not included because both time and intervention did not have signi cant statistical effects.

a Time was coded as 0 ¼ baseline and 1 ¼ 3-month post-intervention follow-up.

b Intervention was coded as 0 ¼ standard of care control and 1 ¼ Family-Centered Advance Care Planning Intervention. * p < .05.

from 92% to 100%) [18] . Differences may have arisen because adolescents in the FACE-HIV study were medically stable out- patients, whereas many adolescents in the FACE-TC study were not, so the discussion was not hypothetical. These ndings may not generalize beyond this study site, although participants represent an ethnically diverse group in an urban children s hospital. A larger, multisite, randomized, controlled trial is needed. Restriction of range in depression scores at recruitment may have decreased the chance of nding signi cant differences between groups. However, depressed in- dividuals should not participate in ACP because of the likelihood of impaired decision making (e.g., suicidality). There was no active control group for time and attention, which may have magni ed differences. The control arm may also have caused bias in the other direction by bringing up the topic of pACP, which is often not addressed, especially in children < 18 years of age. Enrolling sick adolescents in an active control group was judged to be too burdensome. Families ready to discuss pACP agreed to participate, which created sample selection bias. Nevertheless, almost three fourths of eligible dyads approached participated, a higher rate than in comparable adult studies. Furthermore, the intention of pACP is to individualize the process, so adolescents who prefer to have their doctor or their families make decisions for them do so. Sustainability of FACE-TC is unknown and the study was not blinded, which potentially created bias. Although the sample size was small, it was large compared with other pACP studies, and bene ted from being a randomized clinical trial with an intent-to-treat design. Fully powered, longitudinal, blinded studies are needed. This study moves the eld of pACP forward, implementing structured conversations in practice. Adolescents with cancer

recognized pACP as a process, with 75% preferring to discuss pACP early or throughout the illness trajectory, which con rms earlier ndings with adolescents [3,28] and supports the Amer- ican Academy of Pediatrics recommendation of the concurrent use of PC and curative care for children from the time of diag- nosis of a life-threatening illness [37] . Safeguards [38] were put in place that represent effective ACP: (1) an iterative process of adapting a structured, disease-speci c protocol to be develop- mentally appropriate and culturally sensitive; (2) the patient and family/surrogate decision maker were present; (3) conversations rst addressed adolescents values and goals of care before dis- cussing treatment preferences; (4) trained or certi ed facilitators conducted these conversations; (5) oncologists received e-mail communication about the process and pACP documents; and (6) advance directives were readily available in the medical record. These ndings should inform research directions and clinical care for teens with cancer and other serious illnesses.


In addition to the co-authors, the following individuals contributed signi cantly to this work. The authors thank their families for participation and the community for help in devel- oping this protocol, especially Jessica Gaines and her mother, Cynthia Brawner-Jackson. The authors extend their gratitude to the research assistants who worked with the families: Jessica Gaines, Joy Lere, Jessica Greenberg, and Brittney Lee. The authors thank Jennie Perez for help with institutional review board submissions and modi cations. They thank Saeid Goudarzi for help with data management and Annie Fulton for help with creating Teleforms. They thank their consultants, Kathleen

M.E. Lyon et al. / Journal of Adolescent Health 54 (2014) 710 e717


Ennis-Durstine for spiritual/religious consultations and Tomas Silber for ethics consultations. They thank the health care pro- viders for their assistance, especially Debbie LaFond and Gene Hwang. Data on the development of the protocol were presented as posters at the following: 16th Kathleen Foley Palliative Care Retreat in Park City, Utah, October 24, 2012; American Psycho- logical Annual Convention, Orlando, Florida (ind120615), August 2012; Pediatric Academic Societies Annual Meeting, Boston, Massachusetts (Publication 3854.701, Session: 3854 Hematology and Oncology, April 30, 2012; Overcoming barriers to recruit- ment in a dyadic study of adolescents with cancer, Society of Adolescent Health and Medicine Meeting, New Orleans, Louisi- ana, March 15, 2012; Advocating for adolescents: Family- Centered (FACE) advance care planning,The Science of Compassion: Future Directions in End-of-Life and Palliative Care, National Institute of Nursing Research/National Institutes of Health, Bethesda, Maryland, August 11, 2011. Data on the 3- month outcomes were presented at the Fourth International Advance Care Planning and End of Life Conference in Melbourne, Australia, on May 9 and 11, 2013.

Funding Sources

This pilot study was funded by Grant PEP-10-171-01-PCSM from the American Cancer Society; the development and adap- tation phase was funded by the Research Advisory Award from the Children s National Medical Center; and both ethics consul- tations and space for meeting with families were funded by Clinical and Translational Science Institute-Children's National (CTSI-CN) Grant UL1TR000075.

Supplementary Data

Supplementary data related to this article can be found at .





946 e 50 .

733 e 42 .


445 e 51 .

32 e 5 .


Lyon ME. Lyon Family Centered Advance Care Planning Survey e Adolescent Version and Patient Version, ; 2006.


[30] Towey J. The ve wishes. Aging with Dignity: a service program developed through a grant from the Robert Wood Johnson Foundation. Available at:

49 e 58 .

351 e 7 .