Journal of Adolescent Health 54 (2014) 710e717

www.jahonline.org
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
Childrens National Health System, Division of Adolescent and Young Adult Medicine, Childrens Research Institute, Center for Translational Science and George Washington
University School of Medicine and Health Sciences, Washington, DC
b
Childrens National Health System, Division of Oncology/Childrens 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
Childrens National Health System, Childrens 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

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

A B S T R A C T

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 14e20years)/family dyads (N 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 outcomes as measured by the Pediatric Quality of Life Inventory 4.0 Generic Core Scale, the Pediatric
Quality of Life Inventory 4.0 Cancer-Specic Module, the Beck Depression and Anxiety Inventories,
the Spiritual Well-Being Scale of the Functional Assessment of Chronic Illness TherapyeIV, and
advance directive completion.
Results: Acceptability was demonstrated with enrollment of 72% of eligible families, 100% attendance 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%e82%). Adolescents anxiety decreased signicantly from baseline to 3 months postintervention in both groups (b 5.6; p .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 8.1; p .0296) were signicantly higher
among FACE-TC adolescents versus controls. The FACE-TC adolescents endorsed the best time to

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

IMPLICATIONS AND
CONTRIBUTION

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

facilitators to use the Respecting Choices tool. This training is provided by the
non-for-prot 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, Childrens Research Institute, Childrens National Health System, Room
M7658, 111 Michigan Avenue, N.W., Washington, DC 20010-2970.
E-mail address: maureen.lyon2222@gmail.com (M.E. Lyon).

1054-139X/$ e see front matter 2014 Society for Adolescent Health and Medicine. All rights reserved.
http://dx.doi.org/10.1016/j.jadohealth.2013.10.206

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

711

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, indepth 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 medicalized. 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 directives. Adolescents want a voice in choices for their end-of-life
(EOL) care, if they cannot speak for themselves [3e5]. Avoidance
of EOL conversations can lead to miscommunication and disagreements, which may result in families charged with neglect,
court battles, and even legislative intervention [6]. The Institute
of Medicines 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 insufcient to ensure that
patients treatment preferences are honored at EOL [8]. Building
on these ndings, research identied the critical importance of
including the surrogate decision maker in structured conversations 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 reections on death and spirituality.
Only three structured pediatric ACP (pACP) programs are reported for adolescents with complex chronic conditions. First is
FOOTPRINTS, a model of pACP and coordination specic to patients with muscular dystrophy (age 8e21 years) [17]. Second is
Family-Centered (FACE) Advance Care Planning, a model specic
to adolescent patients with human immunodeciency virus
(HIV) (age 14e21 years) and their family. This was the rst protocol 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 conict, 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 understanding of death as those aged 18 years [7]. Unique challenges
also arise from the adolescentefamily dynamic if there are disagreements 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 adolescents [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 cancers 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 efcacy 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 inuence of
pACP on spirituality. Predetermined benchmarks for acceptability 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 signicantly
more likely to engage in a robust pACP session, followed by
completion of an advance directive accessible in their medical chart, compared with controls.
(3) Adolescents in the intervention group will have signicantly
increased spirituality compared with controls.

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

712

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

dyads were computer-randomized to either study arm, using an

intention-to-treat design.
Participants
Data were collected from 30 adolescents and 30 family
members. Inclusion criteria for adolescents were that they knew
the diagnosis, were age 14e21 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.
Procedures
Potential participants were identied 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
patients 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-Specic
Advance Care Planning Interview [29]; Visit 4 (Session 3), the
Five Wishes [30]; and Visit 5: 3-month post-intervention followup. Family-Centered Advance Care Planning for Teens With
Cancer consists of three weekly, 60-minute sessions in a dyadic
format with a trained or certied 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 1e3, and 3month follow-up).

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 17e80 years.
We extended downward to age 14 years to allow for consistency
of data collection. Clinical score interpretation of levels of anxiety
is: 0e7, minimal; 8e15, mild; 16e25, moderate; and 26e63,
severe.
The Beck Depression InventoryeII [32] is a 21-item scale selfreport 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: 0e13, minimal; 14e19, mild; 20e28, moderate;
and 29e63, 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-Specic Module measured cancer symptoms. Higher
scores indicate better quality of life.
The Spiritual Well-Being Scale of the Functional Assessment
of Chronic Illness TherapyeVersion 4 [34], a 23-item scale, focuses 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 1e8 [e.g., I feel peaceful,
I have a reason for living]) and Faith (Items 9e12 [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 certied facilitators
Facilitators were four part-time, ethnically diverse, female
graduate students. Facilitators were trained or certied through
the Respecting Choices competency-based Disease-Specic 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 competency 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.

Measures
Analysis
Demographic data and diagnosis were collected from patient
report and conrmed by chart review. The second author then
conrmed 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 pilottested 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 expression 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.

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 signicant, 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 effect between time and intervention.
Data were missing for two dyads for the 3-month postintervention 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) 710e717

713

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

SurveyeAdolescent 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 beliefs and goals with regard to death and dying
2. To prepare the family to be able to fully represent the adolescents wishes

Stage 1 assesses the adolescents 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 immunodeciency
syndrome and related treatment choices
Stage 5 summarizes the discussion or need for
future discussions as situations or preferences
change. Gaps in information are identied and
referrals are made.

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 persons needs:
medical, personal, emotional, and spiritual.

1. Which person the teen wants to make health

care decisions for him or her
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
5. What teen wants loved ones to know
6. Any spiritual or religious concerns teens may
have

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 identied 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

EOL 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 TherapyeVersion 4 Sp-EX.

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

Results

We evaluated and compared the Satisfaction Questionnaire

[18] scores immediately after Sessions 1e3 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: Im 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 aboutdgood thought process. Comments from families included I think this is very, very good and
this is a way to bring up difcult topics to the table. This was
a relief . I always want to know what my son thinks. The

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 signicant differences in baseline
characteristics between study arms.
Feasibility
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

Satisfaction/worthwhile

714

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

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

Adolescent

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

Family

FACE

SOC

Total

FACE

SOC

Total

16.5
16
14e21

16.0
14
14e20

16.3
14
14e21

45.4
46
36e62

42.1
38
22e56

43.9
46.0
22e62

12 (71)
5 (29)

6 (46)
7 (54)

18 (60)
12 (40)

1 (6)
16 (94)

1 (8)
12 (92)

2 (7)
28 (93)

1 (6)
7 (41)
9 (53)

1 (8)
6 (46)
6 (46)

2 (7)
13 (43)
15 (50)
0

1 (8)
6 (46)
6 (46)
0

1
13
15
1

(3)
(43)
(53)
(3)

8 (62)

14
2
6
8

0
6
2
4
5

(35)
(12)
(24)
(29)

14 (82)
1 (6)
2 (12)
0

0
2 (15)
3 (23)
10 (77)
1 (8)
2 (15)
0

7 (41)
9 (53)
1 (6)

(47)
(7)
(20)
(27)

24 (80)
2 (7)
4 (13)
0

1
1
10
5

(6)
(6)
(59)
(29)

0
4 (31)
3 (23)
6 (46)

1
5
13
11

(3)
(16)
(43)
(36)

2
1
6
5
3

(12)
(6)
(35)
(29)
(18)

2
1
0
9
1

4
2
6
14
4

(13)
(7)
(20)
(47)
(13)

(15)
(8)
(69)
(8)

FACE 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 adolescents dropped signicantly (b 5.6; p .0212) in both groups
because time effect does not depend on intervention assignment
(i.e., timeeintervention interaction is not statistically signicant
[b 1.4; p .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 signicantly

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

Standard of care
Family-Centered
advance care planning control (N 13)
for teens with cancer
(N 17)
Baseline

3 month

Baseline

3 month

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

Beck Anxiety Inventory
(range, 0e126)
Adolescent
Surrogates own mood
Beck Depression Inventory
(range, 0e63)
Adolescent
Surrogates own mood

6.8 (8.2)
3.4 (3.4)

2.6 (2.2)
4.0 (5.1)

5.5 (4.8)
5.4 (6.6)

6.3 (5.3)
5.3 (7.7)

Higher scores represent higher symptom levels.

SD standard deviation.

9.8 (10.0) 4.0 (3.2)

4.3 (8.6) 3.5 (8.7)

10.9 (8.1)
5.8 (5.8)

7.4 (4.3)
5.3 (8.0)

from baseline. With respect to change in the anxiety score of

surrogates, the time effect depends on intervention assignment,
because the timeeintervention interaction is statistically signicant (b 2.0; p .039). That is, the anxiety score dropped
signicantly (b 1.2; p .0314) in the control group but
increased in families in the intervention group (1.2 2.0 .8).
The FACE-TC adolescents had a signicantly lower depression
score (b 5.4; p .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 support and wanted to be in the study.
Quality of life
There were no signicant differences for adolescents in
quality of life by group, controlling for baseline levels by intervention 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 signicantly (b 15.9; p .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 signicant differences in symptoms by group, nor
were there signicant 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) 710e717

Table 4
Descriptive statistics of Pediatric Quality of Life Inventory: General and Cancer
Symptom; and Functional Assessment of Chronic Illness TherapyeSpiritual WellBeing Scale, Version 4 at baseline and 3-months post-intervention
Quality of life

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
adolescents
Peace
Faith
Total score

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

Standard of care control

(N 13)

Baseline

3 month

Baseline

3 month

Mean (SD)

Mean (SD)

Mean (SD)

Mean (SD)

71.9
70.2
55.3
75.0
85.0

(17.4)
(23.7)
(25.2)
(19.9)
(11.3)

77.2
77.9
60.4
79.7
86.9

(13.4)
(19.8)
(17.6)
(17.7)
(11.5)

68.7
62.5
58.1
82.3
75.8

(17.4)
(25.4)
(21.6)
(17.9)
(17.3)

76.2
69.0
70.6
84.1
84.5

(10.4)
(23.0)
(15.1)
(14.3)
(10.8)

68.9
65.3
65.9
73.5
75.0

(18.9)
(29.4)
(22.2)
(19.7)
(17.3)

74.7
73.6
73.5
72.2
80.6

(15.8)
(21.5)
(18.9)
(21.0)
(14.6)

61.7
55.3
50.0
69.6
76.5

(16.3)
(27.8)
(17.8)
(14.6)
(17.5)

66.9
57.6
60.1
72.5
81.4

(11.1)
(20.7)
(22.8)
(16.0)
(14.7)

75.7
82.6
81.4
86.3
735.
66.3
82.4

(25.2)
(17.5)
(22.7)
(16.4)
(26.9)
(25.2)
(23.4)

78.1
89.1
79.7
88.0
67.7
71.9
87.5

(28.7)
(12.3)
(20.9)
(20.4)
(29.3)
(20.4)
(16.9)

74.0
70.8
88.5
91.7
71.2
76.9
88.8

(30.0)
(19.9)
(19.7)
(12.8)
(14.3)
(16.7)
(13.8)

79.2
76.3
84.0
91.7
70.1
74.2
91.7

(22.8)
(20.9)
(17.9)
(13.8)
(17.6)
(17.4)
(15.5)

83.3 (15.9)

84.4 (15.2)

89.1 (18.4)

94.4 (8.9)

28.2 (3.8)
13.2 (4.0)
78.9 (13.1)

27.6 (3.6)
12.2 (4.4)
78.2 (8.1)

24.4 (5.5)
11.8 (3.7)
70.8 (7.8)

25.4 (4.0)
9.9 (4.9)
67.2 (14.3)

Higher scores indicate better quality of life.

SD standard deviation.

standard of care (100% vs. 0%) and the advance directive documents were easily found in the medical record at 3-month postintervention. 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 3), At diagnoses (19%; n 3),
When rst hospitalized (0%); When dying (25%; n 4), or all
of the above (38%; n 6). Only one adolescent reported ever
talking to anyone about wishes for care at EOL before the study.
Spirituality
Total score of adolescents in the FACE-TC group was signicantly higher (b 8.1; p .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 adolescents in the FACE-TC group were signicantly higher (b 3.9;
p .0239) at baseline and at 3-month follow-up, compared with
controls. Faith subscale scores of adolescents at 3-month followup dropped signicantly from baseline in both groups, because
the time effect is statistically signicant (b 2.2; p .0466) and

715

the effect does not depend on intervention assignment (the

timeeintervention interaction is not statistically signicant [b
1.20; p .3654]).
Discussion
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% attendance 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 certied facilitators.
Family-Centered Advance Care Planning for Teens With Cancer demonstrated that participating in a pACP study alleviated
adolescents anxiety signicantly. In contrast, intervention families were statistically more anxious than controls at 3-month
post-intervention. However, these differences are clinically
insignicant, 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 conrm our HIV pilot results,
which also found that pACP decreased adolescents anxiety, yet
increased families anxiety, but to clinically insignicant levels
[13]. This rise in anxiety is not surprising because the death of
ones child is one of lifes 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. Generalized increases in communication as the result of a pACP in-depth
conversation merit further study.
As hypothesized, adolescents experience of reection 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 insignicant degree. Future research should explore this
phenomenon.
Family-Centered Advance Care Planning for Teens With Cancer demonstrated acceptability to African-American adolescents
with cancer and their families, an underserved population who
composed almost half of this sample. Our ndings also conrm
that seriously ill adolescents understand treatment decisions [3e5,25,35,36], 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 [3e5,19,35,36].
Contrary to the HIV pilot, adolescents with cancer were less
satised with the FACE-TC intervention (ranging from 75% in
Session 1%e82% in Session 3) than adolescents with HIV (ranging

716

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

Table 5
Selected results of Generalized Estimating Equation model
Outcome
Psychological score
Beck Anxiety Inventory
Adolescent
Surrogates own mood
Beck Depression InventoryeII
Adolescent
Surrogates own mood
Quality of life
Adolescent
Total
Physical
School
Emotional
Social
Surrogate perception of adolescents quality of life
Total
Physical
School
Emotional
Social
Quality of life: spirituality of adolescents
Peace subscale
Faith subscale
Total

Main effect of Timea

(p value)

Main effect of
Interventionb (p value)

9.8 (.0002)
4.3 (.0609)

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

3.1 (.3542)
.9 (.6973)

1.4 (.6490)
2.0 (.0392)*

10.9 (<.0001)
5.8 (.0001)

3.0 (.1007)
.9 (.5375)

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

3.8 (.0642)
1.2 (.5007)

98.7
62.5
58.1
82.3
75.8

(<.0001)
(<.0001)
(<.0001)
(<.0001)
(<.0001)

5.9
3.7
13.6
.2
7.2

(.1123)
(.4087)
(.1085)
(.9731)
(.0630)

3.1
7.7
2.7
7.3
9.2

(.6132)
(.3784)
(.7472)
(.2741)
(.0830)

1.5
3.1
9.3
3.6
6.2

(.7339)
(.6043)
(.3674)
(.5193)
(.1552)

61.7
55.3
50.0
69.6
76.5

(<.0001)
(<.0001)
(<.0001)
(<.0001)
(<.0001)

3.8
.3
9.8
2.3
4.0

(.2435)
(.9643)
(.1594)
(.6373)
(.3912)

7.2
10.0
15.9
3.9
1.5

(.2475)
(.3249)
(.0287)*
(.5179)
(.8037)

.8
6.4
2.0
4.3
1.1

(.8741)
(.4889)
(.8445)
(.5586)
(.8464)

Intercept (p value)

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)*

Interaction between time and

intervention (p value)

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 signicant statistical effects.
a
Time was coded as 0 baseline and 1 3-month post-intervention follow-up.
b
Intervention was coded as 0 standard of care control and 1 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 outpatients, 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 childrens hospital. A larger, multisite, randomized,
controlled trial is needed. Restriction of range in depression
scores at recruitment may have decreased the chance of nding
signicant differences between groups. However, depressed individuals 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
magnied 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 beneted 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 conrms
earlier ndings with adolescents [3,28] and supports the American Academy of Pediatrics recommendation of the concurrent
use of PC and curative care for children from the time of diagnosis 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-specic protocol to be developmentally 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 discussing treatment preferences; (4) trained or certied 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.
Acknowledgments
In addition to the co-authors, the following individuals
contributed signicantly to this work. The authors thank their
families for participation and the community for help in developing 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 modications. 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) 710e717

Ennis-Durstine for spiritual/religious consultations and Tomas

Silber for ethics consultations. They thank the health care providers 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 Psychological 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 recruitment in a dyadic study of adolescents with cancer, Society of
Adolescent Health and Medicine Meeting, New Orleans, Louisiana, March 15, 2012; Advocating for adolescents: FamilyCentered (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 3month 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 adaptation phase was funded by the Research Advisory Award from
the Childrens National Medical Center; and both ethics consultations 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
http://dx.doi.org/10.1016/j.jadohealth.2013.10.206.
References
[1] Feudtner C, Feinstein JA, Satchell M, et al. Shifting place of death among
children with complex chronic conditions in the United States, 1989-2003.
JAMA 2007;297:2725e32.
[2] Keele L, Keenan HT, Sheetz J, Bratton SL. Differences in characteristics of dying
who receive and do not receive palliative care. Pediatrics 2013;132:72e8.
[3] Lyon ME, McCabe MA, Patel K, et al. What do adolescents want? An
exploratory study regarding end-of-life decision making. J Adolesc Health
2004;35:529.e1e6.
[4] Pousset G, Bilsen J, De Wilde J, et al. Attitudes of adolescent cancer survivors toward end-of-life decisions for minors. Pediatrics 2009;124:
e1142e8.
[5] Wiener L, Zadeh S, Battles H, et al. Allowing adolescents and young adults
to plan their end-of-life care. Pediatrics 2012;130:897e905.
[6] Mercurio MR. An adolescents refusal of medical treatment: Implications of
the Abraham Cheerix case. Pediatrics 2007;120:1357e8.
[7] Field MJ, Behrman RE, eds. When Children Die: Improving Palliative and
end-of-life Care for children and their families. Washington, DC: Institute
of Medicine, National Academy Press; 2002.
[8] SUPPORT Principal Investigators. A controlled trial to improve care for
seriously ill hospitalized patients: The Study to Understand Prognoses and
Preferences for Outcomes and Risks of Treatments (SUPPORT). JAMA 1995;
274:1591e8.
[9] Romer AL, Hammes BJ. Communication, trust, and making choices:
Advance care planning four years on. J Palliat Med 2004;7:335e40.
[10] Solomon MZ, Sellers DE, Heller KS, et al. New and lingering controversies in
pediatric end-of-life care. Pediatrics 2005;116:872e83.

717

[11] Hinds PS, Burghen EA, Pritchard M. Conducting end-of-life studies in pediatric oncology. West J Nurs Res 2007;29:448e65.
[12] Robinson CA. Our best hope is a cure: Hope in the context of advance
care planning. Palliat Suppor 2012;1:1e8.
[13] Lyon ME, Garvie PA, Briggs L, et al. Is it safe? Talking to teens with HIV/AIDS
about death and dying: A 3-month evaluation of family centered (FACE)
advance care planningdanxiety, depression quality of life. HIV/AIDS
(Auckland, NZ) 2010;2:27.
[14] Kirchhoff KT, Hammes BJ, Kehl KA, et al. Effect of a disease-specic advance
care planning intervention on end-of-life care. J Am Geriatr Soc 2012;60:
946e50.
[15] Temel JS, Greer JA, Muzikansky A, et al. Early palliative care for patients with metastatic non-small lung cancer. N Engl J Med 2010;363:
733e42.
[16] Detering KM, Hancock AD, Reade MC, Silvester W. The impact of advance
care planning on end of life care in elderly patients: Randomized
controlled trial. BMJ 2010;340:c1345.
[17] Toce S, Collins MA. The FOOTPRINTSTM model of pediatric palliative care.
J Palliat Med 2003;6:989e1000.
[18] Lyon ME, Garvie PA, Briggs L, et al. Development, feasibility and acceptability of the Family-Centered (FACE) advance care planning intervention
for adolescents with HIV. J Palliat Med 2009;12:363e72.
[19] Lyon ME, Garvie PA, McCarter R, et al. Who will speak for me? Improving
end-of life decision-making for adolescents with HIV and their families.
Pediatrics 2009;123:e1e8.
[20] Lyon ME, Jacobs S, Briggs L, et al. Family centered advance care planning for
teens with cancer. JAMA Pediatr 2013;167:460e7.
[21] Lorenz KA, Gordon HS. Missed opportunities for advance care planning
communication during outpatient clinic visits. J Gen Intern Med 2012;27:
445e51.
[22] Giedd JN, Blumenthal J, Jeffries NO, et al. Brain development during
childhood and adolescence: A longitudinal MRI study. Nat Neurosci 1999;
2:861e3.
[23] Weithorn LA, Campbell SB. The competency of children and adolescents to
make informed treatment decisions. Child Dev 1982;53:1589e98.
[24] Kochanek KD, Xu J, Murphy SL, et al. Deaths: Final data for 2009. Table 9.
Death rates, by age and age-adjusted death rates for the 15 leading causes
of death in 2009: United States, 1999-2009. Natl Vital Stat Rep 2011;60:
32e5.
[25] Klopfenstein KJ, Hutchinson C, Clark C,, et al. Variables inuencing end-oflife care in children and adolescents with cancer. J Pediatr Hematol Oncol
2001;23:481e6.
[26] Wolfe J, Grier HE, Klar N, et al. Symptoms and suffering at the end of life in
children with cancer. N Engl J Med 2000;342:326e33.
[27] Lyon ME. Lyon Family Centered Advance Care Planning SurveyeAdolescent
Version and Patient Version, mlyon@childrensnational.org; 2006.
[28] Garvie PA, He J, Wang J, et al. An exploratory survey of end-of-life attitudes,
beliefs and experiences of adolescents with HIV/AIDS and their families.
J Pain Symptom Manage 2012;44:373e385.e29.
[29] Hammes BJ, Briggs L. Disease-specic next steps facilitator manual. La
Crosse (WI): Gundersen Lutheran Medical Foundation; 2008.
[30] Towey J. The ve wishes. Aging with Dignity: a service program developed
through a grant from the Robert Wood Johnson Foundation. Available at:
http://www.agingwithdignity.org/. Accessed November 27, 2013.
[31] Beck AT, Steer RA. Beck Anxiety Inventory manual. San Antonio (TX):
Psychological Corporation, Harcourt Brace & Company; 1993.
[32] Beck AT, Steer RA, Brown GK. Beck Depression Inventory manual. 2nd ed.
San Antonio (TX): Psychological Corporation, Harcourt Brace & Company;
1996.
[33] Varni JW, Burwinkle TM, Seid M, Skarr D. The PedsQLTM 4.0 as a pediatric
population health measure: Feasibility, reliability, and validity. Ambul
Pediatr 2003;3:329e41.
[34] Peterman AH, Fitchett G, Brady MJ, et al. Measuring spiritual well-being
in people with cancer the Functional assessment of chronic illness
TherapyeSpiritual Well-Being Scale (FACIT-Sp). Ann Behav Med 2002;24:
49e58.
[35] Himelstein BP, Hilden JM, Boldt AM, et al. Pediatric palliative care. N Engl J
Med 2004;350:1752e62.
[36] Feudtner C. Collaborative communication in pediatric palliative care: A
foundation for problem solving and decision-making. Pediatr Clin North
Am 2007;54:583e607.
[37] American Academy of Pediatrics. Committee on Bioethics and Committee
on Hospital Care: Palliative care for children. Pediatrics 2000;106(2 pt 1):
351e7.
[38] Billings JA. The need for safeguards in advance care planning. J Gen Intern
Med 2012;27:595e600.