A Comparison of Cognitive-Behavioral Therapy,

Sertraline, and Their Combination for

Adolescent Depression
GLENN A. MELVIN, PH.D., BRUCE J. TONGE, M.D., NEVILLE J. KING, PH.D.,
DAVID HEYNE, PH.D., MICHAEL S. GORDON, M.B., B.S., AND ESTER KLIMKEIT, D.PSYCH.

ABSTRACT
Objective: To evaluate cognitive-behavioral therapy, antidepressant medication alone, and combined CBT and
antidepressant medication in the treatment of depressive disorders in adolescents. Method: Seventy-three adolescents
(ages 12Y18 years) with a primary diagnosis of DSM-IV major depressive disorder, dysthymic disorder, or depressive
disorder not otherwise specified were randomly allocated to one of three treatments. Treatment outcome measures were
administered before and after acute treatment, and at a 6-month follow-up. Depression diagnosis was the primary
outcome measure; secondary measures were self- and other report and clinician rating of global functioning. The trial was
conducted at three community-based clinics between July 2000 and December 2002. Data analyses used an intent-
to-treat strategy. Results: Following acute treatment, all treatment groups demonstrated statistically significant
improvement on outcome measures (depressive diagnosis, Reynolds Adolescent Depression Scale, Revised Children`s
Manifest Anxiety Scale, Suicidal Ideation Questionnaire), and improvement was maintained at follow-up. Combined
cognitive-behavioral therapy and antidepressant medication was not found to be superior to either treatment alone.
Compared with antidepressant medication alone, participants receiving cognitive-behavioral therapy alone demonstrated
a superior acute treatment response (odds ratio = 6.86; 95% confidence interval 1.12Y41.82). Although cognitive-
behavioral therapy was found to be superior to antidepressant medication alone for the acute treatment of mild to
moderate depression among youth, this may have stemmed from the relatively low dose of sertraline used. Conclusions:
All treatments led to a reduction in depression, but the advantages of a combined approach were not evident. J. Am. Acad.
Child Adolesc. Psychiatry, 2006;45(10):1151Y1161. Key Words: depression, cognitive-behavioral therapy, sertraline.

Depressive disorders occur in È3% to 8% of ado-
lescents (Lewinsohn et al., 1993). Depressed teenagers
Accepted May 11, 2006.
Drs. Melvin, Tonge, Gordon, and Klimkeit are with the Centre for
Developmental Psychiatry and Psychology and Dr. King is with the Faculty of
Education, Monash University, Australia; and Dr. Heyne is with Develop-
mental and Educational Psychology, Leiden University, The Netherlands.
This research was supported by grants from beyondblue, Australia`s Depression
Initiative, Premier`s Youth Suicide Taskforce, Department of Human Services
Victoria, and Australian Rotary Health Research Fund. Pfizer Pharmaceuticals
provided medication free of charge but did not provide other financial support or
contribution to the experimental design or study protocol.
Correspondence to Dr. Glenn A. Melvin, Monash University Centre for
Developmental Psychiatry and Psychology, Monash Medical Centre, 246
Clayton Road, Clayton, Victoria 3168, Australia; e-mail: glenn.melvin@med.
monash.edu.au.
0890-8567/06/4510Y1151Ó2006 by the American Academy of Child
and Adolescent Psychiatry.
DOI: 10.1097/01.chi.0000233157.21925.71
experience difficulties in family, social, and academic
functioning (Lewinsohn et al., 1998) and are at elevated
risk of attempting and completing suicide (Lewinsohn
et al., 1994; Shaffer et al., 1996). Timely and effective
treatment is highly important. Of the psychological
treatments, cognitive-behavioral therapy (CBT) has
demonstrated short-term efficacy relative to waitlist
(e. g., Clarke et al., 1999; Lewinsohn et al., 1990) and
comparison treatments (e.g., Brent et al., 1997). Most
studies report outcomes of e1 year. Strategies
employed to improve CBT response, for example,
involving parents (Lewinsohn et al., 1990) or includ-
ing booster sessions (Clarke et al., 1999), lack
empirical support. Of the selective serotonin reuptake
inhibitors, fluoxetine is the only medication consid-
ered to be of established efficacy for adolescent
depression given support from two randomized,

J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 45:10, OCTOBER 2006 1151

Copyright @ 2006 American Academy of Child and Adolescent Psychiatry. Unauthorized reproduction of this article is prohibited.
MELVIN ET AL.
controlled trials (Emslie et al., 1997, 2002). Single
trials support the efficacy of paroxetine (Keller et al.,
2001), citalopram (Wagner et al., 2004), and sertraline
(Wagner et al., 2003). More recently, concerns have
been raised about the emergence of suicidality during
antidepressant treatment of pediatric depression.
Although both pharmacological and psychological
treatments have been demonstrated to be efficacious, a
significant minority of adolescents do not respond to
the current best available treatments, and relapse and
recurrence are common (Emslie et al., 1998; Vostanis
et al., 1998). Hence, a need exists to investigate ap-
proaches that may boost treatment response. Combin-
ing pharmacological and psychosocial treatments has
been demonstrated to be more useful than mono-
therapy with severely depressed adults (Thase et al.,
1997). The first evidence of the efficacy of combined
treatments with depressed adolescents was provided by
the Treatment for Adolescent Depression Study Team
(TADS). They compared fluoxetine, CBT, their com-
bination, and placebo in a sample of 439 adolescents
ages 12 to 17 years (Treatment for Adolescents with
Depression Study [TADS] Team, 2004). The com-
bined treatment was found to be superior to placebo
and CBT treatments on main outcome measures and
superior to fluoxetine on some but not all measures.
The present study, the Time for a FutureVAdolescent
Depression Program, aimed to evaluate sertraline,
CBT, and their combination for adolescents with
depression. Sertraline was chosen at the time because of
some evidence of its therapeutic benefit (Ambrosini
et al., 1999; McConville et al., 1996) and its relatively
wide use in Australia (Hickie et al., 1999). First, it was
hypothesized that referred adolescents with a depressive
disorder would show remission following treatment
with CBT, antidepressant medication alone (MED), or
their combination (COMB). Second, those who re-
ceived COMB were expected to experience significantly
greater rates of remission of mood disorders after
3 months of treatment and at a 6-month follow-up than
those who received either CBT or MED. Third, those
who received COMB were also expected to experience
significantly fewer self-reported depressive symptoms
after 3 months of treatment and at a 6-month follow-up
than those who received CBT or MED. Fourth, it was
hypothesized that all adolescents would show general
improvement in mental health and functioning as
measured by self-reported anxiety, suicidal ideation,
1152
Copyright @ 2006 American Academy of Child and Adolescent Psychiatry. Unauthorized reproduction of this article is prohibited.
self-efficacy, family functioning, and mother-reported
problem behaviors.
METHOD
Subjects
Between July 2000 and December 2002, 168 adolescents were
referred by physicians or school counselors for assessment of
possible depression to one of three clinics (two in suburban
Melbourne and one in a regional city of Victoria) colocated with
public child and adolescent mental health services. Recruitment was
achieved by providing information to physicians and school
counselors and inviting them to refer. Telephone screening excluded
66 participants. One hundred two young people and their parents
were individually interviewed to determine a DSM-IV (American
Psychiatric Association, 1994) depressive diagnosis, and they also
completed questionnaires. Seventy-three adolescents ages 12 to
18 years were included based on a DSM-IV primary diagnosis of
major depressive disorder (MDD), dysthymic disorder (DD), or
depressive disorder not otherwise specified (DDNOS). The
remaining 29 adolescents were excluded because they did not
pursue referral, failed to meet diagnostic criteria, were outside the
age range (12Y18 years), or met other exclusion criteria (major
physical illness or epilepsy, bipolar disorder, organic brain
syndrome, intellectual disability of sufficient severity to preclude
participation in therapy, psychotic disorder, primary diagnosis of
substance abuse disorder, active suicidality or other severe
psychiatric disturbance that required acute hospital admission,
pregnancy or breast-feeding, or current antidepressant or psycho-
tropic medication treatment; Fig.1). The final cohort consisted of
48 females (female-to-male ratio: 2:1). Mean age was 15.3 years
(SD = 1.5) and the modal age was 16 years. Sixty-six of the
adolescents (90%) were born in Australia, three in the United
Kingdom, and one each in Iran, Japan, New Zealand, and Serbia.
More than half of the sample of 73 experienced MDD (60.3%), and
of these, the majority were diagnosed with either mild (54.5%) or
moderate (43.2%) MDD and one case was severe (2.3%) according
to DSM-IV criteria (American Psychiatric Association, 1994). The
remainder of the sample was diagnosed with DD (23.3%) or
DDNOS (16.4%). Comorbidity with depression was common;
69% were diagnosed with at least one comorbid disorder or
condition, and 22% with between two and four comorbid diagnoses.
Ten adolescents, at least 16 years old, did not have parents involved
in the program because they lived independently (n = 4), refused
parental involvement (n = 4), or parents refused involvement
(n = 2; Table 1).
Measures
To determine treatment outcome, multiple outcome measures
were administered at pretreatment, immediately following acute
treatment, and 6 months following acute treatment. The pre-
determined primary outcome measure was depressive diagnosis; all
other measures were secondary. Measures were chosen given their
sound psychometric properties and previous use in similar trials.
Diagnosis. The Schedule for Affective Disorders and Schizo-
phrenia for School Age Children-Lifetime Version was administered
to determine a DSM-IV (American Psychiatric Association, 1994)
depressive diagnosis. The Schedule for Affective Disorders and
Schizophrenia for School Age Children has good psychometric
J. AM. ACAD. CH ILD ADOLESC. PSYCHIATRY, 45:10, OCTOBER 2006

Fig. 1 Program overview. CBT = cognitive-behavioral therapy.
properties including interrater and test-retest reliability and
criterion and predictive validity (Ambrosini, 2000).
Clinician Ratings of Global Functioning. The Global Assessment
of Functioning (GAF) Scale (American Psychiatric Association,
1994) is a psychometrically sound (Hilsenroth et al., 2000), 100-
point hypothetical continuum of mental health illness used by the
clinicians to rate current overall level of functioning.
The Global Assessment of Relational Functioning Scale (American
Psychiatric Association, 1994), a 100-point hypothetical continuum
was used by clinicians to rate current, overall family functioning and
has satisfactory psychometric properties (Hilsenroth et al., 2000).
Self-Report Measures. The Reynolds Adolescent Depression Scale
(RADS), a reliable and valid 30-item measure of the severity of
depressive symptomatology was used (Reynolds, 1986; Reynolds
and Mazza, 1998). Toward the end of the study, the RADS-II was
released. Given that only the psychometric properties were revised
but not the items, the new cutoff score for the RADS-II was used.
J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 45:10, OCTOBER 2006
Copyright @ 2006 American Academy of Child and Adolescent Psychiatry. Unauthorized reproduction of this article is prohibited.
TREATMENT OF ADOLESCENT DEPRESSION
The psychometrically sound Revised Children`s Manifest
Anxiety Scale (RCMAS) self-report measure of the nature and
level of trait anxiety in children and adolescents ages 6 to 19 years
(Reynolds and Richmond, 1985) was used.
The Suicidal Ideation Questionnaire-Junior High School Version
is a 15-item self-report measure of current suicidal ideation (Reynolds,
1988). It possesses good psychometric properties (Winters et al., 2002).
The Self-Efficacy Questionnaire for Depressed Adolescents, a 12-
item self-report instrument, was developed to measure perceived
ability to cope with symptoms of MDD and associated difficulties,
such as dealing with negative thinking. Scores range between 12 and
60 with higher scores being more adaptive. Initial evaluation of the
psychometric properties of the Self-Efficacy Questionnaire for De-
pressed Adolescents indicates that the measure has good reliability
and validity (Tonge et al., 2005).
Parent-Report Measure. The mother or primary caregiver completed
the Child Behavior Checklist, a widely used and psychometrically
1153
MELVIN ET AL.

TABLE 1
Sample Demographics and Diagnostic Information
CBT MED COMB Total
(n = 22) (n = 26) (n = 25) (N = 73)
Mean age 15.0 15.5 15.3 15.3
Male/female 7/15 7/19 11/14 25/48
Occupation
School 20 25 23 68
Employed 0 0 1 1
Unemployed 2 1 1 4
Parental education (mother/father)
Secondary school incomplete 8/9 9/7 13/11 30/27
Secondary school complete 4/2 3/7 4/3 11/12
Tertiary education (nondegree) 3/4 10/4 2/6 15/14
Tertiary education (degree) 7/5 2/3 6/5 15/13
Parental noninvolvement in program 2 6 2 10
Diagnosis
MDD mild/moderate/severe 9/4/1 10/5/0 5/10/0 24/19/1
Dysthymic disorder 2 8 7 17
Depressive disorder NOS 6 3 3 12
Comorbid disordera
Anxiety disorders 8 9 10 27
Dysthymic disorder 1 2 3 6
Conduct disorder/ODD 2 3 1 6
Body dysmorphic disorder 1 0 0 1
Adjustment disorder with anxiety 0 1 0 1
Enuresis 1 0 0 1
Reading disorder 0 1 0 1
Cannabis-related disorder NOS 0 1 0 1
ParentYchild relational problem 5 6 8 19
Sibling relational problem 1 2 3 6

Note: CBT = cognitive-behavioral therapy; MED = antidepressant medication alone; COMB = combined
cognitive-behavioral therapy and antidepressant medication; MDD = major depressive disorder; NOS =
not otherwise specified; ODD = oppositional defiant disorder.
a
Some adolescents experienced more than one comorbid disorder.

sound measure of psychopathology in children and adolescents ages
4 to 18 years (Achenbach, 1991).
Family-Report Measure. The mother and the adolescent completed
the Family Assessment Device General Functioning Scale (Epstein
et al., 1983). This is a 12-item global assessment of family function-
ing that has satisfactory psychometric properties (Byles et al., 1988).
Design
institutional review board, the Southern Health Human Research
Ethics Committee. The institutional review board would not allow
use of a placebo or other control group given that suicidal depressed
teenagers (who did not require hospitalization) were included in the
trial and efficacious treatments were available. Informed written
consent for participation was required from the parents and ado-
lescents. Parental or guardian consent was also required for ado-
lescents younger than 16 years.

The study was a group comparison design in which subjects were

randomly allocated by an independent statistician using a computer-
generated assignment to CBT, MED, or COMB. Pretreatment
assessment was conducted during a 2-week period. Allocation for
those eligible for the trial was concealed to all until after pretreatment
assessment. Acute treatment comprised 12 sessions of CBT, 12 weeks
of medication, or 12 weeks of a combination of the two. Postacute
treatment assessment was then conducted, followed by maintenance
treatment of three monthly Bbooster^ sessions for each treatment.
Participants completed a follow-up assessment 6 months after the end
of acute treatment. The study was approved by the relevant
Treatment Procedures
CBT. The CBT program was designed as a developmentally
sensitive, manual-based treatment comprising individual therapy for the
adolescents and their parents. Its efficacy has been previously established
in relation to a supportive counseling intervention control group (Heyne
et al., 2001). The program consisted of twelve 50-minute weekly
individual sessions for the adolescents and companion sessions for the
parents if they were participating. Sessions were scheduled weekly, and
parent and adolescent sessions typically ran concurrently. Two family

1154 J. AM. ACAD. CH ILD ADOLESC. PSYCHIATRY, 45:10, OCTOBER 2006

Copyright @ 2006 American Academy of Child and Adolescent Psychiatry. Unauthorized reproduction of this article is prohibited.

sessions were scheduled (sessions 6 and 10), which included the parents
and adolescent together. The treatment was composed of modules that
were selected for their potential to treat symptoms of adolescent
depression. Module content was drawn primarily from the Adolescent
Coping with Depression Course (Clarke et al., 1990) and adapted for
use in individual therapy (available at http://www.kpchr.org/public/
acwd/acwd.html). The adolescent therapy modules consisted of goal
setting, psychoeducation, affective education, self-monitoring, relaxa-
tion training, communication skills training, problem-solving skills
training, social skills training, pleasant events scheduling, cognitive
therapy, and life goals planning. The parents received two components,
adolescent-focused parent therapy that consisted of goal-setting,
psychoeducation, behavior management, communication skills train-
ing, problem-solving skills training, and a parent-focused therapy that
consisted of goal setting, relaxation training, and cognitive therapy. Each
module was implemented over one to three sessions, depending on the
adolescent`s individual needs and developmental level. A standardized
procedure was used for the delivery of each module. The procedure
included introduction of the skills with a rationale and explanation,
teaching of the skills with in-session practice, and scheduling and review-
ing of home tasks. This acute treatment was followed by three 50-minute
monthly Bbooster^ sessions for both the adolescent and parents.
Antidepressant Medication
A flexible-dose design was used, with the dose being adjusted
according to clinical response and tolerability. Adolescents commenced
sertraline at 25 mg/day for 1 week. The dose was taken in the morning
unless it was found to have a sedating effect, in which case, a switch to
evenings was made. At the first review appointment after 1 week, the
dose was increased to 50 mg/day, depending on clinical response and
adverse events. Adverse events for those receiving sertraline were
monitored by the use of a questionnaire of commonly experienced
selective serotonin reuptake inhibitor adverse events. If significant
adverse events were evident after 1 week, the dose was reduced to
12.5 mg/day for several days to 1 week, after which time it was
increased to 25 mg/day, if possible. The dose was further increased
in 25-mg increments to a maximum of 100 mg/day, depending on
clinical response and tolerability. A maximum of 100 mg/day was
used based on research suggesting good response to a lower dose of
fluoxetine (20 mg) (Emslie et al., 1997). Once the optimal dose
that balanced therapeutic and adverse events was established, ado-
lescents were reviewed every 2 to 3 weeks. Review sessions did not
include any CBT techniques, but typically included education about
depression and the opportunity for the adolescent to report on his
or her progress. To increase treatment fidelity, the medical prac-
titioners were interviewed mid-treatment and at the completion of
treatment about the content of their consultations. Adolescents
receiving the combined treatment received both CBT and antidepres-
sant medication as per the protocol for each individual therapy.
Treatment Integrity
Treatment integrity was enhanced through the use of a manual-
based CBT treatment program and a medication protocol. The
clinicians providing CBT were seven registered psychologists, a
supervised probationary psychologist, two general medical practi-
tioners, and a social worker with 1 to 5 years` experience in providing
CBT for adolescent depression. Before their first case, clinicians
received intensive training in the assessment and treatment protocols
from chief investigators. Each new clinician`s first case was conducted
together with an experienced therapist. Clinicians received weekly to
J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 45:10, OCTOBER 2006
Copyright @ 2006 American Academy of Child and Adolescent Psychiatry. Unauthorized reproduction of this article is prohibited.
TREATMENT OF ADOLESCENT DEPRESSION
twice-weekly supervision with an expert therapist and peer supervision
was held weekly. The medical practitioners were interviewed regarding
their management of 81% (21/26) of the adolescents allocated to the
MED condition. None reported using CBT strategies with their
antidepressant medication treatment. No significant differences
existed in the number of CBT sessions between participants treated
with CBT (mean sessions 10.91) versus COMB (mean sessions 11.32)
or in the number of medication review appointments for MED (mean
sessions 6.92) versus COMB (mean sessions 8.64). Attendance of
booster sessions was poor. On average, less than one of three sessions
was attended. Total average length of treatment (acute and booster
sessions) was not significantly different for CBT (CBT 22.27 weeks,
COMB 20.08 weeks) or antidepressant medication (MED
18.36 weeks, COMB 18.64 weeks). Twenty participants received
additional treatment (14 antidepressant medication, 6 psychological
treatment) elsewhere after the end of the treatment protocol and
before the follow-up assessment (Fig. 1).
Statistical Analyses
Data were analyzed using an intent-to-treat strategy to counter
any possible overestimation of treatment outcomes, using the last
observation carried forward method (Nelson, 1996). For MDD,
three outcomes were defined: full remission, partial remission
(reduction in symptoms or no symptoms for less than 8 weeks
[American Psychiatric Association, 1994]), and no remission. Two
dichotomies were created: full or partial remission versus no
remission (Bresponse^ dichotomy) and full remission versus partial
or no remission (Bremission^ dichotomy). For DD and DDNOS,
the dichotomy was remission versus no remission because no partial
remission criteria are specified in DSM-IV (American Psychiatric
Association, 1994). Repeated-measures analyses of variance
(ANOVAs) were used to check for the presence of treatment 
time interactions on clinician-rated and self-report and other report
measures. The main analysis was by longitudinal regression, which is
able to take account of within-individual changes in explanatory
variables over time. Treatment  time interaction terms were not
included in the regression models because there was no evidence that
the shape of response or remission time paths differed between
treatment groups. Data were analyzed using Statistical Package for
Social Sciences Version 11 and STATA Version 9.
RESULTS
Pretreatment Comparison of Groups
Pretreatment differences between groups on demo-
graphic and outcome variables were tested using # 2 tests
or Fisher exact test for categorical variables, and one-way
ANOVAs for continuous variables. No difference
between treatment groups was detected on any of the
variables, suggesting that the randomization procedure
was effective.
Treatment Outcome
Diagnostic Status. The proportions of adolescents
with MDD who responded to each treatment over time
are presented in Figure 2A. The greatest proportion
1155
MELVIN ET AL.
Fig. 2 Proportion of cognitive-behavioral therapy (CBT), antidepressant
medication, and combined treatment groups (A) responding to treatment
from major depressive disorder (B) in full remission from major depressive
disorder (C), remitted from dysthymic disorder or depressive disorder not
otherwise specified.
responding to treatment by postacute treatment is
observed in the group receiving CBT (86%) and the
lowest proportion responding to treatment in the group
receiving MED (46%). Rate of response over time is
greater between pre- and posttreatment assessment than
between posttreatment and follow-up assessments. At
posttreatment assessment, the COMB group did not
differ from either CBT or MED in the odds of
1156
Copyright @ 2006 American Academy of Child and Adolescent Psychiatry. Unauthorized reproduction of this article is prohibited.
depression (CBT vs. COMB: odds ratio [OR] = 0.19,
CI 0.03Y1.16; MED vs. COMB: OR = 1.31, CI 0.31Y
5.48). The CBT group had significantly lower odds of
depressive disorder at posttreatment compared with
MED, as modeled by logistic regression (MED vs.
CBT: OR = 6.86, CI 1.12Y41.48). Longitudinal regres-
sion comparing the rate of response over time from
posttreatment to follow-up assessments was performed.
An interaction term was not included in the analyses
because ANOVA had shown no evidence of treatment 
time interaction. The regression models detected no
significant difference in the odds of depression between
treatments (OR = 6.46; CI 0.89Y46.77). In addition,
change in the proportion depressed over time (post-
treatment to follow-up assessment) was not significant
(CBT vs. COMB: OR = 0.02, CI 0Y3.45; MED vs.
COMB: OR = 2.66, CI 0.07Y108.11; MED vs. CBT
OR = 84.94, CI 0.83Y8,718.04).
Treatment groups were compared on the proportion
of participants with MDD who reached full remission
(8 weeks asymptomatic; Fig. 2B). The proportion of
participants in full remission by posttreatment assess-
ment ranged from 7% (COMB) to 14% (CBT). By
follow-up assessment, COMB led to the greatest
proportion of fully remitted adolescents (60%) but
the odds of being fully remitted did not differ
significantly from CBT or MED (CBT vs. COMB:
OR = 2.7, CI 0.60Y12.14) or MED (MED vs. COMB:
OR = 3.0, CI 0.68Y13.31).
The proportion of adolescents remitting from DD or
DDNOS over time is presented in Figure 2C. Change
over time is observed to be similar for each treatment
group. The odds of depression at posttreatment assess-
ment did not differ between treatment groups as
modeled by logistic regression (CBT vs. COMB: OR =
0.71, CI 0.10Y5.12; MED vs. COMB: OR = 1.14,
CI 0.17Y7.60; MED vs. CBT: OR = 1.6, CI 0.23Y11.27).
The odds of depression for those with DD or DDNOS
decreased significantly between postacute treatment and
follow-up assessment (OR = 8.52, CI 2.58Y28.15) in
contrast with the odds for those with MDD.
Continuous Measures. Means and SDs for the main
continuous outcome measures at each assessment are
presented in Table 2. Pretreatment self-report measures
indicate that on average each treatment group was
experiencing clinical levels of depression (mean Rey-
nolds Adolescent Depression Scale score) and anxiety
(mean Revised Children`s Manifest Anxiety Scale
J. AM. ACAD. CH ILD ADOLESC. PSYCHIATRY, 45:10, OCTOBER 2006
 TABLE 2
Secondary Outcome Measures: Adolescent and Mother Report on Adolescent Functioning
Pre Post Follow-up

Cutoff CBT MED COMB CBT MED COMB CBT MED COMB
Score Mean SD Mean SD Mean SD Mean SD Mean SD Mean SD Mean SD Mean SD Mean SD
RADS Q76 83.77 13.80 84.92 11.20 83.96 15.01 66.00 15.93 72.92 16.84 71.64 18.28 60.05 18.10 67.08 20.25 63.32 17.88
RCMAS Q60 62.00 9.63 65.54 7.74 62.21 13.00 53.77 11.39 57.73 11.83 54.96 12.43 50.59 12.83 56.27 12.55 50.92 13.86
SIQ-Jr Q31 26.05 19.93 29.42 27.24 30.64 24.42 19.41 19.64 24.23 26.90 23.20 20.24 13.50 9.09 20.96 26.12 19.28 17.72
SEQ-DA N/A 34.41 8.18 31.85 7.14 33.25 5.94 41.41 8.82 37.27 9.37 39.88 7.65 44.31 9.66 37.42 9.11 42.67 8.00

J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 45:10, OCTOBER 2006

FAD-GF Q2.17 2.48 0.45 2.54 0.38 2.45 0.66 2.22 0.52 2.37 0.57 2.28 0.61 1.99 0.50 2.35 0.51 2.12 0.71
CBCL T scores
TotalBehavior Q60 68.25 11.50 64.78 12.66 61.35 8.25 61.38 13.25 59.30 13.77 54.72 11.88 60.90 12.11 55.35 14.32 53.33 11.11
Externalizing Q60 63.05 13.39 58.83 13.83 54.94 9.02 58.71 12.90 55.70 12.97 53.17 12.27 57.00 12.73 52.26 13.32 51.61 10.05
Internalizing Q60 70.05 10.28 66.74 11.82 65.94 9.94 62.33 11.88 60.26 13.60 56.94 13.09 62.38 12.36 56.87 13.37 56.28 11.68
Note: Pre = pretreatment; Post = posttreatment; CBT = cognitive-behavioral therapy; MED = antidepressant medication alone; COMB = combined treatment; RADS = Reynolds
Adolescent Depression Scale; RCMAS = Revised Children`s Manifest Anxiety Scale; SIQ-Jr = Suicidal Ideation Questionnaire-Junior, SEQ-DA = Self-Efficacy Questionnaire for
Depressed Adolescents; FAD-GF = Family Assessment Device General Functioning; CBCL = Child Behavior Checklist.

Copyright @ 2006 American Academy of Child and Adolescent Psychiatry. Unauthorized reproduction of this article is prohibited.
1157
TREATMENT OF ADOLESCENT DEPRESSION
MELVIN ET AL.

TABLE 3
Secondary Outcome Measures Change in Units/Month for All Participants
Longitudinal Regression (Change/Month)
PreYFU PreYPost PostYFU
Global Assessment of Functioning 1.04* 3.27* 0.15
Global Assessment of Relational Functioning 0.49* 0.79* 0.03
Reynolds Adolescent Depression Scale j2.12* j4.62* j1.12*
Revised Children`s Manifest Anxiety Scale j1.07* j2.57* j0.48*
Self-Efficacy Questionnaire for Depressed Adolescents 0.82* 2.10* 0.31*
Suicidal Ideation Questionnaire-Junior j1.10* j2.13* j0.69*
CBCL Total Behavior Subscale j0.84* j2.09* j0.34
CBCL Externalizing Subscale j0.60 j1.11 j0.38
CBCL Internalizing Subscale j0.88* j2.52* j0.24
Family Assessment Device GF Adolescent* j0.03* j0.05* j0.02
Family Assessment Device GF Mother j0.003 j0.02* j0.06
Note: CBCL = Child Behavior Checklist; GF = General Functioning; Pre = pretreatment; Post = posttreatment; FU = follow-up.
*Significant at p > .05.

score). The Suicidal Ideation Questionnaire-Junior assessment with the exception of mother-rated external-
average score for each treatment group approached izing behavior (Child Behavior Checklist) and family
the clinical cutoff, and mean GAF scores were in the function (Family Assessment Device General Function-
range of Bmoderate symptoms with difficulty in ing). Greater change (per month) was evident between
functioning^ (American Psychiatric Association, pre- and posttreatment assessments compared with
1994). Mean Global Assessment of Relational Func- between posttreatment and follow-up assessments on
tioning Scale scores indicated that relationships were all measures. Four measures did not show significant
somewhat unsatisfactory. change between posttreatment and follow-up assess-
Repeated-measures ANOVAs with interaction term ments, namely, GAF, Global Assessment of Relational
(treatment by time) were performed on all continuous Functioning scale, Child Behavior Checklist internaliz-
outcome measures (self-report, other report, and ing subscale, and mother-rated Family Assessment
clinician rating), and no interaction effects were Device General Functioning.
found. This indicates that change over time in each of
the continuous measures occurred at a similar rate for Adverse Events
each treatment group. Estimates of the rate and Data on serious adverse events were available for all
direction of change over time (pretreatment, posttreat- participants and data on adverse events were available
ment, and follow-up) for the whole sample were then
modeled using longitudinal regression for continuous TABLE 4
outcome measures (Table 3). The significance (p value) Adverse Events for Participants Receiving Sertraline
of the rate of change over time between three time With or Without CBT
frames (pretreatment to follow-up assessment, pre- to Adverse Event % (n = 45)
posttreatment assessment, and posttreatment to follow-up Fatigue 31.1
assessment) is presented along with an estimate of the Concentration 24.4
direction and magnitude of change in outcome measure in Insomnia 22.2
Drowsiness 17.8
units per month. Units per month were used to enable the Restlessness 13.3
comparison of change over the acute treatment phase Suicidal ideation 11.1
(3 months) and the follow-up phase (6 months). In Headache 11.1
addition, an estimate of the direction and magnitude of Yawning 11.1
change from pretreatment to follow-up is presented. All Increased appetite 8.9
Nausea 4.4
measures of adolescent functioning showed significant
improvement between pretreatment and follow-up Note: CBT = cognitive-behavioral therapy.

1158 J. AM. ACAD. CH ILD ADOLESC. PSYCHIATRY, 45:10, OCTOBER 2006

Copyright @ 2006 American Academy of Child and Adolescent Psychiatry. Unauthorized reproduction of this article is prohibited.

for 45 of the 51 (88%) adolescents treated with
medication. One participant experienced a serious
adverse event. The participant had been treated with
COMB for 4 weeks (50 mg sertraline stable for 3 weeks)
and received an inpatient admission for several hours
because of suicidality. Suicidality had previously been
reported at pretreatment assessment. Treatment accord-
ing to the protocol was then continued.
Table 4 lists the adverse events reported by more
than 2% of adolescents treated with medication. The
most common adverse events were fatigue, concentra-
tion problems, and insomnia. Three (6%) of the
medicated adolescents discontinued medication be-
cause of adverse events. None were considered to be
serious or required medical attention. The adverse
events comprised slurred speech and dizziness (MED),
feeling agitated and restless after argument with parent
(COMB), and diarrhea (COMB). Five adolescents
(11%; one COMB, four MED) attended treatment ses-
sions with high levels of suicidality, but all of them were
able to continue treatment within the protocol. These
adolescents also reported suicidality at pretreatment
assessment. No participant required cessation or
reduction in medication because of emergence or exac-
erbation of suicidality or manic conversion.
DISCUSSION
The BTime for a Future^ program was designed to
evaluate CBT, sertraline, and their combination in the
community-based treatment of adolescent depression.
The sample was experiencing mild to moderate levels of
depression and high levels of comorbidity. The
depressed youths were clinical referrals who have
previously been found to be less responsive to treatment
than study participants recruited by advertisement
(Brent et al., 1998). Therefore, it is likely that the
findings of this study can be generalized to the
community. Following acute treatment, all three
treatment groups demonstrated statistically significant
improvements on outcome measures, and these
improvements were maintained 6 months later. Of
those with MDD, up to 71% achieved partial remission
by postacute treatment (CBT = 71.4%; MED = 33.3%;
COMB = 46.7%); however, given the diagnostic
criteria, it is not surprising that few reached the full
remission criterion of 8 weeks asymptomatic at this
J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 45:10, OCTOBER 2006
Copyright @ 2006 American Academy of Child and Adolescent Psychiatry. Unauthorized reproduction of this article is prohibited.
TREATMENT OF ADOLESCENT DEPRESSION
stage. Of those with DD or DDNOS, little variability
was observed between groups with respect to remission
at postacute treatment. Significant diagnostic improve-
ment was evident in those with DD or DDNOS
between postacute treatment and follow-up assessments
in contrast with response to treatment for MDD, which
only maintained gains beyond postacute treatment.
Sertraline was generally well tolerated by the partici-
pants. In no case was sertraline ceased or reduced
because of the emergence or exacerbation of suicidal
ideation, self-injurious behavior, or mania. The two
main reasons for discontinuation of treatment were
improvement in symptoms (n = 4) and minor adverse
events (n = 3).
These findings supported the first hypothesis that all
three treatment groups would show remission following
treatment. However, there was no evidence to support
the second or third hypothesis that COMB was
superior to either treatment alone. The study revealed
that for those with MDD, CBT led to a significantly
superior response at the posttreatment assessment
compared with sertraline alone. This difference may
indicate an advantage to CBT relative to MED;
however, this conclusion cannot be drawn because it
is possible that the lower dose and slower titration
schedule used in the MED treatment (in comparison
with other studies [e.g., Wagner et al., 2003]) may
explain the difference. There are no studies systemat-
ically comparing low and high dosing effects, high-
lighting the need for further research to clarify this
issue. The fourth hypothesis was supported with
secondary outcome measures demonstrating significant
improvement in adolescents` self-reported anxiety, self-
efficacy, and suicidality, suggesting the broad effect of
depression treatment on associated psychopathology.
The findings of the present study are consistent with
those of previous research demonstrating the efficacy of
CBT (Brent et al., 1997; Lewinsohn et al., 1990) and
sertraline (Wagner et al., 2003) in the treatment of
adolescent depression, at least in the short term. At the
same time, the results differ from those of the TADS
study, which demonstrated a superior response to a
combined treatment of CBT and fluoxetine, particu-
larly in comparison with CBT (Treatment for
Adolescents with Depression Study [TADS] Team,
2004). There are a number of possible explanations for
why COMB was not found to be superior to either
MED or CBT in the present study. The relatively small
1159
MELVIN ET AL.
sample size may not have provided sufficient statistical
power to detect a difference, if one existed. With a
longer term follow-up, differences may have been
evident. Weisz and Jensen (1999) have argued that
combined treatments offer the promise of longer-term
benefits, but monotherapy is equally effective in the
short term. The majority of participants in the TADS
study were recruited by advertisement, whereas in this
study adolescents were referred by clinicians, which
may bias treatment response. It is possible that for those
adolescents who received COMB, the effect of selective
serotonin reuptake inhibitors on cognition delayed the
cognitive response to CBT, analogous to the suggestion
that benzodiazepines impair cognition and response to
CBT for panic disorder (Foa et al., 2002).
Limitations
First, the study did not include a placebo-control
condition because of ethical considerations given the
seriousness of adolescent depression and suicide risk.
Second, even though the sample experienced high levels
of functional impairment (mean GAF = 51Y60) and
elevated levels of suicidal ideation at pretreatment, few
participants with severe depression were recruited.
Thus, the findings may not be applicable to adolescents
with severe depression who probably require inpatient
treatment. Combined treatment may be more effica-
cious for severe adolescent depression, as it is for adults
with severe depression (Thase et al., 1997). Neverthe-
less, the findings of this study are relevant to the
treatment of depressed youth in community settings.
Third, compared with other studies (e.g., McConville
et al., 1996; Wagner et al., 2003), this study used lower
doses and a slower titration schedule. This may have
reduced treatment response in MED and COMB
groups, but systematic evaluation of optimal dose and
titration schedule is not available. Fourth, independent
review of treatment integrity may have been enhanced
with videotaping. Fifth, independent raters blind to
treatment allocation were not used because of resource
limitations but may have reduced the risk of experi-
menter bias in assessments. However, independently
rated self-report and parent report measures produced
broadly consistent findings with diagnostic interviews.
Sixth, although no CBT participant experienced a
serious adverse event, monitoring of this group for
adverse events (including suicidal ideation) may have
1160
Copyright @ 2006 American Academy of Child and Adolescent Psychiatry. Unauthorized reproduction of this article is prohibited.
provided useful information to inform the current
debate about emergent suicidality during psychotherapy.
Clinical Implications
Taking the study limitations into account, CBT and
sertraline are equally recommended for the treatment
for adolescents with depression, each demonstrating an
equivalent response. The findings suggest that there
may be different trajectories of treatment response for
those experiencing DD or DDNOS because this group
continued to improve beyond acute treatment com-
pared with those with MDD whose treatment response
plateaued following the end of acute treatment. There
remains a large minority of participants who do not
respond to these most empirically validated treatments
and a sizable group who completed the treatment
protocol and then sought further treatment elsewhere.
This highlights the need for further investigation into
methods of enhancing treatment response and predict-
ing treatment response and relapse.
Disclosure: The authors have no financial relationships to disclose.
REFERENCES
Achenbach TM (1991), Manual for the Child Behavior Checklist/4Y18 and
1991 Profile Burlington. Burlington: University of Vermont, Depart-
ment of Psychiatry
Ambrosini PJ (2000), Historical development and present status of the
Schedule for Affective Disorders and Schizophrenia for School-Age
Children (K-SADS). J Am Acad Child Adolesc Psychiatry 39:49Y58
Ambrosini PJ, Wagner KD, Biederman J et al. (1999), Multicenter open-
label sertraline study in adolescent outpatients with major depression.
J Am Acad Child Adolesc Psychiatry 38:566Y572
American Psychiatric Association (1994), Diagnostic and Statistical Manual
of Mental Disorders, 4th Edition (DSM-IV). Washington, DC: American
Psychiatric Association
Brent DA, Holder D, Kolko D et al. (1997), A clinical psychotherapy trial
for adolescent depression comparing cognitive, family, and supportive
therapy. Arch Gen Psychiatry 54:877Y885
Brent DA, Kolko DJ, Birmaher B et al. (1998), Predictors of treatment
efficacy in a clinical trial of three psychosocial treatments for adolescent
depression. J Am Acad Child Adolesc Psychiatry 37:906Y914
Byles J, Byrne C, Boyle MH, Offord DR (1988), Ontario child health study:
reliability and validity of the general functioning subscale of the
McMaster family assessment device. Fam Process 27:97Y104
Clarke GN, Lewinsohn PM, Hops H (1990), Adolescent Coping With Depression
Course: Leader`s Manual for Adolescent Groups. Eugene, OR: Castalia
Clarke GN, Rohde P, Lewinsohn PM, Hops H, Seeley JR (1999),
Cognitive-behavioral treatment of adolescent depression: efficacy of
acute group treatment and booster sessions. J Am Acad Child Adolesc
Psychiatry 38:272Y279
Emslie GJ, Rush AJ, Weinberg WA et al. (1997), A double-blind,
randomized, placebo controlled trial of fluoxetine in children and
adolescents with depression. Arch Gen Psychiatry 54:1031Y1037
Emslie GJ, Rush AJ, Weinberg WA, Kowatch RA, Carmody T, Mayes TL
(1998), Fluoxetine in child and adolescent depression: acute and
maintenance treatment. Depress Anxiety 7:32Y39
J. AM. ACAD. CH ILD ADOLESC. PSYCHIATRY, 45:10, OCTOBER 2006
 TREATMENT OF ADOLESCENT DEPRESSION

Emslie GJ, Heiligenstein JH, Wagner KD et al. (2002), Fluoxetine for acute
treatment of depression in children and adolescents: a placebo-controlled
randomized clinical trial. J Am Acad Child Adolesc Psychiatry 41:1205Y1214
Epstein NB, Baldwin LM, Bishop DS (1983), The McMaster Family
Assessment Device. J Marital Fam Ther 9:171Y180
Foa EB, Franklin ME, Moser J (2002), Context in the clinic: how well do
cognitive-behavioral therapies and medications work in combination?
Biol Psychiatry 52:987Y997
Heyne D, King N, Tonge B, Burdett L (2001), Evaluation of cognitive-
behavioral therapy for adolescent depression: the Berriga House Project.
Paper presented at the 3rd World Congress of Behavioral and Cognitive
Therapies, Vancouver, Canada, July 17Y21
Hickie IB, Scott EM, Davenport TA (1999), Are antidepressants all the
same? Surveying the opinions of Australian psychiatrists. Aust N Z J
Psychiatry 33:642Y649
Hilsenroth MJ, Ackerman SJ, Blagys MD et al. (2000), Reliability and
validity of DSM-IV axis V. Am J Psychiatry 157:1858Y1863
Kaiser Permanente Centre for Health Research (2006), Youth Depression
Treatment and Prevention Programs. Available at: http://www.kpchr.org/
public/acwd/acwd.html. Accessed May 8, 2006
Keller MB, Ryan ND, Strober M et al. (2001), Efficacy of paroxetine in the
treatment of adolescent major depression: a randomized, controlled trial.
J Am Acad Child Adolesc Psychiatry 40:762Y772
Lewinsohn PM, Clarke GN, Hops H, Andrews JA (1990), Cognitive-
behavioral treatment for depressed adolescents. Behav Ther 21:385Y401
Lewinsohn PM, Hops H, Roberts RE, Seeley JR, Andrews JA (1993),
Adolescent psychopathology: I. Prevalence and incidence of depression
and other DSM-III-R disorders in high school students. J Abnorm
Psychol 102:133Y144
Lewinsohn PM, Rohde P, Seeley JR (1994), Psychosocial risk factors for
future adolescent suicide attempts. J Consult Clin Psychol 62:297Y305
Lewinsohn PM, Rohde P, Seeley JR (1998), Major depressive disorder in
older adolescents: prevalence, risk factors, and clinical implications. Clin
Psychol Rev 18:765Y794
McConville BJ, Minnery KL, Sorter MT, West SA, Friedman LM, Christian K
(1996), An open study of the effects of sertraline on adolescent major
depression. J Child Adolesc Psychopharmacol 6:41Y51
Nelson C (1996), Importance of intent-to-treat response rates for the
clinician. Psychiatr Ann 26:8Y18
Reynolds WM (1986), Reynolds Adolescent Depression Scale-Professional
Manual. Odessa, FL: Psychological Assessment Resources
Reynolds WM (1988), Suicidal Ideation Questionnaire: Professional Manual.
Odessa, FL: Psychological Assessment Resources
Reynolds WM, Mazza JJ (1998), Reliability and validity of the Reynolds
Adolescent Depression Scale with young adolescents. J School Psychol
36:295Y312
Reynolds CR, Richmond BO (1985), Revised Children`s Manifest Anxiety
Scale (RCMAS) Manual. Los Angeles: Western Psychological Services
Shaffer D, Gould MS, Fisher P et al. (1996), Psychiatric diagnosis in child
and adolescent suicide. Arch Gen Psychiatry 53:339Y348
Thase ME, Greenhouse JB, Frank E et al. (1997), Treatment of major
depression with psychotherapy or psychotherapy-pharmacotherapy
combinations. Arch Gen Psychiatry 54:1009Y1015
Treatment for Adolescents with Depression Study (TADS) Team (2004),
Fluoxetine, cognitive behavioral therapy and their combination for
adolescents with depression. JAMA 292:807Y820
Tonge B, King N, Klimkeit E, Melvin G, Heyne D, Gordon M (2005), The
self-efficacy questionnaire for depression in adolescents (SEQ-DA):
development and psychometric evaluation. Eur Child Adolesc Psychiatry
14:357Y363
Vostanis P, Feehan C, Grattan E (1998), Two-year outcome of children
treated for depression. Eur Child Adolesc Psychiatry 7:12Y18
Wagner KD, Ambrosini P, Rynn M et al. (2003), Efficacy of sertraline in the
treatment of children and adolescents with major depressive disorder:
two randomised controlled trials. JAMA 290:1033Y1047
Wagner KD, Robb AS, Findling RL, Jin J, Gutierrez MM, Heydorn WE
(2004), A randomized, placebo controlled trial of citalopram for the
treatment of major depression in children and adolescents. Am J
Psychiatry 161:1079Y1083
Weisz JR, Jensen PS (1999), Efficacy and effectiveness of child and adolescent
psychotherapy and pharmacotherapy. Ment Health Serv Res 1:125Y157
Winters NC, Myers K, Proud L (2002), Ten-year review of rating scales. III:
scales assessing suicidality, cognitive style, and self-esteem. J Am Acad
Child Adolesc Psychiatry 41:1150Y1162

Do Bullied Children Get Ill, or Do Ill Children Get Bullied? A Prospective Cohort Study on the Relationship Between
Bullying and Health-Related Symptoms Minne Fekkes, MSc, PhD, Frans I.M. Pijpers, MD, PhD, A. Miranda Fredriks, MD,
PhD, Ton Vogels, MSc, S. Pauline Verloove-Vanhorick, MD, PhD

Objectives: A number of studies have shown that victimization from bullying behavior is associated with substantial adverse effects
on physical and psychological health, but it is unclear which comes first, the victimization or the health-related symptoms. In our
present study, we investigated whether victimization precedes psychosomatic and psychosocial symptoms or whether these
symptoms precede victimization. Design: Six-month cohort study with baseline measurements taken in the fall of 1999 and
follow-up measurements in the spring of 2000. Setting: Eighteen elementary schools in the Netherlands. Participants: The study
included 1118 children aged 9 to 11 years, who participated by filling out a questionnaire on both occasions of data collection.
Outcome Measures: A self-administered questionnaire measured victimization from bullying, as well as a wide variety of
psychosocial and psychosomatic symptoms, including depression, anxiety, bedwetting, headaches, sleeping problems, abdominal
pain, poor appetite, and feelings of tension or tiredness. Results: Victims of bullying had significantly higher chances of developing
new psychosomatic and psychosocial problems compared with children who were not bullied. In contrast, some psychosocial, but
not physical, health symptoms preceded bullying victimization. Children with depressive symptoms had a significantly higher
chance of being newly victimized, as did children with anxiety. Conclusions: Many psychosomatic and psychosocial health
problems follow an episode of bullying victimization. These findings stress the importance for doctors and health practitioners to
establish whether bullying plays a contributing role in the etiology of such symptoms. Furthermore, our results indicate that
children with depressive symptoms and anxiety are at increased risk of being victimized. Because victimization could have an
adverse effect on children_s attempts to cope with depression or anxiety, it is important to consider teaching these children skills
that could make them less vulnerable to bullying behavior. Pediatrics 2006;117:1568Y1574.

J. AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 45:10, OCTOBER 2006 1161

Copyright @ 2006 American Academy of Child and Adolescent Psychiatry. Unauthorized reproduction of this article is prohibited.