Paula Barrett : long-term effectiveness of the FRIENDS Program
PLEASE SCROLL DOWN FOR ARTICLE
This article was downloaded by: [University of Queensland]
On: 23 February 2010
Access details: Access Details: [subscription number 907688116]
Publisher Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-
41 Mortimer Street, London W1T 3JH, UK
Journal of Clinical Child & Adolescent Psychology
Publication details, including instructions for authors and subscription information:
http://www.informaworld.com/smpp/title~content=t775648094
Long-Term Outcomes of an Australian Universal Prevention Trial of
Anxiety and Depression Symptoms in Children and Youth: An Evaluation
of the Friends Program
Paula M. Barrett a; Lara J. Farrell a; Thomas H. Ollendick b; Mark Dadds c
a University of Queensland, Pathways Health and Research Centre, Brisbane, Australia. b Virginia
Polytechnic Institute and State University, Blacksburg. c University of New South Wales, Sydney,
Australia.
To cite this Article Barrett, Paula M., Farrell, Lara J., Ollendick, Thomas H. and Dadds, Mark(2006) ‘Long-Term Outcomes
of an Australian Universal Prevention Trial of Anxiety and Depression Symptoms in Children and Youth: An Evaluation
of the Friends Program’, Journal of Clinical Child & Adolescent Psychology, 35: 3, 403 — 411
To link to this Article: DOI: 10.1207/s15374424jccp3503_5
URL: http://dx.doi.org/10.1207/s15374424jccp3503_5
Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf
This article may be used for research, teaching and private study purposes. Any substantial or
systematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply or
distribution in any form to anyone is expressly forbidden.
The publisher does not give any warranty express or implied or make any representation that the contents
will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses
should be independently verified with primary sources. The publisher shall not be liable for any loss,
actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly
or indirectly in connection with or arising out of the use of this material.
Long-Term Outcomes of an Australian Universal Prevention Trial
of Anxiety and Depression Symptoms in Children and Youth:
An Evaluation of the Friends Program
Paula M. Barrett
University of Queensland, Pathways Health and Research Centre, Brisbane, Australia
Lara J. Farrell
University of Queensland, Pathways Health and Research Centre, Brisbane, Australia
Thomas H. Ollendick
Virginia Polytechnic Institute and State University, Blacksburg
Mark Dadds
University of New South Wales, Sydney, Australia
This study evaluated the long-term effectiveness of the FRIENDS Program in reducing
anxiety and depression in a sample of children from Grade 6 and Grade 9 in comparison
to a control condition. Longitudinal data for Lock and Barrett’s (2003) universal
prevention trial is presented, along with data from 12-month follow-up to 24-
and 36-month follow-up. Results of this study indicate that intervention reductions in
anxiety reported in Lock and Barrett were maintained for students in Grade 6, with
the intervention group reporting significantly lower ratings of anxiety at long-term
follow-up. A significant Time Ă— Intervention Group Ă— Gender Effect on Anxiety was
found, with girls in the intervention group reporting significantly lower anxiety at
12-month and 24-month follow-up but not at 36-month follow-up in comparison to
the control condition. Results demonstrated a prevention effect with significantly
fewer high-risk students at 36-month follow-up in the intervention condition than in
the control condition. Results are discussed within the context of prevention research.
Emotional disturbances in children and youth occur
at alarmingly high rates, are associated with a number
of negative life consequences, and come at a tremendous
cost to society. A Year Book of Australia report
indicated that 20% of children between the ages of
12 and 16 had a significant mental health problem
(Stanley, 2002). Anxiety disorders are the most frequently
experienced mental health disorder in childhood
and adolescence, with studies estimating a point
prevalence of 5% to 10% and a lifetime prevalence of
approximately 20% (Essau, Condradt, & Petermann,
2000; Shaffer, Fischer, Dulcan, & Davies, 1996). Research
has demonstrated that anxiety and depressive
symptoms are highly related in child and youth populations
(e.g., Dobson, 1985; Tannenbaum, Forehand, &
Thomas, 1992). Estimates indicate approximately 2%
to 5% of children and adolescents will suffer a major
depressive disorder of clinical severity (Kashani et al.,
1987; Lewinsohn, Clarke, & Rohde, 1994). Beyond
the high prevalence rates, these emotional disorders are
associated with a wide range of psychosocial impairments,
tend to be chronic and unremitting in course,
and are associated with significant risk for other psychological
disorders if left untreated (e.g., Cole,
Peeke, Martin, Truglio, & Seroczynski, 1998; Kashani
& Orvaschel, 1990; Last, Hansen, & Franco, 1997;
Orvaschel, Lewinsohn, & Seeley, 1995).
There has been a recent surge in the field of prevention
research for children and youth. Given the potential
of such approaches to impact on the incidence and
prevalence of childhood emotional disorders, the need
for evidence-based prevention is strong. Primary preventive
interventions can be defined as either universal,
selected, or indicated and targeted (Mrazek &
Haggerty, 1994). Universal interventions target whole
population groups, selective interventions involve
children and youth identified as at risk of psychological
problems, and indicated interventions target individuals
identified with mild to moderate symptoms of
a disorder (Mrazek & Haggerty, 1994). Universal prevention
interventions conducted in the school context
have many advantages, including reducing recruitment,
screening, and attrition difficulties; reaching a
broad range of children and adolescents with varying
Journal of Clinical Child and Adolescent Psychology
2006, Vol. 35, No. 3, 403–411
Copyright © 2006 by
Lawrence Erlbaum Associates, Inc.
403
Correspondence should be addressed to Lara Farrell, School
of Applied Psychology, Mt. Gravatt Campus, Griffith University,
Queensland, Australia 4111. E-mail: l.healy@griffith.edu.au
Downloaded By: [University of Queensland] At: 03:16 23 February 2010
levels of risk for psychopathology; reducing stigmatization;
enhancing peer support; and reducing psychosocial
difficulties within the classroom—thus promoting
learning and healthy development (Evans, 1999;
Kubiszyn, 1999). Support for preventative outcomes
based on the delivery of cognitive behavioral interventions
in schools is growing.
The Queensland Early Intervention and Prevention
of Anxiety Project (Dadds, Spence, Holland, Barrett &
Laurens, 1997) represented the first cognitive behavioral
trial for prevention of childhood anxiety. This
study targeted children (ages 7 to 14 years) who were
disorder-free but exhibited anxious symptomatology
(indicated prevention; Mrazek & Haggerty, 1994) as
well as children who met criteria for an anxiety disorder
but were in the less severe range (early intervention;
Mrazek & Haggerty, 1994). Screening identified
128 eligible participants who were randomly allocated
to either an intervention or a control condition. Diagnostic
status was used as an outcome measure, and results
were favorable. Both groups demonstrated improvement
immediately postintervention; however, by
6-month follow-up, the improvement was maintained
in the intervention group only. No differences between
groups were evident at 12-month follow-up; however,
at the 2-year follow-up, intervention effectiveness was
demonstrated through the reduction of existing rates of
anxiety disorder and prevention of the onset of new
anxiety disorders (Dadds, Spence, Laurens, Mullins,&
Barrett, 1999). Consistent with prior research (Last,
Perrin, Hersen, & Kazdin, 1996), regardless of intervention
status, participants in this study showed a general
improvement across time; results further revealed
gender (female), parental anxiety and pretreatment severity
predicted poor response to intervention (Barrett,
Dadds, & Rapee, 1996; Cobham, Dadds, & Spence,
1998). This study demonstrated that anxiety disorders
and the number of children at risk of anxiety can be
successfully reduced through selected school-based
cognitive–behavioral intervention. An interesting outcome
was putative delay in intervention effects, which
is consistent with the results of a similar prevention
trial for adolescent depression (Jaycox, Reivich, Gillham,
& Seligman, 1994).
Barrett and Turner (2001) evaluated the effects of a
universal cognitive behavioral intervention for the prevention
of internalizing symptoms in children in Grade
6 (ages 9 to 10 years). Ten schools in the Brisbane region
participated in the project, which involved all
children participating in the FRIENDS program during
class time (Barrett, 2004, 2005; Barrett, Lowry-Webster,
& Turner, 1999a, 1999b). This study evaluated
a “train-the-trainer” model of intervention, whereby
children were assigned to one of three conditions: (a)
psychologist-led intervention, (b) teacher-led intervention
(following a standardized teacher training workshop),
or (c) standard curriculum (control condition).
Barrett and Turner (2001) trained classroom teachers
and psychologists to implement the 12-session
FRIENDS program as part of the standard classroom
curriculum. Parents were invited to attend four parent
evenings, which involved psychoeducation and parenting
strategies. Evaluation of children’s self-report
measures at postintervention indicated preventive effects,
with participants reporting significant reductions
in anxiety symptoms across psychologist and teacher
intervention conditions. This study provided preliminary
evidence for the effectiveness of the FRIENDS
program delivered by teachers at a school-based population
level, integrated within the standard school
curriculum.
Lowry-Webster, Barrett, and Dadds (2001) further
examined the effectiveness of the FRIENDS program
as a universal strategy for prevention of childhood anxiety.
In total, 594 students, ages 10 to 13 years, were allocated
to either an intervention or a control condition
on the basis of class. At postassessment, all children reported
significant reductions in anxiety, although these
decreases were significantly greater in the intervention
group compared to the control condition. A significant
reduction in depression was found for the intervention
group only. Further analysis of changes in risk status
showed positive findings. Of the children in the intervention
group at risk at preintervention, 75.3% were no
longer at risk at postintervention, compared to 54.8%
of high-risk children in the control group. Lowry-Webster,
Barrett, and Lock (2003) reported on outcomes at
12-month follow-up for this sample. Results indicated
that prevention effects were maintained up to the 12-
month follow-up for children who received the program.
The intervention group evidenced lower scores
on anxiety self-report measures, and the high-anxiety
children from the intervention condition reported reductions
in both anxiety and depression scores. Diagnostic
interview data demonstrated that 85% of children
in the intervention group who were scoring above
the clinical cutoff for anxiety and depression were
diagnosis-free at 12-month follow-up, compared to
only 31.2% of children in the control group. This follow-
up study demonstrated clinically and statistically
significant reductions in anxiety symptoms and disorders
from pretest to 12-month follow-up following the
FRIENDS universal program.
Most recently, Lock and Barrett (2003) presented
the results of a longitudinal school-based study of universal
prevention using the FRIENDS program across
two distinct age groups. This study involved a group of
733 children enrolled in Grade 6 (n = 336; ages 9 and
10 years) and Grade 9 (n = 401; ages 14 to 16 years)
from seven socioeconomically diverse schools in the
metropolitan area of Brisbane, Australia. Schools were
randomly assigned to either an intervention condition
404
BARRETT, FARRELL, OLLENDICK, DADDS
Downloaded By: [University of Queensland] At: 03:16 23 February 2010
(FRIENDS) or a control condition (standard curriculum),
and all students completed self-report measures
of anxiety, depression, and coping. Students identified
as high risk based on elevated scores on an anxiety
measure were interviewed using a structured diagnostic
interview. As with previous research (e.g., Dadds
et al., 1997, 1999; Lowry-Webster et al., 2001, 2003),
this study found general reductions in anxiety across
time regardless of intervention condition; however,
these reductions were significantly greater for students
in the intervention condition at both posttest and
12-month follow-up. In terms of age differences, children
in Grade 6 reported significantly higher levels of
anxiety prior to the intervention and at postintervention
yet greater reductions in anxiety at 12-month followup,
as well as lower levels of depression across time
compared to Grade 9 children. This finding suggested
that the optimal time for preventing anxiety may be in
late childhood (9 to 10 years of age) versus early adolescence.
This study also examined gender differences
and found that girls were more likely to be at risk of an
anxiety disorder and tend to report higher levels of anxiety
than boys over time. Moreover, Grade 6 girls
were most responsive to the intervention, as they reported
greater reductions in anxiety compared to girls
in Grade 9 and boys across grades.
Lock and Barrett (2003) also examined the effects
of the intervention on depressive symptoms. Results
indicated that there significant reductions in depression;
however, this effect was only apparent at 12-
month follow-up, suggesting a delayed “prevention”
effect for depression. This finding of a delayed effect is
consistent with the finding from the Queensland Early
Intervention project (Dadds et al., 1997) and is also
consistent with Jaycox and colleagues (1994) prevention
trial for depression.
These preliminary studies have indicated significant
promise for the effectiveness of the FRIENDS
program as a selective, indicated, and universal prevention
program for schools. However, as the true preventive
impact of an intervention can better be determined
over a longer term, it is important to examine outcomes
beyond the 12-month follow-up. This study aims to
evaluate the long-term prevention outcomes at 24-
month and 36-month follow-up, compared to existing
data at 12-month follow-up reported in Lock and
Barrett’s (2003) longitudinal study. Students involved
in Lock and Barrett’s study were followed up again at
24 months and 36 months. This study evaluates outcomes
on measures of self-reported anxiety and depression
across intervention and control schools from
12-month follow-up to 24-month follow-up and 36-
month follow-up. It was hypothesized that intervention
gains from Lock and Barrett’s study would be maintained;
that is, there would be significantly lower
scores on anxiety and depression self-report across the
follow-up for children and youth in the intervention
condition in comparison to students in the control condition.
It was further hypothesized that there would be
significantly fewer students at high risk for anxiety
and depression in the intervention condition compared
with the control condition at each follow-up point.
Based on Lock and Barrett’s findings, it was anticipated
that prevention effects (maintenance of gains vs.
increases in symptoms over time) would be strongest
for children in the Grade 6 group compared with students
in the Grade 9 group and for girls compared to
boys.
Method
Participants
Participants included in this follow-up study were
669 children and youth from Lock and Barrett’s (2003)
original sample of 737 Grade 6 (age 10 to 11 years) and
Grade 9 (age 13 to 14 years) students. One school from
the original sample (control school, n = 68) withdrew
from the longitudinal study following the 12-month
follow-up. The sample at 12-month follow-up consisted
of 334 students in Grade 7 and 335 students in
Grade 10. These students were followed up again at 24
months (Grade 8 and Grade 11) and 36 months (Grade
9 and Grade 12). All participants were students from
one of six coeducational schools extending from preschool
to Year 12 in the metropolitan area of Brisbane,
Australia. Schools, rather than participants, were selected
as the unit of random assignment; with schools
randomly assigned to either an intervention condition
or a control condition. The schools remaining in this
study were three intervention schools and three control
schools. At 12-month follow-up, there were a total of
379 participating students (54% female) within the intervention
condition and a total of 290 participating
students within the control condition (48% female).
Socioeconomic status (SES) was based on paternal
occupation and was coded using the 9-point Australian
Standard Classification of Occupations Dictionary
(Australian Bureau of Statistics and Department of Education,
Training and Youth Affairs, 1997). The average
SES rating across schools was 5.67 (SD = 1.21),
typical of the SES distribution of Australia in general.
This value is indicative of middle SES on average (e.g.,
skilled, clerical, and trade occupations are coded as 5
and semiprofessional occupations as 6) and is broadly
consistent with the average SES reported in other Australian
studies (e.g., Spence, Barrett, & Turner, 2003).
The intervention and control schools did not differ
from each other on SES ratings. The majority of intervention
(89%) and control students (84%) were born in
Australia, with the remainder coming from a wide vari-
405
UNIVERSAL SCHOOL-BASED ANXIETY PREVENTION PROGRAM
Downloaded By: [University of Queensland] At: 03:16 23 February 2010
ety of ethnic backgrounds, as is typical of the Australian
population.
Procedure
Informed consent and assignment to experimental
conditions. All schools participating in the project
were from the independent education sector, and
initial consent was obtained from the principal of each
school to invite students, their parents, and their teachers
to participate in a longitudinal research project.
Schools were matched in pairs based on geographical
location, and one school from each pair was randomly
assigned to either an intervention or a control condition.
All parents of students were sent an information
sheet describing the project and an informed consent
form to be completed and returned by parents. A good
consent rate was obtained for each grade level: Grade
6, 79.36%, and Grade 9, 77.62%. See Lock and Barrett
(2003) for further details.
Program evaluation—Long-term follow-up. In
this study, students were assessed at two time intervals:
24-month follow-up and 36-month follow-up. This data
will be compared with Lock and Barrett’s (2003) 12-
month follow-up data to evaluate change over time. All
questionnaire assessments were completed within class
groups, within normal school hours.Apostgraduate research
assistant read the instructions and questionnaires
aloud to all students while other project staff walked
around the classroom assisting students who required
help or indicated they did not understand one of the
questions. Students were informed that all questionnaire
responses were confidential. Questionnaires were
presented in a counterbalanced order within the assessment
package, with each school receiving a different ordering
of questionnaires, across each data collection
point.
Program implementation. The FRIENDS intervention
(discussed later) was implemented within
schools assigned to the intervention condition as part
of the curriculum within the subject areas of health and
physical education or social and personal development.
The intervention consists of 10 sessions of approximately
70 min each, with one session scheduled
per week over a 10-week term. There are two booster
sessions in the program, which were implemented in
the following term. Two developmentally tailored versions
of the program were implemented: FRIENDS for
Children (Barrett et al., 1999a) was offered to Grade
6 intervention participants and FRIENDS for Youth
(Barrett et al., 1999b) was implemented with Grade 9
intervention participants. The FRIENDS program is a
brief cognitive–behavioral intervention designed and
validated as an individual or group-based treatment for
clinically anxious children (Barrett et al., 1996; Shortt,
Barrett, & Fox, 2001). The program, described in detail
by Barrett (1999), assists children and youth in
learning important skills and techniques that help them
to cope with and manage anxiety and emotional distress
through the application of learned coping and
problem-solving skills. The FRIENDS program was
implemented by teachers with the assistance of clinically
trained psychology postgraduate students following
a 1-day standardized teacher-training workshop.
Teachers were the main facilitator for the sessions,
with the psychology students assisting children when
needed. An earlier study reported that there were no
differences in postintervention outcomes for students
who received the intervention led by teachers and students
who received the intervention led by psychologists
(see Barrett & Turner, 2001). The program also
incorporates four evening sessions for parents, which
are scheduled at regular intervals throughout the 10
weeks of the program. These psychoeducational sessions
provided parents with an opportunity to learn
about the program and to discuss parenting and reinforcement
strategies. FRIENDS is an acronym for the
different skills taught (F = feeling worried; R = relax
and feel good; I = inner helpful thoughts; E = explore
plans; N = nice work, reward yourself; D = don’t forget
to practice; and S = stay calm for life!).
Prior to implementing the program, group leaders
were each given a program fidelity checklist to record
whether they completed each activity within the session.
Of the 18 teachers who implemented the intervention,
only 5 returned fidelity checklists. Adherence
to the intervention content ranged from 72.3% to
91.66%.
Measures
Spence Children’s Anxiety Scale (SCAS). The
SCAS (Spence, 1997) is a 44-item scale assessing anxiety
symptoms. Six subscales, corresponding to Diagnostic
and Statistical Manual of Mental Disorders (4th
ed.; American Psychiatric Association, 1994) anxiety
disorders are calculated from the anxiety items and a
Total Anxiety score (used in this study) is obtained by
summing all subscales. Participants rate each symptom
on a 4-point scale corresponding to the frequency with
which they experience each symptom, and higher
scores reflect a greater number of anxiety symptoms.
Spence (1998) and Spence et al. (2003) reported high
internal consistency in community child (Cronbach’s
α = .92) and adolescent populations (Cronbach’s α =
.92), respectively, and 6-month test–retest reliability of
.60 for children (Spence, 1998) and 12-week test–retest
reliability of .63 for youth (Spence et al., 2003).
Good convergent and discriminant validity was also reported.
Analysis of internal consistency was conducted
for this sample at 12-month follow-up, with results in-
406
BARRETT, FARRELL, OLLENDICK, DADDS
Downloaded By: [University of Queensland] At: 03:16 23 February 2010
dicating high internal reliability (Cronbach’s α = .89)
for this measure.
Revised Children’s Manifest Anxiety Scale
(RCMAS). The RCMAS (Reynolds & Richmond,
1985) provides a measure of anxiety symptomatology.
The questionnaire contains 37 items, 9 of which form a
Lie scale. For each item, the child is asked to respond
“yes” or “no.” This measure has been found to have
high internal consistency and test–retest reliability, as
well as showing convergent and divergent validity
(Reynolds & Richmond, 1985). The internal consistency
for this measure was high, based on responses
from this sample at 12-month follow-up, with a
Cronbach’s α of .85.
Children’s Depression Inventory (CDI). Depressive
symptoms were assessed using the 27-item
CDI (Kovacs, 1981). Items assess depressive symptoms
such as sadness, self-blame, loss of appetite, interpersonal
relationships, and school adjustment. For
each item, participants choose a statement from three
response alternatives, with each increasing in symptom
severity. Higher scores reflect more severe symptomatology.
In this study, and in line with other research
studies (e.g., Hannon, Rapee, & Hudson, 2000;
Shochet et al., 2001; Weiss et al., 1991), one item pertaining
to suicidal ideation was omitted due to concerns
expressed by school personnel and parent
groups. In comparing children’s scores in samples with
and without the suicide item, Weiss et al. (1991) reported
that deletion of the suicide item did not significantly
alter CDI scores. Cole, Hoffman, Tram, and
Maxwell (2000) reported high internal consistency in a
community sample of children and youth (Cronbach’s
α = .90) and 6-month test–retest reliability of .66. Internal
consistency for this measure based on this sample
at 12-month follow-up indicated high internal reliability
(Cronbach’s α = .89).
Results
Attrition and Missing Data
Patterns of missing data from 12-month follow-up
to 24-month follow-up and 36-month follow-up were
examined to determine dropout and absenteeism rates
to assess potential influences of these factors on the
long-term outcomes. At 24-month follow-up, there
were no differences in the frequency of missing data
across between grade (34% missing Grade 6, 33%
missing Grade 9), gender (34% missing male, 33%
missing female), or students at high risk (32% missing)
and in the healthy range (34% missing). There were
significant differences between the intervention and
the control conditions, with the control condition having
significantly more missing data (44%) than the intervention
group (25%), χ2(1, 669) = 26.09, p < .001.
At 36-month follow-up there were no differences in
the frequency of missing data between grade (47%
missing Grade 6, 46% missing Grade 9), gender (49%
missing male, 43% missing female), or students at high
risk (44% missing) and in the healthy range (47%
missing). There were significant differences between
the intervention and the control conditions, with the
control condition having significantly more missing
data (54%) than the intervention group (41%), χ2(1,
669) = 10.98, p < .01. Reasons for attrition at each time
point were absenteeism from school on the day of
assessment, students leaving the school, and absenteeism
from class due to extracurricular activities that
were occurring at the time of assessments (i.e., sports
meetings, music classes, learning assistance). Control
schools were less likely to provide additional opportunities
to screen children who were missing from class.
Single imputation via expectation maximization
method in SPSS was used for missing data in all analyses.
Risk Group Status
Participants were classified as high risk based on elevated
scores (above a specified cutoff) on either the
SCAS or the CDI. Given there were significant grade
and gender differences within this sample on the SCAS
at preintervention, this study used separate cutoffs for
grade and gender. The cutoffs used reflected the score
above which 10% of the normative sample scored
(Spence, 2005), which closely paralleled the mean
scores of the top 10% of this sample. The cutoff for the
CDI was 14 across all students given that there were no
significant grade or gender differences on the CDI at
preintervention in this sample. Table 1 presents the frequency
and percentage of students at high risk across
grade and intervention conditions at each time point.
Chi-square tests revealed significant differences between
the intervention and control conditions at 36-
407
UNIVERSAL SCHOOL-BASED ANXIETY PREVENTION PROGRAM
Table 1. Frequency and Percentage of Students at High
Risk Across Intervention Conditions for Each Follow-Up
Time Point
12 Months 24 Months 36 Months
n % n % n %
Total Intervention 59 16 49 17 22 12a
Female 34 58 29 59 13 59
Male 25 42 20 41 9 41
Total Control 62 21 40 25 26 31a
Female 31 50 24 60b 13 42
Male 31 50 16 40b 13 50
aSignificance between intervention and control groups at 36 months,
p < .001. bSignificance between females and males within the control
condition at 24 months, p < .05.
Downloaded By: [University of Queensland] At: 03:16 23 February 2010
month follow-up, χ2(1, 261) = 13.57, p < .001, with
significantly more students in the control condition at
high risk. There were equal girls to boys at high risk
within the intervention and control conditions at each
time point, except at 24-month follow-up, when there
were significantly more girls at high risk within the
control condition, χ2(1, 162) = 3.65, p < .05.
Long-Term Follow-Up
of Universal Prevention Effects
Prior to evaluating prevention effects across time
and intervention condition, a three-level (schools, students,
and occasions) multilevel analysis was conducted
across the dependent variables (CDI, SCAS,
RCMAS) to examine whether there was a clustering
effect of schools. The results indicated that the
“schools” level of data accounted for less than 5%
of total variance across dependent measures; hence it
was concluded that there was no clustering effect of
schools. Further analyses were conducted using univariate
analysis of variance covarying preintervention
group differences (analyses of covariance) on all dependent
measures, given that these group differences
were significant and not controlled for in Lock and
Barrett’s study (2003). To evaluate the long-term prevention
effects of the FRIENDS program on measures
of anxiety and depression, 2 (group: intervention or
control) Ă— 2 (grade: Grade 6 or Grade 9) Ă— 2 (gender) Ă—
2 (risk: high risk or healthy) Ă— 3 (time: 12-month follow-
up, 24-month follow-up, 36-month follow-up) repeated-
measures analyses of covariance were conducted.
As a result of multiple comparisons being
examined, Bonferroni adjustments, based on student’s
t statistic, were used in SPSS for the repeated-measures
analyses to adjust the observed significance level and
control for experimentwise error due to the multiple
comparisons.
Results of these analyses found no significant outcomes
on the CDI. On the SCAS, however, there was a
significant Group Ă— Grade interaction, F(1, 193) =
9.25, p < .005, partial eta squared = .05. Simple effects
analyses of group differences within each grade level
demonstrated that there were significant intervention
group differences on the SCAS for Grade 6 students,
F(1, 96) = 7.48; p < .01. Grade 6 students within the intervention
condition scored significantly lower on the
SCAS (M = 9.53, SE = 0.88) across time compared to
Grade 6 students in the control condition (M = 17.07,
SE = 2.61). There were no significant intervention
group differences for Grade 9 students on the SCAS.
On the RCMAS, there was also a significant Group
Ă— Grade interaction, F(1, 193) = 8.63, p < .005, partial
eta squared = .04. Simple effects analyses of group differences
within each grade level demonstrated that
there were significant intervention group differences
for Grade 6 students, F(1, 96) = 4.43; p < .05. Across
time, Grade 6 students within the intervention condition
scored significantly lower on the RCMAS (M =
4.67, SD = 0.45) than Grade 6 students in the control
condition (M= 7.63, SD = 1.34). There were no significant
intervention group differences on the RCMAS for
the Grade 9 students. There was also a significant Time
Ă— Group Ă— Gender effect on the RCMAS measure of
anxiety, F(2, 386) = 7.37; p < .005, partial eta squared
= .04. To examine this interaction further, univariate
analyses were conducted separately for girls and boys.
Results revealed a significant Time Ă— Group interaction
for girls on the RCMAS, F(2, 186) = 4.07; p <
.04. Pairwise comparisons demonstrated that girls in
the intervention group scored significantly lower than
girls in the control group at 12-month follow-up, p <
.001, and 24-month follow-up, p < .05; however, girls
were not significantly different on RCMAS scores
across the intervention or control groups at 36-month
follow-up. There was no significant Time Ă— Group interaction
for boys. Table 2 presents the means and standard
deviations for all measures, across time, intervention
group, grade, and gender.
Discussion
This study examined the long-term prevention effects
of the FRIENDS program within the context of a
randomized controlled universal school-based prevention
trial. The FRIENDS intervention was offered to
both primary and secondary school students and was
evaluated against a control condition in Lock and Barrett’s
(2003) longitudinal study. This study aimed
to evaluate the longer term prevention effects of the
FRIENDS program across two grades, with long-term
follow-up data from 12-month, to 24-month, to 36-
month follow-up. It was anticipated that intervention
gains reported in Lock and Barrett would be maintained,
with students from within the intervention condition
self-reporting significantly lower scores on anxiety
and depression in comparison to the control
condition. Furthermore, it was expected that there
would be significantly fewer students within the highrisk
range on measures of anxiety and depression in the
intervention condition compared with the control condition.
Based on Lock and Barrett’s results, it was
further expected that there would be an Intervention
Group Ă— Grade effect, with students in Grade 6 reporting
more positive gains than students in Grade 9, and a
possible gender effect, with girls reporting the strongest
reductions in anxiety and depression.
The results of this study were generally supportive
of the study predictions. Consistent with Lock and
Barrett (2003), there was an Intervention Group Ă—
Grade effect, whereby students in Grade 6 appeared to
benefit more from the intervention than students in
Grade 9 on both measures of anxiety. For students in
408
BARRETT, FARRELL, OLLENDICK, DADDS
Downloaded By: [University of Queensland] At: 03:16 23 February 2010
Grade 6, there were significant interventions group differences
at all time points across both the SCAS and
RCMAS, with students who received the FRIENDS
program reporting significantly lower anxiety scores
than students in the control condition. There were
no significant group differences for students in Grade
9. This finding strengthens Lock and Barrett’s suggestion
that intervening with prevention in Grade 6 may
be an optimal time for reducing risk for anxiety and
depression.
Consistent with Lock and Barrett’s (2003) study,
there was a significant Time Ă— Intervention Group Ă—
Gender effect; however, this was on anxiety (RCMAS)
only. Univariate analyses revealed that girls in the intervention
group scored significantly lower than girls
in the control group at 12-month and 24-month follow-
up; however, there were no significant differences
between girls in each group at 36-month follow-up.
This finding suggests that there is a prevention effect
for girls on anxiety symptoms up to 24-month follow-
up, however, this effect washed out by 36-month
follow-up for girls, who are more at risk for anxiety
than boys, based on the self-report measures of this
study.
Based on previous findings, it seems that gender
may play an important role in predicting both risk for
anxiety and intervention outcome. Results from Lock
and Barrett (2003) suggest that girls, across age
groups, tend to be at higher risk for anxiety than boys
but also tend to be most responsive to an intervention
up to 12-months follow up when the intervention is delivered
in Grade 6. This finding, taken together with
the results of this long-term outcome study, suggests
that intervening during primary school years can provide
very positive short-term outcomes for girls at a
universal level. However, given that being female is associated
with elevated anxiety and risk for anxiety, results
suggest that prevention programs need to be extended
across education, in that students—particularly
female students—receive such interventions every 2
to 3 years during their schooling. Results from the
Queensland Early Intervention and Prevention of Anxiety
Project (Dadds et al., 1997), which examined a targeted
intervention for high-risk and clinical students,
found that being female was a predictor for poorer response
to the targeted intervention. This finding further
highlights that female students are at higher risk and,
when experiencing subclinical and clinical levels
of anxiety (such as the sample examined in Dadds et
al.), school-based interventions may need to be coupled
with more intensive treatment interventions in the
community.
In terms of students at high risk (that is, elevated
scores on either the SCAS or CDI), there were proportionally
more students at high risk in the control condition
at each time point in comparison to the intervention
condition, and this effect was significant at
36-month follow-up. The frequency of students at high
risk in the intervention condition remained relatively
stable over time (i.e., 16%, 17%, 12%), whereas there
was a substantial increase in the frequency of high-risk
students in the control condition over time (i.e., 21%,
25%, 31%).
The results of this study did not demonstrate any
significant interaction outcomes for depression over
time. Examination of the depression means across gen-
409
UNIVERSAL SCHOOL-BASED ANXIETY PREVENTION PROGRAM
Table 2. Means and Standard Deviations for SCAS, RCMAS, and CDI Across Grade, Condition, Gender, and Follow-Up Point
CDI SCAS RCMAS
12 24 36 12 24 36 12 24 36
M SD M SD M SD M SD M SD M SD M SD M SD M SD
Intervention
Grade 6
Female 4.75 4.97 6.43 8.59 4.75 5.34 14.61 9.86 14.59 13.20 8.56 6.12 6.15 5.41 7.11 6.00 4.93 4.96
Male 5.68 5.72 6.81 7.36 5.44 6.16 10.51 10.30 10.52 13.37 6.81 8.72 4.95 5.29 4.65 4.97 3.52 3.80
Total 5.18 5.34 6.61 8.01 5.06 5.71 12.71 10.25 12.66 13.40 7.55 7.73 5.59 5.37 5.94 5.65 4.30 4.51
Grade 9
Female 10.72 8.12 9.97 8.75 7.63 5.31 21.36 13.04 21.35 17.77 17.09 12.92 10.95 6.52 8.63 6.18 8.14 5.99
Male 8.60 6.93 7.75 6.68 6.86 5.05 15.10 12.83 14.38 16.61 11.62 7.09 8.99 5.82 6.83 5.63 5.78 3.47
Total 9.76 7.66 9.03 7.99 7.38 5.21 18.54 13.28 18.37 17.56 15.14 11.45 10.06 6.27 7.87 6.00 7.37 5.40
Control
Grade 6
Female 7.69 6.78 7.96 7.23 8.14 6.16 20.75 12.78 16.48 11.94 21.50 4.95 9.92 5.79 7.42 6.85 7.57 4.20
Male 8.00 6.68 9.84 8.07 8.04 8.37 17.94 13.33 20.05 17.24 12.00 12.15 7.99 6.14 8.86 5.94 7.76 5.72
Total 7.86 6.70 9.08 7.74 8.08 7.57 19.23 13.10 18.61 15.31 13.46 11.74 8.88 6.03 8.30 6.30 7.69 5.17
Grade 9
Female 9.25 7.14 11.73 8.51 8.87 6.20 18.51 12.04 17.84 11.08 12.27 8.43 11.11 6.07 9.33 5.40 7.87 5.14
Male 9.13 8.01 7.82 6.66 8.94 6.89 14.33 11.34 11.02 10.04 14.19 18.99 7.99 6.04 5.88 4.50 7.32 6.65
Total 9.19 7.57 9.81 7.87 8.91 6.52 16.40 11.84 14.50 11.08 13.33 15.11 9.54 6.23 7.64 5.25 7.60 5.92
Note: SCAS = Spence Children’s Anxiety Scale; RCMAS = Revised Children’s Manifest Anxiety Scale; CDI = Children’s Depression Inventory.
Downloaded By: [University of Queensland] At: 03:16 23 February 2010
der and grade reveal consistently lower scores for children
in the intervention group in Grade 6 compared to
students in the control group; however, this trend was
not apparent within the Grade 9 cohort. Although these
results were not significant, they suggest that intervention
reduction in depression at 12-month follow-up
(see Lock & Barrett, 2003) may be maintained over
time.
This follow-up study provides evidence for the durability
of prevention effects for children who received
this program in Grade 6 up to 3 years following a brief
cognitive–behavioral intervention, delivered by classroom
teachers, within the school curriculum. For girls,
however, who reported the highest scores of anxiety at
preintervention and who reported the largest reductions
in anxiety up to 12-month follow-up (Lock &
Barrett, 2003), it seems that prevention effects are only
durable up to 24-month follow-up. Future research and
school-based prevention initiatives should aim to examine
a multilevel approach to prevention for children
and youth, such that universal prevention is coupled
with indicated programs for students at elevated risk,
effectively strengthening the dose of intervention and
potentially increasing prevention outcomes. Longterm
prevention outcomes would also be greatly improved
by maximizing booster sessions and ongoing
exposure to the life-skills taught in such a program,
whereby students are taught these skills throughout
their schooling from preschool through to primary and
high school. It would be of great interest and benefit to
examine whether prevention outcomes could be enhanced
for high school students if they were previously
exposed to these strategies during primary school.
This study is the first to present long-term outcomes
for a universal prevention program targeting anxiety
and associated depression. Strengths of this study that
serve to increase the generalizability of findings include
random assignment of schools to intervention
conditions, the presence of a control group, relatively
large sample size, intervention fidelity checklists, implementation
of an evidence-based protocol involving
students and parents, teacher delivery of the program,
and the use of highly reliable and valid measures of assessment.
These findings are limited due to the absence
of diagnostic data and the absence of multi-informant
self-report measures (i.e., teacher and parent report).
This study has provided evidence for the worth in
pursuing universal prevention in reducing the incidence
of anxiety and depression and the subsequent
burden of suffering associated with these disorders.
The development of a strong evidence base in treating
and preventing emotional distress is crucial. The
FRIENDS program has a promising evidence base at
every level of intervention—from treatment to universal
prevention. As highlighted by the WHO project
summary on the prevention of mental illnesses (Hosman,
2004), it is the political, ethical, and professional
obligation of policymakers, educators, researchers,
and consumers to persist with efficacy and effectiveness
research using evidence-based programs. Further
replication studies demonstrating efficacy of long-term
preventative outcomes are warranted and will assist in
driving effective prevention initiates and programs
such as the FRIENDS program into schools.
References
American Psychiatric Association. (1994) Diagnostic and statistical
manual of mental disorders (4th ed.). Washington, DC:
Author.
Australian Bureau of Statistics and Department of Education, Training
and Youth Affairs. (1997). Australian Standard Classification
of Occupations (ASCO) (2nd ed.). Statistical Concepts Library,
Australian Bureau of Statistics (Cat. No. 1220.0).
Barrett, P. M. (1999). Interventions for child and youth anxiety disorders:
Involving parents, teachers, and peers. Australian Educational
and Developmental Psychologist, 16, 5–24.
Barrett, P. M. (2004). FRIENDS for life! for children. Participant
workbook and leaders manual. Brisbane, Australia: Australian
Academic Press.
Barrett, P. M. (2005). FRIENDS for life! for youth. Participant workbook
and leaders manual. Brisbane, Australia: Australian Academic
Press.
Barrett, P. M., Dadds, M. R., & Rapee, R. (1996). Family treatment
of childhood anxiety: A controlled trial. Journal of Consulting
& Clinical Psychology, 64, 333–342.
Barrett, P. M., Lowry-Webster, H., & Turner, C. M. (1999a).
FRIENDS for children group leader manual. Brisbane, Australia:
Australian Academic Press.
Barrett, P. M., Lowry-Webster, H., & Turner, C. M. (1999b).
FRIENDS for youth group leader manual. Brisbane, Australia:
Australian Academic Press.
Barrett, P. M., & Turner, C. M. (2001). Prevention of anxiety symptoms
in primary school children: Preliminary results from a universal
school-based trial. British Journal of Clinical Psychology,
40, 399–410.
Cobham,V. E., Dadds, M. R.,&Spence, S. H. (1998). The role of parental
anxiety in the treatment of childhood anxiety. Journal of
Consulting & Clinical Psychology, 66, 893–905.
Cole, D. A., Hoffman, K., Tram, J. M., & Maxwell, S. E. (2000).
Structural differences in parent and child reports of children’s
symptoms of depression and anxiety. Psychological Assessment,
12, 174–185.
Cole, D. A., Peeke, L. G., Martin, J. M., Truglio, R., & Seroczynski,
A. D. (1998). A longitudinal look at the relation between depression
and anxiety in children and adolescents. Journal of
Consulting & Clinical Psychology, 66, 451–460.
Dadds, M. R., Spence, S. H., Holland, D., Barrett, P. M., & Laurens,
K. (1997). Prevention and early intervention for anxiety disorders:
A controlled trial. Journal of Consulting & Clinical Psychology,
65, 627–635.
Dadds, M. R., Spence, S. H., Laurens, K., Mullins, M., & Barrett, P.
M. (1999). Early intervention and prevention of anxiety disorders
in children: Results at a 2-year follow-up. Journal of Consulting
& Clinical Psychology, 67, 145–150.
Dobson, K. (1985). The relationship between anxiety and depression.
Clinical Psychology Review, 5, 307–324.
Essau, C. A., Conradt, J., & Petermann, F. (2000). Frequency, comorbidity,
and psychosocial impairment of depressive disorders
in adolescents. Journal of Adolescent Research, 15,
470–481.
410
BARRETT, FARRELL, OLLENDICK, DADDS
Downloaded By: [University of Queensland] At: 03:16 23 February 2010
Evans, S. W. (1999). Mental health services in schools: Utilization,
effectiveness, and consent. Clinical Psychology Review, 19,
165–178.
Hannon, A. P., Rapee, R. M.,&Hudson, J. L. (2000). The prevention
of depression in children: A pilot study. Behaviour Change, 17,
78–83.
Hosman, C. (2004). WHO’s project in the prevention of mental illness.
Paper presented at the meeting of the Promotion of Mental
Health and Prevention of Mental and Behavioural Disorders,
Auckland, New Zealand.
Jaycox, L. H., Reivich, K. J., Gillham, J., & Seligman, M. E. P.
(1994). Prevention of depressive symptoms in school children.
Behaviour Research & Therapy, 32, 801–816.
Kashani, J. H., Carlson, G. A., Beck, N. C., Hoeper, E.W., Corcoran,
C. M., McAllister, J. A., et al. (1987). Depression, depressive
symptoms, and depressed mood among a community sample of
adolescents. American Journal of Psychiatry, 144, 931–934.
Kashani, J. H., & Orvaschel, H. (1990). A community study of anxiety
in children and adolescents. American Journal of Psychiatry,
147, 313–318.
Kovacs, M. (1981). Rating scales to assess depression in school-age
children. Acta Paedopsychiatrica, 46, 305–315.
Kubiszyn, T. (1999). Integrating health and mental health services in
schools: Psychologists collaborating with primary care providers.
Clinical Psychology Review, 19, 179–198.
Last, C. G., Hansen, C., & Franco, N. (1997). Anxious children in
adulthood: Prospective study of adjustment. Journal of American
Academy Child & Adolescent Psychiatry, 36, 645–652.
Last, C. G., Perrin, S., Hersen, M., & Kazdin, A. E. (1996). A prospective
study of childhood anxiety disorders. Journal of
the American Academy of Child & Adolescent Psychiatry, 35,
1502–1510.
Lewinsohn, P. M., Clarke, G. N., & Rohde, P. (1994). Psychological
approaches to the treatment of depression in adolescents. InW.
M. Reynolds & H. E. Johnston (Eds.), Handbook of depression
in children and adolescents (pp. 3–17). New York: Plenum.
Lock, S., & Barrett, P. M. (2003). A longitudinal study of developmental
differences in universal preventive intervention for child
anxiety. Behaviour Change, 20, 183–199.
Lowry-Webster, H., Barrett, P. M.,&Dadds, M. R. (2001). A universal
prevention trial of anxiety and depressive symptomatology
in childhood: Preliminary data from an Australian study. Behaviour
Change, 18, 36–50.
Lowry-Webster, H., Barrett, P., & Lock, S. (2003). A universal prevention
trial of anxiety symptomatology during childhood: Results
at one-year follow-up. Behaviour Change, 20, 25–43.
Mrazek, P. J.,&Haggerty,R. J. (1994). Reducing risks for mental disorders:
Frontiers for preventive intervention research. Washington,
DC: National Academy Press.
Orvaschel, H., Lewinsohn, P. M., & Seeley, J. R. (1995). Continuity
of psychopathology in a community sample of adolescents.
Journal of the American Academy of Child & Adolescent Psychiatry,
34, 1525–1535.
Reynolds, C. R., & Richmond, B. O. (1985). The Revised Children’s
Manifest Anxiety Scale: Manual. New York: Western Psychological
Services.
Shaffer, D., Fisher, P., Dulcan, M. K., & Davies, M. (1996). The
NIMH Diagnostic Interview Schedule for Children Version 2.3
(DISC 2.3): Description, acceptability, prevalence rates, and
performance in the MECA study. Journal of the American
Academy of Child & Adolescent Psychiatry, 35, 865–877.
Shochet, I. M., Dadds, M. R., Holland, D., Whitefield, K., Harnett, P.
H., & Osgarby, S. (2001). The efficacy of a universal schoolbased
program to prevent adolescent depression. Journal of
Clinical Child Psychology, 30, 303–315.
Shortt, A., Barrett, P. M.,&Fox, T. (2001). Evaluating the FRIENDS
program: A cognitive–behavioral group treatment for anxious
children and their parents. Journal of Clinical Child Psychology,
30, 525–535.
Spence, S. H. (1997). The structure of anxiety symptoms among
children: A confirmatory factor analytic study. Journal of Abnormal
Psychology, 106, 280–297.
Spence, S. H. (1998). A measure of anxiety symptoms among children.
Behavior Research & Therapy, 36, 545–566.
Spence, S. H., Barrett, P. M., & Turner, C. M. (2003). Psychometric
properties of the Spence Children’s Anxiety Scale with young
adolescents. Manuscript submitted for publication.
Spence, S. H. (2005). [Australian normative data for the Spence
Children’s Anxiety Scale]. Unpublished raw data.
Stanley, F. (2002). Year book Australia: Health centenary article—
Child health since federation. Brisbane, Australia: Australian
Bureau of Statistics.
Tannenbaum, L. E., Forehand, R., & Thomas, A. M. (1992). Adolescent
self-reported anxiety and depression: Separate constructs
or a single entity? Child Study Journal, 22, 61–72.
Weiss, B., Weisz, J., Politano, M., Carey, M., Nelson, W., & Finch,
A. (1991). Developmental differences in the factor structure
of the Children’s Depression Inventory. Psychological Assessment,
3, 38–45.
Received December 21, 2004
Accepted March 6, 2006
411
UNIVERSAL SCHOOL-BASED ANXIETY PREVENTION PROGRAM
Downloaded By: [University of Queensland] At: 03:16 23 February 2010