Journal of Consulting and Clinical Psychology Copyright 1996 by the Am an Psychological Association, Inc.

19%, Vol. 64, No. 2, 333-342 0022-006X/96/$3.00

Family Treatment of Childhood Anxiety: A Controlled Trial

Paula M. Barrett Mark R. Dadds

Griffith University Griffith University

Ronald M. Rapee
Macquarie University

A family-based treatment for childhood anxiety was evaluated. Children (n = 79) aged 7 to 14
who fulfilled diagnostic criteria for separation anxiety, overanxious disorder, or social phobia were
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

randomly allocated to 3 treatment conditions: cognitive-behavioral therapy (CBT), CBT plus fam-
This document is copyrighted by the American Psychological Association or one of its allied publishers.

ily management (CBT -f FAM), and waiting list. The effectiveness of the interventions was evaluated
at posttreatment and at 6 and 12 months follow-up. The results indicated that across treatment
conditions, 69.8% of the children no longer fulfilled diagnostic criteria for an anxiety disorder, com-
pared with 26% of the waiting-list children. At the 12-month follow-up, 70.3% of the children in the
CBT group and 95.6% of the children in the CBT + FAM group did not meet criteria. Comparisons
of children receiving CBT with those receiving CBT + FAM on self-report measures and clinician
ratings indicated added benefits from CBT + FAM treatment. Age and gender interacted with treat-
ment condition, with younger children and female participants responding better to the CBT + FAM
condition.

Anxiety difficulties are among the common psychological behavioral treatment (CBT) of childhood anxiety, successfully us-
problem reported by children (Mattison, 1992), and there is ing a combination of behavioral techniques (e.g., in vivo exposure,
growing evidence that many anxious adults report their prob- relaxation and contingency management) and cognitive coping
lems to have originated in childhood (Rapee & Barlow, 1993). skills (e.g., self-instructional training), have mainly concentrated
For most children, various fears occur as part of normal child- on school fears, nighttime fears, and fears of medical procedures
hood development. However, for some, these fears intensify and (Dadds, etal., 1991; Kendall etal., 1992).
persist over time, preventing the child from enjoying age-related The only randomized treatment study of the general anxiety
activities and, thereby, warranting clinical attention. For exam- disorders in children (Kendall, 1994) has shown a CBT inter-
ple, Klein and Last (1989) reported that anxiety has a negative vention to be superior to a waiting-list condition, with improve-
effect on children's general social adjustment. Anxiety prob- ment across measures being maintained at the 12-month fol-
lems occur in 10% to 20% of school-age children; the more gen- low-up. In that study, 27 children (the waiting-list control group
eral anxiety disorders that have a major impact on children's consisted of 20 children) with a diagnosis of overanxiety, sepa-
functioning, such as overanxious disorder, separation anxiety, ration anxiety, or avoidant disorder were taught over 16 sessions
and social phobias, are found in approximately in 5% to 10% of
to develop realistic expectations, to develop coping self-talk, and
children (Werry, 1986).
to self-evaluate performance as well as modeling, exposure, and
Research is scarce with respect to the treatment of childhood
relaxation training. A multimethod assessment strategy was
anxiety disorders and the majority of published treatment studies
used, including a variety of child, parent, and teacher self-report
have focused on single-case designs used in the treatment of simple
measures, as well as behavioral observations.
phobias and specific fears in children (Dadds. Heard, & Rapee,
A point raised by Kendall (1994) concerned parental in-
1991; King, Hamilton, & Ollendick, 1988). Studies of cognitive-
volvement and the treatment of anxious children. Research has
not directly addressed this issue, and research is needed to eval-
uate the incorporation of structured family intervention in the
Paula M. Barrett, School of Applied Psychology, Griffith University,
Nathan, Australia; Mark R. Dadds, Faculty of Health and Behavioral treatment of anxious children. Indirect evidence points to the
Sciences, School of Applied Psychology, Griffith University, Nathan, potential importance of the family. A recent review by Rutter et
Australia; Ronald M. Rapee, Department of Psychology, Macquarie al. (1990) of genetic transmission studies indicated that a fa-
University, Sydney, Australia. milial loading is evident for adult anxiety disorders. However, it
This research was supported by grants from The National Health and is difficult to discern the weight of genetic versus environmental
Medical Research Council of Australia, and The Myer Foundation of
factors. Anxious children are more likely to have parents with a
Australia. We thank the families, all the anxiety project staff, and Candi
variety of disturbance and anxiety problems (Rutter et al.,
Peterson for her helpful feedback on a draft of this article.
Correspondence concerning this article should be addressed to Paula 1990). There is growing evidence that anxiety in children is sig-
M. Barrett, School of Applied Psychology, Griffith University, Gold nificantly related to frequent negative feedback and parental re-
Coast Campus, PMB50 Gold Coast Mail Centre, Queensland, 4217 striction (Krohne & Hock, 1991). Moreover, recent experi-
Australia. mental studies have provided evidence for the support of a

333
 334 BARRETT, DADDS, AND RAPEE

childhood anxiety model based on the development of an anx- Therapy was provided by five registered clinical psychologists (three
ious cognitive style in the context of anxiety supporting family women, two men) in the Behavior Research and Therapy Centre of the
processes; more specifically, family processes have been shown University of Queensland, Australia. One female therapist treated 38
to provoke an enhancement of avoidant responses in anxious participants (1 dropout), and each of the other male and female thera-
pists treated 4 participants each (4 dropouts among the remaining
children (Barrett, Rapee, Dadds, & Ryan, in press; Dadds, Bar-
therapists). Each therapist treated equal numbers of children in each
rett, Rapee, & Ryan, in press).
active treatment condition; that is, for each child treated in the CBT
Some studies have shown parenting training programs to-
group, the same therapist would treat a child in the CBT + FAM group.
gether with cognitive procedures to be beneficial in the treat- Participants were randomly assigned to the therapists so that the first
ment of simple phobias in children (Dadds et al., ] 991) but no treatment participant received CBT, the second received CBT +• FAM,
study has yet evaluated the value of incorporating parent train- the third received CBT, and so on, following alternate assignments both
ing in treatment outcome studies in childhood anxiety. For a to treatment conditions and therapists.
number of other childhood disorders, family interaction has Children who had intellectual or physical disabilities, who were cur-
been identified as playing a major role in the development and rently taking antianxiety or depression medication, or whose parents
were involved in acute marital breakdown (n - 2), were referred else-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

treatment of such disorders (Dadds, 1995) and several sugges-

This document is copyrighted by the American Psychological Association or one of its allied publishers.

where and not included in the study.

tions have been made to improve the treatment outcome of
childhood disorders by expanding the focus of treatment. Train-
ing parents in specific skills to manage the child's problems, Measures
improvements in marital and family problem solving and com- A double reliability diagnosis procedure was used where two clini-
munication, and parent's own psychosocial adjustment, are cians, interviewing parents and child separately, had to reach a final
some of the skills that appear to be associated with improved consensus diagnosis. The overall kappa agreement for the presence of
outcomes in childhood clinical interventions (Miller & Prinz, any anxiety disorder was .70, and kappas for specific anxiety disorder
1990). Thus, for families of anxious children, a comprehensive diagnoses were .63 for overanxiety disorder, .82 for social phobia, and
intervention might include training the parents in skills for .69 for separation anxiety disorder. More details of the diagnostic reli-
ability procedures and results can be found in Rapee, Barrett, Dadds,
managing the child's anxiety and avoidance, helping parents
and Evans (1994). The Anxiety Disorder Interview Schedule for Chil-
deal with any anxiety problems they themselves experience,
dren (ADIS-C) and a parallel version for the parents (ADIS-P) were
and improving family problem solving, as well as working with used (Silverman &. Nelles, 1988). These schedules are in accordance
the child using Kendall's CBT program (1994). with the Diagnostic and Statistical Manual of Mental Disorders (3rd
The present study reports on a randomized clinical trial in- ed., revised; DSM-III-R; American Psychiatric Association, 1987)
vestigating the effectiveness of cognitive-behavioral and family and contains criteria for the screening of other childhood disorders.
management training procedures with childhood anxiety disor- At posttreatment and follow-up, clinicians who were unaware of the
ders. The two treatment conditions were CBT and CBT plus child's treatment condition conducted diagnostic interviews and rated
family anxiety management training (CBT + FAM). It was hy- improvement in the child and family, on the basis of all anxiety disorder
items of the ADIS (an exact copy of all the ADIS questions for each
pothesized that both active treatment conditions would pro-
single anxiety diagnosis), and direct questions about seven dimensions
duce significant change in the dependent variables in contrast
of adjustment: (a) clinical global impression, (b) overall functioning,
to the waiting-list control condition. Improvement was ex-
(c) overall anxiety, (d) avoidant behaviors, (e) family disruption, (f)
pected to be across measures and maintained at 6- and 12- parental perceived ability to deal with child, and (g) child's perceived
month follow-ups. It was further hypothesized that the group ability to deal with feared situations. The clinicians made ratings of
with the added family training component would produce improvement using 0- to 6-point Likert-type scales, where 0 = markedly
greater improvement in the dependent measures in comparison worse, 3 = no improvement, and 6 = marked improvement.
with the cognitive-behavioral intervention.
Self-Report Measures
Method The Revised Children's Manifest Anxiety Scales (RCMASs) contain
37 items, 28 forming an anxiety subscale and 9 forming a lie scale. This
Participants
measure has sound validity and reliability (Reynolds & Richmond,
Seventy-nine children aged 7 to 14 years (45 boys, 34 girls) partici- 1978). The global manifest anxiety score has a mean of 13.8 for school-
pated in the treatment study. Children with one or more anxiety disor- age nonclinic children in Grades 1 through 12.
ders were recruited from referrals from community centers, schools, The Fear Survey Schedule for Children—Revised (FSSC-R) contains
mental health professionals, and medical practitioners, or parents re- 80 items (rated on a 3-point scale) assessing specific fears in children,
ferred them after media releases. The children were randomly allocated and it has Australian norms. This scale can be divided into five sub-
to CBT (« = 28), a CBT + FAM condition (n = 25), and a waiting-list scales, and it has good test-retest reliability and internal consistency.
(WL) condition (n = 26; treated after the waitlist period). We inter- The total fear score mean for girls was 145, and for boys it was 126
viewed children and their parents separately using a structured in- (nonclinic school samples). Matched school-age phobic children ob-
terview schedule and only children with a principal diagnosis of over- tained the means of 175 for girls and 151 for boys (Ollendick, Matson,
anxiety disorder (n = 30), separation anxiety disorder (n = 30), or AHelsel, 1985).
social phobia (n = 19) were included. Six percent of the children were The Children's Depression Inventory (CDI) has 27 items related to
comorbid with depression, 22% with simple phobias, and 2% with op- different depressive symptomatology. Each item has three descriptive
positional disorder. Children with a principal diagnosis bf simple pho- sentences, and children are instructed to select the one that best charac-
bia or other diagnoses were referred to the university clinic for separate terizes them during the previous 2 weeks. This scale has good internal
intervention. Children with a secondary diagnosis of simple phobia or consistency and moderate test-retest reliability (Kovacs, 1981). The
any other of the aforementioned anxiety disorders were included. clinical cutoff is approximately 13.
 TREATMENT OF CHILDHOOD ANXIETY 335

The Child Behavior Checklist (CBCL; Achenbach & Edelbrock, Procedure

1991) is a well-known and researched, psychometrically sound, 118-
item scale that assesses specific child behaviors from the parent's per- After referral, clinic staff contacted parents and made an intake ap-
spective. In the present study, it was used with both mothers and fathers. pointment interview during which consent forms were completed and
The CBCL provides a total behavior problem score, several subscale reliability diagnostic interviews were conducted. Parents and children
scores, and scores on two dimensions of dysfunction: Internalizing (e.g., were given a packet of self-report measures and asked to return them at
anxiety, depression, and withdrawal) and Externalizing (e.g., aggression the second intake (experimental) interview, which occurred within a
and impulsivity). Only the Internalizing and Externalizing scores were week. During the second interview, parents and children completed the
used in this study. For every scale, scores obtained above the 98th per- experimental procedure of interpreting and providing solutions to hy-
centile are considered within the clinical range. pothetical ambiguous situations and the two family discussion tasks.
The Depression Anxiety Stress Scales (DASS; Lovibond & Lovibond, After the intake interviews and establishment of a consensus diagno-
1994) comprise 3 seven-item scales assessing depression, anxiety, and ten- sis, participants were randomly assigned to the 12-week CBT, the 12-
sion-stress. The scales have been extensively developed and have excellent week CBT + FAM, or the 12-week WL condition. Participants in the
psychometric properties. The Depression and Anxiety subscales have been treatment conditions were randomly assigned to therapists and seen on
found to correlate strongly with more widely used measures of depression a weekly basis for 60 to 80 min. Participants assigned to the WL condi-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

and anxiety, respectively, and the Stress scale has been found to discrimi- tion were given the same measures as those in the treatment conditions
This document is copyrighted by the American Psychological Association or one of its allied publishers.

nate myocardialinfarct patients from nonpatients. In the present study, at the beginning and end of the 12-week waiting period. Families who
both mothers and fathers were asked to complete the DASS, and the scores sought alternative treatment during the waiting period (n ~ 2) were
were used only to assess treatment failure predictors. excluded from the analysis. All the WL control participants who con-
tinued to meet diagnostic criteria after the 12-week waiting period were
offered the family intervention treatment, as well as failures at 12
months follow-up in the CBT-only group.
Experimental Measures
We measured treatment integrity by randomly selecting and audio-
taping 60% of the therapy sessions across treatment conditions. A clini-
Threat interpretation and response plans to ambiguity (Barrett et al.,
cian who had not been informed of each family's treatment condition
in press) were measured. Threat and avoidance scores came from the
listened to all the audiotapes and followed a treatment integrity check-
child's interpretation and plans of responses and solutions to 12 ambig-
list to assure treatment adherence to each condition.
uous situations respectively. Mean threat scores and avoidant solutions
were obtained pretreatment and posttreatment for children.
The Family Enhancement of Avoidant Responses (FEAR) effects Treatment Materials
(Barrett et al., in press; Dadds et al., in press) were measured as follows:
FAM (Barrett, Dadds, & Rapee, 1991) was designed to run in paral-
Two of the ambiguous threat situations were selected to be discussed by
lel with an Australian adaptation of Kendall's CBT program (Coping
the family as a whole. The child was instructed to propose a final solu-
Cat Workbook; Kendall, 1990), which specifically targets the child's
tion (what he or she would do in response to the situation) at the end of
anxiety using exposure and cognitive restructuring strategies. Thus, two
the family discussion. The two ambiguous situations used for the family
treatment manuals were used in the FAM condition, and the Coping
discussions were as follows.
Koala Workbook (Barrett, Dadds, & Rapee, 1991) was used on its own
for the 12-session CBT child treatment condition.
Physical situation: On the way to school you (your child) feels All children in the active treatment conditions (i.e., CBT delivered in
funny in the tummy. What do you think is happening? What would the form of individual therapy) received the Coping Koala Workbook,
you (your child) do? which included recognizing anxious feelings and somatic reactions to
anxiety, cognitive restructuring in anxiety-provoking situations, coping
Social situation: You see a group of students from another class
self-talk, exposure to feared stimuli, evaluating performance, and ad-
playing a great game. As you (your child) walk over and want to
ministering self-reinforcement as appropriate. The first four sessions
join in, you notice that they are laughing. What do you think is
were training sessions in which anxiety management procedures were
happening? What would you (your child) do?
introduced, role-played by the therapist and practiced by each child
(identification of positive-negative thoughts that one forms in a variety
The child and her or his parents were asked to interpret and respond of situations and the feelings that one experiences, relaxation training,
to these situations, first alone, and then they were brought together to coping self-talk in anxiety-provoking situations, realistic self-evalua-
discuss the two ambiguous situations for 5 min each. The family was tion, development of self-reward strategies). Throughout the remaining
told that, although the parents could help the child, the final solution eight sessions, each child practiced the previously mentioned anxiety
was to be the child's decision. The child's solutions were recorded for coping skills by using in vivo exposure to feared situations, starting with
comparison with those previously suggested by the child in response to the low-stress situations and gradually increasing to high-stress situa-
the initial, individual protocol of the same ambiguous situations and tions (for more detail on specific treatment procedures, see Kendall,
questions. 1994).
The type of solutions suggested were scored as proactive, aggressive, Family Anxiety Management (FAM) was used in parallel with the
or avoidant; in the present study, just avoidant solutions were taken into Coping Koala Workbook in the CBT + FAM condition; that is, after the
account. Aioidant solutions were defined as those in which the child child completed each of the Coping Koala sessions, the child and par-
removes himself or herself from the situation. Raters were two psychol- ents would have a Family Anxiety Management therapy session.
ogists naive to the diagnostic status of the children, and 100% agreement FAM emphasizes process methods for empowering parents and chil-
occurred on the assignment of child responses to avoidant, aggressive, dren by forming an "expert team" with them, including the open shar-
and prosocial categories. Previous research has shown that this measure ing of information, joint determination of the content and processes of
of avoidance discriminates between anxious, nonclinic, and aggressive therapy, and the identification and reinforcement of family members for
children and that anxious children's tendency to report avoidant solu- any existing areas of expertise they have (Sanders & Dadds, 1993).
tions increases following family discussions (FEAR effect; Barrett, The three specific content aims of the FAM anxiety intervention are
Rapee, Dadds, & Ryan, in press). as follows: First, we provided training to parents in how to reward cou-
 336 BARRETT, DADDS, AND RAPEE

rageous behavior and extinguish excessive anxiety in the child; thus, cated significant differences between the conditions on these
parents were trained in reinforcement strategies including verbal praise, self-report measures at pretreatment.
privileges, and tangible rewards made contingent on facing up to feared
During treatment, there were 3 dropouts in the CBT condi-
situations. Planned ignoring was used as a method for dealing with ex-
tion (completes = 28), 4 in the WL condition (completers =
cessive complaining and anxious behavior; that is, the parent was
23), and 2 in the CBT + FAM condition (completers = 25).
trained to listen and respond empathically to the child's complaints the
first time they occur. However, repetitions were followed by the parent's During follow-up, one family in the CBT group and two fami-
prompting the child to engage in a coping strategy (that was learned lies in the CBT + FAM group moved out of state. For all mea-
in the parallel CBT intervention) and then the parent's withdrawing sures, we found no significant differences between completers
attention until the anxious or complaining behavior ceases. Thus, par- and dropouts using ANOV\ comparisons on all dependent
ents were taught how to use simple contingency management strategies measures.
such as descriptive praise, natural consequences, and planned ignoring Procedures to maximize and measure protocol adherence
to reduce conflict and increase cooperation in the family. During the were taken to ensure that therapists followed each item of the
sessions, parents role-played the contingency management strategies FAM protocol during each of the sessions. We also wanted to
with examples of their child's fearful behaviors.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

rule out the possibility that, if parents in the CBT-only condi-

This document is copyrighted by the American Psychological Association or one of its allied publishers.

Second, we taught parents how to deal with their own emotional up-
tion asked questions about how to manage their child, therapists
sets, gain awareness of their own anxiety responses in stressful situa-
would respond the same way and not provide child management
tions, and model problem-solving and proactive responses to feared sit-
uations. Third, we provided brief training in communication and prob- suggestions characteristic of the FAM condition. If such ques-
lem-solving skills to parents so that they would become better able to tions occurred, therapists followed the standard procedure of
work as a team in solving future problems and maintaining therapeutic only empathizing with the parent and redirecting him or her to
gains once therapy has terminated. This training involves the following: the content of the child's ongoing CBT sessions. Protocol adher-
(a) responding to conflict—parents are trained in skills for reducing the ence was confirmed for both treatment conditions for the 60%
escalation of interparental conflict over child-rearing issues and in being of sessions checked by a clinician who had not been informed
consistent with each other in terms of appropriate handling of their of each participant's treatment allocation. The clinician was
child's fearful behaviors; (b) daily discussions—parents are encouraged
given a checklist of each session's contents in both the CBT and
to set aside time for regular, casual discussions about each other's day,
FAM conditions, which she had to tick, item per item, as she
and brief training in listening skills is provided to increase the effective-
listened to the audiotapes. Hence, within each treatment condi-
ness of these discussions; and (c) problem solving—parents are encour-
aged to schedule weekly problem-solving discussions to aid in the tion, session-by-session contents were checked and adherence to
effective management of child and family problems. Brief training in protocol was confirmed for all sessions with each therapist—
problem-solving skills is provided (Dadds, 1989). The family interven- parents in the CBT did not receive any of the components of
tion is designed to be completed in 12 sessions; 4 sessions are devoted the FAM program.
to each of the discipline, anxiety management, and parental communi- To assess therapist comparability in terms of effectiveness, we
cation sections. conducted ANOVA comparisons on all dependent measures at
Therapist contact time was matched for both treatment conditions. posttreatment and at 6- and 12-month follow-ups using thera-
In the CBT condition, sessions were 60-80 min. In the CBT + FAM
pists as the independent variable. No significant differences
condition, approximately 30 min was spent on CBT and 40 min on the
were found.
family intervention. Treatment manuals were implemented with flexi-
bility to allow for the individuality of each family and child.
Diagnostic Status
Results
At posttreatment, the percentage of children who no longer met
To assure there were no significant demographic differences DSM-HI-R criteria for a current anxiety disorder was signifi-
across treatment conditions at pretreatment, we performed cantly less for the treatments (37 of 53 children, or 69.8%) than
one-way analyses of variance (ANOY\s) or chi-square tests for the WL (6 of 23 children, or 26.0%), X 2 ( 1, ff = 76) = 10.16,
comparing both treatments and WL conditions. There were no p < .05. There was also a significant difference between the two
significant differences across conditions for child's sex, mother's treatments (CBT: 16 of 28 children, or 57.1 %; CBT + FAM: 21 of
age and father's age, number of siblings, socioeconomic status, 25 children, or 84.0%), x 2 ( 1, N = 53) = 4.43,p< .05.
or marital status. There was a significant difference across treat- At the 6-month follow-up, the difference between the two treat-
ment conditions for child's age (CBT, M = 9.7, SD = 2.5; CBT ment conditions in the proportion of participants who no longer
+ FAM, M = 10.1, SD= 1.9; WL, M= 8.2, SD = 1.9), F(2, met diagnostic status was not significant (CBT: 20 of 28 children,
76) = 5.43, p < .01. A follow-up Tukey honestly significant or 71.4%; CBT + FAM: 21 of 25, or 84.0%), * 2 ( 1, AT= 53) = 1.1,
difference (HSD) test showed the difference to be between the ns. At the 12-month follow-up, the difference between the two
CBT + FAM and WL conditions. To determine whether post- treatment conditions on diagnostic status was again significant
treatment diagnosis results were different for child's age, we cre- (CBT 19 of 27 children, or 70.3%; CBT + FAM: 22 of 23, or
ated a dummy independent variable by rating each child as ei- 95.6%), X 2 ( 1, JV= 50) = 5.27,p < .05.
ther diagnosis free or having an anxiety diagnosis at posttreat-
ment and then comparing these two conditions on child's age.
Clinical Evaluations
The difference across conditions was not significant, F( 1, 51) =
0.40, ns. All dependent measures (self-report measures for both Table 1 presents means for the seven clinical evaluation
children and parents) were compared across both treatment scales, each on a scale ranging from 0 to 6 (0 = markedly worse,
conditions and the WL condition. None of the ANOVAs indi- 3 = no improvement, 6 = marked improvement). Means center
 TREATMENT OF CHILDHOOD ANXIETY 337

Table 1
Mean Independent Clinician Ratings and Standard Deviations of Improvement

Posttreatment 6-month follow-up 12-month follow-up

Scale CBT CBT + FAM CBT CBT + FAM CBT CBT-I-FAM

Clinical Global Impression

M 4.67' 5.20 4.82' 5.28 4.96' 5.52
SD 0.77 0.86 0.72 0.73 0.89 0.73
Overall Functioning
M 4.71 4.96 4.71' 5.24 4.85" 5.65
SD 0.85 0.93 0.71 0.78 0.86 0.65
Overall anxiety
M 4.50- 5.24 4.57° 5.32 4.77' 5.52
SD 0.79 0.88 0.79 0.69 0.97 0.79
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Avoidant Behaviors
This document is copyrighted by the American Psychological Association or one of its allied publishers.

M 4.89 5.20 4.67" 5.28 4.74" 5.47

SD 0.78 0.86 0.86 0.89 1.02 0.84
Family"
M 3.96' 5.00 4.00- 5.16 4.48' 5.65
SD 1.07 1.10 0.86 0.94 1.01 0.57
Family Skill0
M 4.14" 5.44 4.03' 5.28 4.37" 5.65
SD 0.97 1.00 0.84 0.84 1.11 0.57
Child Skill*1
M 4.78" 5.28 4.53" 5.36 4.67' 5.65
SD 0.87 0.93 0.88 0.57 1.14 0.49

Note. Univariate comparisons were made between the two conditions using ANOVAs, p < .05. A rating of
0 indicates markedly worse, 3 indicates no improvement, and 6 indicates marked improvement. CBT =
cognitive-behavioral treatment; CBT + FAM = cognitive behavioral treatment plus family anxiety man-
agement intervention.
• The mean is different from the corresponding mean. * Change of Family Disruption by the Child's Be-
havior. ° Change of Parent's Perception of Own Ability to Deal With Child's Behaviors. " Change of
Child's Ability to Deal With Difficult Situations.

around 5.0 for CBT and around 5.4 for CBT + FAM, and they F( 1, 51) = 6.2, p < .05; Change of Family Disruption by the
generally increase from posttreatment to the 6-month follow- Child's Behavior, F( 1, 51) = 21.9, p < .01; Change of Parent's
up and to the 12-month follow-up, indicating an overall im- Perception of Own Ability to Deal With Child's Behaviors, F( 1,
provement with time. As the clinical evaluations represent 51) = 28.9, p < .01, and Change of Child's Ability to Deal With
seven interrelated dimensions, we first tested for treatment Difficult Situations, F( 1, 51) = 15.9, p < .01.
group differences at posttreatment and at 6- and 12-month fol- At the 12-month follow-up, the MANOVA indicated a sig-
low-ups, using a series of multivariate analyses of variance nificant difference between treatments, F( 7,42) = 4.8, p < .01.
(M ANOVAs [repeated measures]). Where significant differ- Follow-up ANOVAs showed that the CBT + FAM treatment
ences were found, follow-up univariate ANOVAs were group was again significantly superior to the CBT group on all
conducted. seven clinical evaluation scales: Clinical Global Impression,
At posttreatment, an overall MANOVA indicated a signifi- F( 1, 48) = 5.6, p < .05; Overall Functioning, F( 1, 48) = 13.3,
cant difference between treatments, F(1, 44) = 5.28, p < .01. p < .01; Overall Anxiety, F(l, 48) = 8.5, p < .01; Avoidant
All means were higher in the CBT + FAM condition, and fol- Behaviors, F( 1, 48) = 7.5, p < .01; Change of Family Disrup-
low-up ANOVAs showed that the CBT + FAM treatment group tion by the Child's Behavior, F( 1, 48) = 24.0, p < .01; Change
was significantly superior to the CBT group on three of seven of in Parent's Perception of Own Ability to Deal with Child's Be-
the clinical evaluation scales: Clinical Global Impression, F( 1, haviors, F( 1,48) = 24.7, p< .01; and Change of Child's Ability
51) = 5.37, p< .05; Change of Family Disruption by the Child's to Deal With Difficult Situations, F( 1,48) = 14.7, ,p < .01.
Behavior, F( 1, 51) = 22.8, p < .01; and Change in Parent's Per-
ception of Own Ability to Deal With Child's Behaviors, F( 1,
Self-Report Measures
51) = 22.8,p<.01.
At the 6-month follow-up, the MANOVA indicated a signifi- Tables 2 and 3 show means and standard deviations for the
cant difference between treatments, F(7, 45) = 5.79, p < .01. self-report measures at pretreatment, at posttreatment, and at
Follow-up ANOVAs showed that the CBT + FAM treatment 6- and 12-month follow-ups for the CBT, CBT + FAM, and WL
group was significantly superior to the CBT group on all seven treatment conditions. Results for each self-report measure were
clinical evaluation scales: Clinical Global Impression, F( 1, 51) analyzed in two stages. First, we used a 3 (condition: CBT, CBT
= 5.37, p < .05; Overall Functioning, F( 1, 51) = 6.5, p < .05; + FAM, WL) X 2 (phase: pre- vs. posttreatment) ANOVA or
Overall Anxiety, F( 1, 51) = 13.3,p < .01; Avoidant Behaviors, MANOVA to examine immediate treatment effects. Second, we
 338 BARRETT, DADDS, AND RAPEE

Table 2
Means (and Standard Deviations) of Child Self-Report Measures for the Three Treatment Conditions Prelreatment, Posttreatment,
and Follow-Up (6-Month and 12-Month)

Pretreatment Posttreatment 6-month follow-up 1 2-month follow-up

Measure CBT CBT + FAM WL CBT CBT + FAM WL CBT CBT + FAM CBT CBT + FAM

RCMAS
M 12.7 11.5 13.3 9.0 6.6 11.6 5.6 3.8 4.9 4.3
SD 5.8 6.2 6.5 6.8 4.6 6.0 4.4 4.0 4.2 4.1
FSSCR
M 131.4 127.1 141.6 119.9 114.2 134.3 105.5 100.8 103.8 87.9
SD 23.5 26.1 21.4 26.0 20.2 32.6 22.0 19.4 22.4 28.6
CDI
M 8.3 8.6 8.4 4.5 4.1 6.8 3.7 2.4 2.8 2.2
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

SD 6.3 6.1 5.9 3.8 4.8 5.3 3.1 2.5 2.7 2.5
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Note. CBT = cognitive-behavioral treatment; CBT + FAM = cognitive-behavioral treatment plus family anxiety management; WL = waiting list;
RCMASs = Revised Children's Manifest Anxiety Scales; FSSCR = Fear Survey Schedule for Children—Revised; CDI = Children's Depression
Inventory.

used a 2 (treatment: CBT, CBT + FAM) X 3 (phase: posttreat- vealed a time effect only, F( 1, 71) = 25.49, p < .01, showing an
ment, 6-month follow-up, 12-month follow-up) ANOVA or overall reduction of self-reported anxiety for both treatments
MANOVA to examine the durability of treatment effects for and the WL condition. The ANOVA comparing treatment con-
the two active treatments. MANOVAs were used for measures ditions at posttreatment and at 6- and 12-month follow-ups also
having multiple subscales. Where a significant interaction oc- revealed phase effects only, F( 2,427 ) = 8.55, p < .01, indicating
curred between treatment and time, time effects were examined continued improvement for both treatments.
within each treatment group, and treatment conditions were From pre- to posttreatment for the FSSC-R, the ANOVA re-
compared at the relevant time. As an estimate of the clinical vealed phase and treatment main effects, F( 1, 72) = 11.10, p <
significance of change on self-report measures, percentages of .01, andf(2, 72) = 4.18, p < .05, respectively. The time effect
children scoring in the normal range (T score < 70) were cal- indicated an overall reduction of self-reported fears for all con-
culated for posttreatment and 12-month follow-up scores on ditions. The treatment effect appeared inconsistent with the
mothers' CBCL Internalizing scale, the measure that has the earlier analysis showing that the conditions were not different
best established discriminant validity for children in the age on any of the measures at pretreatment. Post hoc univariate
group used. analyses comparing the three conditions confirmed no signifi-
From pre- to posttreatment for the RCMAS, the ANOVA re- cant differences on the FSSC-R at pretreatment, but differences

Table 3
Means and Standard Deviations of Parent Self-Report Measures for the Three Treatment Conditions at Pretreatment,
Posttreatment, and 12-Month Follow-Up

Pretreatment Posttreatment 6-month follow-up 12-month follow-up

Parent and measure CBT CBT + FAM WL CBT CBT + FAM WL CBT CBT + FAM CBT CBT + FAM

Mother
CBCL-I
M 67.1 66.3 68,6 58.0 56.0 66.6 55.4 49.1 52.4 45.8
SD 7.5 7.3 6.3 8.3 8.1 9.5 8.6 9.4 9.6 7.6
CBCL-E
M 57.1 53.5 58.1 51.5 48.2 55.2 49.4 46.0 46.9 40.2
SD 8.9 7.9 8.3 7.1 8.9 8.6 7.8 9.7 7.8 8.2
Father
CBCL-I
M 66.3 64.9 64.3 57.9 57.2 64.1 53.9 47.0 52.9 44.2
SD 7.3 8.2 8.7 8.6 8.0 10.1 6.7 10.7 9.1 8.7
CBCL-E
M 58.7 54.7 57.9 54.8 49.2 55.4 49.6 45.5 47.3 39.8
SD 9.2 8.1 8.6 7.6 8.4 6.2 6.5 9.8 8.1 9.4

Note. CBT = cognitive-behavioral treatment; CBT + FAM = cognitive-behavioral treatment plus family intervention; WL = waiting list; CBCL-
I = Child Behavior Checklist, Internalizing subscale; CBCL-E = Child Behavior Checklist, Externalizing subscale.
 TREATMENT OF CHILDHOOD ANXIETY 339

were evident at posttreatment, F(2, 72) = 3.64, p < .05, despite CBT + FAM had significantly lower Internalizing scale scores
the lack of a Treatment X Time interaction. A post hoc Tukey's than children in the WL group.
HSD test showed that at posttreatment, CBT + FAM had sig- From pre- to posttreatment for father's scores on the Exter-
nificantly lower fear scores than the WL group. The ANOVA nalizing scale, time effects were evident for the CBT, F( 1,21) =
comparing treatment conditions at posttreatment and at 6- and 5.30, p < .05, CBT + FAM, F(l, 22) = 12.91, p < .01, and
12-month follow-ups produced a Time X Treatment interac- WL conditions, F( 1, 16) = 6.50, p < .05, indicating a general
tion, F(2, 46) = 3.87, p < .05, and a time effect, F(2, 46) = improvement in Externalizing scale scores for all three condi-
1 1 .53, p< .0 1 . Univariate comparisons of the conditions at the tions. The ANOVA at posttreatment was significant, /•"( 2,61) =
6- and 1 2-month follow-ups showed that CBT + FAM had lower 4.32, p < .05, and a follow-up Tukey showed that CBT + FAM
fear scores than CBT at 12-month follow-up, F( 1, 47) = 4.70, had significantly lower Internalizing scale scores than both the
p < .05, but not at the 6-month follow-up. WL and CBT conditions.
From pre- to posttreatment for the CDI, the ANOVA revealed The MANOVA comparing father's CBCL scores for the two
a time effect only, F(\, 73) = 25.27, p < .01, showing an im- treatment conditions at posttreatment and at 6-month and 12-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

provement in terms of overall reduction of self-reported depres- month follow-ups, showed significant time, F(4, 30) = 6.56, p
This document is copyrighted by the American Psychological Association or one of its allied publishers.

sion for the three conditions. The same applied for the ANOVA < .01, and treatment main effects, F( 2, 32) = 7.07, p < .01, but
comparing the treatment conditions at posttreatment, at 6- a nonsignificant Treatment X Time interaction. Thus, as for
month follow-up and at 12-month follow-up, F(2, 46) = 3.76, mothers, both treatment conditions continued to improve
p < .05, indicating that both treatment conditions continued to through follow-up, with CBT + FAM maintaining consistently
show reductions in CDI scores. lower Internalizing and Externalizing scale scores than CBT.
From pre- to posttreatment for mother's scores on the CBCL In terms of the clinical significance of CBCL results, there
Internalizing and Externalizing scales, the MANOVA revealed were no significant differences between treatment conditions in
a significant Treatment X Time interaction, F(4, 140) = 4.06, the number of children scoring above the CBCL-Internalizing
p< .0 1 , as well as treatment and time main effects. For the In- clinical cutoff at pretreatment. At posttreatment, 7.7% of the
ternalizing scale, there was a time effect associated with reduc- CBT group, 4.4% of the CBT + FAM group, and 39.1% of the
tions in Internalizing scale scores for both CBT, F(\, 22) = WL were still in the clinical range. At the 12-month follow-up,
21.19,p<.01,andCBT + FAM,F(l,23) = 80.08,p<.01,but 8.3% of the CBT, and 4.5% of the CBT + FAM group scored in
not for WL, F(l, 22) = 1.94, ns. An ANOVA revealed signifi- the clinical range. Differences between the CBT and CBT +
cant differences across conditions on the Internalizing scale at FAM groups were not significant at either posttreatment or the
posttreatment, F(2, 71) = 10.1, p < .01. A follow-up Tukey 12-month follow-up.
HSD test showed that children in both the CBT and CBT +
FAM conditions had significantly lower internalizing scores
than participants in the WL condition. Age, Sex, and Diagnosis Effects
From pre- to posttreatment for mother's scores on the Exter-
Numbers of participants did not provide sufficient power to
nalizing scale, time effects associated with reductions in Exter-
analyze all possible interactions of age, sex, and specific diagno-
nalizing scale scores were evident for CBT, F( 1 , 22 ) = 11 .24, p
sis on treatment outcome using all measures. Thus, we re-
< .01, and CBT + FAM, F( 1, 23) = 80.1, p < .01, but not the
stricted our analyses here to examination of main effects of age
WL, F( 1, 22) = 1.94, ns. The ANOVA at posttreatment was
(coded as younger [7-10 years] and older [11-14 years]), sex,
significant, F(2, 73) = 4.21, p < .05, and a follow-up Tukey
and pretreatment diagnosis (overanxiety, separation anxiety
showed that only CBT + FAM had significantly lower External-
disorder, social phobia) on treatment outcome at posttreatment
izing scale scores than the WL group.
and at the 12-month follow-up using rates of being diagnosis-
The MANOVA comparing mother's CBCL scores for the two
free as the dependent measure.
treatments at posttreatment, 6- and 12-month follow-ups,
For male participants, there were no significant differences
showed significant time, F(4, 35) = 9.83, p < .01, and treat- across treatment conditions both at posttreatment (diagnosis-
ment main effects, F( 2, 37 ) = 4.22, p < .05, but a nonsignifi- free: 65% CBT, 84% FAM), X 2 U , N = 33) = 1.5, ns, and at
cant Treatment X Time interaction. Thus, both treatments con- the 12-month follow-up (diagnosis-free: 75% CBT, 92% FAM),
tinued to improve through follow-up, with the CBT + FAM X 2 ( 1 , J V = 3 3 ) = 1.5, ns. For female participants, there were
group maintaining consistently lower internalizing and exter- significant differences across treatment conditions, indicating a
nalizing scores than the CBT-only group. superiority of the CBT + FAM intervention both at posttreat-
From pre- to posttreatment for father's scores on the CBCL ment (diagnosis-free: 37% CBT, 83% FAM), X 2 ( 1, N = 20) =
Internalizing and Externalizing scales, the MANOVA revealed 4.43, p < .05, and at the 12-month follow-up (diagnosis-free:
a significant Treatment X Time interaction, F(4, 120) = 3.42, 57% CBT, 1 0 0 % F A M ) , x 2 ( l , A r = 17) = 5.20,p< .05.
p < .05, as well as treatment and time main effects. For the In- At posttreatment, there were no significant differences rela-
ternalizing scale, there was a time effect associated with reduc- tive to absence of diagnoses, when participants' status was com-
tions in Internalizing scores for both the CBT, F( 1 , 2 1 ) = 29.36, pared across pretreatment diagnosis groups, x 2 (2, N = 53) =
0.99, ns (percentages of participants who were diagnosis free
the WL, F( 1, 16) = 1.45, ns. An ANOVA revealed significant when considering pretreatment diagnosis: overanxiety disorder,
differences across conditions at posttreatment, F( 2, 6 1 ) = 3.44, = 68.2%; separation anxiety disorder, 77.8%; social phobia =
p < .05. A follow-up Tukey HSD test showed that children in 61.5%). Identical results applied at the 12-month follow-up
 340 BARRETT, DADDS, AND RAPEE

(overanxiety disorder, 75.0%; separation anxiety, 94.1%; social TableS

phobia, 76.9%), * 2 (2, N= 50) = 2.58, ns. Percentage of Children in Each Treatment Group Choosing
In terms of age, younger children (7 through 10 years) had Avoidant Solutions Pre- 'and Post-Family Discussions,
significant higher rates of diagnosis-free participants at post- Before and After Treatment
treatment ( 100%) in the CBT + FAM condition in comparison
Pretreatment Posttreatment
with those ( 55.6% ) in the CBT condition, x 2 ( 1 , N = 33 ) = 8.8,
Treatment
p < .003. Older children ( 1 1 through 14 years, 60% diagnosis- group Pre Post Pre Post
free for both CBT and FAM; x 2 [l, N = 20] = 0, ns) did not
show significant differences across CBT and CBT + FAM treat- Waiting list 23.3 71.7 30.3 67.8
ment conditions at posttreatment. The same effect was observed CBT 38.4 88.5 15.9 3.9
CBT and family 25.0 88.5 15.9 0.0
for both age groups at the 12-month follow-up, with only the
younger group showing a significant higher proportion of par- Note. Pre = prediscussion; Post = postdiscussion; CBT = cognitive-
ticipants diagnosis-free in the CBT + FAM group, x 2 ( 1, N = behavioral treatment.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Experimental Measures tions similarly increased their avoidant plans of action after
family discussions. At posttreatment, children in both the CBT
Table 4 shows the children's means and standard deviations for
and the CBT + FAM groups decreased their avoidant plans after
threat interpretations and avoidant solutions across the three treat-
the family discussions, compared with the WL group children
ment conditions. For these scores (children, mothers, and fathers),
who maintained similar percentages of avoidant responses.
ANOVAs ( repeated measures) comparing pre- and posttreatment
For specific details on methodology and results of this section,
indicated significant Treatment X Phase interactions. Hence, we
please refer to Barrett et al. (in press).
dismantled the interactions by conducting a series of one-way AN-
OVAs comparing conditions at each phase. At pretreatment there
Discussion
were no significant differences across conditions both for threat
interpretations, F(2, 77) = 2.3, ns, and avoidant solutions, F(2, Both active treatment conditions produced significant change
77) = l.l,ns. At posttreatment, both active treatment conditions in the dependent variables in contrast to the WL control condi-
showed less threat interpretations and avoidant responses. Post tion. Improvement occurred across measures and was main-
hoc tests indicated that, for threat interpretations, the CBT + tained at the 6- and 12-month follow-ups. Furthermore, the
FAM group had lower mean threat scores in comparison with both condition with the added family training component showed
the CBT group and the WL group, F(2, 75) = 12.1,/><0.01. For significant improvement on a number of measures in compari-
avoidant responses, post hoc tests indicated that the CBT + FAM son with the CBT condition.
group had lower scores than those of the WL group, F(2, 75) = Similar to Kendall's (1994) study, the present study demon-
3.3,p<.05. strated the effectiveness of using cognitive-behavioral proce-
Table 5 shows the percentage of children choosing avoidant dures with anxious children. Kendall's controlled treatment
solutions before and after family discussions, pretreatment and study showed that 64% of children who had received CBT in-
posttreatment. At pretreatment, children in the three condi- tervention no longer met diagnostic criteria at posttreatment.
In the present study, 57% of children who had received the
shortened CBT (12 instead of 16-20 sessions used in Kendall's
intervention) no longer met diagnostic status at posttreatment.
Table4 For the same group, the proportion of children who were diag-
Child Threat Scores andstooidant Solutions Pre- and
nosis free increased at the 6-month follow-up to 71%, and this
Posttreatment
was maintained at the 12-month follow-up (70%). For the chil-
Threat scores and dren in the CBT + FAM condition at posttreatment, 6-month
avoidant solutions CBT CBT and family Waiting list follow-up, and 12-month follow-up, 84%, 84%, and 95% of chil-
dren were diagnosis-free, respectively. The success of both treat-
Threat scores ments and the extra benefits associated with the adjunctive fam-
Pretreatment
ily intervention applied equally to each specific anxiety group
M 6.3 4.8 5.8
SD 2.5 2.3 2.5 (overanxious, separation anxiety, and social phobia).
Posttreatment On the independent clinical evaluation scales, both the CBT
M 4.0 2.0 5.8 and the CBT + FAM groups showed improvement at posttreat-
SD 3.1 1.4 3.2
ment that was maintained at all follow-ups. Participants in the
Avoidant solutions
Pretreatment latter group reported continued significant improvement on all
M 3.8 3.3 2.8 clinical evaluation scales both at the 6- and the 12-month fol-
SD 3.5 3.2 2.8 low-ups, supporting the usefulness of a family component in the
Posttreatment maintenance and generalization of therapeutic gains.
M 2.2 0.8 3.5
Different treatment success rates by gender and age in re-
SD 4.0 1.2 4.6
sponse to the CBT and CBT + FAM treatment conditions were
Note. CBT = cognitive-behavioral treatment. evident. Younger children (7- to 10-year olds) responded better
 TREATMENT OF CHILDHOOD ANXIETY 341

to the CBT + FAM condition, but for older children (11- to 14- interpretations and avoidant plans at posttreatment, in compar-
year-olds), there was no significant difference across treatment ison with children in the WL and CBT conditions. However,
conditions both at posttreatment and at follow-up. Enhancing children in both active treatment conditions showed a reduction
parenting skills may be important for younger children, but for in the FEAR effect (Barrett et al., in press; Dadds et ah, in
older children individual child cognitive work and exposure to press) in comparison with WL participants. We would have ex-
feared stimuli may be sufficient to produce improvement in pected that the CBT + FAM treatment group would differ from
anxiety problems. the CBT-alone group on this specific experimental measure of
Regarding gender, some interesting findings were also ob- family interaction. The results show otherwise, perhaps because
served when we considered diagnostic status at posttreatment parents of the CBT children were aware of the exposure tech-
and at the 12-month follow-up: Female participants responded niques and the emphasis on substituting avoidance with proac-
better in the CBT + FAM treatment condition, but male partic- tive behaviors that was taught to their children during therapy.
ipants did equally well in both treatment conditions. We are Other possible explanations are that no differences were found
reticent to attempt an interpretation of this finding, given the because of a ceiling effect for change on the measure being
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

small numbers of female participants involved in the compari- reached or that a structured intervention at any level of a system
This document is copyrighted by the American Psychological Association or one of its allied publishers.

sons. However, it is interesting to note that previous research (parents and child vs. child alone) will induce changes in the
has found gender differences in the way parents interact with system's ability to problem solve in a proactive manner, at least
anxious children (Krohne & Hock, 1991), and further studies in front of the experimental team in the clinic.
are needed for examination of the interaction between gender The family intervention we evaluated incorporated a combi-
and family in the development and treatment of anxious nation of three components, and future research would benefit
children. from assessing which specific family treatment components en-
For the parent self-report measures, both CBT and CBT + hanced effectiveness and whether families with identifiable
FAM showed improvement at posttreatment in comparison with characteristics respond differentially to the different compo-
the WL group. For both the Internalizing and Externalizing scales nents. Future studies could also be improved by the inclusion of
of the CBCL for mothers and fathers, both treatment conditions a WL group in which equal therapist unstructured play contact
improved at posttreatment and continued to improve through fol- would be provided so that possible relationship effects could be
low-up, with CBT + FAM maintaining consistently lower Interna- ascertained.
lizing and Externalizing scale scores than the CBT-only group. The Other considerations have to be taken into account when in-
greater improvement in externalizing problems for the children in terpreting the results of the added FAM treatment components.
CBT 4- FAM may point to the importance of teaching parents The benefits documented in this study could simply be due to
contingency management skills to help their children deal with the adding of multiple treatments; the FAM benefits could be
feared stimuli and maximize the benefit of exposure techniques. interpreted with extra confidence if the study included another
Overall, the parent report measures support the extra benefits of condition in which an adjunctive treatment was provided to
the CBT + FAM intervention. CBT that did not produce additional gains. It is possible that
The children's self-report measures also showed the benefit parents in the FAM treatment were more motivated to provide
of active treatment, with limited support for the superiority of a favorable rating of the outcomes associated with the extra
the CBT + FAM condition. In the case of the FSSC-R, the CBT family condition because of the extra effort they contributed in
+ FAM group produced significantly lower FSSC-R scores at that condition.
the 12-month follow-up than the CBT group did. For the CDI, In conclusion, the provision of a structured intervention for
scores were below the clinical level for all conditions, so it was parents improved outcomes in the CBT of anxious children.
not surprising that no differences were found between WL and Further research is needed to establish the specificity of effect of
active treatment conditions at posttreatment and follow-ups. this intervention as a whole and in terms of its specific compo-
However, one needs to be cautious when interpreting child self- nents with regard to different ages and gender of children, and
report measures, because previous research has questioned with regard to parental characteristics such as psychosocial
their validity in discriminating anxious from nonclinic children adjustment.
(Perrin & Last, 1992). It would also be beneficial to implement controlled treatment
Another issue relates to the length and number of therapeutic studies with long-term follow-ups (throughout childhood and
sessions necessary for long-term successful results and the im- adolescence) to ascertain maintenance effects and the value of
plementation of cost-effective strategies. The present study's developing prevention programs targeting children with identi-
CBT program (twelve 60- to 80-min sessions) showed similar fied vulnerability for anxiety disorders in childhood.
success to Kendall's (1994) CBT intervention with sixteen 50-
to 60-min sessions. Perhaps future research could further ex- References
amine the issue of optimum intervention length for anxious
Achenbach, T. M., & Edelbrock, C. S. (1991). [Link] the Child
children. With respect to demands for therapist's expertise,
Behavior Checklist and Profile. Burlington: University of Vermont.
family interventions require specific parent management skills
American Psychiatric Association. (1987). Diagnostic and statistical
and sound knowledge of family processes during therapy. Fu-
manual of mental disorders (3rd ed., revised). Washington, DC:
ture studies implementing behavioral family programs need to Author.
take these considerations into account. Barrett, P. M., Dadds, M. R., & Rapee, R. M. (1991). Coping koala
On the experimental tasks, children in the CBT + FAM workbook. Unpublished manuscript. School of Applied Psychology,
group showed significant reductions in the number of threat Griffith University, Nathan, Australia.
 342 BARRETT, DADDS, AND RAPEE

Barrett, P. M., Rape*, R. M., Dadds, M. R., & Ryan, S. M. (in press). criteria and clinical assessment (pp. 179-202). Hillsdale, NJ:
Family enhancement of cognitive style in anxious and aggressive chil- Erlbaum.
dren: Threat bias and the FEAR effect. Journal of Abnormal Child Miller, G. E., & Prinz, R. J. (1990). Enhancement of social learning
Psychology. family interventions for childhood conduct disorder. Psychological
Dadds, M. R. (1989). Partner support training. Unpublished manu- Bulletin, 108,291-307.
script. School of Applied Psychology, Griffith University, Nathan, Ollendick, T. H., Matson, J. L., & Helsel, W. J. (1985). Fears in children
Australia. and adolescents: Normative data. Behavior Research and Therapy,
Dadds, M. R. (1995). Families, children and the development of dys- 23, 465-467.
function. London: Sage. Pen-in, S., & Last, C. G. (1992). Do childhood anxiety measures mea-
sure anxiety? Journal of Abnormal Child Psychology, 20, 567-577.
Dadds, M. R., Barrett, P. M., Rapee, R. M., & Ryan (in press). Family
Rapee, R. M., & Barlow, D. H. (1993). Generalized anxiety disorder,
process and child psychopathology: An observational analysis of the
panic disorder and the phobias. In P. B. Suther & H. E. Adams (Eds.),
FEAR effect. Journal of Abnormal Child Psychology.
Comprehensive handbook of psychopathology (2nd edition). New
Dadds, M. R., Heard, P. M., & Rapee, R. M. (1991). Anxiety disorders
York: Plenum.
in children. International Review of Psychiatry, 3,231-241.
Rapee, R. M., Barrett, P. M., Dadds, M. R., & Evans, L. (1994). Reli-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Kendall, P. C. (1994). Treating anxiety disorders in children: A con- ability of the DSM-III-R Childhood Anxiety Disorders using struc-
This document is copyrighted by the American Psychological Association or one of its allied publishers.

trolled trial. Journal of Consulting and Clinical Psychology, 62, 100- tured interview: Interrater and parent-child agreement. Journal of
110. the American Academy of Child and Adolescent Psychiatry, 33, 984-
Kendall, P. C., Chansky, T. E., Kane, M., Kim, R., Kortlander, E., Ro- 992.
nan, K., Sessa, F., & Siqueland, L. (1992). Anxiety disorders in youth: Reynolds, C. R., & Richmond, B. O. (1978). A revised measure of chil-
Cognitive-behavioral interventions. Needham Heights, MA: Allyn & dren's manifest anxiety scale. Journal of Abnormal Child Psychology,
Bacon. 6,271-280.
King, N. J., Hamilton, D. L, & Ollendick, T. H. (1988). Children 'sfears Rutter, M., McDonald, H., LeCouteur, A., Harrington, R., Bolton, P.,
and phobias: A behavioral perspective. Chichester, England: Wiley. & Bailey, A. (1990). Genetic factors in child psychiatric disorders-
Klein, R. G., & Last, C. G. (1989). Anxiety disorders in children. Lon- H. Empirical findings. Journal of Child Psychology and Psychiatry,
don: Sage. 31, 39-83.
Kovacs, M. (1981). Rating scales to assess depression in school aged Sanders, M. R., & Dadds, M. R. (1993). Behavioral'family intervention.
New York: Pergamon Press.
children. Acta Paedopsychiatrica, 46, 305-315.
Silverman, W. K., & Nelles, W. B. (1988). The Anxiety Disorders In-
Krohne, H. W., & Hock, M. (1991). Relationships between restrictive
terview Schedule for Children. Journal of the American Academy of
mother-child interactions and anxiety of the child. Anxiety Research,
Child and Adolescent Psychiatry, 27,772-778.
4, 109-124.
Werry, J. S. (1986). Anxiety disorders of childhood. New York: Guilford
Lovibond, P., & Lovibond, S. (1994). Self-report scales (DASS)forthe
Press.
differentiation of depression, anxiety and stress. Manuscript submit-
ted for publication. Received August 15, 1994
Mattison, R. E. (1992). Anxiety disorders. In S. R. Hooper, G. W. Revision received May 12, 1995
Hynd, & R. E. Mattison (Eds.), Child psychopathology: Diagnostic Accepted June 12, 1995 •