Effectiveness of CBT for Children and Adolescents With Anxiety Disorders Influenced by Type, Severity and Total Symptom Score
Masterthesis Orthopedagogiek Faculty of Social and Behavioral Sciences Graduate School of Childhood Development and Education Student name: A.M. van der Meijden Student number: 10514392 Supervisor and first assessor: dr. F.J.A. van Steensel Second assessor: G.E. Telman MSc Amsterdam, February 2016
Abstract
Objective: This study examined the influence of the characteristics type of anxiety disorder,
severity and total symptom score on the effectiveness of cognitive behavior therapy (CBT) for anxiety disorders in children and adolescents. Method: All participants with specific phobia, generalized anxiety disorder, separation anxiety disorder or social phobia (n = 60, 27 boys;
Mage = 12.20; range = 7-17 years) and their parents were referred to 1 of 7 mental health care centers and received the same CBT. Anxiety disorder was measured by the ADIS-C/P and anxiety symptoms by the SCARED-71 at pretest, posttest, 3 months, 1 year and 2 years after treatment (all parent reports). Results: Multiple regression analyses showed that CBT was more effective for those with separation anxiety disorder (at posttest, follow-up 1, follow-up 3) and specific phobia (at posttest and follow-up 1) as compared to those with social phobia. CBT was more effective for those with more severe anxiety disorders and higher total symptom scores as compared to those with mild anxiety disorders and lower total symptom scores. Conclusion: This study stresses the importance of considering the above-mentioned characteristics when treating anxiety disorder with CBT, allowing all participants an equal chance at optimal and permanent improvement.
Keywords: CBT, anxiety disorder, severity, type, symptom score, children,
adolescents
Samenvatting
Doel: Deze studie onderzocht de invloed van de kenmerken type angststoornis, ernst en totale
symptoomscore op de effectiviteit van cognitieve gedragstherapie (CGT) voor kinderen en adolescenten met een angststoornis. Methode: Alle participanten met een specifieke fobie, gegeneraliseerde angststoornis, separatieangststoornis of een sociale fobie (n = 60, 27 jongens; Mleeftijd = 12.20; range = 7-17 jaar) en hun ouders werden verwezen naar 1 van de 7 GGZ-centra en ondergingen dezelfde vorm van CGT. Voorafgaand aan de behandeling (pretest), direct erna (posttest), 3 maanden, 1 jaar en 2 jaar na de behandeling werd de angststoornis gemeten met de ADIS-C/P en de angstsymptomen met de SCARED-71 (gerapporteerd door ouders). Resultaten: Multipele regressieanalyses laten zien dat CGT effectiever is voor degenen met separatieangst (op posttest, 3 maanden en 2 jaar na
behandeling) en specifieke fobie (op posttest en 3 maanden na behandeling) dan voor degenen met sociale angst. CGT blijkt effectiever voor degenen met een ernstigere angststoornis en
hogere totale symptoomscores dan voor degenen met een mildere angststoornis en lagere totale symptoomscores. Conclusie: Deze studie benadrukt dat bij de behandeling van
angststoornissen rekening moet worden gehouden met de bovengenoemde kenmerken, zodat alle deelnemers een gelijke kans hebben op optimale en blijvende vooruitgang.
Sleutelwoorden: CGT, angststoornis, ernst, type, symptoomscore, kinderen,
Effectiveness of CBT for Children and Adolescents With Anxiety Disorders Influenced by Type, Severity and Total Symptom Score
Many children and adolescents experience anxiety, which can be seen as normal age-appropriate anxiety (Ramsawh, Chavira & Stein, 2010) that generally dissolves over a short period of time (Muris, 2007). For example, separation anxiety is common among children from 12 to 18 months, fear of thunder or lighting is common at 2 to 4 years (Beesdo, Knappe & Pine, 2009), and social anxiety is common during adolescence (Westenberg, Gullone, Bokhorst, Heyne & King, 2007). Because of this, many child psychiatrists initially thought that anxiety in children and adolescents was not worth paying close attention to. This led to little research attention for his phenomena. However, this has changed over the last decade in which there has been a rise in research aimed at understanding the development and treatment of anxiety in children and adolescents (Cartwright-Hatton, McNicol & Doubleday, 2006). Anxiety becomes pathological (i.e. an anxiety disorder) when it is persistent and severe, and causes clinically significant impairment or distress (Ramsawh et al., 2010). Research shows that anxiety disorder is the most prevalent psychiatric disorder (Merikangas, Nakamura & Kessler, 2009), with a life time prevalence of 28.8% in adults above 18 years (Kessler et al., 2005). Pine and Klein (2011) have estimated the life time prevalence in children and adolescents to be between 5 and 10%. Anxiety disorders are often comorbid among themselves and without treatment they are found to predict anxiety disorders and depression in adulthood as well as substance use problems, suicide attempts, hospitalization (Kendall et al., 2010) and poor school performance (Moreno, 2010), later in childhood. There are several treatment options for anxiety disorders, of which cognitive behavior therapy (CBT), behavior therapy and pharmacotherapy are proven to be effective (Pine & Klein, 2011). Results of a meta-analysis conducted by Ishikawa, Okajima, Matsuoka and Sakano (2007) strongly supports the effectiveness of CBT for anxiety in children and adolescents. The main focus of CBT is acquiring adaptive coping skills and to establish a sense of power over anxiety symptoms or situations that are associated with distress and impairment (Connolly & Bernstein, 2007). Although research has shown that CBT is effective for the treatment of anxiety disorder, less research has focused on identifying specific factors, such as the type and severity of the anxiety disorder and total symptom score, that could predict the effectiveness of CBT. See Appendix A for a summary of treatment studies. A study with a focus on the severity of the anxiety disorder, conducted by Southam-Gerow, Kendall and Weersing (2001), involved 107 children and adolescents, aged 7-15
years, who underwent at least 12 sessions of CBT and were divided into two groups after treatment: (1) good treatment response versus (2) poor treatment response. The selection of good versus poor treatment response was based on whether the children and adolescents met criteria for any of the Diagnostic and Statistical Manual of Mental Disorders [DSM-III-R or DSM-IV; American Psychiatric Association, 1987, 1994] anxiety disorders, reported by their parents at posttreatment and 1-year follow-up. At posttreatment, 41% of the children and adolescents were classified as good treatment responders. At 1-year follow-up this has
increased to 61% of children and adolescents. The study showed that poor treatment response was predicted by higher levels of child symptoms at baseline, measured by the Child
Behavior Checklist (CBCL) scale withdrawn behavior and the Teacher’s Report Form (TRF) scale anxiety/depressed.
Consistent with these findings is the research from Compton et al. (2014) who used data involving 488 children and adolescents aged 7-17 years from a large randomized
controlled trial (RCT), the Child/Adolescent Anxiety Multimodal Study (CAMS; Compton et al., 2010). Treatment response was measured by the Clinical Global Impression Scale–
Improvement (CGI-I). Overall results showed that more severe and impairing anxiety was associated with poorer treatment outcomes. In addition, higher baseline symptom severity, as measured by the Pediatric Anxiety Ratings Scale (PARS), was one of the most consistent predictors of poorer treatment response.
Finally, research of Liber et al. (2010) showed contrasting results. A higher level of baseline severity predicted poorer treatment outcome in 124 children, aged 8-12 years, as reported by children, parents and clinicians. However, children with a higher level of baseline severity needed greater reduction of symptoms in order to reach a subthreshold level of symptoms. Therefore, Liber et al. (2010) also calculated pre- to posttest Reliable Change Indexes, which demonstrated that the Reliable Change Index (RCI) was not predicted by child reported symptom severity. However, parent reported severity of internalizing symptoms did predict a greater RCI.
With regard to the type of anxiety disorder, Compton et al. (2014) and Hudson et al. (2015) found that children and adolescents who had a primary diagnosis of social phobia were associated with poorer treatment outcomes. In contrast with this finding is research from Layne, Bernstein, Egan and Kushner (2003). In this study, 41 anxious and depressed adolescents aged 12-18 years, underwent 8 weeks of CBT combined with medication (i.e. imipramine) or placebo. Layne et al. showed that having a separation anxiety disorder or an
avoidant disorder1 leads to poorer treatment response.
Research from Hirshfeld-Becker et al. (2010), involving 64 children with anxiety disorders, aged 4-7 years, showed that children who participated in the CBT intervention showed a significantly larger decrease in anxiety symptoms (effect size [ES] .55) than children in the control-group (i.e. wait list condition). In addition, children treated for social phobia, separation anxiety disorder and specific phobia showed significantly better Clinical Global Improvement (CGI; ES respectively .95, .82 and .78) than children treated for generalized anxiety disorder or agoraphobia.
In contrast, research of Shortt, Barrett and Fox (2001), involving 71 children, aged 6-10 years, showed no significant difference in treatment effects at posttreatment between generalized anxiety disorder, separation anxiety disorder and social phobia. Respectively 71%, 73% and 56% of these children were diagnosis-free at posttreatment.
The research described above showed that CBT is effective for the treatment of anxiety disorders and that higher severity of the anxiety disorder and higher levels of symptoms at pretest predict poorer treatment response. Findings regarding the type of the anxiety disorder are inconsistent. Research of Compton et al. (2014) and Hudson et al. (2015) showed that having a social phobia leads to poorer treatment outcomes, while other research showed that poorer treatment outcomes were predicted by having a separation anxiety disorder, avoidant disorder (Layne et al., 2003), generalized anxiety disorder or agoraphobia (Hirshfeld-Becker et al., 2010). Finally, research of Shortt et al. (2001) showed no significant differences in treatment outcomes between different types of anxiety disorders. Because of these inconsistent findings and the scarcity of research focusing on identifying specific factors that could predict the effectiveness of CBT, such as the type and severity of the anxiety disorder and total symptom score, more research is needed.
This study aimed to identify these factors and therefore focused on the relationship between treatment effectiveness and child and adolescent anxiety characteristics at pretest (i.e. type and severity of the anxiety disorder and total symptom score). It is hypothesized that CBT is more effective for children and adolescents with mild anxiety disorders and lower total symptom scores as compared to children and adolescents with more severe anxiety disorders and higher total symptom scores. With regard to the type of anxiety disorder,
1
In the revision of DSM-III-R (American Psychiatric Association, 1987) to DSM-IV (American Psychiatric Association, 1994) avoidant disorder was combined with specific anxiety disorder (i.e. social phobia), given their similarity.
it is cautiously hypothesized that CBT is less effective for children and adolescents with social phobia as compared to children and adolescents with other anxiety disorders.
Method Participants
The original study of van Steensel and Bögels (2015) included 95 children and adolescents. The current study contained 60 children and adolescents of their sample, due to the decision to only include children and adolescents with specific phobia (SP), generalized anxiety disorder (GAD), separation anxiety disorder (SAD) and social phobia (SOP) who participated in at least the pre- and posttest assessments.
Participants were 60 children and adolescents (27 boys; Mage = 12.20; SD = 2.66; range = 7-17 years) who received CBT for treating anxiety disorders at a secondary mental health care center in the Netherlands. See Table 1 for a summary of primary anxiety
diagnoses in children and adolescents at pretest, as well as comorbid anxiety disorders within each group. Their parents were also asked to participate in the study. Inclusion criteria were: (a) children/adolescents were diagnosed with at least one anxiety disorder (which was confirmed with the Anxiety Disorder Interview Schedule-Child/Parent version, ADIS-C/P; Silverman & Albano, 1996), (b) child/adolescent and at least one parent had to be willing to engage in the study and (c) children/adolescents had to be medication-free or medication use was stable for at least three months before and during CBT. Exclusion criteria were: (a) an IQ lower than 70, (b) untreated psychotic disorder, (c) acute suicidal risk, or (d) present physical or sexual abuse. All children and adolescents were Caucasian. Parental participation was higher among mothers (100%) than among fathers (76,7%). The mean age of the parents was 42.32 years (mothers) and 44.91 years (fathers). Their mean educational level (measured on a scale from 1 [no education] to 9 [university level]) was 5.43 (mothers) and 5.78 (fathers). Measures
Anxiety disorders. The presence or absence of anxiety disorders was determined with the ADIS-C/P (Silverman & Albano, 1996), which is based on DSM-IV criteria. The
interview started with exploring symptoms of anxiety disorders in children/adolescents. Then, the child/adolescent and their parents were requested to rate the impairment for daily
functioning on a 0-8 point scale. A score of 4 or higher means that a diagnosis of an anxiety disorder is given. The severity of the primary anxiety disorder was determined using the clinician rating. The ADIS-C/P has good psychometric qualities (e.g., Silverman, Saavedra & Pina, 2001), and has been used in several studies evaluating treatment effectiveness for
Table 1
Anxiety Diagnoses in Children and Adolescents at Pretest Reported by Parents
Primary anxiety Comorbid anxiety
disorder (%, n) disorder
SP GAD SAD SOP
SP 21.7 (13) - 15.4 (2) 0 (0) 7.7 (1)
GAD 20 (12) 66.7 (8) - 66.7 (10) 33.3 (4)
SAD 25 (15) 66.7 (10) 66.7 (10) - 26.7 (4)
SOP 33.3 (20) 55 (11) 30 (6) 5 (1) -
Note. SP = specific phobia; GAD = generalized anxiety disorder; SAD = separation anxiety disorder; SOP = social phobia.
anxiety in children and adolescents (Compton et al., 2014; Liber et al., 2010; Rapee, Abbott, Lyneham & La Greca, 2006). Interrater reliability within the original sample is found to be high, with kappa’s ranging from .73 to 1.00 (van Steensel & Bögels, 2015).
In addition to the ADIS-C/P, the Screen for Child Anxiety Related Emotional Disorders-71 (SCARED-71; Bodden, Bögels & Muris, 2009) was completed by children, adolescents and their parent(s). The SCARED-71 consists of 71 descriptions of anxiety symptoms of which a total score can emerge, as well as seven subscales; specific anxiety (15 items), separation anxiety (12 items), social anxiety (9 items), generalized anxiety (9 items), panic disorder (13 items), obsessive-compulsive disorder (9 items), and post-traumatic stress disorder (4 items). The total symptom scores were used in the analyses. Children/adolescents and their parents were asked to rate each item in terms of how often a particular symptom is endorsed (0 = almost never; 1 = sometimes; 2 = often). Bodden et al. (2009) established cut-off scores for both child and parent reports of the SCARED-71. See Appendix B for a summary of children and adolescents who fall within the clinical range, as reported by their mothers. Furthermore, Bodden et al. (2009) found that psychometric properties of the
SCARED-71 were good. Cronbach’s alpha for the current sample across assessments ranged between .82 and .84 for parent reports.
Procedure
Inclusion of the children and adolescents who participated in the study began in 2006 and ended in 2010. The final follow-up assessment (two years after CBT) was conducted at the beginning of 2013. All children and adolescents were referred to one of seven
participating mental health care centers due to anxiety problems. After having followed intake procedure at the mental health care centers, a multidisciplinary team of psychologists,
therapists, social workers and psychiatrists established whether children and adolescents met criteria for a clinical DSM-IV-TR anxiety diagnosis, based on interviews with the
child/adolescent and parent(s), observations of child-parent interactions and/or school
observations, diagnostic assessments and psychiatric consults. When a child met criteria for at least one DSM-IV-TR anxiety disorder and the family agreed to participate in the study, a first assessment was conducted by a psychologist or diagnostician who worked and/or conducted research within one of the mental health care centers. These psychologists or diagnosticians were independent of the staff that designated the anxiety diagnoses and of the staff that treated the children.
At the first assessment, the ADIS-C/P was conducted and questionnaires were completed. If ADIS-C/P criteria for at least one anxiety disorder were met, children were enrolled in the study. Medical ethical approval for the study and informed consent was obtained. All assessments were administered at the mental health care center or at the
child’s/adolescent home at five occasions: (1) before treatment (pretest), (2) immediately after treatment (posttest), (3) three months after treatment (follow-up 1), (4) one year after
treatment (follow-up 2), and (5) two years after treatment (follow-up 3). Intervention
For this study existing data from the study of van Steensel and Bögels (2015) was used. The study concerned a longitudinal study examining the effectiveness of CBT for anxiety disorders in children and adolescents with and without an autism spectrum disorder. This study focused solely on children and adolescents with anxiety disorders.
The intervention consisted of a combined version of a family and individual CBT-program (Bodden et al., 2008) called ‘Discussing + Doing = Daring’, developed for treating anxiety disorders in developing children. Previous research, among a large Dutch sample, proved that this treatment is highly effective to treat anxiety disorders in children and adolescents. Effect sizes reached 1.4, and 73% of the children and adolescents were free of their primary anxiety disorder at the 3 months follow-up (Bodden et al., 2008).
Therapists had to be cognitive-behavioral therapists or registered mental health care psychologists to provide the CBT-program and were trained by professor doctor Bögels, who is the developer of the CBT-program. During the intervention, weekly supervision group meetings and bi-annual meetings were organized to discuss protocol compliance and to provide additional training.
Research design and analyses
When focused on children and adolescents reports, there appeared to be significant differences between the number of children per anxiety group. Because these differences were much smaller for parent reports it is decided to solely include parent reports. Since 25% of the fathers failed to fulfill the SCARED-71, it is decided to solely include maternal reports in all analyses regarding the SCARED-71. For missing data the method of “last observation carried forward’’ - assuming no change has occurred - was used, following Walkup et al. (2008), for cases that did not participated at some time points.
To measure the effectiveness of the CBT treatment the differences between pretest scores and all follow-up scores on the ADIS-C/P and the SCARED-71 were calculated (e.g. pretest score minus posttest score). Furthermore, anxiety severity scores (ADIS-C/P) and total symptom scores (SCARED-71) were used in repeated measures analyses of variance
(ANOVA’s) to examine decreases per anxiety disorder over time and calculate an effect size (partial eta squared) per anxiety disorder. The Greenhouse-Geisser correction was used to correct for violations of the sphericity assumption (Field, 2009).
To investigate which anxiety characteristics predict treatment effectiveness, multiple regression analyses, using sequential regression, were applied for the ADIS-C/P and the SCARED-71. These analyses were used to examine a possible positive or negative
relationship between the independent variables (severity of the anxiety disorder and type of the anxiety disorder for ADIS-C/P and total symptom score for SCARED-71) and the
dependent variable (effectiveness of the CBT treatment). Gender differences between groups were checked using a Chi-square test of independence. There appeared to be no significant differences between child gender and type of the anxiety disorder (χ2 (3, n = 60) = .08, p = .994). Age differences between groups were checked using a one-way ANOVA. There appeared to be significant differences between child age and type of the anxiety disorder (respectively F(3, 59) = 3.32, p = .026). Therefore, analyses were run with and without age as a predictor. Since there were no differences, age was not included in the final analyses. In the first step of the multiple regression analyses the independent variables were entered. SAD, GAD, and SP, were converted into dummy variables and compared to SOP (i.e. the reference group). The second step of the analyses included the interaction between the independent variables.
Results CBT and severity of anxiety disorders (ADIS-C/P)
A repeated measures ANOVA was conducted for each of the anxiety disorders to determine if severity scores decreased over time and to compare effect sizes between anxiety disorders. Statistically significant results were found for all anxiety groups: SP (F(4, 48) = 60.73, p = .000), GAD (F(1.40, 15.35) = 27.91, p = .000), SOP (F(3.12, 59.23) = 33.24, p = .000), and SAD (F(1.92, 26.93) = 166.52, p = .000). CBT had a large effect on all four anxiety disorders: SP (η2p= .95), GAD (η2p= .80), SOP (η2p= .82), and SAD (η2p= .95). See Figure 1 for a visual summary. Post-hoc pairwise comparisons, using a Bonferroni correction, showed that CBT had a significant effect in reducing anxiety severity scores between time
points for all anxiety disorders. Severity scores for all anxiety disorders significantly differed between pretest and posttest and all follow-up assessments. In addition, severity scores for SOP significantly differed between posttest and follow-up 2. See Appendix C for a summary of the results.
CBT and anxiety symptoms (SCARED-71) A repeated measures ANOVA was conducted for each of the anxiety disorders to determine if total symptom scores decreased over time and to compare effect sizes between anxiety disorders. Statistically significant results were found for all anxiety groups: SP (F(4, 48) = 7.12, p = .000) GAD (F(1.80, 19.76) = 8.37, p = .003), SOP (F(4. 76) = 12.28, p = .000), and SAD (F(2.22, 28.82) = 21.51, p = .000). CBT had a large effect on all four anxiety disorders: SP (η2p= .63), GAD (η2p= .53), SOP (η2p= .67), and SAD (η2p= .78). See Figure 2 for a visual summary. Post-hoc pairwise comparisons, using a Bonferroni correction, showed that CBT had a significant effect in reducing total symptom scores between time points for all anxiety disorders. Total symptom scores for SP significantly differed between pretest and all follow-up assessments. For SAD and SOP, total symptom scores significantly differed between
0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 AD IS -C /P A n x iet y S ev er ei ty S co res Specific phobia
Generalized anxiety disorder Separation anxiety disorder Social phobia
Figure 1. ADIS-C/P: Anxiety Severity Scores for Pretest (1), Posttest (2), FU1 (3), FU2 (4) and FU3 (5).
pretest and posttest and all follow-up
assessments. In addition, total symptom scores for SAD significantly differed between posttest and follow-up 3. Total symptom scores for GAD significantly differed between pretest and
follow-up 2. See Appendix D for a summary of the results.
CBT and severity and type of anxiety disorders and total symptom score
ADIS-C/P. Multiple regression analyses were conducted to test if the type and severity of the anxiety disorder at pretest predicted the effectiveness of CBT. Analyses showed that the type and severity of the anxiety disorder
predicted up to 34% of the variance in effectiveness of CBT. Severity of the anxiety disorder at pretest did significantly predict the effectiveness of CBT at all time points. SP and SAD did significantly predict the effectiveness of CBT at posttest and follow-up 1. Including the interaction between type of the anxiety disorder and severity of the anxiety disorder at pretest did not result in an increase in variance at all time points. No interaction effect was found between type of anxiety disorder and severity of the anxiety disorder. See Table 2 for a summary of the results.
SCARED-71. Multiple regression analyses were conducted to test if the type of the anxiety disorder and total symptom score at pretest predicted the effectiveness of CBT.
Analyses showed that the type of the anxiety disorder and total symptom score predicted up to 35% of the variance in the effectiveness of CBT. Total symptom score at pretest did
significantly predict the effectiveness of CBT at all time points. SAD did significantly predict the effectiveness of CBT at follow-up 1 and follow-up 3. Including the interaction between type of the anxiety disorder and total symptom score at pretest did not result in an increase in variance at all time points. No interaction effect was found between the type of anxiety disorder and total symptom score. See Table 3 for a summary of the results.
0 10 20 30 40 50 60 70 0 1 2 3 4 5 S CA RE D -71 T ot al S y m pt o m S cor es Specific phobia
Generalized anxiety disorder Separation anxiety disorder Social phobia
Figure 2. SCARED-71: Total Symptom Scores for Pretest (1), Posttest (2), FU1 (3), FU2 (4) and FU3 (5).
Table 2
Multiple Regression Analyses: Effectiveness of CBT Predicted by the Type and Severity of the Anxiety Disorder at Pretest
Pre-Post Pre-FU1 Pre-FU2 Pre-FU3
B SE ß B SE ß B SE ß B SE ß Step 1 Constant -.75 1.82 .73 1.75 1.53 1.66 1.55 1.49 SP 3.06 .86 .45* 1.87 .82 .30* .22 .78 .04 1.32 .70 .24 GAD .61 .90 .09 -.07 .86 -.01 -1.27 .82 -.21 -.96 .73 -.17 SAD 2.74 .82 .42* 2.29 .79 .38* 1.04 .75 .18 1.29 .67 .25 Severity .63 .25 .28* .57 .24 .28* .63 .23 .33* .59 .21 .33* R² .34* .28* .26* .32* F 7.11* 5.40* 4.79* 6.57* Step 2 Constant -1.49 1.93 -.29 1.83 .51 1.73 .68 1.56 SP 4.84 1.77 .70* 4.31 1.67 .68* 2.67 1.59 .45 3.40 1.43 .61* GAD 3.83 2.93 .54 4.35 2.78 .67 3.16 2.63 .52 2.79 2.40 .49 SAD .95 1.76 .15 -.17 1.67 -.03 -1.43 1.58 -.25 -.80 1.42 -.15 Severity 1.25 .59 .56* 1.42 .56 .69* 1.48 .53 .77* 1.31 .48 .73* TAD*Severity -.25 .22 -.66 -.35 .21 -.98 -.35 .20 -1.05 -.30 .18 -.95 R² .36 .32 .30 .36 ∆R² .02 .04 .04 .03 F 1.33 2.79 3.12 2.77 ∆F .255 .101 .083 .102
Note. Type of anxiety disorder was represented as three dummy variables with social anxiety disorder serving as a reference group (constant). Pre = pretest; Post = posttest; FU1 = follow-up 1; FU2 = follow-up 2; FU3 = follow-up 3; SP = specific phobia; GAD = generalized anxiety disorder; SAD = separation anxiety disorder; TAD = type of anxiety disorder. *p = < .05.
Table 3
Multiple Regression Analyses: Effectiveness of CBT Predicted by the Type of the Anxiety Disorder and Total Symptom Score at Pretest
Pre-Post Pre-FU1 Pre-FU2 Pre-FU3
B SE ß B SE ß B SE ß B SE ß Step 1 Constant -3.43 6.00 -1.98 6.00 -.01 5.98 .07 6.17 SP 6.33 5.61 .16 5.10 5.61 .12 1.05 5.59 .03 3.26 5.77 .07 GAD -.35 5.59 -.01 -3.84 5.59 -.09 -5.59 5.57 -.13 -.12 5.75 -.00 SAD 9.18 5.45 .24 12.82 5.45 .31* 8.29 5.43 .20 14.07 5.60 .32* TSS .30 .11 .38* .36 .11 .42* .41 .11 .48* .40 .11 .44* R² .21* .32* .33* .35* F 3.54* 6.26* 6.53* 7.17* Step 2 Constant -5.19 6.54 -4.88 6.49 -4.89 6.32 -1.32 6.75 SP 8.79 6.67 .22 9.17 6.62 .22 7.89 6.45 .19 5.20 6.88 .12 GAD 5.75 10.46 .14 6.22 10.38 .14 11.33 10.11 .26 4.69 10.79 .10 SAD 5.23 7.92 .14 6.31 7.85 .15 -2.68 7.65 -.07 10.96 8.16 .25 TSS .46 .26 .57 .62 .25 .72* .86 .25 .99* .52 .26 .58 TAD*TSS -.06 .09 -.28 -.10 .09 -.43 -.17 .09 -.72 -.05 .09 -.19 R² .22 .33 .37 .35 ∆R² .01 .02 .05 .00 F .48 1.32 3.94 .28 ∆F .49 .26 .05 .60
Note. Type of anxiety disorder was represented as three dummy variables with social anxiety disorder serving as a reference group (constant). Pre = pretest; Post = posttest; FU1 = follow-up 1; FU2 = follow-up 2; FU3 = follow-up 3; SP = specific phobia; GAD = generalized anxiety disorder; SAD = separation anxiety disorder; TSS = total symptom score; TAD = type of anxiety disorder. *p = < .05.
Discussion
This study focused on the relationship between treatment effectiveness and child and adolescent anxiety characteristics at pretest (i.e. type and severity of the anxiety disorder and total symptom score). Results showed that type and severity of the anxiety disorder and total symptom score predicted the effectiveness of CBT. Findings did not fully support the hypotheses.
First, findings showed that CBT appeared to be highly effective for the treatment of all four anxiety disorders and led to a decrease of total symptom scores. Second, findings based on the ADIS-C/P showed that, at posttest and follow-up 1, CBT appeared to be more effective for children and adolescents with SAD as compared to children and adolescents with SOP. This is consistent with findings based on the SCARED-71, in which having a SAD leads to better treatment outcomes at follow-up 1 and follow-up 3 as compared to having SOP. Furthermore, findings based on the
ADIS-C/P also showed that, at posttest and follow-up 1, children and adolescents with SP benefit more from CBT than children and adolescents with SOP. Because poorer treatment outcomes for children and adolescents with SOP were not found in comparison to all four types of anxiety disorders, these findings only partially correspond with the hypothesis and research of Compton et al. (2014) and Hudson et al. (2015), which showed that having SOP leads to poorer treatment outcomes. Children and adolescents with GAD benefit just as much from CBT as those with SOP.
Third, contrary to expectations, CBT appeared to be more effective for children and adolescents with more severe anxiety disorders and higher total symptom scores at pretest as compared to children and adolescents with mild anxiety disorders and lower total symptom scores at pretest. This finding conflicts with various research which showed the opposite finding (Compton, 2014; Liber et al., 2010; Southam-Gerow et al., 2001). A possible explanation for this contrasting finding might be that children and adolescents with more severe anxiety disorder and higher total symptom scores were more committed to CBT and therefore gained more improvement (Glenn et al., 2013). Another explanation might be that having a more severe anxiety disorder and higher total symptom score provided more room for improvement (Salemink, van den Hout & Kindt, 2009). Finally, for children and
adolescents who did not participate at all follow-up assessments, the use of the method ‘’last-observation-carried-forward’’, led to the inability to improve and could have influenced this finding (Streiner, 2008).
This study contains several strengths. The first strength is the clinical nature of the sample, consisting of children and adolescents all referred to a mental health care center. The second strength is the inclusion of long term follow-up assessments. The third strength is the use of a structural interview (i.e. the ADIS-C/P) as a scientific basis to determine the presence or absence of an anxiety disorder.
Despite these strengths, this study contains some limitations. The first limitations are the small sample size and small differences between the number of participants per anxiety group. This results in a lack of statistical power which leads to a limited generalizability. Therefore, results should be interpreted with caution. Another limitation is the inability to include child and adolescent reports. Results are now solely based on parents reports, which were subjective interpretations of their child’s behavior and therefore could be biased (Najman et al., 2001; Shortt et al., 2001). Furthermore, this study solely focused on the primary anxiety disorder of children and adolescents, while comorbid anxiety disorders were often present in this sample. For example, half of the children and adolescents with SOP had a
comorbid diagnosis of SP. It is unclear to what extent this has influenced the results. Another limitation that is described by van Steensel and Bögels (2015) and applies to the current sample, is the fact that all participants were recruited via community mental health care centers not specifically specialized in anxiety disorders. Although this benefits the generalizability of the findings, findings could be less generalizable to inpatient settings where an important proportion of children and adolescents with severe anxiety disorders may be admitted. Finally, all children and adolescents were Caucasian and therefore it is unsure to what extent these findings apply to children and adolescents with other origins.
The above limitations result in suggestions for further research. First, further research in larger and ethnographic diverse samples is needed to confirm findings and increase
generalizability. Second, further research should control for comorbid anxiety disorders, because these could possibly influence treatment effectiveness. Third, further research is needed to determine how best to effectively treat children and adolescents who are predicted to benefit less, or less quickly, from CBT. An option could be the addition of booster sessions (Hirshfeld-Becker, 2010, van Steensel & Bögels, 2015).
In conclusion, the current findings suggest that CBT is highly effective for the treatment of children and adolescents with anxiety disorders. Children and adolescents with more severe anxiety disorders and higher total symptom scores benefit more from CBT than those with mild anxiety disorders and lower symptom scores. In addition, children and adolescents with SP benefit more quickly from CBT than those with SOP. Finally, children and adolescents with SAD benefit more and more quickly from CBT than those with SOP. This study shows that the pace and extent to which children and adolescents with anxiety disorders benefit from CBT depends on the characteristics type and severity of the anxiety disorder and total symptom score. This stresses the importance of considering these characteristics when treating with CBT, allowing all children and adolescents with anxiety disorders to have an equal chance at optimal and permanent improvement.
References
Beesdo, K., Knappe, S., & Pine, D. S. (2009). Anxiety and anxiety disorders in children and adolescents: Developmental issues and implications for DSM-V. Psychiatric Clinics
of North America, 32, 483-524. doi:10.1016/j.psc.2009.06.002
Bodden, D. H. M., Bögels, S. M., & Muris, P. (2009). The diagnostic utility of the Screen for Child Anxiety Related Emotional Disorders-71 (SCARED-71). Behaviour Research
and Therapy, 47, 418–425. doi:10.1016/j.brat.2009.01.015
Bodden, D. H. M., Bögels, S. M., Nauta, M. H., De Haan, E., Ringrose, J., Appelboom, C., ... Appelboom-Geerts, K. C. (2008). Child versus family cognitive-behavioral therapy in clinically anxious youth: An efficacy and partial effectiveness study. Journal of the
American Academy of Child & Adolescent Psychiatry, 47, 1384–1394.
doi:10.1097/CHI.0b013e318189148e
Cartwright-Hatton, S., McNicol, K., & Doubleday, E. (2006). Anxiety in a neglected population: Prevalence of anxiety disorder in pre-adolescent children. Clinical
Psychology Review, 26, 817-833. doi:10.1016/j.cpr.2005.12.002
Compton, S. N., Peris, T. S., Almirall, D., Birmaher, B., Sherrill, J., Kendall, P. C., … Gosch, J. S. (2014). Predictors and moderators of treatment response in childhood anxiety disorders: Results from the CAMS trial. Journal of Consulting and Clinical
Psychology, 82, 212-224. doi:10.1037/a0035458
Compton, S. N., Walkup, J. T., Albano, A. M., Piacentini, J. C., Birmaher, B., Sherrill, J. T., … March, J. S. (2010). Child/adolescent anxiety multimodal study (CAMS):
Rationale, design, and methods. Child and Adolescent Psychiatry and Mental Health,
4. doi:10.1186/1753-2000-4-1
Connolly, S. D., & Bernstein, G. A. (2007). Practice parameter for the assessment and treatment of children and adolescents with anxiety disorders. Journal of the American
Academy of Child and Adolescent Psychiatry, 46, 267-283.
doi:10.1097/01.chi.0000246070.23695.0
Field, A. (2009). Discovering statistics using IBM SPSS statistics : And sex and drugs and rock 'n' roll. (3rd ed.). Los Angeles: Sage.
Glenn, D., Golinelli, D., Rose, R., Roy-Byrne, P., Stein, M., Sullivan, G., … Nezu, A. M. (2013). Who gets the most out of cognitive behavioral therapy for anxiety disorders? The role of treatment dose and patient engagement. Journal of Consulting and Clinical
Hirshfeld-Becker, D. R., Masek, B., Henin, A., Blakely, L. R., Pollock-Wurman, R. A., McQuade, J., … Biederman, J. (2010). Cognitive behavioral therapy for 4- to 7-year- old children with anxiety disorders: A randomized clinical trial. Journal of Consulting
and Clinical Psychology, 78, 498-510.
Ishikawa, S., Okajima, I., Matsuoka, H., & Sakano, Y. (2007). Cognitive behavioral therapy for anxiety disorders in children and adolescents: A meta-analysis. Child and
Adolescent Mental Health, 12, 164–172. doi:10.1111/j.1475-3588.2006.00433.x
Hudson, J. L., Keers, R., Roberts, S., Coleman, J. R. I., Breen, G., Arendt, K., … Eley, T. C. (2015). Clinical predictors of response to cognitive-behavioral therapy in pediatric anxiety disorders: The genes for treatment (GxT) study. Journal of the American
Academy of Child & Adolescent Psychiatry, 54, 454-463.
doi:10.1016/j.jaac.2015.03.018
Kendall, P. C., Compton, S. N., Walkup, J. T., Birmaher, B., Albano, A. M., & Sherrill, J., … Piacentini, J. (2010). Clinical characteristics of anxiety disordered youth. Journal of
Anxiety Disorders, 24, 360-365. doi:10.1016/j.janxdis.2010.01.009
Kessler, R. C., Berglunnd, P., Demler, O., Jin, R., Merikangas, K. R., & Walter, E. E. (2005). Lifetime prevalence and age-of-onset distributions of DSM-IV disorders in the
national comorbidity survey replication. Archives of General Psychiatry, 62, 593-602. doi:10.1001/archpsyc.62.6.593
Layne, A. E., Bernstein, G. A., Egan, E. A., & Kushner, M. G. (2003). Predictors of treatment response in anxious-depressed adolescents with school refusal. Journal of the
American Academy of Child and Adolescent Psychiatry, 42, 319-326.
doi:10.1097/01.CHI.0000037026.04952.96
Liber, J. M., Van Widenfelt, B. M., Van der Leeden, A. J. M., Goedhart, A. W., Utens, E. M. W. J., Treffers, P. D. A. (2010). The relation of severity and comorbidity to
treatment outcome with cognitive behavioral therapy for childhood anxiety disorders. Journal of Abnormal Child Psychology, 5, 683-694. doi:10.1007/s10802-010-9394-1
Merikangas, K. R., Nakamura, E. F., & Kessler, R. C. (2009). Epidemiology of mental disorders in children and adolescents. Dialogues in Clinical NeuroSciences, 11, 7-20. Moreno, M. A. (2010). Anxiety disorders in children and adolescents. Archives of Pediatric
Adolescents Medicine, 164, 984. doi:10.1001/archpediatrics.2010.179
Muris, P. (2007). Normal and abnormal fear and anxiety in children and adolescents. Rotterdam: Elsevier.
Najman, J., Williams, G., Nikles, J., Spence, S., Bor, W., O'Callaghan, M., … Shuttlewood, G. (2001). Bias influencing maternal reports of child behaviour and emotional state.
Social Psychiatry and Psychiatric Epidemiology, 36, 186-194.
doi:10.1007/s001270170062
Pine, D. S., & Klein, R. G. (2011). Anxiety disorders. In M. Rutter, D. Bishop, D. Pine, S. Scott, J. S. Stevenson, E. A. Taylor, & et al. (Ed.), Rutter's child and adolescent
psychiatry (pp. 822-840). New York: John Wiley and sons.
Ramsawh, H. J., Chavira, D. A., & Stein, M. B. (2010). Burden of anxiety disorders in pediatric medical settings: Prevalence, phenomenology, and a research agenda. Archives of Pediatric Adolescents Medicine, 164, 965-972.
doi:10.1001/archpediatrics.2010.170
Rapee, R., Abbott, M., Lyneham, H., & La Greca, A. M. (2006). Bibliotherapy for children with anxiety disorders using written materials for parents: A randomized controlled trial. Journal of Consulting and Clinical Psychology, 74, 436-444. doi:10.1037/0022- 006X.74.3.436
Salemink, E., Van den Hout, M., & Kindt, M. (2009). Effects of positive interpretive bias modification in highly anxious individuals. Journal of Anxiety Disorders, 23, 676- 683. doi:10.1016/j.janxdis.2009.02.006
Shortt, A. L., Barrett, P. M., & Fox, T. L. (2001). Evaluating the FRIENDS program: A cognitive-behavioral group treatment for anxious children and their parents. Journal of
Clinical Child Psychology, 30, 525.
Silverman, W. K., & Albano, A. M. (1996). Anxiety disorders interview schedule for DSM– IV child version, child interview schedule. San Antonio, TX: The Psychological Corporation.
Silverman, W. K., Saavedra, L. M., & Pina, A. A. (2001). Test–retest reliability of anxiety symptoms and diagnoses with the Anxiety Disorders Interview Schedule for DSM–IV: Child and parent versions. Journal of the American Academy of Child & Adolescent
Psychiatry, 40, 937– 944. doi:1097/00004583-200108000-00016
Southam-Gerow, M. A., Kendall, P. C., & Weersing, R. (2001). Examining outcome variability: Correlates of treatment response in a child and adolescent anxiety clinic. Journal of Clinical Child and Adolescent Psychology, 30, 422-436,
doi:10.1207/S15374424JCCP3003_13
Streiner, D. (2008). Missing data and the trouble with LOCF. Evidence Based Mental Health,
van Steensel, F. J. A., & Bögels, S. M. (2015). Cognitive-behavioral therapy for anxiety disorders in children with and without ASD. Journal of Consulting and Clinical
Psychology, 83, 512-523. doi:10.1037/a0039108
Walkup, J., Albano, A., Piacentini, J., Birmaher, B., Compton, S., Sherrill, J., … Kendall, P. (2008). Cognitive behavioral therapy, sertraline, or a combination in childhood anxiety. The New England Journal of Medicine, 359, 2753-2766.
doi:10.1056/NEJMoa0804633
Westenberg, P. M., Gullone, E., Bokhorst, C. L., Heyne, D. A., & King, N. J. (2007). Social evaluation fear in childhood and adolescence: Normative developmental course and continuity of individual differences. British Journal of Developmental Psychology, 25, 471-483. doi:10.1348/026151006X173099
Footnotes
1 In the revision of DSM-III-R (American Psychiatric Association, 1987) to DSM-IV (American Psychiatric Association, 1994) avoidant disorder was combined with specific anxiety disorder (i.e. social phobia), given their similarity.
Appendix A
Table 1
Summary of Treatment Studies
Authors, Year n Age Type of anxiety disorder Results
Compton et al., 2014
488 7-17 SAD, GAD, SOP More severe and
impairing anxiety and higher baseline symptom severity predicted poorer treatment outcomes. A primary diagnosis of social phobia was associated with poorer treatment outcomes Hudson et al., 2015 1519 5-18 SAD, GAD, SOP, SP A primary diagnosis
of social phobia was associated with poorer treatment outcomes Hirshfeld-Becker et
al., 2010
64 4-7 SAD, SOP, AD,
GAD, agoraphobia, SP
Treated children with SP, SAD or SOP showed greater improvement. Layne, Bernstein,
Egan and Kushner, 2003
41 12-18 SAD, OAD, AD,
PTSD, simple phobia, SOP, agoraphobia, panic disorder Having SAD of AD leads to poorer treatment outcomes.
Liber et al., 2010 124 8-12 SAD, GAD, SOP, SP Higher levels of baseline severity, as reported by parents, predicted poorer treatment outcomes. Shortt, Barrett and
Fox, 2001
71 6-10 SAD, GAD, SOP No significant
differences in treatment effects between various types
of anxiety disorders. Southam-Gerow,
Kendall and Weersing, 2001
107 7-15 SAD, GAD, OAD,
SOP, AD Poor treatment response was predicted by higher levels of child symptoms at baseline.
Note. SAD = separation anxiety disorder; GAD = generalized anxiety disorder; OAD = overanxious disorder; SOP = social phobia; AD = avoidant disorder; SP = specific phobia; PTSD = posttraumatic stress disorder.
Appendix B
Table 2
Children and Adolescents Within the Clinical Range at Pretest: n and % (SCARED-71)
M Cut-off score n %
SP 31.85 25 9 69.20
GAD 52.25 25 10 83.30
SAD 57.43 25 14 100
SOP 45.55 25 18 90
Note. SP = specific phobia; GAD = generalized anxiety disorder; SAD = separation anxiety disorder; SOP = social phobia; M = mean total symptom score within each anxiety group.
Appendix C
Table 3
ADIS-C/P Anxiety Severity Scores Between Time Points: Pairwise Comparisons
SP GAD SAD SOP
MD MD MD MD Pre Post 6.77* 3.75* 6.40* 3.60* FU1 6.62* 4.25* 7.00* 4.65* FU2 6.23* 4.25* 7.00* 5.90* FU3 7.08* 4.33* 7.00* 5.65* Post Pre -6.77* -3.75* -6.40* -3.60* FU1 -.15 .50 .60 1.05 FU2 -.54 .50 .60 2.30* FU3 .31 .58 .60 2.05 FU1 Pre -6.62* -4.25* -7.00* -4.65* Post .15 -.50 -.60 -1.05 FU2 -.39 .00 .00 1.25 FU3 .46 .08 .00 1.00 FU2 Pre -6.23* -4.25* -7.00* -5.90* Post .54 -.50 -.60 -2.30* FU1 .39 .00 .00 -1.25 FU3 .85 .08 .00 -.25 FU3 Pre -7.08* -4.33* -7.00* -5.65* Post -.31 -.58 -.60 -2.05 FU1 -.46 -.08 .00 -1.00 FU2 -.85 -.08 .00 .25
Note. SP = specific phobia; GAD = generalized anxiety disorder; SAD = separation anxiety disorder; SOP = social phobia. Pre = pretest; Post = posttest; FU1 = follow-up 1; FU2 = follow-up 2; FU3 = follow-up 3. A Bonferroni correction was used. *p = < .05.
Appendix D
Table 4
SCARED-71 Total Symptom Scores Between Time Points: Pairwise Comparisons
SP GAD SAD SOP
MD MD MD MD Pre Post 12.46 11.92 23.00* 10.25* FU1 14.54* 12.92 31.43* 14.35* FU2 14.15* 15.92* 31.93* 18.75* FU3 16.00* 20.75 37.00* 18.20* Post Pre -12.46 -11.92 -23.00* -10.25* FU1 2.08 1.00 8.43 4.10 FU2 1.69 4.00 8.93 8.50 FU3 3.54 8.83 14.00* 7.95 FU1 Pre -14.54* -12.92 -31.43* -14.35* Post -2.08 -1.00 -8.43 -4.10 FU2 -.39 3.00 .50 4.40 FU3 1.46 7.83 5.57 3.85 FU2 Pre -14.15* -15.92* -31.93* -18.75* Post -1.69 -4.00 -8.93 -8.50 FU1 .39 -3.00 -.50 -4.40 FU3 1.85 4.83 5.07 -.55 FU3 Pre -16.00* -20.75 -37.00* -18.20* Post -3.54 -8.83 -14.00* -7.95 FU1 -1.46 -7.83 -5.57 -3.85 FU2 -1.85 -4.83 -5.07 .55
Note. SP = specific phobia; GAD = generalized anxiety disorder; SAD = separation anxiety disorder; SOP = social phobia. Pre = pretest; Post = posttest FU1 = follow-up 1; FU2 = follow-up 2; FU3 = follow-up 3. A Bonferroni correction was used. *p = < .05.