Predictors of treatment outcome in an effectiveness trial of cognitive behavioral therapy for children with anxiety disorders

Behaviour Research and Therapy 76 (2016) 1e12

Contents lists available at ScienceDirect

Behaviour Research and Therapy journal homepage: www.elsevier.com/locate/brat

Predictors of treatment outcome in an effectiveness trial of cognitive behavioral therapy for children with anxiety disorders Gro Janne H. Wergeland a, b, c, *, Krister W. Fjermestad a, d, Carla E. Marin e, Ingvar Bjelland b, f, Bente Storm Mowatt Haugland a, c, Wendy K. Silverman e, € a, g, h, i, Jon Fauskanger Bjaastad a, c, j, Kristin Oeding a, Odd E. Havik a, g, € ran Ost Lars-Go Einar R. Heiervang a, k, l a

Anxiety Research Network, Haukeland University Hospital, Bergen, Norway Department of Child and Adolescent Psychiatry, Haukeland University Hospital, Bergen, Norway c Regional Centre for Child and Youth Mental Health and Child Welfare, Uni Research Health, Bergen, Norway d Frambu Resource Centre for Rare Disorders, Siggerud, Norway e Child Study Center, Yale University School of Medicine, New Haven, CT, USA f Department of Clinical Medicine, Faculty of Medicine and Dentistry, University of Bergen, Norway g Department of Clinical Psychology, Faculty of Psychology, University of Bergen, Bergen, Norway h Department of Clinical Neuroscience, Karolinska Institutet, Stockholm, Sweden i Department of Psychology, University of Stockholm, Stockholm, Sweden j Division of Psychiatry, Stavanger University Hospital, Stavanger, Norway k Institute of Clinical Medicine, Faculty of Medicine, University of Oslo, Oslo, Norway l Division of Mental Health and Addiction, Oslo University Hospital, Oslo, Norway b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 7 April 2015 Received in revised form 29 October 2015 Accepted 3 November 2015 Available online 5 November 2015

A substantial number of children with anxiety disorders do not improve following cognitive behavioral therapy (CBT). Recent effectiveness studies have found poorer outcome for CBT programs than what is typically found in efficacy studies. The present study examined predictors of treatment outcome among 181 children (aged 8e15 years), with separation anxiety, social phobia, or generalized anxiety disorder, who participated in a randomized, controlled effectiveness trial of a 10-session CBT program in community clinics. Potential predictors included baseline demographic, child, and parent factors. Outcomes were as follows: a) remission from all inclusion anxiety disorders; b) remission from the primary anxiety disorder; and c) child- and parent-rated reduction of anxiety symptoms at post-treatment and at 1-year follow-up. The most consistent findings across outcome measures and informants were that child-rated anxiety symptoms, functional impairment, a primary diagnosis of social phobia or separation anxiety disorder, and parent internalizing symptoms predicted poorer outcome at post-treatment. Child-rated anxiety symptoms, lower family social class, lower pretreatment child motivation, and parent internalizing symptoms predicted poorer outcome at 1-year follow-up. These results suggest that anxious children with more severe problems, and children of parents with elevated internalizing symptom levels, may be in need of modified, additional, or alternative interventions to achieve a positive treatment outcome. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Anxiety Predictors Effectiveness Cognitive behavior therapy Children

Up to 40% of children with anxiety disorders do not show significant symptom reduction or diagnostic recovery following cognitive behavioral therapy (CBT) (James, James, Cowdrey, Soler, &

* Corresponding author. Department of child and adolescent psychiatry, Haukeland University Hospital, Bergen, Norway. E-mail address: [email protected] (G.J.H. Wergeland). http://dx.doi.org/10.1016/j.brat.2015.11.001 0005-7967/© 2015 Elsevier Ltd. All rights reserved.

Choke, 2013; Silverman, Pina, & Viswesvaran, 2008). Recent community clinic studies have found even poorer outcome for standard CBT programs than what is typically found in university clinic research studies (Bodden et al., 2008; Southam-Gerow et al., 2010; Wergeland et al., 2014). With the dissemination of CBT to community clinics, identifying predictors of treatment outcome is important, as this knowledge may be critical to adapt CBT to enhance treatment effectiveness (March & Curry, 1998; Weisz,

2

G.J.H. Wergeland et al. / Behaviour Research and Therapy 76 (2016) 1e12

Ugueto, Cheron, & Herren, 2013). Current knowledge about outcome predictors for CBT for children with anxiety disorders is derived mainly from studies in university-based child anxiety research clinics. Reviews have found some support for type and severity of child psychopathology at baseline (e.g., symptom severity, impairment, comorbidity, principal disorder of social phobia) and parent internalizing psychopathology (e.g., anxiety and depressive symptoms) being related to poorer treatment outcome (Compton et al., 2014; Knight, McLellan, Jones, & Hudson, 2014; Lundkvist-Houndoumadi, Hougaard, & Thastum, 2014; Nilsen, Eisemann, & Kvernmo, 2013). Only two studies have examined predictors of treatment outcome of CBT for children with anxiety disorders in community clinics (Bodden et al., 2008; Nauta, Scholing, Emmelkamp, & Minderaa, 2003). Bodden et al. (2008) reported that higher child age and the presence of parental anxiety disorder negatively influenced treatment outcome, whereas duration of anxiety complaints, but not age or gender, was associated negatively with treatment outcome in a study by Nauta et al. (2003). These two studies examined only a restricted numbers of predictors (i.e., age, gender, duration of anxiety and parent anxiety disorders) and applied different definitions of treatment outcome (e.g., diagnostic remission vs. symptom change). Thus, there is limited knowledge about predictors for CBT outcome in children with anxiety disorders when delivered in community clinics, and more research is needed in this area. Generalizability of findings from university research clinics to community clinics is uncertain, given differences that may exist between samples of children with anxiety disorders in these two settings. For example, higher levels of anxiety symptoms and externalizing problems, higher impairment, and more frequent comorbid conduct disorders have been found among children with anxiety in community clinics. Furthermore, these patients report higher levels of life stressors and more frequently come from lowincome, ethnically diverse, and single-parent families (EhrenreichMay et al., 2010; Southam-Gerow, Chorpita, Miller, & Gleacher, 2008; Southam-Gerow, Weisz, & Kendall, 2003; Villabø, Cummings, Gere, Torgersen, & Kendall, 2013). Such differences in child and family characteristics may influence the generalizability of effects and the aptness of CBT programs developed in universitybased child anxiety clinics when transported to community clinics (Weisz, Ng, Rutt, Lau, & Masland, 2013). Some of the clinical child factors (e.g. symptom severity, impairment, comorbidity) that distinguish the patients in the two settings have also been related to poorer treatment outcome in university research trials (Compton et al., 2014; Knight et al., 2014) and thus may contribute to poorer treatment response in community clinics (Weisz, Donenberg, Han, & Weiss, 1995; Weisz, Ugueto, et al., 2013). Certain family characteristics, such as lower family social class, single-parent family status, higher levels of family stresses, and parent mental health challenges have been associated with lower caregiver involvement and support in treatment and poorer treatment adherence in youth psychotherapy (Weisz, Ugueto, et al., 2013). It is important to examine whether these clinical child factors and family characteristics adversely influence treatment outcome of CBT for anxiety disorders when delivered in community clinics, as this would have implications for treatment delivery. The inclusion of factors beyond demographic, child and parent symptom and diagnosis variables could improve our understanding of factors associated with treatment outcome for children with anxiety disorders. It is functional impairment that usually brings clients to treatment (Angold, Costello, Farmer, Burns, & Erkanli, 1999), and current level of distress has been found to be one of the strongest motivators for treatment engagement (Lyneham,

Rapee, & Hudson, 2014). However, children are most often referred for treatment at the initiative of adults (DiGiuseppe, Linscott, & Jilton, 1996). Child treatment motivation, defined as acknowledgment of problems, perceived distress, and willingness to change (Keijsers, Schaap, Hoogduin, Hoogsteyns, & de Kemp, 1999), may therefore be relevant for treatment outcome. It may be difficult to engage children in treatment who do not acknowledge personal distress, and motivation may be particularly relevant for treatment of anxious children expected to face demanding tasks such as exposure aspects of CBT protocols (Kendall et al., 2009). Examining whether motivation influences treatment outcome is important, because motivation may be amenable to change before and during the early phase of treatment (Greenberg, Constantino, & Bruce, 2006). Therefore, a measure of motivation was included as a predictor in the present study. The aim of the present study was to examine predictors of treatment outcome in children with anxiety disorders treated with manualized CBT in community clinics. Predictors were selected on the basis that they should be representative of anxious children in community clinics. Beyond the variables already identified in the reviews, we included variables found to differ between samples of children with anxiety disorders in university research clinics and community clinics, as well as treatment motivation as mentioned above. The resulting set of predictors was organized into the following categories: demographic factors (i.e. age, gender, singleparent family status, family social class), child factors (i.e. internalizing and externalizing symptoms, impairment from these symptoms on daily functioning, nonanxiety comorbid disorders, principal anxiety disorder, and child motivation), and parent factors (i.e. parent self-rated internalizing symptoms and family stresses). Because our sample comprised at least 90.7% Caucasian, we were unable to include ethnicity as a predictor. Outcome was defined as complete remission (i.e. the remission of all three principal anxiety disorders; separation anxiety disorder, SAD; social phobia, SOP; generalized anxiety disorder, GAD); remission of the principal anxiety disorder; and anxiety symptom improvement (i.e. decrease in the child and parent ratings of child anxiety symptoms), allowing a broad evaluation of treatment outcome. We predicted that high baseline levels of child symptoms (both internalizing and externalizing), higher impairment from symptoms, nonanxiety comorbidity, a principal disorder of SOP, and parent internalizing symptoms would be associated with poorer treatment outcome, whereas higher child motivation would be associated with better outcome. Furthermore, we expected singleparent family status, low family social class, and elevated levels of family stresses to be associated with poorer treatment outcomes. The predictors were examined at post-treatment and at 1-year follow-up to test for stability of the observed relationships with outcome. 1. Methods This study was part of a randomized waitlist-controlled effectiveness trial of CBT for children with anxiety disorders. The main aims of the trial were to investigate the effectiveness of CBT in community clinics and to compare the relative effectiveness of individual CBT (ICBT) and group CBT (GCBT) for children with anxiety disorders. The procedures for recruitment of participants, training of therapists and assessors, and randomization have been reported elsewhere (Wergeland et al., 2014) and, therefore, are not presented in detail here. The study was approved by the Regional Committee for Medical and Health Research Ethics of Western Norway. The trial is registered at www.clinicaltrials.gov

G.J.H. Wergeland et al. / Behaviour Research and Therapy 76 (2016) 1e12

(NCT00586586). 1.1. Participants The trial initially enrolled 182 children (ages 8e15 years) recruited from referrals to seven public mental health outpatient clinics from 2008 to 2010. Informed written consent from all parents and assent from children above the age of 12 years were obtained. Parents provided consent for patients under 12 years of age. Inclusion criteria were a principal disorder of SAD, SOP, or GAD according to DSM-IV (American Psychiatric Association, 1994). Exclusion criteria included pervasive developmental disorder, psychotic disorder, or mental retardation. Youth on psychotropic medication (a total of 6.0% for Attention-Deficit/Hyperactivity Disorder (AD/HD), anxiety, or depression) were included if the dosage had been stable for the last 3 months before study entry and kept constant during the treatment period. Participants were assigned randomly in blocks of six children at each clinic from the younger (age 8e12 years) or the older agegroup (age 12e15 years) to either ICBT (n ¼ 77), GCBT (n ¼ 67), or a 10-week waitlist (n ¼ 38). Children on the waitlist were randomized subsequently to ICBT or GCBT, with the exception of one child who improved and no longer met inclusion criteria. This resulted in a final sample of 181 children (mean age 11.5 years, SD 2.1, 47.0% boys). Twenty-six participants (14.4%) did not complete the treatment program, including two children who dropped out before the start of the treatment. Thus, 88 children participated in the group treatment, and 91 children were treated individually. Most participants were Caucasian (90.7%), 1.6% were Asian, whereas ethnicity was not reported for 7.7%. The majority (57.5%) lived in two-parent households; 19.9% lived with a single parent; 13.3% lived with one biological parent and one step-parent; and 1.6% lived in foster families. Family composition was not reported for 7.7% of the participants. The ranking of family social class was based on parent occupation in accordance with the Registrar General Social Class coding scheme (Currie et al., 2008) and the family social class was defined by the highest ranking parent. Family social class was high for 30.4%, medium for 51.4%, low for 7.7%, and not reported for 10.5%. This pattern of ethnicity, family composition, and social class matches closely with what has been reported for other Norwegian community clinic sample groups (Brøndbo et al., 2011; Nilsen, Handegard, Eisemann, & Kvernmo, 2015). The principal diagnoses of the children were SOP (n ¼ 84, 46.4%), SAD (n ¼ 59, 32.6%), and GAD (n ¼ 38, 21.0%). A total of 141 (77.9%) had at least one comorbid disorder; 125 (69.1%) met criteria for more than one inclusion anxiety disorder (i.e. SAD, SOP, GAD); 25 (13.8%) had other specified anxiety disorders, mainly specific phobia; and 21 (11.6%) had comorbid depression. Sixteen children (8.8%) had comorbid externalizing disorders (oppositional defiant disorder and/or ADHD), and 12 (6.6%) had comorbid tic disorders. 1.2. Measures 1.2.1. Diagnostic interviews for study entry Anxiety Disorder Interview Schedule for DSM-IV, child and parent versions (ADIS-C/P; Silverman & Albano, 1996); SAD, SOP, and GAD modules were used to assess inclusion diagnoses. Children and parents were interviewed separately, and diagnoses and clinician severity ratings (CSR) were assigned based on the combined parent and child report. When multiple inclusion diagnoses were present, the diagnosis causing the highest interference was considered principal. A CSR of four or above (on a 0e8 scale) was required for inclusion. The ADIS-C/P has demonstrated excellent inter-rater reliability, retest reliability, and concurrent validity

3

(Silverman, Saavedra, & Pina, 2001; Wood, Piacentini, Bergman, McCracken, & Barrios, 2002). In the current study, a rescoring of 20% of the interviews by blinded evaluators gave k values for the presence of an inclusion anxiety diagnosis of 0.84 (ADIS-C) and 0.86 (ADIS-P). For CSR, the intraclass correlation (ICC) was 0.82 (ADIS-C and ADIS-P). Development and Well-Being Assessment (DAWBA; Goodman, Ford, Richards, Gatward, & Meltzer, 2000) was completed as part of the routine intake procedure at the participating clinics. The DAWBA is a web-based diagnostic interview, combining structured questions on symptoms and impairment with open-ended questions for detailed accounts of symptoms and impairment. Parents and children from the age of 11 years completed the interview online. Parent and child DAWBA information was used to assess comorbid disorders, other than SAD, SOP, and GAD, and demographic data. Parents provided information on ethnicity, family composition, parental occupation, and family stresses (nine items covering socioeconomic and housing stresses, work pressure, and physical and mental health stresses). Each of the family stress items is rated on a three-point scale (0 ¼ none, 2 ¼ major), yielding a maximum score of 18. The DAWBA has been found to discriminate well between community and clinic samples of youth (Goodman et al., 2000) and to give realistic estimates of prevalence of psychiatric illnesses (Heiervang et al., 2007). Good to excellent reliability between rating clinicians has been reported (Ford, Goodman, & Meltzer, 2003; Heiervang et al., 2007). In the current study, the first 50 interviews were rated jointly by the first and last authors who were psychiatric residents and consultant child psychiatrists, respectively. Inter-rater reliability was based on independent ratings of the next 44 interviews, and agreement on the presence/absence of a disorder gave k values ranging from satisfactory to excellent (k ¼ 0.58 for other anxiety, 0.66 for ADHD, 0.72 for specific phobia, 0.77 for depression, and 1.00 for oppositional defiant and tic disorders). Subsequently, the last 87 interviews were rated by the first author only. For the current study, nonanxiety comorbidity was defined as the co-occurrence of one of or more than the three inclusion anxiety disorders and one or more nonanxiety disorders (Angold, Costello, & Erkanli, 1999). Nonanxiety comorbidity included mood, externalizing, and tic disorders as reported by the DAWBA. Family social class was categorized as high, medium, or low but, for the current analyses, dichotomized into low versus medium and high. 1.2.2. Internalizing problems (anxiety and depression symptoms) Spence Children's Anxiety Scale (SCAS; Spence, 1998), child and parent versions (SCAS-c, SCAS-p), were used to assess child anxiety symptoms. The SCAS comprises 38 items, rated on a fourpoint scale (from 0 ¼ never to 3 ¼ always), yielding a maximal score of 114. Internal consistency for the SCAS in the current sample was good to excellent (a: parent ¼ 0.85, child ¼ 0.91). Short Mood and Feelings Questionnaire (SMFQ; Angold, Costello, Messer, & Pickles, 1995), child and parent versions, were used to assess child's depressive symptoms. The SMFQ comprises 13 items rated on a three-point scale (from 0 ¼ not true to 2 ¼ true), yielding a maximum score of 26. In the current sample, internal consistency was good (a: parent ¼ 0.86, child ¼ 0.88). Depression Anxiety Stress Scale (DASS; Lovibond & Lovibond, 1995) was used to assess parental self-rated depression, anxiety, and stress. The DASS comprises 42 items; each rated on a four-point scale (from 0 ¼ hardly ever to 3 ¼ almost always). Internal consistency of the DASS in the current sample was excellent (a ¼ 0.95). 1.2.3. Externalizing problems and functional impairment Strengths and Difficulties Questionnaire (SDQ, Goodman, 1997), parent version, was used to assess the child's externalizing

4

G.J.H. Wergeland et al. / Behaviour Research and Therapy 76 (2016) 1e12

problems and functional impairment. The SDQ comprises 25 items divided into five subscales of five items each: Emotional symptoms, Conduct problems, Hyperactivity-inattention, Peer problems, and Prosocial behavior (Goodman, 1997; 1999). Items are scored on a three-point scale (from 0 ¼ not true to 2 ¼ true), yielding a maximal score of 10 for the subscales. The extended version of SDQ also includes an Impact scale, covering severity of difficulties, overall distress to the child, and the impairment from symptoms in the child's daily life (Goodman, 1999). The parent version of the Impact scale comprises five items, giving a possible score range of 0e10. The psychometric properties of the SDQ are strong (Stone, Otten, Engels, Vermulst, & Janssens, 2010). In the current study, the Conduct problems and Hyperactivity-inattention subscales were combined into an externalizing problem scale (Goodman, Lamping, & Ploubidis, 2010). Internal consistency for the externalizing problem (SDQ-e) scale and the impact (SDQ-i) scale was good (SDQ-e a: parent ¼ 0.77; SDQ-i a: parent ¼ 0.80). 1.2.4. Motivation Nijmegen Motivation List (NML; Keijsers et al., 1999) was developed to assess treatment motivation in adults and has later been modified for use with children (Ollendick et al., 2009). The NML-child version (NML-c) consists of 15 items, scored on a threepoint scale (from 0 ¼ not at all true to 2 ¼ true). The NML-c has demonstrated acceptable internal consistency (a ¼ 0.73), with children aged 7e16 years (Ollendick et al., 2009). Internal consistency of the NML-c in the current sample was good (a ¼ 0.86). 1.3. Procedures Assessments of the child and parent at baseline, after waitlist, at post-treatment, and at one-year follow-up were conducted separately (one parent per child, 92.0% mothers). Children and parents were interviewed with the ADIS-C/P (Silverman & Albano, 1996) and completed symptom measures (see Measures). At pretreatment, children and parents at all sites were assessed with the routine assessment procedure, including the DAWBA (Goodman et al., 2000). Parents and children received no fees for participation and were informed that they would be offered standard treatment in the clinic if they declined participation or withdrew from the study.

parent sessions in GCBT were held with all parents in the treatment group present. During the parent sessions, program content was explained in detail. Up to three missed appointments were rescheduled as individual sessions in both conditions, whereas children absent from more than three sessions were considered dropouts. Twenty-six participants (14.4%) did not complete the treatment protocol and were considered dropouts (Wergeland et al., 2015). 1.5. Study setting, therapists, and assessors The study was conducted in seven public child and adolescent mental health outpatient clinics, covering both rural and urban areas in western Norway. Each clinic covered a given catchment area. The clinics were not specialized with regard to therapy orientation or specific diagnoses, and the staff represented different professional backgrounds (e.g. clinical psychologists, clinical pedagogues, clinical social workers). Children were referred to the clinics by their general practitioners or less often by child welfare services. The clinics are part of the public health system, and services are free of charge. There is only minimal use of private mental health care for youth in Norway. All therapists were regular employees at the participating clinics, who volunteered for the study and conducted the treatments as part of their ordinary workload. Seventeen therapists (mean age ¼ 48.2 years, SD 11.0, range 30e63, 94.1% females) conducted the treatment. Therapists had on average 10.8 (SD 6.3) years of clinical experience. Five of the therapists had completed a formal two-year postgraduate CBT training, whereas the other had little or no formal training in CBT before the study. All therapists attended a 2-day workshop on CBT and childhood anxiety disorders and a 2-day workshop on the Friends program. They treated two pilot cases approved by the supervisors before the start of the study and received supervision by one of two experienced CBT therapists and licensed Friends trainers throughout the treatment phase of the study. Assessors (N ¼ 16) were experienced clinicians also employed at the participating clinics. They also attended workshops focusing on CBT and anxiety disorders, received training for the ADIS-C/P in a 2day workshop with licensed ADIS-C/P raters, and met regularly with the first and last authors to discuss interview administration and scoring.

1.4. Treatment 1.6. Ratings of adherence and competence The treatment was the Friends for Life Program, 4th ed. (Barrett, 2004), which is a 10-week manual-based CBT program addressing cognitive, physiological, and behavioral components that have been found to contribute to development or maintenance of anxiety. The program is skill-based (relaxation, identifying, and challenging anxious thoughts, problem-solving skill training, social support training), and exposure exercises are negotiated as specific weekly home tasks from midtreatment onward. Children are instructed to follow individual exposure plans, and rewards are planned to aid the exposure. The same manual was used for ICBT and GCBT, and the therapists were instructed to complete the same agenda and session tasks in both formats. There are two versions of the program that are adjusted for developmental level. The child version of the program was used with 8e12-year-olds (n ¼ 120), and the adolescent version was used for 12e15-year-olds (n ¼ 61). The 12year-old adolescents could be included in either age-group based on clinician evaluation of maturity level. Parents attended 2 of the 10 child sessions and the last 15 min of the remaining eight sessions. There were also two separate parent sessions before sessions 1 and 6. Parents in ICBT had individual parent sessions, whereas

All therapy sessions were videotaped; 20% was randomly selected for adherence and competence rating using an 11-item scale developed for the study (Bjaastad et al., 2015). Adherence was rated from 0 (“none”) to 6 (“thorough”), and competence was rated from 0 (“poor skills”) to 6 (“excellent skills”) for each item. The measure assessed CBT structure (agenda setting, homework task, progress, and structure, as well as parental involvement); process and relational skills (positive reinforcement, cooperation, and flexibility); and specific goals of each session. The selection of sessions to be rated was stratified on early (2e5) and late (6e9) sessions, and all therapists were represented. The ratings were made by two experienced CBT and Friends therapists and two psychology graduate students. Using the criteria recommended by Cicchetti (1994), inter-rater reliability was excellent (ICC ¼ 0.83) for adherence and good (ICC ¼ 0.64) for competence. The mean score across treatments for the individual therapists ranged from 3.83 to 5.43 (mean ¼ 4.57, SD ¼ 0.94) for adherence and 3.44 to 5.25 (mean ¼ 4.30, SD ¼ 0.91) for competence. The range across all therapy sessions was 0.43e6.00 for adherence (mean of seven items) and 1.00 to 6.00 (mean of four items)

G.J.H. Wergeland et al. / Behaviour Research and Therapy 76 (2016) 1e12

regarding competence. All 17 therapists obtained a mean score above 3.0 for both adherence and competence, which was set as the minimum criterion for adequate therapist adherence and competence. 1.7. Data analysis Results are based on data from the sample of 181 children included in the study, because no significant differences in outcomes (diagnostic recovery or clinically significant symptom change at post-treatment or at 1-year follow-up) were observed between ICBT and GCBT (Wergeland et al., 2014). This finding is in line with a recent review indicating comparable effect for ICBT and GCBT for children with anxiety disorders (James et al., 2013). Therefore, data from ICBT and GCBT were combined for the current predictor analyses. In addition, separate analyses were performed on the sample of children who completed the treatment (n ¼ 155). Treatment outcome was defined as diagnostic remission and symptom improvement, measured at post-treatment and at 1-year follow-up. Diagnostic remission was defined as either a) remission of all inclusion anxiety disorders or b) remission of the principal anxiety disorder, according to ADIS-C/P. Analyses were conducted separately for these two remission criteria. Treatment response was defined as improvement in anxiety symptoms, as assessed by the SCAS-c/p. Prior to the analyses, we estimated ICC for the outcome variables post treatment and at follow-up to determine the percent of variance both at the between-site and at the between-group level. The between-site ICC ranged from 0.01 to 0.11, and the betweengroup ICC ranged from 0.12 to 0.22. Although this between group variance can be considered borderline and of less relevance (Guo, 2005), multilevel modeling was still applied in Mplus because the analyses of interest represent interactions between nested levels n & Muthe n, 2011). (Guo, 2005; Muthe The GCBT approach comprised a total of 16 separate treatment groups whereas children in the ICBT approach were grouped as one cluster at each of the seven clinics, giving a total of 23 clusters. The design was therefore partially clustered, and all models were thus adjusted for potential clustering effects (Baldwin, Bauer, Stice, & Rohde, 2011; Bauer, Sterba, & Hallfors, 2008). Missing data at the item and measure level were examined using the missing value analysis in SPSS 20 (IBM Statistics, Chicago, IL, USA). The proportion of missing values of the predictors did not exceed 8.4% for any measure across informants, with the exception of parent self-report on the DASS, where 11.6% of the values was missing. Missing outcome data originated mainly from treatment dropouts (14.4%) and from participants lost to follow-up (4.4%). As indicated by Little's missing completely at random test, missing data on the measure level occurred completely at random. Missing data were accommodated using full information maximum likelihood (FIML) missing data methodology (Wothke, 2000) in Mplus. Thus, a missing data point did not result in deletion of the participant. Analyses were run with and without outliers present. Inclusion of the outliers did not alter results; therefore, they were included in all of the analyses. Nonnormality was evident in several variables. To account for the nonnormality present in the data, analyses were pursued in Mplus with the MLR estimator shown to be robust to violations of normality based on the Huber-White algorithm n & Muthe n, 2011). (Muthe To investigate potential independent baseline predictors of diagnostic recovery, variables were grouped into demographic predictors, child and parent predictors, including symptom measures and comorbidity, principal anxiety disorder, and motivation. The predictors were entered in four groups, one at a time, into

5

logistic regression analyses separately post-treatment and at 1-year follow-up. Demographic variables were entered in first, followed by child- and parent-rated symptoms and comorbidity. Type of principal anxiety disorder, using GAD as the reference, was entered in the third group. Lastly, motivation was entered as group four. A similar approach was used to investigate potential predictors of treatment response in rated anxiety symptoms. Latent growth curve modeling (LGM), a structural equation modeling technique, was used separately on child- and parent-rated anxiety to model treatment response in anxiety symptoms over time, both at the group and individual level, accommodating individually varying times of observation and with random intercepts and slope values. In a latent growth curve model, the results indicate the effect of the independent variable on the slope or rate of symptom change during the treatment and follow-up periods. At post-treatment, with two measurement points, this is considered a difference score model (Duncan, Duncan, & Strycker, 2006). In the follow-up model, three measurement points were included. The slope values were regressed onto the potential predictors to assess their influence on symptom change. Model fit was evaluated using the log likelihood value, Akaike's information criteria (AIC), and the Bayesian information criteria (BIC). To control for multiple testing (four analyses), predictors were considered significant at a Bonferroni corrected ɑ-level of 0.0125.

2. Results At post-treatment, 22.7% of the children was free from all inclusion anxiety disorders, and 35.2% was free from their principal anxiety disorder. At 1-year follow-up, the corresponding numbers were 36.5 and 45.9%. The observed mean change in child-rated anxiety symptoms at post-treatment was 8.51 (SD 14.40) and at 1-year follow-up, it was 11.89 (SD 14.39). For parent-rated child anxiety symptoms, the corresponding means were 8.16 (SD 12.08) and 11.19 (SD 13.22), respectively. Child- and parent-rated child anxiety was significantly, but small to largely correlated at all assessment points, with r ¼ 0.26 at pretreatment, r ¼ 0.43 at posttreatment, and r ¼ 0.54 at follow-up. Consequently, child- and parent-rated child anxiety symptoms were treated as separate outcome variables. Demographic factors, baseline means, and proportions of potential predictors are presented in Table 1. Children who did not experience a complete remission had significantly higher symptom impact (SDQ-i) at baseline, and their parents had significantly higher self-rated internalizing symptoms (DASS). These differences were significant both at post-treatment and at 1-year follow-up.

2.1. Remission from all anxiety disorders All potential predictors were entered into a logistic regression model, with remission from all-inclusion anxiety disorders as outcome at post-treatment (Table 2), as described in the Methods section. In this model, a principal diagnosis of SAD, but not SOP, significantly reduced the relative odds for diagnostic recovery compared with GAD. Also, higher parent self-rated internalizing symptoms significantly reduced the odds for diagnostic recovery. The full model explained 46% of the outcome variance (p < 0.05). For outcome at 1-year follow-up, higher parent-self-rated internalizing symptoms at baseline significantly reduced the odds for diagnostic recovery, whereas higher child baseline motivation increased the odds for diagnostic recovery. The proportion of the outcome variance explained by the full model was 30% (p < 0.05).

6

G.J.H. Wergeland et al. / Behaviour Research and Therapy 76 (2016) 1e12

Table 1 Demographic, symptom and diagnostic variables as predictors of remission from all inclusion anxiety disorders. Predictor variable

Post-treatment

One-year follow-up

Recovered % Age Gender Male Female Single parent status FSC High Medium Low Family stresses SCAS-c SMFQ-c NML-c SCAS-p SMFQ-p SDQ Externalizing problem scale Impact scale Non-anxiety comorbidity DASS

Not recovered M

SD

11.44

%

Recovered

M

2.15

SD

11.50

%

Not recovered M

2.04

SD

11.02

%

M

2.08

51.2 48.8 17.1

46.0 54.0 21.2

50.0 50.0 15.2

45.0 55.0 25.0

37.5 55.0 7.5 75.0

34.0 56.6 9.4 82.0

34.9 57.1 7.9 81.5

35.5 53.9 10.5 79.5

2.10 33.18 7.68 19.73 32.34 6.24

2.09 14.21 5.65 5.04 11.66 4.22

5.88 3.34

3.36 2.67*

6.46

5.62*

29.3

2.55 37.63 7.20 20.06 35.84 8.05

2.64 17.73 5.13 5.97 12.48 5.24

6.16 4.55

3.65 2.86

12.67

14.62

38.1

2.57 33.02 7.09 20.09 32.53 6.60

2.46 14.04 5.31 5.10 8.21 5.01

5.70 3.61

3.46 2.60*

8.64

0.99*

31.8

SD

11.81

2.06

2.42 38.44 7.69 19.71 36.36 8.05

2.88 17.94 5.24 6.39 13.56 4.76

6.25 4.51

3.53 2.82

12.96

15.37

36.2

Note. Observed proportions, means, and standard deviation of child- and parent reported predictors of treatment success (i.e., recovery from all inclusion anxiety disorders). Column proportions and means are compared using logistic or linear univariate regression analysis. FSC ¼ Family social class; SCAS ¼ Spence Child Anxiety Scale; SMFQ ¼ Short Mood and Feelings Questionnaire; NML ¼ Nijmegen Motivation List; c ¼ child; p ¼ parent; SDQ ¼ Strengths and Difficulties Questionnaire; DASS ¼ Depression, Anxiety and Stress Scales. * Significant difference after Bonferroni correction.

Table 2 Predictors of remission from all inclusion anxiety disorders. Predictor variable

Demographics Age Gender Single parent status FSC Symptom measures and comorbidity FS SCAS-c SMFQ-c SCAS-p SMFQ-p SDQ-e SDQ-i DASS Non-anxiety CoDa Principal anxiety disorder Principal GADb Principal SAD Principal SOP Motivation NML-c

Post-treatment

One-year follow-up

OR

95% CI for OR

p

1.04 0.64 0.98 0.76

[0.82, [0.34, [0.27, [0.18,

1.30] 1.21] 3.57] 3.18]

0.770 0.170 0.970 0.702

0.98 0.95 1.12 1.02 0.89 1.10 0.82 0.93 0.49

[0.85, [0.91, [1.02, [0.97, [0.76, [0.95, [0.68, [0.88, [0.19,

1.13] 0.99] 1.23] 1.08] 1.05] 1.28] 0.99] 0.98] 1.31]

0.772 0.042 0.014 0.454 0.163 0.221 0.036 0.004 0.154

1 0.17 0.37

e [0.05, 0.52] [0.11, 1.32]

e 0.002 0.126

1.06

[0.95, 1.17]

0.315

R2

DR2

0.008

0.287

0.427

0.460

p

OR

95% CI for OR

p

0.83 0.76 0.70 0.88

[0.67, [0.30, [0.26, [0.24,

1.03] 1.94] 1.92] 3.29]

0.090 0.571 0.488 0.854

1.10 0.96 1.03 0.99 0.98 0.98 0.90 0.95 0.87

[0.95, [0.93, [0.95, [0.95, [0.88, [0.83, [0.79, [0.92, [0.45,

1.27] 1.00] 1.12] 1.03] 1.09] 1.14] 1.03] 0.98] 1.67]

0.218 0.047 0.447 0.616 0.700 0.750 0.123 0.005 0.675

1 0.40 0.16

e [0.12, 1.28] [0.15, 1.77]

e 0.123 0.293

1.10

[1.03, 1.17]

0.005

0.520

0.279

0.140

0.033

R2

DR2

0.066

0.015

0.004

0.002

p 0.169

0.209

0.143

0.035

0.235

0.026

0.030

0.299

0.064

0.009

Note. FSC ¼ Family Social Class; FS ¼ Family Stresses; SCAS ¼ Spence Child Anxiety Scale; SMFQ ¼ Short Mood and Feelings Questionnaire; SDQ ¼ Strengths and Difficulties Questionnaire (e ¼ externalizing problems; i ¼ impact); DASS ¼ Depression, Anxiety and Stress Scales; CoD ¼ Comorbid disorders. GAD ¼ Generalized Anxiety Disorder; SAD ¼ Separation Anxiety Disorder; SOP ¼ Social Phobia; NML ¼ Nijmegen Motivation List; c ¼ child; p ¼ parent. a Based on combined information from children and parents for participants 11 years or older, otherwise on parents' reports only. b Reference category.

2.2. Remission from principal anxiety disorder When remission from the principal anxiety disorder was used as outcome in the logistic regression analysis, results indicated that, at posttreatment, both SAD and SOP significantly reduced the odds for recovery compared with GAD (Table 3). The full model explained 27% of the outcome variance (p < 0.05). At 1-year follow-up, higher baseline child-rated anxiety symptoms reduced, whereas

motivation increased the odds for remission. The proportion of the outcome variance explained by the full model was 32% (p < 0.001).

2.3. Child reported anxiety symptoms Using anxiety symptom as outcome, all potential predictors were entered into a LGM model separately for child and parent ratings. Child-rated symptom change at post-treatment was

G.J.H. Wergeland et al. / Behaviour Research and Therapy 76 (2016) 1e12

7

Table 3 Predictors of remission from primary inclusion anxiety disorder. Predictor variable

Post-treatment

Demographics Age Gender Single parent status FSC Symptom measures and comorbidity FS SCAS-c SMFQ-c SCAS-p SMFQ-p SDQ-e SDQ-i DASS Non-anxiety CoDa Principal anxiety disorder Principal GADb Principal SAD Principal SOP Motivation NML-c

One-year follow-up

OR

95% CI for OR

p

1.03 0.57 1.53 0.81

[0.87, [0.27, [0.55, [0.24,

1.12] 1.20] 4.30] 2.74]

0.764 0.139 0.417 0.731

0.88 0.98 1.09 1.02 0.91 1.00 0.91 0.99 0.70

[0.78, [0.95, [1.01, [0.98, [0.80, [0.90, [0.80, [0.95, [0.29,

0.98] 1.01] 1.17] 1.07] 1.03] 1.11] 1.02] 1.03] 1.71]

0.023 0.196 0.025 0.275 0.138 0.970 0.092 0.569 0.433

1 0.14 0.16

e [0.04, 0.54] [0.05, 0.51]

e 0.004 0.002

1.00

[0.92, 1.09]

0.935

DR2

R2

p

0.019

0.126

0.260

0.271

OR

95% CI for OR

p

0.83 1.01 0.60 0.39

[0.70, [0.44, [0.28, [0.10,

0.97] 2.31] 1.28] 1.51]

0.017 0.977 0.186 0.170

1.08 0.95 1.08 0.99 0.95 1.03 0.90 0.98 1.51

[0.94, [0.92, [0.98, [0.96, [0.84, [0.87, [0.78, [0.95, [0.48,

1.23] 0.99] 1.19] 1.02] 1.03] 1.21] 1.03] 1.01] 4.75]

0.281 0.007 0.104 0.541 0.213 0.774 0.127 0.247 0.482

1 0.43 0.39

e [0.10, 1.86] [0.09, 1.63]

e 0.258 0.197

1.12

[1.04, 1.21]

0.004

0.474

0.107

0.134

0.011

R2

DR2

0.088

0.066

0.023

0.021

p 0.039

0.216

0.128

0.005

0.237

0.021

0.008

0.317

0.080

0.000

Note. FSC ¼ Family Social Class; FS ¼ Family Stresses; SCAS ¼ Spence Child Anxiety Scale; SMFQ ¼ Short Mood and Feelings Questionnaire; SDQ ¼ Strengths and Difficulties Questionnaire (e ¼ externalizing problems; i ¼ impact); DASS ¼ Depression, Anxiety and Stress Scales; CoD ¼ Comorbid disorders. GAD ¼ Generalized Anxiety Disorder; SAD ¼ Separation Anxiety Disorder; SOP ¼ Social Phobia; Disorder; NML ¼ Nijmegen Motivation List; c ¼ child; p ¼ parent. a Based on combined information from children and parents for participants 11 years or older, otherwise on parents' reports only. b Reference category.

Table 4 Predictors of child rated anxiety symptom change. Post-treatment

Demograpics Age Gender Single parent status FSC Symptom measures FS SCAS-c SMFQ-c SCAS-p SMFQ-p SDQ-e SDQ-i DASS-p Non-anxiety CoDb Principal anxiety disoder GAD SAD SOP Motivation NML-c

One-year follow-up

ba

SE (b)

p

0.028 0.101 0.080 0.245

0.018 0.051 0.051 0.162

0.109 0.046 0.112 0.130

0.003 0.016 0.001 0.001 0.014 0.009 0.019 0.008 0.136

0.009 0.004 0.007 0.003 0.007 0.014 0.012 0.002 0.059

0.734 0.001 0.861 0.779 0.036 0.513 0.126 0.001 0.021

c

0.077 0.134

e 0.152 0.080

e 0.614 0.095

0.007

0.006

0.220

LL, df 1902, 27

5546, 119

5722, 152

6197, 170

AIC 3858

11,331

11,748

12,734

BIC

ba

SE (b)

p

0.007 0.030 0.047 0.184

0.010 0.031 0.027 0.072

0.480 0.333 0.833 0.011

0.002 0.001 0.010 0.001 0.000 0.002 0.009 0.002 0.039

0.010 0.005 0.006 0.002 0.004 0.004 0.006 0.001 0.029

0.880 0.845 0.069 0.779 0.927 0.604 0.102 0.302 0.175

c

0.025 0.004

e 0.043 0.040

e 0.555 0.914

0.008

0.004

0.058

LL, df 2419, 30

3945

11,711

12,234

13,277

AIC

BIC

4899

4994

6060, 122

12,366

12,755

6237, 155

12,784

13,279

6714, 173

13,774

14,327

Note. FSC ¼ Family Social Class; FS ¼ Family Stresses; SCAS ¼ Spence Child Anxiety Scale; SMFQ ¼ Short Mood and Feelings Questionnaire; SDQ ¼ Strengths and Difficulties Questionnaire (e ¼ externalizing problems; i ¼ impact); DASS ¼ Depression, Anxiety and Stress Scales; CoD ¼ Comorbid disorders. GAD ¼ Generalized Anxiety Disorder; SAD ¼ Separation Anxiety Disorder; SOP ¼ Social Phobia; NML ¼ Nijmegen Motivation List; c ¼ child; p ¼ parent. a b values indicate the effect of the predictor variable on weekly anxiety symptom level change. b Based on combined information from children and parents' for participants 11 years or older, otherwise on parents reports only. c Reference category.

significantly predicted by baseline child-rated anxiety symptom level and parent internalizing symptoms (Table 4). Higher baseline child anxiety levels were associated with greater symptom change, whereas higher parent internalizing symptoms were associated with less symptom change. At 1-year follow-up, the only significant predictor of child-rated symptom change was family social class, where lower family social class was associated with less symptom

change.

2.4. Parent reported anxiety symptoms Parent-rated symptoms at post-treatment (Table 5) were predicted by baseline parent-rated child anxiety symptoms and impact. Higher baseline anxiety levels were associated with greater

8

G.J.H. Wergeland et al. / Behaviour Research and Therapy 76 (2016) 1e12

Table 5 Predictors of parent rated symptom change. Post-treatment

Demograpics Age Gender Single parent status FSC Symptom measures FS SCAS-c SMFQ-c SCAS-p SMFQ-p SDQ-e SDQ-i DASS-p Non-anxiety CoDb Principal anxiety disoder GAD SAD SOP Motivation NML-c

One-year follow-up

ba

SE (b)

p

0.014 0.042 0.101 0.017

0.017 0.050 0.066 0.065

0.041 0.396 0.126 0.794

0.032 0.004 0.008 0.018 0.016 0.002 0.052 0.005 0.053

0.017 0.002 0.005 0.004 0.009 0.009 0.014 0.002 0.065

0.054 0.034 0.139 0.001 0.077 0.816 0.001 0.042 0.418

c

0.023 0.096

e 0.100 0.080

e 0.819 0.232

0.008

0.006

0.187

LL, df 1855, 27

5515, 119

5691, 152

6165, 170

AIC 3764

11,269

11,686

12,671

BIC

ba

SE (b)

p

0.011 0.032 0.010 0.027

0.010 0.033 0.052 0.093

0.273 0.332 0.855 0.771

0.006 0.002 0.006 0.007 0.008 0.002 0.011 0.000 0.065

0.011 0.003 0.008 0.012 0.004 0.006 0.014 0.003 0.055

0.598 0.595 0.441 0.547 0.044 0.694 0.435 0.936 0.232

c

0.032 0.000

e 0.049 0.044

e 0.511 0.992

0.001

0.004

0.831

LL, df 2351, 30

3851

11,649

12,171

13,214

AIC

BIC

4763

4858

6001, 122

12,247

12,635

6175, 155

12,661

13,155

6648, 173

13,643

14,194

Note. FSC ¼ Family Social Class; FS ¼ Family Stresses; SCAS ¼ Spence Child Anxiety Scale; SMFQ ¼ Short Mood and Feelings Questionnaire; SDQ ¼ Strengths and Difficulties Questionnaire (e ¼ externalizing problems; i ¼ impact); DASS ¼ Depression, Anxiety and Stress Scales; CoD ¼ Comorbid disorders. GAD ¼ Generalized Anxiety Disorder; SAD ¼ Separation Anxiety Disorder; SOP ¼ Social Phobia; NML ¼ Nijmegen Motivation List; c ¼ child; p ¼ parent. a b values indicate the effect of the predictor variable on weekly anxiety symptom level change. b Based on combined information from children and parents' for participants 11 years or older, otherwise on parents reports only. c Reference category.

symptom change, while higher impact was associated with less symptom change. At 1-year follow-up, no significant predictors of parent-rated anxiety symptoms were identified. Although both child- and parent-rated baseline anxiety symptom levels were significant predictors of anxiety symptom improvement at post-treatment (i.e. higher baseline anxiety level predicted larger change), children with high baseline anxiety symptoms levels still had higher anxiety symptoms at posttreatment compared to children with low baseline symptom levels (SCAS-c: b ¼ 0.48, 95% CI [0.27, 0.69], p < 0.001; SCAS-p: b ¼ 0.49, 95% CI [0.28, 0.70], p < 0.001), and at follow-up (SCASc: b ¼ 0.50, 95% CI [0.29, 0.70], p < 0.001; SCAS-p: b ¼ 0.36, 95% CI [0.21, 0.50], p < 0.001).

2.5. Supplementary analyses To evaluate the impact of each of the inclusion anxiety disorders on diagnostic outcome, we repeated our analyses by replacing the inclusion anxiety disorder with dummy-coded variables that indicated whether or not a disorder of SAD, SOP, or GAD were present, regardless of their position in the ADIS-C/P diagnostic profile. In this analysis, the presence of SOP anywhere in the diagnostic profile (OR 0.35; 95% CI [0.15, 0.80], p ¼ 0.013), not SAD (OR 0.81; 95% CI [0.29, 2.26], p ¼ 0.69), reduced the odds for remission from the primary inclusion anxiety disorders. Likewise, the presence of SOP anywhere in the diagnostic profile reduced the odds for remission from all-inclusion anxiety disorders (SOP: OR 0.40; 95% CI [0.16, 0.99], p ¼ 0.048; SAD: OR 0.38; 95% CI [0.11, 1.25], p ¼ 0.11). All predictor analyses of diagnostic recovery and child- and parent-rated anxiety symptoms were repeated on the sample group of treatment completers (n ¼ 155). In these analyses, the significant predictors as found in the intention-to-treat sample group were replicated, with the exception of the analysis of childrated anxiety symptom improvement at post-treatment where no significant predictors were found (data not shown).

3. Discussion The present study examined potential predictors of CBT outcome in children with anxiety disorders who were treated in community clinics with a manualized CBT program. Treatment outcomes were evaluated across different methods focusing on different aspects of outcome. The most consistent findings were that higher child anxiety levels, higher functional impairment, higher levels of parent internalizing symptoms at baseline, and primary diagnoses of social phobia or separation anxiety were associated with a less favorable outcome at post-treatment. Higher child anxiety level, higher levels of parent internalizing symptoms, lower family social class, and lower pretreatment child motivation at baseline were associated with less favorable outcome at 1-year follow-up. The finding that higher baseline anxiety levels predicted greater post-treatment anxiety symptom change is in line with previous studies (Kerns, Read, Klugman, & Kendall, 2013; Kley, Heinrichs, Bender, & Tuschen-Caffier, 2012; Liber et al., 2010) and may be attributed to the greater room for improvement for those with higher baseline anxiety levels, and/or the interdependence of the predictor and outcome variable. Our results indicate that higher baseline anxiety symptom levels also were associated with reduced odds for recovery from the primary anxiety disorder and higher anxiety symptom levels at 1-year follow-up. Thus, although children with higher baseline symptom severity benefit from CBT, they end treatment with a higher symptomatic level. Larger impairment was associated with less post-treatment anxiety symptom change as rated by the parent. Overall, our finding that severity and impairment negatively influence treatment outcome is in line with recent reviews of predictors of outcome in child anxiety treatment (Compton et al., 2014; Knight et al., 2014; Lundkvist-Houndoumadi et al., 2014; Nilsen et al., 2013). These results indicate that children with more severe anxiety and impairment may need additional treatment in order to achieve sufficient symptom relief.

G.J.H. Wergeland et al. / Behaviour Research and Therapy 76 (2016) 1e12

Parent level of internalizing symptoms at baseline was related both to diagnostic recovery and to symptom improvement. Models explaining the development and maintenance of child anxiety disorders emphasize the influence of parent internalizing psychopathology (Beidel & Turner, 1997; Lieb et al., 2000; Messer & Beidel, 1994; Rapee, 1997; Rapee, Schniering, & Hudson, 2009), and parental depression and anxiety may influence the child's treatment response through genetic, family environment factors, and/or rearing behavior (McLeod, Wood, & Weisz, 2007). Furthermore, parental depression and anxiety may interfere with parents' ability to support the child during treatment. Thus, there could be several different mechanisms behind the poorer treatment outcome for these children. Associations between parent internalizing symptoms and treatment outcome have been found in several studies (for review see Knight et al., 2014; Lundkvist-Houndoumadi et al., 2014). Assessing and targeting parent internalizing psychopathology or focusing on parenting behavior related to the parent's own anxiety could be important additions to the treatment to enhance child treatment outcome from CBT (e.g., Lundkvist-Houndoumadi et al., 2014). Having a principal disorder of SAD or SOP was associated with poorer treatment outcome in terms of remission from the principal anxiety disorder. A principal disorder of SAD also reduced the probability of remission from all anxiety disorders. These findings cannot be explained by baseline anxiety symptoms, nonanxiety symptoms and comorbidity, or parent internalizing symptoms as the effect was evident after controlling for these variables. Our finding is in line with recent studies identifying a principal diagnosis of SOP as a predictor of poorer outcome (e.g. Compton et al., 2014; Kerns et al., 2013). However, other studies have not found inferior treatment outcome for children with SAD. Our results indicate that presence of SOP anywhere in the clinical profile is the main risk factor for poorer diagnostic recovery. It is possible that SOP requires more targeted treatment than provided in the current manual. A recent meta-analysis reported larger effects for disorderspecific compared with generic programs for anxiety disorders, particularly evident for SOP (Reynolds, Wilson, Austin, & Hooper, 2012). Future studies should examine whether the practice of applying a generic CBT manual across these anxiety disorders should be sustained or whether enhanced or more differentiated treatments are needed to provide optimal recovery. Treatment outcome was not predicted by demographic factors such as age and gender, consistent with a recent meta-analysis (Bennett et al., 2013) and two recent reviews (Knight et al., 2014; Lundkvist-Houndoumadi et al., 2014). Low family social class was related to poorer outcome for child-rated anxiety symptoms at follow-up. Socioeconomic status has not been associated with CBT treatment outcome in past university clinic studies of child anxiety disorders (Knight et al., 2014). Because the relationship between family social class and symptom change was evident for only one outcome measure at follow-up, this finding should be interpreted with caution and would need replication before firm conclusions can be made. There could also be other, unobserved mechanisms or factors not captured in our assessments, that contribute to the poorer outcome in youth from low social class families (GordonHollingsworth et al., 2015). Contrary to our hypothesis, single parent status and family stresses were not related to treatment outcome. However, our study population was relatively homogenous regarding demographic factors, which may have limited the potential to demonstrate such associations. Also, contrary to our hypotheses, nonanxiety comorbidity and externalizing symptoms did not predict outcome over and above all the other predictors included in the models. Past reviews have shown inconsistent effects regarding the effect of nonanxiety comorbidity on treatment outcome (Knight et al., 2014). However,

9

none of the studies have evaluated this in community clinics, in which higher rates of comorbid symptoms and disorders are to be expected (Southam-Gerow et al., 2008, 2003). Our non-significant findings may be explained by the lower frequency of nonanxiety comorbid disorders in our sample compared with other community clinic samples (Bodden et al., 2008; Southam-Gerow et al., 2010). Likewise, externalizing symptoms, i.e. conduct problems and hyperactivity-inattention, were not associated with treatment outcome. An explanation may be that 9 of the 16 children with externalizing disorders in our sample group were on stable medication. This may have reduced externalizing symptoms and thus their potential effect on outcome. Furthermore, child depressive symptoms were not associated with poorer outcome. Our finding may be related to the low level of depressive symptoms in our sample, and as such, the result may not hold up for children with higher levels of depressive symptoms. Higher child motivation was associated with diagnostic recovery at follow-up. The relationship between treatment motivation and outcome is in line with previous findings among youth with behavioral, emotional, and learning problems (Adelman, KaserBoyd, & Taylor, 1984), and from adults with anxiety disorders (Keijsers, Schaap, & Hoogduin, 2000). Motivation has been associated with higher levels of treatment adherence and engagement (Drieschner, Lammers, & van der Staak, 2004). Thus, a possible explanation for the finding is that more motivated children engage more actively in the treatment, practice the skills learned during treatment more and continue doing so also after the treatment has ended. Strengths of the present study include the use of a large sample of clinically referred children with anxiety disorders, assessment of diagnoses, and symptoms from both the child and the parent perspective, the inclusion of end point diagnoses as well as symptom change, and post and follow-up assessments. Still, the findings need to be interpreted in light of some limitations. First, our study was not primarily designed to evaluate predictors. The sample size, although relatively large, allowed only for a limited number of predictors to be evaluated. In addition, we focused on baseline predictors of treatment outcome for the total sample and not interactions between treatment format and predictors of outcome, as this would have required an even larger sample. Second, we were unable to examine ethnicity as a predictor of outcome because of the low rate of non-Caucasian participants enrolled. However, the small number of non-Caucasian participants is typical for most Norwegian community clinics (Nilsen et al., 2015). Third, parent internalizing symptoms were assessed with a self-report questionnaire, as opposed to a diagnostic interview. The use of self-report for parental internalizing symptoms may have led to under-reporting. The use of the total DASS score may also have been too global a measure of internalizing symptoms. However, the DASS subscales were highly correlated (r ¼ 0.65e0.69). To further capture the impact of parental psychopathology for child outcomes, a diagnostic interview may have provided more robust information. Furthermore, in our study, parent self-reported measures and parent reported child symptom measures were mainly reported by mothers. Future research should also include both maternal and paternal ratings, as studies have found paternal psychopathology to be as important as maternal psychopathology for child treatment outcome (Liber et al., 2008). Fourth, we used only the SAD, SOP, and GAD modules of the ADIS-C/P, whereas comorbidity was assessed by the DAWBA already used routinely at the clinics. This meant that differential diagnoses and comorbidity based on the full ADIS-C/P could not be assessed. The use of the ADIS-C/P modules was chosen to reduce the burden on the participants, because both DAWBA and ADIS-C/P are extensive and time-consuming interviews. Conducting a full diagnostic interview twice was considered too taxing

10

G.J.H. Wergeland et al. / Behaviour Research and Therapy 76 (2016) 1e12

on the families. Fifth, the low prevalence of mood and externalizing disorders in the sample reduces the power to detect a potential association with outcome. Examining nonanxiety comorbidity as a predictor of outcome in future community clinics studies will be important to determine whether adaptions or modifications of standard CBT programs are needed for children with nonanxiety comorbidity. Sixth, as we focused on pretreatment predictors, we did not consider process factors such as therapeutic alliance, therapist adherence or competence in relation to outcome. Future studies should examine predictors active during treatment. Finally, results can be generalized to children with a principal diagnosis of SAD, SOP, or GAD, but not necessarily to children with other anxiety disorders. 3.1. Comparability with other child anxiety community clinic studies To place our sample in the context of similar research, we compared the demographic and baseline clinical characteristics for our sample with other samples of children with anxiety disorders from community clinics. The number of children living in singleparent families and the ethnic background of the study participants were comparable with other European community clinic samples of children with anxiety disorders (Bodden et al., 2008; Nauta et al., 2003). The proportion of medium or high social class families was slightly higher compared with a study population of children with anxiety disorders in a Dutch community clinic (Nauta et al., 2003) but in line with another Norwegian community clinic sample of children with anxiety disorders (Villabø et al., 2013). The sample is also comparable with other community clinic samples with regard to level of anxiety symptoms (Barrington, Prior, Richardson, & Allen, 2005; Nauta et al., 2003), depressive symptoms (Nauta et al., 2003), and number of children with externalizing symptoms in the clinical range (Villabø et al., 2013). 3.2. Clinical implications Overall, our results indicate that children in community clinics presenting with severe anxiety or impairment, a principal disorder of SOP and SAD, and low motivation, have parents with elevated level of internalizing symptoms, or who come from lower social class families are at risk for poorer treatment outcome. Careful assessment is necessary to identify these children and families as they most likely will need enhanced or adapted treatment to achieve a more optimal treatment outcome. This suggests the need for a flexible application of treatment manuals while still maintaining fidelity to the evidence-based treatments (Kendall & Beidas, 2007). Adaptations of treatments may involve longer treatment duration, an increased number of sessions (Ishikawa, Okajima, Matsuoka, & Sakano, 2007), additional components added to standard programs (Compton et al., 2014), use of modular CBT (Chorpita, Daleiden, & Weisz, 2005; Weisz et al., 2012), or additional therapeutic interventions such as pharmacotherapy (Walkup et al., 2008). Adapted CBT programs may also be organized in a stepped-care model in which additional or higher intensity treatment is offered after a standard program when outcome monitoring indicates inadequate response (Bower & Gilbody, 2005). For the principal anxiety disorders SOP, SAD, and GAD, enhanced treatment better adjusted to the individual case formulation may improve recovery (Lundkvist-Houndoumadi, Thastum, & Hougaard, 2015). Efforts to increase child motivation early in treatment may also help to improve the long-term effects of CBT in children with anxiety disorders. Clinicians could measure motivation at treatment onset, and intervene to enhance motivation if low. Child motivation could be improved via psychoeducation and

treatment engagement strategies from motivational interviewing (Westra, Arkowitz, & Dozois, 2009; Westra & Dozois, 2006). Motivational interviewing (MI) aims to help clients to explore and resolve their ambivalent feelings towards treatment. MI has been found to be associated with improved CBT outcome in adult clients with anxiety disorders (Westra et al., 2009). Providing a program that targets parent internalizing psychopathology or focusing on parenting behavior related to parental anxiety could also be an important step to enhance child treatment outcome (Cobham, Dadds, Spence, & McDermott, 2010; Creswell, Willetts, Murray, Singhal, & Cooper, 2008). Finally, to appropriately address the needs of children experiencing multiple stressors (e.g. financial hardships, reality-based anxiety about social and environmental problems, stressful life events), clinicians need to apply manualized CBT programs with flexibility, and adaptations of treatment frequency, duration, and content may be needed (Ginsburg, Becker, Drazdowski, & Tein, 2012; Ginsburg, Becker, Kingery, & Nichols, 2008; Mifsud & Rapee, 2005). In summary, our findings demonstrate that some pretreatment child and parent factors are associated with a less favorable treatment outcome following a standard CBT program delivered in community clinics. This information may be important for decision-making in a clinical context to decide whether a child should be offered standard CBT program or an adapted or augmented treatment program. There is a need to further explore the impact of these client factors on the transportability of evidence-based treatment programs to community clinics, and future research should aim to evaluate the different adaptations of CBT to improve outcome. Conflicts of interest Wergeland, Gro Janne H. reports no conflicts of interest. Fjermestad, Krister W. reports no conflicts of interest. Marin, Carla E. reports no conflicts of interest. Bjelland, Ingvar reports no conflicts of interest. Haugland, Bente Storm-Mowatt reports no conflicts of interest. Silverman, Wendy K. reports no conflicts of interest. € Lars-Go € ran reports no conflicts of interest. Ost, Bjaastad, Jon Fauskanger reports no conflicts of interest. Oeding, Kristin reports no conflicts of interest. Havik, Odd E. reports no conflicts of interest. Heiervang, Einar R. reports no conflicts of interest. Acknowledgments The study received support from the Western Norway Regional Health Authority through project numbers 911366 and 911253. The study received additional financial support from the Meltzer Research Foundation at the University of Bergen, Norway; Josef and Haldis Andresen Foundation, Solveig and Johan P. Sommer Foundation for promotion of research on clinical psychiatry, Gidske and Petter Jacob Sørensen Foundation, and Risteigen Foundation. The authors are grateful to Professor Paula Barrett for contribution to CBT training, to Professor Robert Goodman for contribution to planning and data collection, and to Professor Torbjørn Torsheim and Dr. Rolf Gjestad for contribution to data handling. They also thank clinicians, administrative staff, and management at the participating community mental health clinics. Finally, The authors wish to express their sincere gratitude to children and parents for participation in the study. References Adelman, H. S., Kaser-Boyd, N., & Taylor, L. (1984). Children's participation in

G.J.H. Wergeland et al. / Behaviour Research and Therapy 76 (2016) 1e12 consent for psychotherapy and their subsequent response to treatment. Sum, 1984. Journal of Clinical Child Psychology, 13, 170e178. http://dx.doi.org/10.1080/ 15374418409533186. American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders: DSM-IV (4th ed.) (Washington, D.C.: Author). Angold, A., Costello, E. J., & Erkanli, A. (1999a). Comorbidity. The Journal of Child Psychology and Psychiatry, 40, 57e87. Angold, A., Costello, E. J., Farmer, E. M., Burns, B. J., & Erkanli, A. (1999b). Impaired but undiagnosed. Journal of the American Academy of Child and Adolescent Psychiatry, 38, 129e137. Angold, A., Costello, E. J., Messer, S. C., & Pickles, A. (1995). Development of a short questionnaire for use in epidemiological studies of depression in children and adolescents. International Journal of Methods in Psychiatric Research, 5, 237e249. Baldwin, S. A., Bauer, D. J., Stice, E., & Rohde, P. (2011). Evaluating models for partially clustered designs. Psychological Methods, 16, 149e165. Barrett, P. M. (2004). Friends for life e Group leader's manual (Vol. 4). Brisbane: Australian Academic Press. Barrington, J., Prior, M., Richardson, M., & Allen, K. (2005). Effectiveness of CBT versus standard treatment for childhood anxiety disorders in a community clinic setting. Behaviour Change, 22, 29e43. Bauer, D. J., Sterba, S. K., & Hallfors, D. D. (2008). Evaluating group-based interventions when control participants are ungrouped. Multivariate Behavioral Research, 43, 210e236. Beidel, D. C., & Turner, S. M. (1997). At risk for anxiety: I. Psychopathology in the offspring of anxious parents. Journal of the American Academy of Child and Adolescent Psychiatry, 36, 918e924. http://dx.doi.org/10.1097/00004583199707000-00013. Bennett, K., Manassis, K., Walter, S. D., Cheung, A., Wilansky-Traynor, P., DiazGranados, N. … Wood, J. J. (2013). Cognitive behavioral therapy age effects in child and adolescent anxiety: an individual patient data meta analysis. Depression and Anxiety, 30, 829e841. http://dx.doi.org/10.1002/da.22099. Bjaastad, J. F., Haugland, B. S. M., Fjermestad, K. W., Torsheim, T., Havik, O. E., € Heiervang, E. … Ost, L.-G. (2015). Competence and adherence scale for cognitive behavioral therapy (CAS-CBT) for anxiety disorders in youth: Psychometric properties (Submitted manuscript). €gels, S. M., Nauta, M. H., de Haan, E., Ringrose, J., Bodden, D. H., Bo Appelboom, C. … Appelboom-Geerts, K. C. (2008). Child versus family cognitivebehavioral therapy in clinically anxious youth: an efficacy and partial effectiveness study. Journal of the American Academy of Child and Adolescent Psychiatry, 47, 1384e1394. http://dx.doi.org/10.1097/CHI.0b013e318189148e. Bower, P., & Gilbody, S. (2005). Stepped care in psychological therapies: access, effectiveness and efficiency. Narrative literature review. British Journal of Psychiatry, 186, 11e17. http://dx.doi.org/10.1192/bjp.186.1.11. Brøndbo, P. H., Mathiassen, B., Martinussen, M., Heiervang, E., Eriksen, M., Moe, T. F. … Kvernmo, S. (2011). The strengths and difficulties questionnaire as a screening instrument for norwegian child and adolescent mental health services, application of UK scoring algorithms. Child and Adolescent Psychiatry and Mental Health, 5, 32. http://dx.doi.org/10.1186/1753-2000-5-32. Chorpita, B. F., Daleiden, E. L., & Weisz, J. R. (2005). Modularity in the design and application of therapeutic interventions. Applied and Preventive Psychology, 11, 141e156. http://dx.doi.org/10.1016/j.appsy.2005.05.002. Cicchetti, D. V. (1994). Guidelines, criteria, and rules of thumb for evaluating normed and standardized assessment instruments in psychology. Psychological Assessment, 6, 284e290. Cobham, V. E., Dadds, M. R., Spence, S. H., & McDermott, B. (2010). Parental anxiety in the treatment of childhood anxiety: a different story three years later. Journal of Clinical Child & Adolescent Psychology, 39, 410e420. http://dx.doi.org/10.1080/ 15374411003691719. Compton, S. N., Peris, T. S., Almirall, D., Birmaher, B., Sherrill, J., Kendall, P. C. … Albano, A. M. (2014). Predictors and moderators of treatment response in childhood anxiety disorders: results from the CAMS trial. Journal of Consulting and Clinical Psychology. http://dx.doi.org/10.1037/a0035458. Creswell, C., Willetts, L., Murray, L., Singhal, M., & Cooper, P. (2008). Treatment of child anxiety: an exploratory study of the role of maternal anxiety and behaviours in treatment outcome. Clinical Psychology & Psychotherapy, 15, 38e44. http://dx.doi.org/10.1002/cpp.559. Currie, C., Molcho, M., Boyce, W., Holstein, B., Torsheim, T., & Richter, M. (2008). Researching health inequalities in adolescents: the development of the health behaviour in school-aged children (HBSC) family affluence scale. Social Science & Medicine, 66, 1429e1436. http://dx.doi.org/10.1016/j.socscimed.2007.11.024. DiGiuseppe, R., Linscott, J., & Jilton, R. (1996). Developing the therapeutic alliance in child-adolescent psychotherapy. Applied & Preventive Psychology, 5, 85e100. http://dx.doi.org/10.1016/S0962-1849%2896%2980002-3. Drieschner, K. H., Lammers, S. M. M., & van der Staak, C. P. F. (2004). Treatment motivation: an attempt for clarification of an ambiguous concept. Clinical Psychology Review, 23, 1115e1137. http://dx.doi.org/10.1016/j.cpr.2003.09.003. Duncan, T. E., Duncan, S. C., & Strycker, L. A. (2006). An introduction to latent variable growth curve modeling concepts, issues, and applications (2nd ed.). Mahwah, N.J: Lawrence Erlbaum Associates. Ehrenreich-May, J., Southam-Gerow, M. A., Hourigan, S. E., Wright, L. R., Pincus, D. B., & Weisz, J. R. (2010). Characteristics of anxious and depressed youth seen in two different clinical contexts. Administration and Policy in Mental Health and Mental Health Services Research. http://dx.doi.org/10.1007/s10488010-0328-6. Ford, T., Goodman, R., & Meltzer, H. (2003). The British child and adolescent mental

11

health survey 1999: the prevalence of DSM-IV disorders. Journal of the American Academy of Child and Adolescent Psychiatry, 42, 1203e1211. http://dx.doi.org/ 10.1097/00004583-200310000-00011. Ginsburg, G. S., Becker, K. D., Drazdowski, T. K., & Tein, J.-Y. (2012). Treating anxiety disorders in inner city schools: results from a pilot randomized controlled trial comparing CBT and usual care. Child & Youth Care Forum, 41, 1e19. http:// dx.doi.org/10.1007/s10566-011-9156-4 22701295. Ginsburg, G. S., Becker, K. D., Kingery, J. N., & Nichols, T. (2008). Transporting CBT for childhood anxiety disorders into inner-city school-based mental health clinics. Cognitive and Behavioral Practice, 15, 148e158. http://dx.doi.org/10.1016/ j.cbpra.2007.07.001. Goodman, R. (1997). The strengths and difficulties questionnaire: a research note. Journal of Child Psychology and Psychiatry, 38, 581e586. http://dx.doi.org/ 10.1111/j.1469-7610.1997.tb01545.x. Goodman, R. (1999). The extended version of the strengths and difficulties questionnaire as a guide to child psychiatric caseness and consequent burden. Journal of Child Psychology and Psychiatry, 40, 791e799. http://dx.doi.org/ 10.1111/1469-7610.00494. Goodman, R., Ford, T., Richards, H., Gatward, R., & Meltzer, H. (2000). The Development and Well-Being Assessment: description and initial validation of an integrated assessment of child and adolescent psychopathology. The Journal of Child Psychology and Psychiatry, 41, 645e655. http://dx.doi.org/10.1111/j.14697610.2000.tb02345.x. Goodman, A., Lamping, D. L., & Ploubidis, G. B. (2010). When to use broader internalising and externalising subscales instead of the hypothesised five subscales on the strengths and difficulties questionnaire (SDQ): data from British parents, teachers and children. Journal of Abnormal Child Psychology, 38, 1179e1191. http://dx.doi.org/10.1007/s10802-010-9434-x 20623175. Gordon-Hollingsworth, A. T., Becker, E. M., Ginsburg, G. S., Keeton, C., Compton, S. N., Birmaher, B. B. … March, J. S. (2015). Anxiety disorders in Caucasian and African American children: a comparison of clinical characteristics, treatment process variables, and treatment outcomes. Child Psychiatry Human Development, 46, 643e655. http://dx.doi.org/10.1007/s10578-014-0507x 25293650. Greenberg, R. P., Constantino, M. J., & Bruce, N. (2006). Are patient expectations still relevant for psychotherapy process and outcome? Clinical Psychology Review, 26, 657e678. http://dx.doi.org/10.1016/j.cpr.2005.03.002. Guo, S. (2005). Analyzing grouped data with hierarchical linear modeling. Children and Youth Services Review, 27, 637e652. http://dx.doi.org/10.1016/ j.childyouth.2004.11.017. Heiervang, E., Stormark, K. M., Lundervold, A. J., Heimann, M., Goodman, R., Posserud, M. B. … Gillberg, C. (2007). Psychiatric disorders in Norwegian 8- to 10-year-olds: an epidemiological survey of prevalence, risk factors, and service use. Journal of the American Academy of Child and Adolescent Psychiatry, 46, 438e447. Ishikawa, S.-i., Okajima, I., Matsuoka, H., & Sakano, Y. (2007). Cognitive behavioural therapy for anxiety disorders in children and adolescents: a meta-analysis. Child and Adolescent Mental Health, 12, 164e172. http://dx.doi.org/10.1111/ j.1475-3588.2006.00433.x. James, A. C., James, G., Cowdrey, F. A., Soler, A., & Choke, A. (2013). Cognitive behavioural therapy for anxiety disorders in children and adolescents. The Cochrane Database of Systematic Reviews, 6, CD004690. http://dx.doi.org/ 10.1002/14651858.CD004690.pub3. Keijsers, G., Schaap, C., & Hoogduin, C. (2000). The impact of interpersonal patient and therapist behavior on outcome in cognitive-behavior therapy. A review of empirical studies. Behavior Modification, 24, 264e297. http://dx.doi.org/10.1177/ 0145445500242006. Keijsers, G. P. J., Schaap, C. P. D. R., Hoogduin, C. A. L., Hoogsteyns, B., & de Kemp, E. C. M. (1999). Preliminary results of a new instrument to assess patient motivation for treatment in cognitive-behaviour therapy. Behavioural and Cognitive Psychotherapy, 27, 165e179. Kendall, P. C., & Beidas, R. S. (2007). Smoothing the trail for dissemination of evidence-based practices for youth: flexibility within fidelity. Professional Psychology: Research and Practice, 38, 13e20. http://dx.doi.org/10.1037/07357028.38.1.13. Kendall, P. C., Comer, J. S., Marker, C. D., Creed, T. A., Puliafico, A. C., Hughes, A. A. … Hudson, J. (2009). In-session exposure tasks and therapeutic alliance across the treatment of childhood anxiety disorders. Journal of Consulting and Clinical Psychology, 77, 517e525. http://dx.doi.org/10.1037/ a0013686. Kerns, C. M., Read, K. L., Klugman, J., & Kendall, P. C. (2013). Cognitive behavioral therapy for youth with social anxiety: differential short and long-term treatment outcomes. Journal of Anxiety Disorders, 27, 210e215. http://dx.doi.org/ 10.1016/j.janxdis.2013.01.009. Kley, H., Heinrichs, N., Bender, C., & Tuschen-Caffier, B. (2012). Predictors of outcome in a cognitive-behavioral group program for children and adolescents with social anxiety disorder. Journal of Anxiety Disorders, 26, 79e87. http:// dx.doi.org/10.1016/j.janxdis.2011.09.002. Knight, A., McLellan, L., Jones, M., & Hudson, J. (2014). Pre-treatment predictors of outcome in childhood anxiety disorders: a systematic review. Psychopathology Review, 1, 77e129. http://dx.doi.org/10.5127/pr.034613. Liber, J. M., Van Widenfelt, B. M., Goedhart, A. W., Utens, E. M., Van der Leeden, A. J., Markus, M. T. … Treffers, P. D. (2008). Parenting and parental anxiety and depression as predictors of treatment outcome for childhood anxiety disorders: has the role of fathers been underestimated? Journal of Clinical Child &

12

G.J.H. Wergeland et al. / Behaviour Research and Therapy 76 (2016) 1e12

Adolescent Psychology, 37, 747e758. http://dx.doi.org/10.1080/ 15374410802359692. 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, 38, 683e694. http://dx.doi.org/10.1007/S10802-010-9394-1. Lieb, R., Wittchen, H. U., Hofler, M., Fuetsch, M., Stein, M. B., & Merikangas, K. R. (2000). Parental psychopathology, parenting styles, and the risk of social phobia in offspring: a prospective-longitudinal community study. Archives of General Psychiatry, 57, 859e866. Lovibond, P. F., & Lovibond, S. H. (1995). The structure of negative emotional states: comparison of the depression anxiety stress scales (DASS) with the beck depression and anxiety inventories. Behaviour Research and Therapy, 33, 335e343. http://dx.doi.org/10.1016/0005-7967(94)00075-U. Lundkvist-Houndoumadi, I., Hougaard, E., & Thastum, M. (2014). Pre-treatment child and family characteristics as predictors of outcome in cognitive behavioural therapy for youth anxiety disorders. Nordic Journal of Psychiatry. http:// dx.doi.org/10.3109/08039488.2014.903295. Lundkvist-Houndoumadi, I., Thastum, M., & Hougaard, E. (2015). Effectiveness of an individualized case formulation-based cbt for non-responding youths with anxiety disorders. Journal of Child and Family Studies. http://dx.doi.org/10.1007/ s10826-015-0225-4. No Pagination Specified. Lyneham, H. J., Rapee, R. M., & Hudson, J. L. (2014). Prevention of anxiety disorders. In T. Ehring, & P. M. G. Emmelkamp (Eds.), The Wiley handbook of anxiety disorders (pp. 625e642). Chichester, West Sussex, UK Malden, MA: John Wiley & Sons Ltd. March, J. S., & Curry, J. F. (1998). Predicting the outcome of treatment. Journal of Abnormal Child Psychology, 26, 39e51. http://dx.doi.org/10.1023/A: 1022682723027. McLeod, B. D., Wood, J. J., & Weisz, J. R. (2007). Examining the association between parenting and childhood anxiety: a meta-analysis. Clinical Psychology Review, 27, 155e172. http://dx.doi.org/10.1016/j.cpr.2006.09.002. Messer, S. C., & Beidel, D. C. (1994). Psychosocial correlates of childhood anxiety disorders. Journal of the American Academy of Child and Adolescent Psychiatry, 33, 975e983. http://dx.doi.org/10.1097/00004583-199409000-00007. Mifsud, C., & Rapee, R. M. (2005). Early intervention for childhood anxiety in a school setting: outcomes for an economically disadvantaged population. Journal of the American Academy of Child & Adolescent Psychiatry, 44, 996e1004. http:// dx.doi.org/10.1097/01.chi.0000173294.13441.87 16175104. n, L. K., & Muthe n, B. O. (2011). Mplus user's guide (Vol. 6). Los Angeles, CA: Muthe n & Muthe n. Muthe Nauta, M. H., Scholing, A., Emmelkamp, P. M. G., & Minderaa, R. B. (2003). Cognitivebehavioral therapy for children with anxiety disorders in a clinical setting: no additional effect of a cognitive parent training. Journal of the American Academy of Child and Adolescent Psychiatry, 42, 1270e1278. http://dx.doi.org/10.1097/ 01.chi.0000085752.71002.93. Nilsen, T. S., Eisemann, M., & Kvernmo, S. (2013). Predictors and moderators of outcome in child and adolescent anxiety and depression: a systematic review of psychological treatment studies. European Child & Adolescent Psychiatry, 22, 69e87. http://dx.doi.org/10.1007/s00787-012-0316-3. Nilsen, T. S., Handegard, B. H., Eisemann, M., & Kvernmo, S. (2015). Evaluating change in symptomatic and functional level of children and youth with emotional disorders: a naturalistic observation study. European Child & Adolescent Psychiatry. http://dx.doi.org/10.1007/s00787-014-0671-3. € €ld, L., Costa, N., Cederlund, R., Ollendick, T. H., Ost, L. G., Reuterskio Sirbu, C. … Jarrett, M. A. (2009). One-session treatment of specific phobias in youth: a randomized clinical trial in the United States and Sweden. Journal of Consulting and Clinical Psychology, 77, 504e516. http://dx.doi.org/10.1037/ a0015158. Rapee, R. M. (1997). Potential role of child rearing practices in the development of anxiety and depression. Clinical Psychology Review, 17, 47e67. Rapee, R. M., Schniering, C. A., & Hudson, J. L. (2009). Anxiety disorders during childhood and adolescence: origins and treatment. Annual Review of Clinical Psychology, 5, 311e341. http://dx.doi.org/10.1146/ Annurev.Clinpsy.032408.153628. Reynolds, S., Wilson, C., Austin, J., & Hooper, L. (2012). Effects of psychotherapy for anxiety in children and adolescents: a meta-analytic review. Clinical Psychology Review, 32, 251e262. http://dx.doi.org/10.1016/j.cpr.2012.01.005. Silverman, W. K., & Albano, A. M. (1996). Anxiety disorders interview schedule (ADISIV) parent interview schedules. Albany, NY: Greywind Publications. Silverman, W. K., Pina, A. A., & Viswesvaran, C. (2008). Evidence-based psychosocial treatments for phobic and anxiety disorders in children and adolescents. Journal of Clinical Child & Adolescent Psychology, 37, 105e130. http://dx.doi.org/ 10.1080/15374410701817907.

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 and Adolescent Psychiatry, 40, 937e944. http://dx.doi.org/10.1097/00004583200108000-00016. Southam-Gerow, M. A., Chorpita, B. F., Miller, L. M., & Gleacher, A. A. (2008). Are children with anxiety disorders privately referred to a university clinic like those referred from the public mental health system? Administration and Policy in Mental Health and Mental Health Services Research, 35, 168e180. http:// dx.doi.org/10.1007/s10488-007-0154-7. Southam-Gerow, M. A., Weisz, J. R., Chu, B. C., McLeod, B. D., Gordis, E. B., & ConnorSmith, J. K. (2010). Does cognitive behavioral therapy for youth anxiety outperform usual care in community clinics? an initial effectiveness test. Journal of the American Academy of Child and Adolescent Psychiatry, 49, 1043e1052. http://dx.doi.org/10.1016/j.jaac.2010.06.009. Southam-Gerow, M. A., Weisz, J. R., & Kendall, P. C. (2003). Youth with anxiety disorders in research and service clinics: examining client differences and similarities. Journal of Clinical Child & Adolescent Psychology, 32, 375e385. http://dx.doi.org/10.1207/S15374424JCCP3203_06. Spence, S. H. (1998). A measure of anxiety symptoms among children. Behavioral Research and Therapy, 36, 545e566. http://dx.doi.org/10.1016/S0005-7967(98) 00034-5. Stone, L. L., Otten, R., Engels, R. C., Vermulst, A. A., & Janssens, J. M. (2010). Psychometric properties of the parent and teacher versions of the strengths and difficulties questionnaire for 4- to 12-year-olds: a review. Clinical Child and Family Psychology Review, 13, 254e274. http://dx.doi.org/10.1007/s10567-0100071-2. Villabø, M. A., Cummings, C. M., Gere, M. K., Torgersen, S., & Kendall, P. C. (2013). Anxious youth in research and service clinics. Journal of Anxiety Disorders, 27, 16e24. http://dx.doi.org/10.1016/j.janxdis.2012.09.003. Walkup, J. T., Albano, A. M., Piacentini, J., Birmaher, B., Compton, S. N., Sherrill, J. T. … Kendall, P. C. (2008). Cognitive behavioral therapy, sertraline, or a combination in childhood anxiety. New England Journal of Medicine, 359, 2753e2766. http://dx.doi.org/10.1056/NEJMoa0804633. Weisz, J. R., Chorpita, B. F., Palinkas, L. A., Schoenwald, S. K., Miranda, J., Bearman, S. K. … Research Network on Youth Mental, H. (2012). Testing standard and modular designs for psychotherapy treating depression, anxiety, and conduct problems in youth: a randomized effectiveness trial. Archives of General Psychiatry, 69, 274e282. http://dx.doi.org/10.1001/archgenpsychiatry.2011.147. Weisz, J. R., Donenberg, G. R., Han, S. S., & Weiss, B. (1995). Bridging the gap between laboratory and clinic in child and adolescent psychotherapy. Journal of Consulting and Clinical Psychology, 63, 688e701. Weisz, J. R., Ng, M. Y., Rutt, C., Lau, N., & Masland, S. (2013a). Psychotherapy for children and adolescents. In M. J. Lambert (Ed.), Bergin and Garfield's handbook of psychotherapy and behavior change (6th ed., pp. 541e586). Hoboken, N.J: Wiley. Weisz, J. R., Ugueto, A. M., Cheron, D. M., & Herren, J. (2013b). Evidence-based youth psychotherapy in the mental health ecosystem. Journal of Clinical Child & Adolescent Psychology, 42, 274e286. http://dx.doi.org/10.1080/ 15374416.2013.764824. Wergeland, G. J., Fjermestad, K. W., Marin, C. E., Haugland, B. S., Bjaastad, J. F., Oeding, K. … Heiervang, E. R. (2014). An effectiveness study of individual vs. group cognitive behavioral therapy for anxiety disorders in youth. Behaviour Research and Therapy, 57, 1e12. http://dx.doi.org/10.1016/j.brat.2014.03.007. Wergeland, G. J., Fjermestad, K. W., Marin, C. E., Haugland, B. S., Silverman, W. K., Ost, L. G. … Heiervang, E. R. (2015). Predictors of dropout from community clinic child CBT for anxiety disorders. Journal of Anxiety Disorders, 31, 1e10. http:// dx.doi.org/10.1016/j.janxdis.2015.01.004. Westra, H. A., Arkowitz, H., & Dozois, D. J. A. (2009). Adding a motivational interviewing pretreatment to cognitive behavioral therapy for generalized anxiety disorder: a preliminary randomized controlled trial. Journal of Anxiety Disorders, 23, 1106e1117. http://dx.doi.org/10.1016/j.janxdis.2009.07.014 19665347. Westra, H. A., & Dozois, D. J. A. (2006). Preparing clients for cognitive behavioral therapy: a randomized pilot study of motivational interviewing for anxiety. Cognitive Therapy and Research, 30, 481e498. http://dx.doi.org/10.1007/s10608006-9016-y. Wood, J. J., Piacentini, J. C., Bergman, R. L., McCracken, J., & Barrios, V. (2002). Concurrent validity of the anxiety disorders section of the anxiety disorders interview schedule for DSM-IV: child and parent versions. Journal of Clinical Child & Adolescent Psychology, 31, 335e342. http://dx.doi.org/10.1207/ S15374424JCCP3103_05. Wothke, W. (2000). Longitudinal and multi-group modeling with missing data. In T. D. Little, K. U. Schnabel, & J. Baumert (Eds.), Modeling longitudinal and multilevel data. Mahwah, N.J: Lawrence Erlbaum.