Aims and hypotheses. The main objective of the present study is to develop a group delivered treatment manual for youth with OCD and evaluate its efficacy relative to (1) a waiting list control condition and (2) outcomes from existing data from a study of individually delivered CBT using a benchmarking strategy. The efficacy of the new treatment manual will be evaluated at the end of 13 weeks of acute treatment and at 3-, 6- , 12-, 24- and 36-month follow-up.
Based on the previous research we hypothesize that: (i) G-CBT will show superiority to the waiting period condition, and (ii) G-CBT will show similar effectiveness compared to I-CBT at the end of acute treatment as well as at longer-term follow-up assessment points.
Secondary aims are to evaluate predictors of treatment response to G-CBT and to identify specific response trajectories, i.e. patterns of symptom fluctuation during treatment.
Based on the previous literature we want to explore whether: (1) youth in families with higher levels of family conflict, parental blame and lower levels of cohesion will benefit less from treatment when compared to youth without these family characteristics, (2) lower levels of initial and intermediate treatment motivation will be associated with lower levels of treatment gains, and (3) specific patient response trajectories possibly associated with the onset of specific treatment ingredients (i.e., initial psychoeducation, in session exposure and response prevention, specific work on obsessions, work meetings with parents, etc.) and/or with patient characteristics, i.e. OCD subtypes, can be identified.
Further, the aim is to develop and/or evaluate relevant Danish versions of clinical instruments for the assessment of pediatric OCD (symptoms and severity). More specifically, one aim is to develop and evaluate a short self-rating of OCD severity for youth with OCD, which can be utilized by therapists during treatment and as a tool for youth's self-monitoring of symptoms.
Procedures. Eligible patients will be recruited as they are consecutively referred to the OCD Clinic at Department of Child and Adolescent Psychiatry, Aarhus University Hospital, Psychiatry, Denmark.
Seventy-two adolescents, 13-17 years old will be included in the study, corresponding to inclusion of approximately 140 adolescents into twenty treatment groups.
The benchmarking sample includes all 13- to 17-year-old Danish participants from the aforementioned Scandinavian pediatric multicenter OCD CBT study, NordLOTS. NordLOTS data provide a unique opportunity to benchmark outcomes from the two different CBT formats delivered at the same clinic. Therefore, the procedures used in the present study are largely similar to the procedures used in NordLOTS (the Committee of Ethical Science file number: 20070140). Patients for the current study will be included as they are consecutively referred to the OCD Clinic and enrolled in the standard clinical assessment and treatment procedures. All assessments will be conducted by experienced external clinician raters (i.e., clinicians from the clinic who are not acting as therapists in the treatment of the assessed patient).
In order to be able to adjust data analyses for possible interdependence of observations due to group membership, patients who are offered G-CBT treatment (i.e. offered G-CBT as a part of the standard clinical procedures) but fall for study exclusion criteria (e.g. previously or presently treated with SSRI's) or patients who do not wish to participate in the extended assessment procedures in the project (i.e. decline participation after presented with project information at the first visit to the clinic) will be asked permission to use the following clinical data according to their clinical record: gender, age, age at OCD-onset, present psychiatric diagnoses, and CY-BOCS scores at pre-treatment, mid-treatment, end of acute treatment, and at 3 months follow-up.
Intervention The G-CBT manual for adolescents has been developed in the OCD-clinic in Aarhus and pilot tested in Aarhus and subsequently in collaboration with clinicians at the OCD Clinic, Queen Silvia Children's Hospital, Sahlgrenska University Hospital, Department of Child and Adolescent Psychiatry Gothenburg, Sweden, who also participated in NordLOTS. The G-CBT manual is based on the NordLOTS I-CBT manual63 in overall content and structure, with a few additional elements added with inspiration from previous individual and group CBT literature12,13,64-67 as well as the author's ideas. The NordLOTS manual is primarily based on a neurobiological model for psychoeducation, and the main therapeutic ingredient is exposure and response prevention. The additional or slightly altered elements in the G-CBT manual include: a) a stronger emphasis on cognitive aspects and goal-setting, b) introduction of formalized supportive "work-meetings" between children/adolescents and parents, c) exercises in thought focus strategies, d) verbalization of parents' recognition of the young patient's work and efforts during the program by a parent speech at the final treatment session, and e) introduction of an OCD farewell letter during the booster sessions. The number of individual family and parent sessions is largely modeled after the Australian OCD Busters manual (albeit, only the published journal paper reporting of this manual was available at the time of the development of the Danish manual)13. The manual is made up of three individual family sessions, 13 weekly adolescent group sessions, three parent group sessions and two adolescent group booster sessions conducted after one and three months. The adolescent group sessions are conducted by two therapists and the parent group sessions are conducted by one therapist. The manual contains detailed guiding of therapists for each session and a comprehensive collection of working sheets for the adolescents to be used during sessions.
Reliability of primary outcome measures and treatment fidelity ratings Following the previous procedures from NordLOTS, reliability ratings of primary outcome assessment will be performed by experienced clinicians on a random sample of 15 % of baseline and week 13 CY-BOCS interviews. All interviews will be videotaped and uploaded to the research database (RedCAP). An adaptation of the NordLOTS Treatment Integrity Scale (TIS)9 will be used to evaluate therapist fidelity to the CBT sessions. All therapy sessions will be videotaped and uploaded to the research database (RedCAP). A random sample of 15 % of all sessions will be scored by experienced semi-independent clinicians. In TIS 3 categories of treatment fidelity are evaluated on a four-point scale ranging from 1 = very poor compliance to 4 = very good compliance: (1) manual adherence, (2) competence in manual delivery, and (3) therapist relational competencies.
Power calculation and statistical analyses In total, we expect to include 72 experimental subjects treated with G-CBT and 45 subjects previously included in NordLOTS treated with I-CBT and completing the program. With reference to NordLOTS we expect a drop-out rate of 10 % during treatment9 and a further 12.5 % drop-out during the one-year post-treatment follow-up period. One-year Outcome for Responders of Cognitive Behavior Therapy for Child and Adolescent Obsessive-Compulsive Disorder.). Expected distribution, standard deviation and differences in response between waiting period condition and CBT calculated based on data from two previous studies showing CY-BOCS change scores of -1.09 (SD = 4.86) and 13.2 (SD = 6.41) respectively, indicates that the response within each subject group is normally distributed with a standard deviation of 5.69 and a change score mean difference between conditions of 14.29. If this mean difference between conditions holds true for the planned study we will be able to reject the null hypothesis i.e. that the means of the two conditions are equal with a probability (power) of 1.000 and a type I error probability of 0.05.
Descriptive statistics for sample characteristics as well as outcome measures will be calculated. In order to assess treatment effect, a multilevel linear mixed model with CY-BOCS severity score as primary outcome dependent variable including random slopes and intercept will be applied. To estimate the treatment effect measured with the CGI-S, a generalized linear effects model will be used. Waiting period vs. treatment condition CY-BOCS change scores will be estimated using a random effects model in order to adjust for differences in exposure length in the two conditions. Secondary outcome measures include KINDL, OBQ-CV, COIS, MFQ, and SCARED.
Out of the 269 patients in NordLOTS, the Aarhus site included 80. Of these, 45 were adolescents between the age of 13-17. The expected inclusion of 72 adolescents in the current study will be sufficient for comparative benchmarking and predictor analyses.
In the benchmarking analyses participant baseline demographics (age, gender, BQ), clinical characteristics (CY-BOCS, CGI, KINDL, C-GAS, CBCL, FAS, COIS, MFQ, SCARED, ASSQ) and scores on outcome measures (primary measures: CY-BOCS, CGI, KINDL; secondary measures: C-GAS, COIS, MFQ, SCARED, OBQ-CV) will be compared between the two samples using independent samples t-tests and Fisher's exact test. In order to adjust for possible interdependence of observations due to group membership, intra-class correlation coefficients (ICCs) will be calculated for primary outcome measures (CY-BOCS): ICC = (MSbetweengroup - MSwithingroup)/(MSbetweengroup + \[n - 1\] MSwithingroup). Pre- to post CY-BOCS change will be estimated as pre-post effect sizes \[ES\] where the ES equals the mean difference between pre-treatment and post-treatment scores divided by the standard deviation of the mean difference, adjusted by the correlation between the pre- and post-scores. Difference between the two interventions will be estimated by calculating the standardized mean difference between the pre-post change scores of the two samples, divided by a pooled estimate of the post-test standard deviations. Analyses will use an intention-to-treat (ITT) approach. A two-sample t-test for proportions will be conducted in order to examine the difference in participant completion rate between the current study and the benchmark study.
