Internet-delivered cognitive behavior therapy for adolescents with obsessive-compulsive disorder

Abstract

Background: About one in 100 adolescents suffers from obsessive-compulsive disorder (OCD). OCD usually causes severe impairment in social, academic, and family functioning. As OCD often has a chronic course and is associated with a high risk for future psychological problems and disability, it is important to detect and treat the symptoms early in time. Cognitive behavior therapy (CBT) is recommended as the first choice treatment. Unfortunately, most sufferers do not have access to CBT, due to different factors such as limited health care resources and geographical distances. Internet-delivered CBT (ICBT) has been proposed as a solution to make effective treatments more available to patients. The feasibility and efficacy of ICBT are well studied for adults with OCD but have not been established for young people with OCD.

Objective: The main objective of the studies included in this thesis was to develop and evaluate an ICBT intervention for adolescents with OCD. Specifically, we aimed to 1) test the feasibility and preliminary efficacy of ICBT for OCD in an open pilot study, 2) test the efficacy of ICBT for OCD in a randomized controlled trial, 3) study the cost-effectiveness of ICBT for OCD, and 4) evaluate predictors of treatment response to ICBT for OCD from baseline patient characteristics.

Methods: We developed a novel ICBT intervention for adolescents with OCD and their parents – named “BiP OCD” – and conducted an open pilot study with 21 adolescents with OCD (study I). In the second study, we randomized 67 adolescents with OCD to either ICBT or a waitlist of equal duration (study II). To evaluate the cost-effectiveness of ICBT, we collected cost data on health care utilization, medication use, school absence, and productivity loss in the same sample alongside the randomized controlled trial (study III). All patients that had received ICBT in the randomized controlled trial, either immediately or after the waitlist, were then pooled together (N=61) and analyzed regarding baseline predictors of treatment response three months after ICBT (study IV). Predictor analyses were conducted with classical linear regression as well as with machine learning methodology.

Results: Large within-group effects of ICBT were observed in studies I and II. Betweengroup comparison with the waitlist group in study II revealed moderately sized treatment effects. Overall, patients were highly satisfied with the intervention. The average clinician time was below 20 minutes per patient and week. Study III identified cost-saving effects of ICBT when compared with untreated patients on the waitlist, indicating that it is more costly for society to not treat those patients than to offer ICBT. In study IV, linear regression analyses could not identify any predictors for ICBT treatment response. However, machine learning algorithms predicted treatment response with high accuracy.

Conclusions: The results presented in this thesis provide support for the feasibility, efficacy, and cost-effectiveness of ICBT for adolescent OCD. Machine learning methodology is a novel, promising statistical approach that could guide us to more accurate predictions of treatment outcome. The results warrant replication in larger clinical samples and in comparison with active control conditions as well as face-to-face CBT.