Latent class analysis of obsessive-compulsive symptoms in a clinical sample

Abstract

Objective: Obsessive-compulsive disorder (OCD) is probably a heterogeneous condition. To date, research investigating the symptom structure of OCD has generally focused on variable-classification approaches, primarily factor analysis. Our aim was to use the latent class analysis, an advanced individual-classification method, in order to define homogeneous sub-groups based on the Yale-Brown Obsessive-Compulsive Scale Check List (Y-BOCS CL) assessments within 193 outpatients with OCD.

Method: Latent class analysis of 13 obsessive-compulsive symptom-clusters measured by the Yale-Brown Obsessive-Compulsive Scale Check List was performed. Associations of demographic and clinical features with latent class membership were tested by using logistic regression models. Differences in severity of obsessive-compulsive symptoms, obsessive beliefs, and depressive symptomotology between latent classes were evaluated with the chi-square likelihood test.

Results: Latent class analysis models of best fit yielded 3 latent classes: "autogenous obsessions", "reactive obsessive-compulsive", and "reactive obsessions". Outpatients in reactive obsessive-compulsive group reported significantly higher scores on the Y-BOCS global, obsessions, and compulsions scales than outpatients in other two sub-groups. Participants in reactive obsessive-compulsive group were predominantly female, single, and had a history of family psychopathology; whereas autogenous and reactive obsessions groups were predominantly male. There were no differences in severity of depression and obsessive belief domains.

Conclusions: The results provide support for the validity of autogenous/reactive distinction for OCD. These three classes may have important implications for research and clinical purposes.