A network approach to understanding obsessions and compulsions

Background: Efforts to understand the constellation of symptoms in obsessive-compulsive disorder (OCD) have typically relied on models where latent variable(s) are assumed to underlie all symptoms. In contrast, a network approach does not assume that there are underlying latent variables and allows for the possibility that clusters of symptoms may mutually reinforce each other. We aimed to determine whether obsessions and compulsions formed a coherent and mutually reinforcing network of symptoms. Method: 400 participants were recruited online and administered the Obsessive-Compulsive Inventory-Revised (OCI-R). A network analysis was computed using an Extended Bayesian Information Criterion estimator. Results: There were five communities of symptoms: 1. A mixed contamination and checking community, 2. An ordering/arranging community, 3. A superstitious/counting/repeating community, 4. A mixed hoarding and checking community, and 5. An intrusive thoughts community. In the accuracy check, edges displayed wide confidence intervals, indicating that edges’ strength could not be interpreted. Additional analyses at the level of OCI-R subscales indicated that checking was significantly more central than other subscales in the network. Conclusions: Obsessions and compulsions may be related in a mutually reinforcing way, thereby constituting OCD as a psychopathological entity. Prospective investigations are needed to ascertain the directionality of relationships in the network.