The presence of obsessive-compulsive disorder (OCD) is routinely assessed in clinical practice using the diagnostic criteria as described in the Diagnostic and Statistical Manual of Mental Disorders (DSM). Despite the significant successes of this categorical classification system of psychopathology, its widespread application has also generated several important problems. Three of these problems form the starting point of the first three studies of this thesis. The empirical studies presented here are all embedded within the Netherlands Obsessive Compulsive Disorder Association study, a multicentre naturalistic cohort study designed to investigate the long-term course and outcome in OCD. The study presented in chapter 2 addressed the problem of the reduction of diagnostic assessments in psychiatry to the conduction of checklist assessments. The main goal of this study was to come to a deeper phenomenological understanding of core obsessive-compulsive symptoms, beyond their usefulness as indicators in the categorical classification system. Using a newly developed questionnaire based on insights and concepts derived from the philosophy of free will, we identified three aspects of free will experience in patients with OCD (“alternative possibilities”, “intentionality” and “ownership”) and related these experiences to core clinical characteristics. The study presented in chapter 3 addressed the problem of lingering clinical heterogeneity within the boundaries of the DSM categories, with a focus on course of illness variations in OCD. The main goal of this study was to identify factors beyond the OCD criteria themselves that might be involved in inducing and in maintaining a chronic course of illness in OCD. Using data from 270 subjects with a current diagnosis of OCD, chronicity status at two-year follow-up was regressed on a selection of baseline predictors. We found that external factors (psychotrauma, recent negative life events and presence of a partner) influenced the risk of becoming chronic, whereas illness-related factors (longer illness duration and higher illness severity) increased the risk of remaining chronic. The study presented in chapter 4 addressed the problem of psychiatric comorbidity. The main goal of this study was to identify and characterize homogeneous subgroups of OCD patients based on their comorbidity profile. In a cohort of 419 subjects with OCD, the lifetime presence of fifteen comorbid disorders was assessed. Latent class analysis was used to identify comorbidity-based subgroups. The study sample could be divided in a large group with a low amount of comorbidity that could be further subdivided into two subgroups: OCD simplex and OCD with lifetime major depressive disorder; and a group with a high amount of comorbidity that could be further subdivided into a general anxiety-related subgroup, an autism/social phobia-related subgroup and a psychosis/bipolar-related subgroup. Group membership was associated with core clinical characteristics. The narrative review presented in chapter 5 provided a theoretical analysis of the phenomenon of psychiatric comorbidity, its links to the structure of the DSM and the problems it has generated over time. We discussed several alternative models of psychopathology that have been developed over time to address these problems. The study presented in chapter 6 applied one of these models, network analysis, to the study of OCD. Using data from 419 patients with a lifetime diagnosis of OCD, we constructed a regularized partial correlation network with the 10 items of the Yale-Brown Obsessive Compulsive Scale as nodes. We then calculated three centrality parameters for each node. We found that the items related to resistance to and control over obsessions and compulsions were particularly important in the dynamics within the OCD network. We concluded that these items might form suitable targets for therapeutic interventions in OCD and that network analysis might provide a promising avenue to develop more dynamic models of OCD pathology.