The effectiveness of Fluvoxamine was compared to that of Cognitive Therapy (CT) in a 12-week randomized controlled trial (RCT) in 48 patients with obsessive-compulsive disorder (OCD), who were treatment-resistant to a previous behavior therapy (BT). A considerable amount of patients did not comply with the assigned treatment and switched treatments. The aim of this study was to identify patient characteristics predictive of assignment compliance and to study whether these characteristics were related to outcome. A logistic model, based on psychological and social patient characteristics, in addition to or in interaction with the assignment, was used for the explanation of compliance with treatment assignment. Especially patients who have a higher score on the Yale-Brown Obsessive Compulsive Scale (Y-BOCS) tend to comply with the effective Fluvoxamine treatment. The same set of variables was related to both compliance and outcome of therapy received. Therefore, the logistic model of compliance could be used to reduce the positive bias of As-Treated analysis (AT). The difference between the results of Fluvoxamine and Cognitive Therapy remained statistically significant after correcting for the positive bias as the result of assignment refusal and after applying the assumption that two drop-out patients needed imputation of lesser results.