The present study investigated the factor structure of the Beck Depression Inventory (BDI), and causal attributions for the development of illness in a sample of 102 inpatients of a thoracic surgery department, with the main objective of examining the power of causal attributions and functional support in predicting different factors derived from the BDI. The results revealed that the BDI clusters into affective/motivational, somatic/vegetative, self-blame and self-punitiveness dimensions. Causal attributions for the development of illness were represented by family conflicts, environmental adversity, finance/health problems, bad luck, and four self-related clusters. Examination of the predictors of the BDI dimensions showed that causal attributions to uncontrollable, unmodifiable variables were predictors for the affective/motivational dimension, whereas illness severity and greater age were related to the somatic/vegetative symptomatology. The results supported the recommendations for the exclusion of such symptoms in assessing depression severity among medical patients.