Exploring the Relations between Depression, Somatization, Dissociation and Alexithymia – Overlapping or Independent Constructs?

Abstract

Background: The aim of this study was to extend our knowledge of associations among the constructs of alexithymia, depression, somatization and dissociation. Sampling and Methods: 924 nonclinical subjects answered questions about depression (21-item Beck Depression Inventory), somatization (13-item somatization part of Symptom Check List-90), dissociation (28-item Dissociative Experiences Scale) and alexithymia (20-item Toronto Alexithymia Scale). In addition, a 12-item General Health Questionnaire (GHQ-12) was administered to detect psychiatric distress among subjects. Results: The results suggested that there was a significant clinical correlation between somatization, dissociation, depression and alexithymia (rho varied from 0.31 to 0.56). The principal component analysis revealed the presence of four components: depression, somatization, dissociation and alexithymia. The use of factor scores diminished the covariance between measures (rho varied from –0.10 to 0.01 between the factor scores). There was almost no correlation between the dissociation factor (rho = 0.06) and alexithymia factor (rho = 0.09) scores and general distress (GHQ-12). Conclusions: This study suggests that while somatization, dissociation, depression and alexithymia are distinct constructs, they correlate to a considerable extent. The use of factor analysis and factor scores should be considered to diminish covariance between the above constructs. Comparing results between factored and unfactored results may prove illuminating. As a case in point, the results suggest that the part of dissociation that coincides with other constructs (overlaps) is associated with distress, whereas the distinct part of dissociation (no shared covariance) is not associated with distress. The same applies to the alexithymia construct. Longitudinal studies are needed to show whether there is a trait such as a relatively stable dissociation component and also whether a separate state-dependent dissociation component exists that is associated with coincident distress, somatization and depression.