Social anxiety is a common problem in psychotic disorders. The Liebowitz Social Anxiety Scale, Self-Rating version (LSAS-SR) is a widely used instrument to capture different aspects of social anxiety, but its psychometric properties have not been tested in this patient group. The aims of the present study were to evaluate the psychometric properties of the LSAS-SR in patients with first episode psychosis, to investigate whether it differentiated between active and passive social withdrawal and to test which clinical factors contributed to current level of social anxiety.
A total of 144 first episode psychosis patients from the ongoing Thematically Organized Psychosis (TOP) study were included at the time of first treatment. Diagnoses were set according to the Structured Clinical Interview (SCID-1) for DSM-IV. A factor analysis was carried out and the relationship of social anxiety to psychotic and general symptomatology measured by the Positive and Negative Syndrome Scale (PANSS) was evaluated. Possible contributors to social anxiety were analyzed using multiple hierarchic regression analysis.
The factor analysis identified three subscales: public performance, social interaction and observation. All three subscales showed satisfactory psychometric properties, acceptable convergent and discriminate properties, and confirmed previous findings in social anxiety samples. Self-esteem explained a significant amount of the variance in social anxiety, even after adjusting for the effects of delusions, suspiciousness and depression.
The study shows that the LSAS-SR can be used in this patient group, that social anxiety is strongly related to both behavioral social avoidance and to self-esteem. The results support the use of this measure in assessment of social anxiety in both clinical settings and in research.
The existence of subtypes of social phobia has been questioned. Although cluster analytic methods have been used to support various subtype models, a continuous model based on total number of feared social situations seems equally plausible. In a community sample, we calculated mean similarity measures for combinations of item pairs of feared social situations. Speaking fear items were significantly more similar to each other than to other items. There was also a trend for interaction fear items to be more similar to each other than to other items. These findings suggest that fear of speaking and interactional situations may represent distinct domains of socially feared situations. They should be considered separately in delineation of treatment response and may, if replicated in clinical samples, help identify meaningful subtypes of social phobia.