The factor structure of clinical symptoms in depressed inpatients with unipolar or bipolar spectrum disorders: a preliminary study.

Factor analysis of symptom structure has proven to be a valuable tool to identify dimensions of symptoms in various psychiatric conditions. This study used exploratory and confirmatory factor analysis to examine symptom structures in depressed inpatients with unipolar (n = 718) or bipolar (n = 134) spectrum disorders who were rated at admission with a psychiatric rating scale. No differences in overall symptom severity on the scale were found in the 2 samples, although different factor structures were detected with exploratory analyses. These models were modified in a confirmatory modeling procedure to improve their fit to the data, resulting in models with good, but not perfect, fits. For people with bipolar disorders, a 5-factor model fit best, with depression loading with anxiety symptoms and in people with unipolar disorders, a 4-factor model with depression loading with vegetative symptoms was found. Our results suggest that similar levels of symptom severity may have different underpinnings in the 2 groups and suggest that more comparative studies of symptoms in these 2 conditions may be useful.