The low- and higher-order factor structure of symptoms in patients with a first episode of psychosis

Background: The phenotype of psychotic disorders is presumed to be heterogeneous, but the best way to describe this heterogeneity remains unclear. Objective: To examine the lower-order and higher-order symptomatic architecture of psychotic disorders by means of factor analysis and the Schmid-Leiman transformation. Methods: Patients experiencing their first episode of psychosis (n = 486) were comprehensively examined for 70 symptomatic variables, which were subjected to principal components analysis followed by a Promax rotation. First-order factors were subjected to second-order factor analysis, and influence of second-order factors on primary factors was removed using the Schmid-Leiman transformation. Results: First-order factor analysis revealed 13 primary factors that were substantially intercorrelated. Second-order factor analysis showed 5 higher-order factors with no substantial intercorrelations. The Schmid-Leiman transformation revealed that whereas the second-order factors accounted for most of the symptom covariance (63.5%), the first-order factors still accounted for an additional 36.5%. According to this transformation, five second-order factors (bipolar negative-mania, disorganization, psychomotor retardation, hallucinations and grandiosity) plus four first-order factors (depression, catatonia, bizarre delusions and paranoid delusions) best explained the factor structure of the symptoms. Conclusions: The phenotype of psychosis is more complex than previously acknowledged as it embraces a multidimensional hierarchical structure organized into nonredundant first-and second-order factors. (C) 2013 Elsevier B. V. All rights reserved.