The component structure of the scales for the assessment of positive and negative symptoms in first-episode psychosis and its dependence on variations in analytic methods. (December 2018)

• Record Type: Journal Article
• Authors: Tibber, Marc S.; Kirkbride, James B.; Joyce, Eileen M.; Mutsatsa, Stanley; Harrison, Isobel; Barnes, Thomas R.E.; Huddy, Vyv
• Is Part Of: Psychiatry research. Volume 270(2018)
• Page Start: 869
• Page End: 879
• Publication Date: 2018-12
• Subjects: Psychosis -- Positive symptoms -- Negative symptoms -- Disorganisation symptoms -- Factor analysis -- Factor structure; Psychiatry -- Periodicals; Psychiatry -- periodicals; Psychiatrie -- Périodiques; 616.89
• DOI: 10.1016/j.psychres.2018.10.046
• Languages: English
• ISSNs: 0165-1781

Abstract

Highlights: Psychotic symptoms show a complex hierarchical symptom structure. Ten+first-order factors were extracted using data reduction methods. Three second-order factors were extracted using data reduction methods. Some reported variance in symptom structure is due to analytic methods used. ABSTRACT: A secondary analysis was undertaken on Scales for the Assessment of Positive and Negative Symptoms (SAPS/SANS) data from 345 first-episode psychosis (FEP) patients gathered in the West London FEP study. The purpose of this study was to determine: (i) the component structure of these measures in FEP (primary analyses), and (ii) the dependence of any findings in these primary analyses on variations in analytic methods. Symptom ratings were exposed to data reduction methods and the effects of the following manipulations ascertained: (i) level of analysis (individual symptom vs. global symptom severity ratings), (ii) extraction method (principal component vs. exploratory factor analysis) and (iii) retention method (scree test vs. Kaiser criterion). Whilst global ratings level analysis rendered the classic triad of psychotic syndromes (positive, negative and disorganisation), symptom level analyses revealed a hierarchical structure, with 11 first-order components subsumed by three second-order components, which also mapped on to this syndrome triad. These results were robust across data reduction but not component retention methods, suggesting that discrepancies in the literature regarding the component structure of the SAPS/SANS partly reflect the level of analysis and component retention method used. Further, they support a hierarchical symptom model, the implications of which are discussed.