Confirmatory factor analysis of competing PANSS negative symptom models: data from OPTiMiSE first-episode...

The negative symptoms of psychosis are heterogeneous, which complicates efforts to understand their pathophysiology and develop effective treatments. Factor analytic studies of the Positive and Negative Syndrome Scale (PANSS) have reported two factorial negative symptom models, expressive deficit and social amotivation, albeit with different compositions. Although models derived from other assessment scales have been directly compared, no study has previously applied this approach to PANSS.

Our objectives were to (a) to establish which negative PANSS-derived factorial model provided the best fit to our data, (b) test its stability and (c) determine its clinical and demographic correlates.

A cohort of medication naive or minimally treated patients with first-episode schizophrenia (n = 446) were assessed using the PANSS scale before and 4 weeks after amisulpride treatment. Confirmatory factor analysis was performed to test five PANSS models. Hierarchical multiple regression was conducted to examine the associations between identified dimensions and clinical and demographic variables.

A nine-item PANSS model comprising social amotivation and expressive deficit dimensions outperformed the other models: comparative fit index = 0.98, goodness of fit index = 0.97, Tucker-Lewis index = 0.97, root mean square error of approximation = 0.06 (CI 90%: 0.04-0.08), Bayesian information criterion = 191.9, Akaike information criterion = 101.7. At baseline, the social amotivation dimension was associated with more severe depression whereas the expressive deficit dimension was associated with younger age. Both dimensions at baseline were associated with poor functioning, but expressive deficit to a lesser extent.

A nine-item PANSS model incorporating social amotivation and expressive deficit dimensions appeared to best reflect the underlying structure of negative symptoms in our sample.