An outcome prediction model for schizophrenia: A structural equation modelling approach

Introduction: Although it is well-known that several factors such as symptoms and cognition are related with functional outcome in schizophrenia, the complex nature of the disorder makes necessary to study their interaction by means of a more analytic method than simple linkages approaches. Material and methods: One hundred and sixty-five patients with schizophrenia underwent a clinical evaluation (including clinical symptoms, insight, affective symptoms and premorbid adjustment). Neurocognition was represented by a 5-factor structure obtained by confirmatory factor analysis from a neurocognitive battery. The estimation for outcome was obtained throughout the DAS-WHO scale, and quality of life with the Quality of Life Scale. Results: Using structural equation modeling (SEM), specifically measured-variable path analysis, a mediational model consisting of neurocognitive capacity linked to clinical symptoms and premorbid functioning showed good fit to the observed data (Satorra-Bentler chi(2) = 604.83; RMSEA= .08; SRMR = .11; NNFI = .96; CFI = .97). Processing speed, verbal memory and premorbid functioning directly predicted outcome. Verbal fluency predicted outcome both directly and indirectly via negative symptoms. Executive functions, insight, affective symptoms, and additional cognitive data did not significantly contribute to the model. Conclusions: Results suggest that negative symptoms and premorbid functioning directly predict outcome, whereas cognitive factors show more complex interactions with negative symptoms and outcome. These results should be considered for new intervention strategies. (C) 2019 SEP y SEPB. Published by Elsevier Espana, S.L.U. All rights reserved.