Do the same factors predict outcome in schizophrenia and non-schizophrenia syndromes after first-episode psychosis? A two-year follow-up study

摘要

Objective: The aim of this two-year longitudinal study was to identify the best baseline predictors of functional outcome in first-episode psychosis (FEP). We tested whether the same factors predict functional outcomes in two different subsamples of FEP patients: schizophrenia and non-schizophrenia syndrome groups. Methods: Ninety-five patients with FEP underwent a full clinical evaluation (i.e., PANSS, Mania, Depression and Insight). Functional outcome measurements included the WHO Disability Assessment Schedule (DAS-WHO), Global Assessment of Functioning (GAF) and Clinical Global Impression (CGI). Estimation of cognition was obtained by a neuropsychological battery which included attention, processing speed, language, memory and executive functioning. Results: Greater severity of visuospatial functioning at baseline predicted poorer functional outcome as measured by the three functional scales (GAF, CGI and DAS-WHO) in the pooled FEP sample (explaining ut to the 12%, 9% and 10% of the variance, respectively). Negative symptoms also effectively contributed to predict GAF scores (8%). However, we obtained different predictive values after differentiating sample diagnoses. Processing speed significantly predicted most functional outcome measures in patients with schizophrenia, whereas visuospatial functioning was the only significant predictor of functional outcomes in the non-schizophrenia subgroup. Conclusions: Our results suggest that processing speed, visuospatial functioning and negative symptoms significantly (but differentially) predict outcomes in patients with FEP, depending on their clinical progression. For patients without a schizophrenia diagnosis, visuospatial functioning was the best predictor of functional outcome. The performance on processing speed seemed to be a key factor in more severe syndromes. However, only a small proportion of the variance could be explained by the model, so there must be many other factors that have to be considered.