Send to

Choose Destination
Psychol Med. 2006 Oct;36(10):1349-62. Epub 2006 Jun 7.

A systematic review of longitudinal outcome studies of first-episode psychosis.

Author information

Schizophrenia Program, Centre for Addiction and Mental Health, Department of Psychiatry, University of Toronto, Toronto, Ontario, Canada.



Existing outcome literature has had an over-representation of chronic patients and suggested a progressive course and poor outcome for schizophrenia. The current study aimed to recombine data of samples from longitudinal studies of first-episode psychosis (FEP) to describe outcome and its predictors.


A literature search (1966-2003) was conducted for prospective studies examining outcome in first-episode non-affective psychosis using the following key words: early, first, incident, episode, admission, contact, psychosis, schizophrenia, psychotic disorders, course, outcome, follow-up, longitudinal, cohort. These were pooled and analyzed using descriptive and regression analyses.


Thirty-seven studies met the inclusion criteria, representing 4100 patients with a mean follow-up of 35.1+/-6.0 months. Studies varied in the categories of outcome used, the most common being 'good' (54% of studies) and 'poor' (34% of studies), variably defined. In studies reporting these categories, good outcomes were reported in 42.2% (3.5%) and poor outcomes in 27.1% (2.8%) of cases. Predictors associated with better outcome domains were: combination of pharmacotherapy and psychosocial therapy, lack of epidemiologic representativeness of the sample, and a developing country of origin. Use of typical neuroleptics was associated with worse outcome. Stratification analyses suggested that populations with schizophrenia only, and those with prospective design, were associated with worse outcome domains.


Outcome from FEP may be more favorable than previously reported, and treatment and methodological variables may be important contributors to outcome. Significant heterogeneity in definitions and methodology limited the comparison and pooling of data. A multi-dimensional, globally used definition of outcome is required for future research.

[Indexed for MEDLINE]

Supplemental Content

Full text links

Icon for Cambridge University Press
Loading ...
Support Center