Cognitive predictors of symptom change for participants in vocational rehabilitation

Background: This study explored the relationship between cognitive impairments and their impact on rate of symptom improvement in a 26 week work therapy program. Method: Data were abstracted from the archives of 157 outpatients with SCID confirmed DSM-IV diagnoses of schizophrenia or schizoaffective disorder who participated in at least 13 weeks of our 6-month rehabilitation research program. A comprehensive neuropsychological assessment was performed at intake. Symptoms were assessed using the PANSS at intake and at 6 months. Work performance on the job was evaluated biweekly using the Work Behavior Inventory. Results: With initial PANSS score and number of weeks of participation treated as covariates, logical memory was the only significant predictor of the slope of symptom change as measured by PANSS total score. The predictive power of logical memory remained when the model included improvement in work performance. Discriminant analysis showed that logical memory correctly classified 79.8% of participants as improvers or non-improvers. Implications: We conclude that participants with better logical memory may have been better able to integrate their positive work experience into a self-narrative of recovery that favorably influenced their symptoms. Implications for further research are discussed.