Green M.F.,University of California at Los Angeles |
Green M.F.,VA Desert Pacific Mental Illness Research |
Green M.F.,SUNY Downstate Medical Center |
Green M.F.,VISN 5 Mental Illness Research |
And 161 more authors.
American Journal of Psychiatry | Year: 2011
Objective: Because reduction of psychotic symptoms in schizophrenia does not result in adequate community functioning, efforts have shifted to other areas, such as cognitive impairment. The U.S. Food and Drug Administration requires that drugs for cognition enhancement in schizophrenia show improvement on two distinct outcome measures in clinical trials: an accepted cognitive performance battery and a functionally meaningful coprimary measure. The authors examined the reliability, validity, and practicality of functionally meaningful measures. Method: In this four-site validation study, schizophrenia patients were assessed at baseline (N=166) and 4 weeks later (N=144) on performance-based (Independent Living Scales, Test of Adaptive Behavior in Schizophrenia [TABS], and UCSD Performance-based Skills Assessment [UPSA]) and interview-based (Cognitive Assessment Interview and Clinical Global Impression Scale for Cognition) candidate coprimary measures. In addition, cognitive performance, community functioning, and clinical symptoms were assessed. Both full and short forms of the performancebased measures were evaluated. Results: All measures were well tolerated by patients, had adequate test-retest reliability, and showed good utility as a repeated measure. Measures differed in their correlation with cognitive performance, with performance-based measures having stronger correlations than interview-based measures. None of the measures had notable floor or ceiling effects or missing data. Conclusions: Among the full-form measures, the UPSA was judged to have the strongest overall properties. Among the short forms, the TABS and UPSA appeared to have the strongest features. Use of the short forms saves time, but at the cost of lower test-retest reliability and weaker correlations with cognitive performance.