RESUMEN
Most neuropsychological studies on schizophrenia suffer from sample selection bias, with male and chronic patients being overrepresented. This probably leads to an overestimation of cognitive impairments. The present study aimed to provide a less biased estimate of cognitive functions in schizophrenia using a population-representative catchment area sample. Schizophrenia patients (N = 89) from the prospective Mannheim ABC cohort were assessed 14 years after disease onset and first diagnosis, using a comprehensive neuropsychological test battery. A healthy control group (N = 90) was carefully matched according to age, gender, and geographic region (city, rural surrounds). The present sample was representative for the initial ABC cohort. In the comprehensive neuropsychological assessment, the schizophrenia patients were only moderately impaired as compared to the healthy control group (d = 0.56 for a general cognitive index, d = 0.42 for verbal memory, d = 0.61 for executive functions, d = 0.69 for attention). Only 33 % of the schizophrenia patients scored one standard deviation unit below the healthy control group in the general cognitive index. Neuropsychological performance did not correlate with measures of the clinical course including age at onset, number of hospital admissions, and time in paid work. Thus, in this population-representative sample of schizophrenia patients, neuropsychological deficits were less pronounced than expected from meta-analyses. In agreement with other epidemiological studies, this suggests a less devastating picture of cognition in schizophrenia.
