Memory impairment in schizophrenia: a meta-analysis.

Abstract

Memory impairment is well documented in schizophrenia. Less is known, however, about the exact magnitude, pattern, and extent of the impairment. The effect of potential moderator variables, such as medication status and duration of illness, is also unclear. In this article, the authors presented meta-analyses of the published literature on recall and recognition memory performance between patients with schizophrenia and normal comparison subjects. Meta-analyses were conducted on 70 studies that reported measures of long-term memory (free recall, cued recall, and recognition of verbal and nonverbal material) and short-term memory (digit span). Tests of categorical models were used in analyses of potential moderators (clinical variables and study characteristics). The findings revealed a significant and stable association between schizophrenia and memory impairment. The composite effect size for recall performance was large. Recognition showed less, but still significant, impairment. The magnitude of memory impairment was not affected by age, medication, duration of illness, patient status, severity of psychopathology, or positive symptoms. Negative symptoms showed a small but significant relation with memory impairment. This meta-analysis documented significant memory impairment in schizophrenia. The impairment was stable, wide ranging, and not substantially affected by potential moderating factors such as severity of psychopathology and duration of illness.