Measurement invariance and cognitive performance of the Cambridge Neuropsychological Test Automated Battery (CANTAB) in patients with early and later phase of the schizophrenia spectrum disorders compared to control subjects

Abstract

The aim of this study was to use selected CANTAB subtests to investigate the structure of cognitive abilities among the early phase of the schizophrenia spectrum disorders (E-SSD), the later phase of the schizophrenia spectrum disorders (L-SSD), and the entire clinical sample compared to the control subjects. A total of 97 E-SSD patients, 102 L-SSD patients and 188 control subjects were administered six CANTAB subtests of cognitive functioning. To use the most appropriate level of analyses, we first identified the factor structure of selected CANTAB subtests. Multi-group CFA (MG-CFA) was used to test the measurement invariance of factors between groups. General linear models (GLMs) were used to evaluate the nature and severity of cognitive dysfunction among the patients compared to the controls. CFA models fitted almost all data well, except for the young controls. In MG-CFAs the subtest scores corresponded to different latent trait levels indicating that the comparison of mean latent scores between most of the subsamples is meaningless. At subtest score level the patients exhibited extensive cognitive dysfunction when compared with the control subjects (p < .001). Current study demonstrates that the patients exhibit widespread cognitive dysfunction and that the structure of underlying cognitive abilities is not the same for the clinical sample compared to the control sample as well as for the E-SSD compared to the younger control subjects.