The Cognitive Screening Scale for Schizophrenia (CSSS). Part 1. Design and structure of the scale

This paper presents the construction of the CSSS - a short screening scale intended for diagnosis of cognitive deficits among people with schizophrenia. 160 persons (124 with schizophrenia and 36 healthy controls) were tested using the initial version of the CSSS scale consisting of 11 subscales. Correlation analysis between the subscales' results was carried out, as well as confirmatory factor analysis, internal consistency analysis of the scale, IRT (item response theory) analysis of the items' difficulty, and analysis of the scale's accuracy as a classifier. One factor (overall cognitive efficiency) explains 37% of the variance of the subscales' results. The scale has satisfactory internal consistency (Cronbach's alfa = 0.83). Subjects with schizophrenia achieved significantly lower scores than healthy subjects. The area under the ROC curve (AUC) for discriminating between subjects with schizophrenia and healthy subjects was 0.83. Cut-off point of 16 raw points is 86% sensitive and has 70% specificity. The form of the tool that has been achieved as a result of presented analyses suggests that this scale has a potential to fulfil the assumed goals, which will be tested during continuing validation studies.