Development and validation of the Rasch-based Depression Screening (DESC) using Rasch analysis and structural equation modelling

Questionnaires for the assessment of depression benefit from modern test construction like item-response-modelling. We developed two parallel 10 item depression questionnaires, the Rasch-based Depression Screening version 1 (DESC-I) and 2 (DESC-II), by combining Rasch analysis and structural equation modelling on patient samples suffering primarily from a mental disorder or from somatic diseases. Both scales base upon a Rasch homogeneous item bank and proved unidimensionality and good model fit. Cut-off scores with good sensitivity and specificity were developed using ROC analyses. Results suggest that DESC may be appropriately used to screen for depression and may be beneficial for repeated measurements.