The structure and dimensionality of the Inventory of Depressive Symptomatology Self Report (IDS-SR) in patients with depressive disorders and healthy controls

Abstract

BackgroundThe Inventory of Depressive Symptomatology Self Report (IDS-SR) is a widely used but heterogeneous measure of depression severity. Insight in its factor structure and dimensionality could help to develop more homogeneous IDS-SR subscales. However previous factoranalytical studies have found mixed results. Therefore, the present study tested which factor structure underlies the IDS-SR and, in addition, if the factors can be used as unidimensional subscales. MethodsConfirmatory factor analysis (CFA) was done to identify the best-fitting factor structure. The study sample consisted of 2600 individuals (mean age 40.5±12.1). We assessed model fit in 4 groups: 957 Major Depressive Disorder (MDD) patients, 450 remitted MDD patients, 570 patients with an anxiety disorder and 623 healthy controls to test the consistency of model fit. Rasch analyses in the full sample were used to evaluate and optimize the unidimensionality and psychometric quality of the factors. ResultsCFA indicated that a 3-factor model fits the IDS-SR data best and is consistent across groups, with a ‘mood/cognition’ factor, an ‘anxiety/arousal’ factor and a ‘sleep’ factor. In addition, Rasch analyses indicated that the ‘mood/cognition’ and ‘anxiety/arousal’ factors could be optimized to be used as unidimensional subscales. LimitationsThe fit of only 4 models was tested, ranging from a 1- to 4-factor model. ConclusionsThe IDS-SR is a heterogeneous instrument with a multifactorial underlying structure. It is possible to measure more homogeneous symptomatology with IDS-SR subscales, which could be useful in clinical practice and scientific research.