Rasch analysis of the Fatigue Severity Scale in Italian subjects with multiple sclerosis.

Abstract

To perform a psychometric analysis of the Fatigue Severity Scale (FSS) using Rasch analysis in a sample of Italian subjects with multiple sclerosis. The 9-item FSS was administered to 156 Italian-speaking subjects with multiple sclerosis. Responses were assessed using Cronbach's alpha, item-remainder correlations, factor analysis and Rasch analysis. The FSS showed good internal consistency (Cronbach's alpha = 0.95), but presented problems in the rating categories and failed to fit the Rasch model. After re-coding and deletion of item 1 (infit mean-square (MnSq) = 1.65; outfit MnSq = 1.72), the 8-item version fitted the latent construct that the scale was intended to measure and showed satisfactory reliability indices. The scale was not optimally targeted to the sample, but no sex or age bias was found. The 8-item FSS shows better psychometric properties than the 9-item version. However, a significant ceiling effect emerged in our sample of Italian adults with multiple sclerosis and, consequently, there is a targeting problem for patients with low-level disability. To consider this a valid tool for use in clinical practice and research, further studies with a larger sample of subjects with multiple sclerosis are needed.