A new look at the Western Ontario and McMaster Universities Osteoarthritis Index using Rasch analysis.

Abstract

The Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) is a valid and widely used instrument for the assessment of osteoarthritis patients. We examined whether applying Rasch analysis, a new item-response theory, supports its metric properties. The metric properties were examined in a cohort of patients with osteoarthritis of the hip or knee. In an exploratory approach, a validated German version of the WOMAC was analyzed using Rasch analysis for its unidimensionality and construct definition. One hundred fifty-eight patients were included in the study. The 3 WOMAC domains--pain, stiffness, and physical function--were unidimensional when examined separately. In an analysis including all items, the stiffness items did not fit the model. This indicates that they are distinct from the pain and function items. Item difficulty analyses revealed redundancy between corresponding pain and function items addressing the same task (e.g., "pain while standing upright" and "difficulty while standing"). Applying Rasch analysis to the 3 WOMAC domains confirms their unidimensionality. However, the pain and function items seem to represent the same construct. Thus, it may be possible to simplify these domains to avoid redundancy. The impact of reducing items must nevertheless be examined.