Use of Rasch analysis to investigate structural validity of a set of movement control tests for the neck

Abstract

BackgroundMovement control abilities are often reduced in persons with neck pain. In physiotherapeutic practice observational tests are frequently used to assess the impaired abilities. Several tests for movement control abilities are available, but no evidence exists on how to combine and interpret them. ObjectiveThe aim was to investigate structural validity of a set of movement control tests with Rasch analysis. DesignCross-sectional study. MethodsThirty persons with and thirty without neck pain were recruited for this study. All persons performed ten movement control tests. A partial credit model was applied to investigate item fit, ordering of the item response functions, dimensionality and hierarchy of the tests. ResultsThe majority of persons with neck pain had moderate disabilities and the mean value in the Neck disability index was 10.7. Functioning of the movement control tests to measure the construct “movement control abilities” was adequate for the majority of tests. Three movement control tests showed considerable misfit. Possible explanations were a reactive movement control instead of an active control and a more challenging test position. Test difficulties and person abilities could be estimated for the complete sample. The most difficult test was “sitting rocking forward” (1.13 logits) and the least difficult test was “lifting the right arm” (−1.30 logits). The highest person ability estimate was 3.61 logits indicating that movement control tests are missing to evaluate persons with moderate neck disabilities. ConclusionModifying the existing set of tests is required to evaluate the complete spectrum of persons with neck pain.