Lies, damn lies and statistics (Mark Twain) and thanks to Rasch analysis

In everyday clinical practice, we are forever asking patients whether they are the same, better or worse, and for everyday practice their answers and our memories often suffice. However, in chronic diseases such as neurology and rehabilitation including corresponding clinical trials, we need measures of the amount of change. Usually disability (activity limitation and participation restriction in modern terminology)1 is the most appropriate aspect of disease to measure because it is more meaningful for patients. It addresses the consequences of an underlying pathology and impairment deficits better than quality-of-life measures. Among the scales which have been developed for measuring disability are the 10-item Barthel index with altogether 30 response categories and the Functional Independence Measure motor scale with 13 items and 91 response categories. Both are widely used, and both have demonstrated their scientific soundness in various diseases with classical test theory-based …