ABSTRACT
While some interventions can use directly ascertained measures such as blood pressure or grip strength, most health outcomes, particularly in the context of long-term conditions such as arthritis or multiple sclerosis, use patient completed questionnaires to ascertain the magnitude of some relevant construct. Both classical and item response theory (IRT) approaches would undertake a confirmatory factor analysis (CFA) on an existing scale prior to further analysis. "Differential item functioning" (DIF) occurs when construct-irrelevant covariates interfere with the relationship between construct levels and item response. Originally referred to as "item bias", in an educational setting and early health settings, it was used to ascertain bias in testing by race or gender. The presence of minimum and maximum scores has always been understood to be problematic and is often reported. The ability to detect change is an important characteristic of most health outcome scales, and comparisons of responsiveness are often made between different scales.
