Using item response theory to develop and refine patient-reported outcome measures.

There is a long tradition of incorporating patient-reported outcomes (PROs) into nursing research and practice. Classical Test Theory (CTT) has been the traditional approach used to develop and refine PROs. Item response theory (IRT) offers promise for addressing measurement problems that have been difficult to solve using CTT. This paper presents foundational concepts in IRT to illustrate how it can be used to improve the development and refinement of PRO measures, with emphasis on interpreting key IRT parameters.