ABSTRACT
BACKGROUND: Whether in the context of monitoring disease progression or in assessing the effects of interventions, a major challenge in dementia research is determining when an individual has undergone meaningful change in symptoms and other relevant outcomes such as cognitive test performance. The challenge lies in differentiating genuine improvement or deterioration from change in scores due to random and systematic error. BODY: In this review, we discuss the advantages and limitations of available methods for assessing individual-level change in the context of key challenges, including imperfect and differential reliability of scores, and practice effects. We discuss indices of reliable change and the use of composite and item response theory (IRT) scores. CONCLUSION: We conclude that IRT-based approaches hold particular promise because they have the flexibility to accommodate solutions to a wide range of issues that influence the accuracy of judgements of meaningful change. We close by discussing the practical implications of adopting IRT-based approaches.
