Facilitating interpretation of cognitive impairment: The development of reliable changes scores for the cognitive measures in the Canadian Longitudinal Study on Aging

The first objective is to develop reliable change methods, which are critical for measurement of change over time on the cognitive measures collected in the CLSA. Interpretation of cognitive assessment data requires some comparison standard, such as a cut-off score for screening tests, or, for neuropsychological tests, comparison of an individual's performance with that of cognitively healthy persons of the same age. Similarly, interpretation of change in cognitive performance over time requires an appropriate comparison standard, but the study of change over time is more complicated; for example, measurement error is unavoidable and is present at each assessment point. Change is assessed by comparing baseline performance versus performance on the same cognitive tests at a follow-up assessment. Reliable change methods detail whether any change was beyond expected fluctuations in cognitive performance resulting from measurement error or other factors, such as practice effects, maturation effects, and/or regression to the mean. Reliable change methods can be compared with external criteria, such as change in self-reported diagnosis, to create minimum clinically important difference scores. Our secondary objective is to create robust norms, which leverage longitudinal assessment to refine cross-sectional normative comparisons that can be used for clinical practice. Development of appropriate comparison standards for individual participant's change in cognitive performance over time will provide a powerful basis for measuring health, aging, and cognition in the CLSA. Reliable change scores for cognitive variables will be a foundation for future work for determining incident dementia at the individual participant level in the CLSA, and will serve as a methodological model for other longitudinal aging studies.