Practical relevance of slowing decline on composite cognitive scales
M. Colin Ard, University of California San Diego  *Steven Edland, UCSD  Nandini Raghavan, Janssen R&D 

Keywords: IRT, composite endpoints

Both planned and recently initiated clinical trials in Alzheimer’s disease will test treatments on subjects with normal cognition using novel composite endpoint cognitive scales composed of a linear combination of those subscales and items from existing cognitive scales that are most sensitive to the earliest stages of disease. Only limited normative data are available for these composite endpoints, limiting our ability to describe the practical relevance of change in scores and for example treatment efficacy in the event of a positive trial. We will use available data from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) cohorts and clinical trial data available from the Alzheimer’s Disease Cooperative Study to more formally characterize stages of progression measured by the proposed composite scales. Location parameters from item response theory models will be used to characterize the order in which individual items are progressively lost as disease advances. Examining these scales in this way will help us better understand and describe clinical efficacy of treatment in the event of a positive clinical trial.