Online Program

Developing a Composite Score Sensitive to Clinical Progression in Early Stages of Alzheimer's Disease (AD) Using Partial Least Squares Regression (302785)

Keywords: Partial least squares (PLS), early stages of Alzheimer’s disease, composite score, clinical efficacy outcome measure, clinical progression, treatment effect

No standard clinical trial endpoints exist that are sensitive to change in the early stages of AD. We have developed a composite score with enhanced sensitivity to clinical progression and further evaluated this score's responsiveness to treatment effects. A partial least squares (PLS) regression model used placebo data from four mild cognitive impairment studies to select the combination of cognitive and functional items most sensitive to change over 12 months. PLS regression coefficients from the model were used to form a weighted composite score. The composite score was compared with the original scales for sensitivity to progression and responsiveness to treatment effects. The new composite score consists of four ADAS-Cog items, two MMSE items, and all six CDR-SB items. The composite score demonstrates improvements in sensitivity to clinical decline (as measured using standardized change from baseline) of 113%, 89%, and 18% over ADAS-cog, MMSE, and CDR-SB, respectively. CDR-SB demonstrated sensitivity to clinical progression, but not to treatment effect. The composite score demonstrated sensitivity to both clinical progression and treatment effect.