Abstract:
|
Despite substantial investments, treatments to slow the progression of Alzheimer's disease have been elusive. Statistical consideration of power and sample size is a central concern, especially as we move to performing trials targeting the earliest stages of disease, when treatments may be most effective, but when symptoms of disease are subtle and treatment effects difficult to discern. To address this concern, statistical methods to derive new outcome measures more sensitive to the earliest stages of disease have been developed. This talk will illustrate by way of example the various techniques that have been developed to construct more sensitive outcome measures, including methods to derive composite endpoints formed by combining scales or subscales from existing assessment batteries, and the use of item response theory and other weighting methods to optimize the performance of new scales given their component elements. Based on application to placebo arm data from completed trials, these methods result in outcome measures with substantively improved signal-to-noise properties and promise to increase the probability that effective treatments of disease will be identified.
|
ASA Meetings Department
732 North Washington Street, Alexandria, VA 22314
(703) 684-1221 • meetings@amstat.org
Copyright © American Statistical Association.