Back to Annual Meeting

Back to Annual Meeting

APHA Scientific Session and Event Listing

Margaret L. Healey, MPH, College of Public Health, University of Kentucky, 385 Redding Rd., Apt. 168, Lexington, KY 40517, 859-227-6712, mslucc2@email.uky.edu

There are currently approximately 4 million Americans with Alzheimer's disease (AD). Projected growth in the prevalence of AD makes it a serious public health concern. There is currently no effective treatment, and the cost of this disease will become significant in the coming years. One estimate projects 14.3 cases in the U.S. by the year 2050. Early detection and prevention of AD has become a focal point of recent research. PREADVISE is a large add-on prevention trial investigating the use of anti-oxidants for the prevention of Alzheimer's disease among men enrolled in an even larger prostate cancer prevention study, SELECT. SELECT uses supplements of vitamin E and selenium in a 2X2 factorial design. PREADVISE was designed using a discrete time hazard model using a framework described by Kryscio et al. Its design incorporates assumptions in AD incidence rates, death rates, drop-ins, drop-outs, adherence, accruals, and length of follow-up. A detectible effect size in the assumed best treatment arm was determined for the desired power and significance level. Analysis will be presented showing the effects on power that could occur if there are changes to the assumptions made during the design of the trial. Possible options to regain the desired power of the trial will also be proposed. This sensitivity analysis of power and possible options to regain it may be useful in the planning of similarly designed prevention studies as well as for the PREADVISE study.

Learning Objectives:

• The participant will be able to recognize how the power of a prevention trial can be affected by changes in assumptions made in the design phase.
• The participant will be able to discuss and develop possible strategies to regain statistical power.

Presenting author's disclosure statement:

Not Answered