Bayesian and Frequentist Blinded Sample Size Adjustment for Risk Differences

• Activity Number: 414
• Type: Contributed
• Date/Time: Tuesday, August 5, 2014 : 2:00 PM to 3:50 PM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Andrew Hartley*+ and Anita Moghe and Savanna Steele
• Companies: PPD and PPD and PPD
• Keywords: Adaptive Designs ; Predictive Power ; Bayesian Analysis ; Sample Size Recalculation

Abstract

Adaptive sample size adjustment (SSA) for clinical trials consists of examining early subsets of on-trial data to adjust estimates of statistical parameters and sample size requirements. Blinded SSA is often preferred because it obviates many logistical complications of unblinded SSA and generally introduces less bias. On the other hand, current blinded SSA methods for binary data offer little to no new information about the treatment effect (TE), ignore uncertainties associated with the population treatment proportions, and/or depend on enhanced randomization schemes that risk partial unblinding. We propose 2 innovative blinded SSA methods (1 bayesian and 1 frequentist) for use when the primary analysis is a non-inferiority test regarding a risk difference. The Bayesian method incorporates evidence about the TE via a Bayesian hierarchical model, while protecting the blind. We compare the methods to established ones, in terms of predictive power, frequentist power, type 1 error rate, the bias of the estimated TE, and the average absolute deviation from the targeted power. We illustrate the use of the proposed methods with an example.