Abstract:
|
The hypothesized treatment effect and estimated nuisance parameters play an important role in a study's sample size and power calculation. Point estimates for these parameters are often calculated using historical data. However, the uncertainty in these estimates is rarely addressed. We present a hybrid classical and Bayesian procedure that formally integrates prior information on the distributions of the hypothesized study parameters into the study's power calculation. Conditional expected power, which averages the traditional power curve using the prior distributions of the unknown parameters as the averaging weight, is used, and the sample size is found that equates the pre-specified frequentist power and the conditional expected power of the trial. A method for sample size re-estimation in which the prior distributions for the nuisance parameters are updated using partial trial data is also discussed. We show that using the proposed method for sample size determination during the design phase helps to protect against misspecification of the nuisance parameters, reducing the probability of an underpowered study.
|