TL24: Challenges in Sample Size Planning for Randomized Clinical Trials
*Roy N. Tamura, Pediatrics Epidemiology Center, University of South Florida  

Keywords: sample size calculation, design parameter, analysis model, missing data

Sample size planning is a critical component that can determine success or failure in a clinical trial. With more and more complex designs and analyses being implemented in clinical trials, balancing sufficient power and costs in sample size planning to avoid recruiting an unnecessarily large number of patients to a trial becomes a challenge. Good planning for sample size needs to consider not only the trial design itself, which can be new and adaptive, but also other important elements concerning dropouts, multiple comparisons, and analysis models.

The purpose of this round table is to share experiences and discuss challenges associated with sample size planning for randomized clinical trials in the context of the following questions: (a) What is your experience in selecting the objective(s) (e.g., considering primary only, or with requirement on key secondary) when sizing a study? (b) How do we account for uncertainty in the values of design parameters? Should we use the estimated treatment effect, standard deviation or the lower/upper bound of C.I. of them based on historical trials?)? (c) Is it appropriate to estimate sample size using a statistical method different from that in the proposed primary analysis? (d) What methods have you used to account for missing data in sample size calculations?