Sample Size Calculations in Clinical Trials with Binary Data: Theoretical and Practical Issues

• Activity Number: 423
• Type: Contributed
• Date/Time: Tuesday, August 2, 2011 : 2:00 PM to 3:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Arminda Siqueira*+
• Companies: Universidade Federal de Minas Gerais
• Keywords: binary data ; clinical trials ; closed-form formula ; sample size

Abstract

Sample size, an important element in the planning stage of clinical trials, can be determined from closed-form formulas, in iterative equation solving procedures or via simulation. While closed-form formulas are practical and convenient, their accuracy varies, therefore simulated solutions are often preferred. For binary response trials, the sample size calculation depends on many factors, such as type of study (superiority, non-inferiority and equivalence), design (parallel, crossover), measure (e.g. difference, odds ratio, relative risk), test (e.g. Wald, score and likelihood ratio), solution (exact, asymptotic) and statistical method (classical, Bayesian). This paper presents several comparative analyses of both analytical and simulation-based approaches, and discusses theoretical and practical issues regarding the appropriate methods of calculating sample size for clinical trials with binary response.