Sample size determination often requires dealing with significant uncertainties such as effect size estimates, model choice and defining our objective of interest.
In this webinar, we examine how to find the appropriate sample size for a pilot study, how to calculate the power for the MaxCombo procedure for non-proportional hazards survival data and how Bayesian methods can provide a more practical framework for equivalence testing.
Pilot studies have been widely used to provide meaningful estimates for effect size and other parameters. However, these estimates are often highly variable and can mislead trialists on matters such as sample size.
Methods have been proposed which can quantify the uncertainty of pilot study estimates and provide a better idea of the pilot study size needed to achieve a trialist’s expectations.
Model selection is often a difficult task when designing a study, especially when there is uncertainty in how the data will actually look during the study.
However, modern methods such as MaxCombo for survival analysis and MCP-Mod for dose-finding allow researchers to evaluate and find the most appropriate model while maintaining error control.
Equivalence studies, which test if two treatments are equal, are widely used for generic medicines and medical devices.
Equivalence is usually tested using confidence intervals or two one-sided tests. But recent innovative Bayesian approaches such as “Regions of Practical Equivalence” (ROPE) could allow easier integration of prior data and a more intuitive framework for understanding equivalence.
So join us for this webinar as we look at strategies to deal with uncertainty in sample size determination.
Duration: 60 minutes
|Explore The Data
||Get The Deck
Download and explore the data yourself. Data files include: