We explore common challenges faced when designing and conducting clinical trials with sample size re-estimation (SSR). Common SSR methods are explored along with their pros and cons. We have also examined some real-world examples of these methods, while taking a look at what advantages their use can bring.
Sample Size Re-Estimation (SSR) is a type of adaptive trial design in which the total sample size can be re-estimated at an interim point in the trial. Its use has become increasingly common and important in the clinical trial landscape. Due to the complex nature of these adaptive trials, some unique challenges are faced.
SSR can largely be split into two areas, blinded SSR (BSSR) and unblinded SSR (USSR). These aim to target different aspects of the uncertainty around the initial sample size determination of the trial.
BSSR is very useful in tackling the possible misspecification of any nuisance parameters, for example the variance. Whereas USSR will instead aim to focus on providing an unblinded estimate for the effect size and so aims to address any uncertainty regarding the magnitude of the initial effect size estimate used in the sample size determination.
In this webinar, we examine both of these areas and discuss some of the popular methods within each area. Utilising some real-world examples to illustrate the challenges and solutions these methods provide.
Download the data as shown in the webinar
|Example 1||Example 2||Example 3||Example 4||Example 5||Example 6|
|Blinded Sample Size
Re-estimaton for Inequality Using Internal Pilot
Blinded Sample Size Re-estimation for Two Sample χ2 Test for Inequality using Internal Pilot
|Conditional Power for Two Means
||Interim Monitoring and Unblinded Sample Size Re-Estimation for Two Means (2)||Interim Monitoring and Unblinded Sample Size
Re-Estimation for Two Means (6)
|Group Sequential Test of Two Means|
Looking for more resources?