Design Clinical Trials with Multiple Endpoints:
Methods and Applications

Primary Contact/Presenter Information:

Ying Yuan

yuan@mdanderson.org
Bettyann Asche Murray Distinguished Professor
Department of Biostatistics
The University of Texas MD Anderson Cancer Center
Houston, TX, 77030, USA

​

Sheng Luo

sheng.luo@duke.edu
Professor
Department of Biostatistics & Bioinformatics
Duke University
Durham, NC, 27705, USA

Session Type:

Pre-Conference Educational Workshop

​

​

Target Audience:

Biostatisticians and clinicians in industry, research institutes and academia. Non-statisticians are expected to have basic knowledge on the concepts of hypothesis testing, type I error and power.

​

​

Technical level of the course:

​

Low to moderate. The course mainly focuses on the concepts and strategies of designing trials with multiple endpoints, illustrated with real-world trials (see below). The statistical theory of controlling multiplicity will not be covered in details.

​

​

Goals of Session:

The core goal is to introduce state-of-the-art methods to design clinical trials with multiple endpoints and their applications using high-profile pivotal trials (e.g., Impassion 130, IMbrave 150, IMpower 133, Keynote 240, keynote 604) published in New England Journal of Medicine.  

 

By the end of the session, participants will be able to:

     (1) understand the features and challenges of multiple endpoints;

     (2) master statistical methods to address the challenges of multiple endpoints;         (3) apply the introduced methods to design trials with multiple endpoints.  

​

​

Session Description:

[2 hours workshop]
Considering multiple endpoints is an effective approach to address the intrinsic high uncertainty associated with drug development and maximize the success rate of drug development. An increasing number of pivotal trials have adopted this approach, including Impassion 130, IMbrave 150, IMpower 133 and Keynote 240, resulting in drugs successfully approval by the FDA. Using multiple endpoints brings many design and statistical challenges, e.g., the regulatory concern the multiplicity issues caused by testing multiple endpoints.

​

This short course introduces the state-of-the-art methods to design clinical trials with multiple endpoints, with compliance to the regulatory guidance. The application of the methods will be illustrated using the pivotal trials. The sample size determination for multiple endpoints will be briefly illustrated using the available software (e.g., PASS).

 

The session will consist of three parts:

1. Introduction of multiple endpoints (20 mins): (YY)

• the hierarchy of families of endpoints;

• type I error rate and multiple endpoints;

• types of multiple endpoints;

2. Common statistical methods for trials with multiple endpoints (50 mins): (YY)

• Prospective alpha allocation;

• Fixed-sequence method;

• Fallback method;

• Gatekeeping testing strategies;

• Holm and Hochberge procedure.

3. Case study and discussion (50 mins)  (SL, YY)

• Impassion 130

• IMbrave 150

• IMpower 133

• Keynote 240