A Comparison of Several Methods for Analyzing Data from Thorough QT Studies

• Activity Number: 62
• Type: Topic Contributed
• Date/Time: Sunday, August 1, 2010 : 4:00 PM to 5:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Hong Tian*+ and Wenqian Qiao and Jaya Natarajan+
• Companies: Johnson & Johnson and Rutgers University and Johnson & Johnson
• Keywords: Thorough QT study ; Intersection union test ; Power ; Size

Abstract

Currently, the intersection union test (IUT) is the standard method for analysis of the data from a thorough QT (TQT) study. In this work, we compared the methods proposed in four recent publications (Eaton et al., 2006; Boos et al., 2007; Cheng et al., 2008; Liu et al., 2009) to IUT using a simulation study. The methods proposed by Cheng and Eaton did not produce considerable improvement over IUT in the power and size of the test. The power of the Boos' method was substantially improved over IUT, but the size of the test was found to be greater than 5%. The performance of Liu's method was model-dependent and sensitive to the sample size. With a small or medium sample size, Liu's method was comparable with IUT; with a large sample size, Liu's method offered significant improvement in the power but failed to maintain the size of the test.