On the Clinical Meaningfulness of a Treatment's Effect on a Time-to-Event Variable

• Activity Number: 72
• Type: Contributed
• Date/Time: Sunday, August 1, 2010 : 4:00 PM to 5:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Qi Jiang*+ and Steven Snapinn
• Companies: Amgen Inc. and Amgen Inc.
• Keywords: Hazard Ratio ; Event Rates ; Median Survival Times

Abstract

The standard analysis of a time-to-event variable often involves the calculation of a hazard ratio based on a survival model such as Cox regression. However, a relative measure of effect such as the hazard ratio is sometimes considered to be a poor measure of clinical meaningfulness. Two absolute measures of effect are often used to assess clinical meaningfulness: 1) the absolute difference in event rates (or its inverse, the number-needed-to-treat) is frequently used in many disease areas; 2) the difference between the median survival times in the two groups is often used in oncology trials. While both of these measures appear reasonable, it doesn't seem to be well-known that they directly contradict each other. In this presentation we'll give a mathematical description of the contradiction, and examples. This contradiction raises questions about the concept of clinical meaningfulness.