One Sensitivity Analysis Approach for Responder Analyses

• Activity Number: 134
• Type: Contributed
• Date/Time: Monday, July 30, 2012 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Hua Guo*+
• Companies: Merck
• Keywords: responder analyses ; risk ; Bayesian ; continuous variables ; dichotomized outcome

Abstract

Clinical trials often involve primary efficacy endpoints measured as continuous variables. However, a dichotomized outcome by classifying individual patients as responders or non-responders using a meaningful clinical difference is also interesting to physicians. This is known as a responder analysis. In this study, I calculate the probability of individual patients falling below a clinically relevant cut-off value, as a measure of risk, directly based on continuous data and continuous distributions in a Bayesian framework. Furthermore, I derive the risk by treatment group and provide estimations for three risk measures (risk difference, relative risk and odds ratio between groups). Through simulation studies, I demonstrate that the new approach based on a continuous model is more powerful than the customary approaches using dichotomous outcomes such as logistic regression and estimators based on a Chi Square distribution. The new approach is intended to serve as a sensitivity analysis approach and is illustrated in one example in patients with Type 2 diabetes.