A Mixed-Effect Logistic Regression Model for Predicting Virologic Response in the Treatment of Hepatitis C

• Activity Number: 134
• Type: Contributed
• Date/Time: Monday, July 30, 2012 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Weiping Deng*+ and Kenneth Koury
• Companies: Merck and Merck
• Keywords: Hepatitis C ; SVR ; Mixed-effect logistic regression

Abstract

In clinical trials that evaluate treatments for eradicating the Hepatitis C Virus (HCV), the primary efficacy endpoint is Sustained Virologic Response, SVR, defined as undetectable HCV-RNA at (follow-up) week 24 post-treatment. The commonly used method for predicting SVR is a logistic regression model with treatment and baseline demographic and disease characteristics as exploratory variables, and SVR as the response variable. During the treatment of patients with HCV infection, HCV-RNA values are also available at earlier time points (e.g., at treatment weeks 4, 8, 12, 24, etc.). The HCV-RNA values obtained during the treatment period are correlated with each other, as well as with SVR status. The proposed mixed-effect logistic regression model uses the on-treatment and follow-up responses as repeated measures and accounts for the heterogeneity between patients. This model is compared to the regular logistic regression model in predicting SVR using simulated data and actual data from clinical trials. It provides an alternative approach to analyzing the virologic response data in the treatment of HCV.