Evaluating Correlation and Some Misuses for Biomarker Qualification Within a Causal Correlation Framework

• Activity Number: 238
• Type: Contributed
• Date/Time: Monday, August 2, 2010 : 2:00 PM to 3:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Yue Wang*+ and Robin Mogg and Peter H. Hu and Jared Lunceford and Peggy Wong
• Companies: Merck & Co., Inc. and Merck & Co., Inc. and Merck & Co., Inc. and Merck & Co., Inc. and Merck Research Laboratories
• Keywords: Biomarker Qualification ; Correlation

Abstract

Earlier clinical decisions facilitated by various biomarkers play an increasing role in the clinical development of new therapeutics. Clinical investigators often look for evidence of association between biomarkers and clinical endpoints for internal decision making. Even though it is widely understood that correlation may not serve as a concrete proof of a causal relationship, estimated correlation to clinical outcomes continues to be used in the evaluation of candidate biomarkers in practice. In this talk, a framework will be introduced where two different types of correlations from a causal perspective are defined. Using this framework, we are able to break down the various components that impact the correlation estimates, and show clearly how some common uses of correlation can be misleading if the components making up the correlation are not accounted for appropriately.