Statistical and Study Design Issues in Clinical and Translational Research

• Activity Number: 567
• Type: Roundtables
• Date/Time: Wednesday, August 12, 2015 : 12:30 PM to 1:50 PM
• Sponsor: Mental Health Statistics Section
• Author(s): Mohammad Rahbar*
• Companies: The University of Texas Health Science Center
• Keywords: Confounding ; Effect modification ; Multicollinearity ; Autism spectrum disorder ; Mental health ; Interaction

Abstract

In recent years, due to extraordinary advancement in computational capabilities and data-driven decision processes, there has been significant growth in the number of statistical techniques and research design methodology that can be used for investigating the etiology of various diseases, including mental health disorders. The lack of awareness of these new developments in biostatistics methodology can be compounded by the tendency for individual clinical and translational (CT) studies to have either too few study subjects, too much random noise in the study data, or too much potential for bias. When the most methodologically robust and efficient design and analysis strategies are under-used, the reliability and reproducibility of research findings suffer. In this roundtable, we will discuss potential areas to which statisticians can contribute to improve the quality of mental health research. The discussion points include lack of controlling for potential confounding, effect modification, and multicollinearity. We will primarily use the autism spectrum disorder for illustration and discussion.