Doctors and Data Analysis: A Dangerous Mix?

• Activity Number: 516
• Type: Topic Contributed
• Date/Time: Wednesday, August 7, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Section on Teaching of Statistics in the Health Sciences
• Author(s): Steven Grambow*+
• Companies: Duke University-Veteran's Affairs
• Keywords: introductory statistics ; reproducible research ; scientific misconduct

Abstract

Recent allegations of scientific misconduct at Duke and other institutions have served to amplify a growing national discussion about rigor and quality in clinical and translational research. An important component of this discussion relates to the content of statistical curricula for physician researchers. Traditional approaches to statistical education often involve a strong focus on common statistical procedures and how to analyze them using standard statistical software. Concerns about whether this approach meets the needs of physician learners, supports NIH goals for the proper conduct of clinical and translational science studies, and leads to robust and reproducible study results is a topic of discussion at Duke and elsewhere. This presentation will explore these issues and provide recommendations consistent with planned changes to the statistical curriculum of the Duke Clinical Research Training Program that emphasize statistical literacy, team based relationships with statisticians, and fundamental principles of reproducible research.