STATISTICAL THINKING, SYSTEMATIC REVIEWS, AND META-ANALYSES FOR BIOMAKERS AND MEDICAL TESTS

• Activity Number: 551 - Statistical Machine Learning and Artificial Intelligence in Multi-Parametric Quantitative Imaging
• Type: Topic Contributed
• Date/Time: Thursday, August 11, 2022 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistical Computing
• Author(s): Arkendra De*
• Companies: Agilent Technologies
• Keywords: Artificial Intelligence; Statistical Thinking; Registration; Real-World Data ; Digital Biomarkers; Clinical Accuracy

Abstract

Motivated by findings from the Artificial Intelligence community, this paper starts with reviewing the meaning of statistical thinking. The meaning of statistical thinking will be used as motivation for discussing key ideas and concepts to consider when deciding whether to combine published clinical studies or different real-world datasets for clinical accuracy assessment of medical tests and biomarkers. The concepts discussed in this paper are meant to appeal to a wide inter-disciplinary audience, including health regulatory agencies, developers of medical tests and biomarkers, and statisticians.