TL29: Propensity Score Model Development – Please Share Your Experience and Lessons Learned
*Jie (Jack) Zhou, FDA CDRH 

Keywords: Propensity score, nonrandomized study

As clinical trial designs get more innovative, nonrandomized studies are frequently used in pivotal trials, due to ethical and other practical reasons. In order to meet the statistical challenges raised by these non-randomized studies, propensity score modeling and adjustment are frequently used in the study design and data analysis. What is your experience in dealing with propensity score? Any comments? Suggestions? Complaints? You are welcome to share with us.