"Does This Treatment Cause That Outcome"? Assessing a treatment effect in modern randomized clinical trials.

Speaker: Dr Stephen J Ruberg

Clinical trials are meant to be scientific experiments to assess the effect of a new or experimental treatment versus a control (e.g., placebo or another active treatment). There are well-accepted approaches for making rigorous scientific and statistical comparisons between experimental and control treatments – blinding, randomization, pre-specified criteria for success. However, individual patients must be (a) treated in a manner that is appropriate for their particular disease circumstances (ethics), and (b) allowed to leave a clinical trial at any time (exercise free will). Consequently, patients may not take their assigned treatment as defined by the clinical trial procedures, and there is often incomplete data on patients. How can we still make inference about whether an experimental treatment provides sufficient efficacy and safety information for regulatory approval (e.g., FDA or European medicines Agency, etc.)?

Dr. Ruberg has spent 40+ years in pharmaceutical drug development. In that time, he has held senior level positions in Statistics and Data Management as well as leadership roles for teams of drug developers. In his last 10 years at Lilly, he was the founder and Distinguished Research Fellow for Advanced Analytics, a group of highly trained and experienced professionals in statistics, mathematics, computational chemistry, econometrics and other quantitative/computing sciences. The group tackled some of the most complex and important problems facing Lilly using leading-edge quantitative and computing approaches.

Sponsored by John Peer