Abstract: In the past ten years, the potential of fully Bayesian trials has been recognized by leading Pharmaceutical companies1, regulatory bodies (such as FDA2) and other institutions concerned with all types of research. However, there is still very limited understanding of the clear advantages which Bayesian trials have to offer. This is especially true for situations where traditional concepts such as power do not provide an adequate approach to study design.This talk will start with a short introduction to Bayesian thinking3 followed by real-world examples of fully Bayesian trials. An emphasis will be put on cases where Bayesian trials release their full potential: dealing with parameter uncertainty4, evidence synthesis5, and prediction. In order to actively engage the audience, a real world problem will be posed and discussed at the end of the talk.
References
1. Kakizume, T. Novartis: Challenging to Accelerate Oncology New Drug Development. http://atdd- frm.umin.jp/slide/11/kakizume.pdf accessed on April 11, 2013.
2. FDA - CDRH. Guidance for Industry and FDA Staff Guidance for the Use of Bayesian Statistics in Medical Device Clinical Trials. http://www.fda.gov/downloads/MedicalDevices/DeviceRegulationandGuidance/GuidanceDocuments/ucm071121.pdf accessed on April 25, 2013.
3. Spiegelhalter, DJ, Abrams, K and Myles, J. Bayesian Approaches to Clinical Trials and Health-Care Evaluation. Wiley.
4. Joseph, L, Gyorkos, T and Coupal, L. Bayesian Estimation of Disease Prevalence and the Parameters of Diagnostic Tests in the Absence of a Gld Standard. AJE 141, 1991.
5. Schmidli H, Wandel S, Neuenschwander B. The network meta-analytic -predictive approach to non-inferiority trials. Stat Meth in Med Res, 2011.
----------------------------------------------------
Valentin Rousson