PSI Webinar: How to use prior knowledge and still give new data a chance?

  • Dates: 19 – 19 Dec, 2018

Time: 15:00 - 17:00 UK time

The paper to be presented appeared earlier this year in Pharmaceutical Statistics (Weber, Hemmings, Koch, 17, 329-341) and is motivated by the opportunities and challenges for using Bayesian methods with informative priors to support drug development and licencing when only a small pool of patients is available, as in the case of rare diseases and paediatric population. With a small study population, meeting the expected level of evidence required for regulatory approval can be challenging. The specific focus of the paper is on the use of data‐based priors in the decision-making process of paediatric extrapolation, and results are presented comparing different frequentist and Bayesian meta-analytic methods for combining adult and paediatric data.

In this webinar, the authors will present their paper, followed by discussion on: 

 - Scientific rationale for borrowing existing data to make inference about drug effects under different settings: rare diseases, extrapolation and use of historical controls.

 - Other methods for combining historical and prospective data beyond standard meta-analytic approaches (e.g., robust MAP priors). 

 - Ways in which the Bayesian strategy could be supplemented or adapted to enable the independent contribution of the new trial data to be formally quantified and assessed. 

We are pleased to announce that the meeting will be chaired by Byron Jones (Novartis)


  • Kristina Weber (Institute for Biostatistics, Hannover Medical School)
  • Rob Hemmings (MHRA)
  • Armin Koch (Institute for Biostatistics, Hannover Medical School)


  • Nicky Best (GSK) (on behalf of the EFSPI/PSI Historical Data SIG)
  • Lisa LaVange (University of North Carolina, Chapel Hill)
 PSI Member  Free
 Non-member  £20 (plus VAT) 

Please click here to register.
Add event to: ICal Outlook Google Calendar