Resolving Prior-Data Conflict

It's well acknowledged that late phase clinical trials in rare and complex diseases can benefit enormously by borrowing from historical data.

While in the Frequentist framework, borrowing amounts to pooling via appropriate matching. In the Bayesian framework this is achieved via the use of informative priors that are constructed using historical data. Importantly, arbitrary borrowing from historical data can have detrimental effects on the operating characteristics of the design.

In this Cytel Trial Design Innovations  on-demand webinar, Rajat Mukherjee presents two ways to control the influence of the prior data on the posterior inference under Prior-Data conflicts.