Utilizing a Bayesian Hierarchical Model to Design Quality into a Clinical Trial and Allow Compliance with ICH E6 R2 Quality Tolerance Limits
Featuring Christine Wells, Senior Statistical Scientist at Roche Products Ltd
Abstract: 'Utilizing a Bayesian Hierarchical Model to Design Quality into a Clinical Trial and Allow Compliance with ICH E6 R2 Quality Tolerance Limits'.
ICHE6 R2 has mandated the use of Quality Tolerance Limits. Roche have utilized a Bayesian Hierarchical Model methodology, inspired by the Bayesian Meta analysis example in Berry et al (2011 ). Fixed parameters specify the prior for all unknown parameters, a conservative prior can be used or it can be informed by historical data and profound medical knowledge. The Prior is combined with the observed data (events and exposure) to define a posterior the parameters, computed using Markov Chain Monte Carlo algorithm. We then use percentiles of the posterior distribution for rate to establish limits which can in turn help to establish QTLs and Secondary Limits along with profound medical knowledge. QTLs and Secondary Limits are used to manage parameter risk at the study level and drive quality at the site level by identifying sites with parameters lying beyond the predefined limits (Trent Alert). This presentation will discuss the methodology and demonstrate the outputs.
About the speaker:
Chris Wells is a Senior Statistical Scientist who has a total of 23 years experience in the industry. For the last 11 years Chris has worked for Roche Products Ltd where for 4 years she led the Statistical Monitoring Team which during the past year has also included the application of Quality Tolerance Limits. More recently her work is involving the implementation of Data Surveillance and Advanced Analytics.