|
Abstract:
|
FDA and ICH recently published several documents that encourage the drug process understanding "from pharmaceutical development studies and manufacturing experience" to "support the establishment of the design space, specifications, and manufacturing control". It is specially emphasized that "quality should be built by design" and risk management, the key component of which is risk assessment. In this presentation, the Bayesian posterior predictive approach developed by Peterson (2004) is utilized to identify design space and a reliable operating region for the production of a pharmaceutical drug substance or their intermediates. This approach not only provides a risk assessment but also accounts for modeling uncertainty and the correlation structure of the multiple responses, which can not be done by the traditional, frequentist approaches.
|
