Using R in a GxP Environment


The Data Science team in Pfizer’s Vaccine Research and Development division (VRD) creates and maintains validated applications used during high-throughput clinical testing that enable advanced analytic and reporting requirements. SAS has long been the de-facto standard for analyzing data in a regulated GxP environment. Web deployment of these applications has been the best approach, and Pfizer VRD has developed several mid-tier applications in Java that submit batch SAS processes on a High Performance Computing grid. Pfizer VRD’s high level approach is the same across different assay platforms data are pulled from a combination of electronic files and Oracle databases and analyzed, results are written back to an Oracle database, and electronic output files are made in various formats (e.g. PDF). The regulated nature of Pfizer VRD’s work and the difficulty in deploying R-based applications over the web have previously been an impediment to the use of R, but new tools such as RStudio’s Shiny Server Pro have helped us overcome those challenges. This presentation focuses on a comparison of the architecture used to deploy our SAS applications and the infrastructure required to deploy R-based applications to meet GxP requirements. Real life examples will be provided to illustrate the usefulness of this platform in a regulated laboratory environment.

Presented at 2018 Conference