New Developments for Using R in the Biopharmaceutical Industry Paul Schuette Scientific Computing Coordinator Office of Biostatistics FDA/CDER/OTS/OB Paul.Schuette@fda.hhs.gov www.fda.gov

Disclaimer This presentation reflects the views of the author and should not be construed to represent the FDA's views or policies.

Statistical Software Clarifying Statement “FDA does not require use of any specific software for statistical analyses, and statistical software is not explicitly discussed in Title 21 of the Code of Federal Regulations [e.g., in 21CFR part 11]. However, the software package(s) used for statistical analyses should be fully documented in the submission, including version and build identification. As noted in the FDA guidance, E9 Statistical Principles for Clinical Trials, ‘The computer software used for data management and statistical analysis should be reliable, and documentation of appropriate software testing procedures should be available.’ Sponsors are encouraged to consult with FDA review teams and especially with FDA statisticians regarding the choice and suitability of statistical software packages at an early stage in the product development process. ” (May 6, 2015) https://www.fda.gov/downloads/forindustry/datastandards/studydatastandards/ucm587506.pdf

R for Data Visualization With packages such as shiny, plotly, r2d3, various html tools, R offers more capability for interactive data visualization than most statistical scripting languages. Can we modernize and improve on traditional tables, figures and listings (TFLs)?

Open-Source Tools for Monitoring Clinical Trial Safety, Jeremy Wildfire DIA-ASA Biopharm Working Group safetyGraphics R package, Liver toxicity (DILI) Nephrotoxity QTc prolongation Bone marrow toxicity Collaboration between clinicians, statisticians, programmers

Open Source Tools, Wildfire continued How do we support the development and maintenance of open source tools? Government Grants (NIH, FDA, etc.) Successful application rate typically low Industry What’s in it for me? Academia Professional Societies Consortium

Developing Powerful Shiny Applications for Life Sciences: Best Practices and Lessons Learned, Eric Nantz Focus on necessary infrastructure: very helpful to share experiences. Shiny interfaces Automation Best Practices (git for version control, etc.) Shiny may not be designed for heavy duty computing, but can be enhanced to provide the necessary utility. Reproducibility (shinymeta)

Use of R for Regulatory Submissions, Min Lee Challenges of using R for regulatory submissions: Fear of failure Legacy software/code Challenges with using R Base R Version updates Package Updates Technical fixes? Docker Package Mangers (Packrat, etc.)

FDA Experiences R is currently used by some OB reviewers for regulatory review work R is widely used in academe, most new reviewers are familiar with R R has great graphics and data visualization tools R graphics have shown up on product labels R Shiny is being used for both internal and external uses R package SABE is in development, supplements guidance on topic R/shiny programmers are sought for a new Analytics and Informatics group in the Office of Biostatistics

CID Pilot Program Complex Innovative Trial Design (CID) Pilot Program Joint CDER, CBER 5 year pilot program PDUFA VI commitment to enhance FDA capability Trial designs include, but are not limited to: Complex adaptive designs, Bayesian designs (including the possibility of an informative prior), Other novel designs

CID Pilot Program continued “Initial priority will be given to trial designs for which analytically derived properties (e.g., type I error) may not be feasible and simulations are necessary to determine operating characteristics.” Federal Register, Vol. 83, No. 169, Thursday, August 30, 2018, Notices pp 44274-44277

CID, continued A detailed simulation plan and report are expected as part of the CID meeting request and meeting package submissions. “Detailed simulation report that includes the following: a. Example trials in which a small number of hypothetical trials are described with different conclusions. b. Description of the set of parameter configurations used for the simulation scenarios, including a justification of the adequacy of the choices. c. Simulation results detailing the simulated type I error probability and power under various scenarios. d. Simulation code that is readable, adequately commented on, and includes the random seeds. The code should preferably be written in widely-used programming languages such as R or SAS to facilitate the simulation review.” I

CID Page https://www.fda.gov/drugs/development-resources/complex-innovative-trial-designs-pilot-program

Summary It’s an exciting time for Statistical Programming and Data Science Interactive Data Visualization for Safety Analyses Shiny can be computationally complex, as well as simple Accepting R means having to overcome fear of failure and fear of change. R is part of FDA efforts to enhance and modernize reviews, and ultimately improve public health

Questions?

Additional Slides/Backup

Product Label https://www. accessdata. fda

Another Product Label R Graphic. Drug for the reduction of elevated intraocular pressure (IOP) in patients with open-angle glaucoma or ocular hypertension https://www.accessdata.fda.gov/drugsatfda_docs/label/2017/208254lbl.pdf

Data Anomaly Detection Use open source software to detect potential data problems Example of CRADA software output DABERS: Data Anomalies in BioEquivalence R Shiny app. Used for PK/PD profiles. Cooperative Research and Development Agreement (CRADA) with CluePoints for detecting anomalous clinical trial sites. www.fda.gov