Wonkyung Byon (1), Phylinda Chan (2), Carol Cronenberger (1), Haiqing Dai (1), Jennifer Dong (1), Steve Riley (1), Ana Ruiz-garcia (1), Mike K Smith (2), Kevin Sweeney (1), Michael A Tortorici (1)
(1) Global Clinical Pharmacology, Pfizer, U.S.A.; (2) Global Clinical Pharmacology, Pfizer, U.K.
Objectives: Population pharmacokinetic (POPPK) modelling has become an integral part of model-based drug development and a standard component of regulatory submissions. Some of the assumptions and choices made by analysts before and during model building can be subjective, and are often influenced by the analyst's experience, expertise, and tools used. Within a large pharmaceutical organization, the development and implementation of a more standardized, objective and transparent procedure for conducting and reporting POPPK analyses would greatly facilitate the consistently and quality of output.
Methods: In our attempt to improve consistency, efficiency, and provide a more systematic approach to model building we developed a guidance for POPPK analyses. We considered aspects pertaining to base, full and final models, model selection and diagnostics within the Pharmacometrics group in order to elucidate some initial recommendations. An internal wiki repository was then created to function as a hub for collating viewpoints across the Pfizer Global Clinical Pharmacology (PGCP) organization. An editorial board was formed to consolidate these recommendations, capture pertinent examples, draft and revise the guidance. The draft guidance was circulated to PGCP and senior management for review and endorsement.
Results: The resultant guidance consists of four areas: 1) Considerations prior to conducting a POPPK analysis; 2) Considerations for base model development; 3) Development of a final model, including covariate model building; and 4) Standardizing graphical/numerical diagnostics. The "standardized practices" included, but were not limited to, the following: 1) Development of an analysis plan; 2) Thorough data checks in advance of analyses to understand data and eliminate potential errors; 3) Inclusion of structural covariate parameters in the base model to ensure model stability when highly influential covariates are known; 4) Incorporation of systematic procedures for covariate model building to improve consistency and harmonization across analyses; 5) Sensitivity analyses to check and challenge assumptions; 6) Development of a population modelling analysis report.
Conclusions: The importance and necessity of implementing a systematic, streamlined, and standardized approach to optimize and harmonize the processes which contribute to the POPPK analysis cannot be overstated. As population modelling is an area of continually evolving science and technology, the guidance will be regularly updated and revised.
Reference: PAGE 22 () Abstr 2737 [www.page-meeting.org/?abstract=2737]
Poster: Other Modelling Applications