Background: Complex pharmacometric analyses raise concerns about cost, time, and reliability of the modelbuilding process (MBP). The goal was to use a model feasibility assessment (MFA) process to improve the performance characteristics of the MBP.
Methods: Literature review provided a basis for a proposed mechanistic model of exenatide effects in type 2 diabetes. A study index database (SID) detailing design characteristics, interventions, and comparators of available studies was assembled and used to generate informatics to facilitate data pooling. Cross-study endpoint databases (CSED) for each endpoint were assembled and used to generate exploratory analyses (EA) of posited model relationships. A gap analysis (GA) performed during the assembly of SID and CSED identified issues regarding study design alignment, data adequacy, and the types and timing of interventions and endpoint measurements that impacted the MBP.
Results: 38 studies were reviewed and were included in the MBP. EA aided in determining functional form, providing initial parameter estimates, and specifying data programming requirements. GA was critical in choosing data for the MBP and generating design recommendations for future studies. The informatics generated during the GA and the discussions to resolve discrepancies enhanced data assembly and accelerated model-building efforts.
Conclusions: MFA provides a systematic approach to facilitate data selection and pooling and improves the performance characteristics of the MBP so that results are available for decision-making.
American Society for Clinical Pharmacology and Therapeutics (ASCPT); Orlando, Florida; April 2008
By B. Cirincione, E. Blase, M. Cummings, M. S. Fineman, Thaddues H. Grasela