A fully automatic tool for development of population pharmacokinetic models

Area of Science:

• Pharmacometrics
• Computational Biology
• Drug Development

Background:

• Population pharmacokinetic (PK) models are crucial for drug development and regulatory evaluation.
• Manual development of PK models is complex, iterative, and time-consuming.
• Existing methods require significant user expertise and effort.

Purpose of the Study:

• To present a novel tool for the fully automatic development (AMD) of population PK models.
• To implement the AMD tool within the open-source Pharmpy library.
• To evaluate the performance and efficiency of the AMD tool.

Main Methods:

• The AMD tool integrates modules for structural, inter-individual variability (IIV), inter-occasional variability (IOV), residual unexplained variability (RUV), covariate, and allometry models.
• Implementation within the Pharmpy open-source library.
• Evaluation using 10 real-world PK datasets, assessing structural, IIV, and RUV models across multiple sequences.

Main Results:

• The AMD tool produced generally consistent structural models across evaluation sequences.
• Final models demonstrated lower Bayesian Information Criterion (BIC) values compared to published models.
• Visual predictive checks indicated comparable model quality, with reasonable run times.

Conclusions:

• The AMD tool is a promising solution for rapid and automated population PK model building.
• It has the potential to enhance the application of modeling and simulation in drug development.
• The tool facilitates efficient PK model development, aiding pharmaceutical companies and regulatory agencies.