Related Experiment Videos
Profile likelihood-based confidence intervals using Monte Carlo integration for population pharmacokinetic
Takashi Funatogawa1, Ikuko Funatogawa, Akifumi Yafune
1Clinical Research Coordination Department, Chugai Pharmaceutical Co., Ltd., Chuo-ku Tokyo, Japan. funatogawatks@chugai-pharm.co.jp
Journal of Biopharmaceutical Statistics
|April 6, 2006
Summary
This study introduces profile likelihood confidence intervals using Monte Carlo integration for population pharmacokinetic (PPK) analysis, improving parameter estimation accuracy over traditional linearization methods.
Area of Science:
- Pharmacometrics
- Statistical Modeling
Background:
- Population pharmacokinetic (PPK) analysis commonly uses nonlinear mixed effects models with first-order linearization, which can be inaccurate.
- Traditional confidence intervals for PPK parameters often rely on asymptotic methods based on Fisher information, which may lack precision.
Purpose of the Study:
- To propose and evaluate profile likelihood-based confidence intervals for PPK parameters.
- To address limitations of linearization methods and asymptotic confidence intervals in PPK analysis.
Main Methods:
- Utilized Monte Carlo integration to avoid linearization issues in nonlinear mixed effects models.
- Developed and implemented profile likelihood methods for confidence interval estimation.
- Conducted simulation studies to assess the performance of the proposed method.
Main Results:
- The proposed profile likelihood method using Monte Carlo integration demonstrated improved accuracy for confidence intervals in PPK analysis.
- Simulation results indicated better performance compared to traditional methods.
Conclusions:
- Profile likelihood-based confidence intervals with Monte Carlo integration offer a more accurate approach for PPK parameter estimation.
- This method provides a valuable alternative for robust pharmacokinetic modeling.