Related Experiment Videos
Use of prior information to stabilize a population data analysis
Per O Gisleskog1, Mats O Karlsson, Stuart L Beal
1Exprimo Consulting LLP, London, United Kingdom.
Journal of Pharmacokinetics and Pharmacodynamics
|June 11, 2003
Summary
Leveraging prior study data stabilizes complex pharmacokinetic/pharmacodynamic (PK/PD) model parameter estimation. Combining data or using penalty functions improved results, with combined data and penalty functions yielding similar parameter estimates.
Area of Science:
- Pharmacometrics
- Pharmacokinetics/Pharmacodynamics (PK/PD) Modeling
Background:
- Complex population PK/PD models may lack sufficient data for parameter estimation.
- Information from previous studies can stabilize parameter estimation in new models.
Purpose of the Study:
- To explore and compare three methods for incorporating prior information into new PK/PD models.
- To assess the impact of these methods on parameter estimation and hypothesis testing.
Main Methods:
- Fixed parameter values using estimates from earlier data.
- Combined current and earlier datasets.
- Augmented the objective function with a penalty function based on earlier data (similar to Bayesian prior).
Main Results:
- All three methods stabilized parameter estimation.
- Methods combining data and using penalty functions yielded similar parameter and standard error estimates.
- Fixing parameters led to problematic hypothesis testing results; penalty functions had less severe issues.
Conclusions:
- Prior data incorporation is effective for stabilizing PK/PD model estimation.
- Combining data or using penalty functions are preferable to fixing parameters for hypothesis testing.
- Penalty functions offer a viable alternative, especially for large datasets or unavailable early data.