Related Experiment Videos
On a hybrid method in dose finding studies.
F Bretz1, J C Pinheiro, M Branson
1Novartis Pharma AG, Basel, Switzerland. frank.bretz@novartis.pharma.com
Methods of Information in Medicine
|February 11, 2005
Summary
This study introduces a hybrid method combining multiple testing and modeling for dose-response studies. It offers robust error control during significance testing and flexible dose estimation, proving powerful for identifying dose-related trends and target doses.
Area of Science:
- Biostatistics
- Pharmacometrics
- Toxicology
Background:
- Dose-response studies are crucial for understanding biological effects.
- Traditional methods often lack flexibility in dose estimation or rigorous error control.
Purpose of the Study:
- To develop and illustrate a hybrid method combining multiple testing and modeling for dose-response analysis.
- To enhance the power of dose-response signal detection and enable precise target dose estimation.
Main Methods:
- Utilized a set of candidate dose-response models and contrast tests with optimized weights.
- Employed familywise error rate control for selecting appropriate models.
- Applied inverse regression techniques for target dose estimation, such as minimum effective dose.
Main Results:
- The proposed hybrid method demonstrated comparable power to existing tests for detecting dose-related trends.
- Successfully estimated target doses of interest, including the minimum effective dose.
- A real data example confirmed the practical advantages of the hybrid approach.
Conclusions:
- The combined approach offers a powerful tool with rigid error control and flexible dose estimation.
- The method can be extended to more complex models, including those with covariates and factorial structures.