Related Experiment Videos
Some remarks about the analysis of active control studies.
1Division of Biometrics I, OB/CDER, Food and Drug Administration, HFD-710, 15B-45, 5600 Fishers Lane, Rockville, MD 20857, USA. lawrencej@cder.fda.gov
Biometrical Journal. Biometrische Zeitschrift
|January 3, 2006
Summary
This study compares two methods for determining if a new treatment is non-inferior to an active control. It examines statistical properties like interim monitoring and level of evidence for these active-controlled trial approaches.
Area of Science:
- Clinical trials methodology
- Statistical analysis in medicine
Background:
- Active-controlled trials (ACTs) traditionally assess non-inferiority by comparing confidence intervals to a pre-specified margin.
- Emerging methods combine current ACT data with historical placebo-controlled data for active controls.
Purpose of the Study:
- To evaluate statistical properties of different non-inferiority assessment methods in ACTs.
- To compare traditional confidence interval methods with a novel approach combining current and historical data.
Main Methods:
- The study discusses statistical properties of two non-inferiority assessment approaches in active-controlled trials.
- It considers methods for combining current study data with historical placebo-controlled data.
- Analysis focuses on interim monitoring boundaries and the level of evidence.
Main Results:
- A fixed margin can be defined for operational equivalence between methods, conditional on nuisance parameters.
- The article explores the statistical implications of these different approaches.
Conclusions:
- The study provides insights into the statistical properties of novel non-inferiority assessment methods in ACTs.
- Understanding these properties is crucial for trial design and interpretation of results.