Related Experiment Videos
Comparing completely and stratified randomized designs in cluster randomized trials when the stratifying factor is
1Department of Preventive and Social Medicine, University of Otago, Dunedin, New Zealand. jlewsey@rcseng.ac.uk
Statistics in Medicine
|March 18, 2004
Summary
Stratified randomization in cluster randomized trials (CRTs) improves balance and power. Stratifying by cluster size enhances statistical power when size predicts outcomes, especially with more clusters.
Area of Science:
- Biostatistics
- Clinical Trials Methodology
Background:
- Stratified randomization enhances balance in cluster randomized trials (CRTs).
- Cluster size can be a crucial factor influencing outcomes in CRTs.
Purpose of the Study:
- To evaluate the power gains from stratifying randomization by cluster size.
- To assess the impact of cluster size-outcome association on CRT power.
Main Methods:
- A simulation study using a UK general practice CRT as a template.
- Comparison of stratified randomization versus completely randomized design.
Main Results:
- Stratified randomization by cluster size significantly increases statistical power.
- Power gains are more pronounced when cluster size is strongly associated with outcome-predictive factors.
- The number of clusters influences the degree of power superiority.
Conclusions:
- Stratifying randomization by cluster size is advantageous in CRTs.
- This design is particularly effective when cluster size is a significant predictor of outcomes.
- Consider cluster size stratification for optimizing power in CRTs.