Related Experiment Videos
A computer program to estimate power and relative efficiency of flexibly matched case-control studies
T Stürmer1, O Gefeller, H Brenner
1Department of Epidemiology, German Centre for Research on Ageing at the University of Heidelberg, Heidelberg, Germany. til.sturmer@post.harvard.edu
Methods of Information in Medicine
|January 10, 2006
Summary
This study presents a method and program to optimize control selection in case-control studies by varying matching factor proportions. It helps researchers determine the best matching strategy for increased statistical power and efficiency.
Area of Science:
- Epidemiology
- Biostatistics
Background:
- Case-control studies are crucial for etiological research.
- Flexible matching strategies can enhance study power and efficiency.
- Varying the proportion of a matching factor in controls is a key aspect of flexible matching.
Purpose of the Study:
- To introduce a method and computer program for calculating power and relative efficiency in case-control studies.
- To evaluate flexible matching strategies with varying proportions of a dichotomous matching factor in controls.
Main Methods:
- The program calculates the expected variance of the combined Mantel-Haenszel odds ratio.
- Power is determined using the standard error of the odds ratio under the null hypothesis.
- The method assesses power and efficiency across all possible proportions (0-100%) of the matching factor in controls.
Main Results:
- The program estimates the optimal prevalence of the matching factor in controls, which may differ from case prevalence.
- It quantifies the loss in power and efficiency associated with suboptimal matching.
- The optimal prevalence of the matching factor in controls can be identified for specific study scenarios.
Conclusions:
- The developed method and program facilitate the assessment of flexible matching strategies.
- These estimations aid in deciding when to cease optimizing the degree of matching during control recruitment.
- The tool supports informed decisions regarding control selection to maximize study power and efficiency.