Related Experiment Videos
Sample size requirements for case-control study designs.
1Division of Clinical Epidemiology, Royal Victoria Hospital, Montreal, Quebec, Canada. michael.edwardes@clinepi.mcgill.ca
BMC Medical Research Methodology
|December 19, 2001
Summary
New formulas and software are available for calculating sample sizes in case-control studies. These tools adjust for confounders and interactions, improving the accuracy of statistical power calculations for disease-exposure associations.
Area of Science:
- Epidemiology
- Biostatistics
Background:
- Existing formulas calculate sample sizes for case-control studies to detect significant disease-exposure odds ratios.
- These formulas typically adjust for a single confounder and may consider interaction effects.
Purpose of the Study:
- To extend existing sample size formulas for case-control studies.
- To accommodate multiple controls per case and multi-category confounders.
- To provide software for computing these adjusted sample sizes.
Main Methods:
- Formulas were extended to include F controls per case.
- Adjustments were made for multi-category confounders in both unmatched and matched designs.
- Interactive FORTRAN programs were developed to compute the sample sizes.
Main Results:
- Extended formulas allow for sample size calculations with varying numbers of controls per case.
- The methodology accounts for multi-category confounders and matched/unmatched designs.
- Software facilitates the exploration of disease-exposure-confounder interactions on sample size.
Conclusions:
- Available software enables the computation of adjusted sample sizes for case-control designs.
- This enhances the precision of sample size determination in epidemiological research.
- Researchers can now more accurately plan studies considering complex confounding and interaction scenarios.