Related Experiment Videos
A simple method for converting an odds ratio to effect size for use in meta-analysis.
1Department of Public Health Sciences, King's College, London, 5th floor, Capital House, 42 Weston Street, London SE1 3QD, UK. sue.chinn@kcl.ac.uk
Statistics in Medicine
|December 13, 2000
Summary
Combining different outcome measures in meta-analysis is possible. A natural logarithm of the odds ratio can be converted to effect size, enabling unified analysis of diverse study data.
Area of Science:
- Biostatistics
- Medical Research Methodology
Background:
- Systematic reviews often encounter studies reporting dichotomous (odds ratios) and continuous (mean differences) outcomes separately.
- Conducting distinct meta-analyses for each outcome type can lead to information loss and potentially biased conclusions.
Purpose of the Study:
- To demonstrate a method for unifying odds ratios and effect sizes in a single meta-analysis.
- To provide guidance on combining different statistical measures within systematic reviews.
Main Methods:
- The study proposes converting the natural logarithm of the odds ratio (ln(OR)) to an effect size (ES) using the conversion factor 1.81.
- Effect size is defined as the estimate of interest divided by the residual standard deviation.
- The validity of this approach relies on comparable variation across studies.
Main Results:
- A ln(OR) can be transformed into an effect size metric.
- This transformation facilitates the pooling of odds ratios and effect sizes in a single meta-analysis when study variations are comparable.
Conclusions:
- Researchers can combine odds ratios and effect sizes in a single meta-analysis if residual standard deviations are consistently reported.
- This unified approach enhances information synthesis and reduces potential biases in systematic reviews.