Meta-analyze dichotomous data: Do the calculations with Log Odds Ratios and report Risk Ratios or Risk Differences
This paper describes a method to convert meta-analytic results in (log) Odds Ratio to either Risk Ratio or Risk Difference. It has been argued that odds ratios are mathematically superior for meta-analysis, but risk ratios and risk differences are shown to be easier to interpret. Therefore, the proposed method enables the calculation of meta-analytic results in (log) odds ratio and to transform them afterwards in risk ratio and risk difference. This transformation is based on the assumption of equal significance of the results. It is implemented Meta-Essentials: Workbooks for meta-analyses.
|Keywords||risk ratio, risk difference, (log) odds ratio, meta-analytic results|
|Note||The authors thank Alan Chang, and contributors to the Second Workshop on Synthesizing Complex Statistical Models at the University of Cologne, Germany, for their helpful comments.|
van Rhee, H.J, & Suurmond, R. (2015). Meta-analyze dichotomous data: Do the calculations with Log Odds Ratios and report Risk Ratios or Risk Differences. Retrieved from http://hdl.handle.net/1765/78636