In a meta-analysis which includes 3 RCTs comparing two chemotherapy strategies , for analyzing incidence of side effects, which one should be used as an effect size: Odds Ratio (OR) or Risk Ratio (RR)? Which one is more accurate? What's the difference exactly?
Dear Sara,
Deeks (2002) gives a useful overview of the pros and cons of various effect sizes (Article Issues in the selection of a summary statistic for meta-anal...
).The bottom line is that to derive the combined effect size in a meta-analysis, the (log) odds ratio is best for statistical reasons. However, after deriving this combined effect size it is useful to transform the odds ratio in to a risk ratio or risk difference because they are easier to interpret.
This procedure is implemented in Meta-Essentials (www.meta-essentials.com). A free Excel-based tool for meta-analysis.
Best regards,
Henk
Because you want to analyze the incidence, RR should be used. Risk Ratio (RR) is a ratio that used to compare the risks (incidence) for the two groups. To calculate RR, you should divide the incidence in exposed group by the incidence in the unexposed group.
The use of odds ratio (OR) in randomized controlled trials (RCTs) needs careful consideration because it might overestimate the effect size, especially if it is incorrectly interpreted as risk ratio (RR), either by the authors or the readers. However, there are occasions when either OR or RR can be used. The paper below gives some insights:
Article Potential Misinterpretation of Treatment Effects Due to Use ...
Since you are working on incidence of occurrence, RR which is the probability of a situation occurring is more appropriate.
Bonjour Mme, il me semble que vous faites un essai clinique pour comparer 2 stratégies de chimiothérapie donc c'est un essai clinique sur 2 groupes parallèles. Le RR est celui qui vous permet de savoir lequel des 2 stratégies donne le plus d'effets secondaires.
Bon courage
Risk ratio also known as relative risk is applicable to longitudinal study, like what u r doing. While odds ratio is the proxy measure of relative risk is applied in case control or cross-sectional studies.
RR for incidence. It also depends on prevalence of a problem/outcome.
There is a great interest in this topic. You may wish to read CONSORT 2010 articles on how to report parallel group randomized trials. They recommend that the odds ratio and risk difference should be reported for such trials
Article CONSORT 2010 statement: updated Guidelines for Reporting Par...
https://www.journal-surgery.net/article/S1743-9191(11)00562-0/abstract
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis