In a systematic review, we usually can apply meta-regression to further analyse outcomes, especially when we try to explain high heterogeneity of included studies.
1. Is there any other condition can we apply meta-regression analysis?
2. Is the target outcome limited to dichotomous outcome?
3. any limitation in study type of included studies? could it be applied to case control or cohort, even case series?
When we are looking at studying the impact of a moderator on a dependent variable that is when we can apply meta regression.For more information mail me at [email protected]
The purpose of meta-regression is to attempt to explain between-study variance in effect size (ES). The moderator variables used must be continuous and/or dichotomous categorical variables. I do not know what you mean by dichotomous outcome. In meta-regression, you are attempting to predict study ES, which is treated as being continuous in nature. You are not attempting to predict an outcome.
Dear Liang-Tseng,
Metaregression analysis, as part of a pooled analysis, is often done to explore the effect of a moderator variable on the outcome / dependent variable - concomitantly exploring sources of heterogeneity on a pooled effect estimate.
It's feasible to do this for both continuous and categorical moderator variables. The outcome would certainly be an effect estimate, hence a continuous variable.
The study design is not a limiting factor to a metaregression analysis.
I hope you find this useful.
Regards,
Godfrey
Meta-regression constitutes an effort to explain statistical heterogeneity in terms of study-level variables, thus summarizing the information not as a single value but as function. Do you mean the target outcome as effect size? If it so, it is inherently continuous.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis