I've been doing meta-analyses on medication use and/or exposure during pregnancy for a while. Because ethical barriers exist regarding RCTs, the best evidence in this field usually comes from prospective cohort and case-control studies both of which may have different types of biases.
My question is; would you combine the case-control and cohort data in the same meta-analysis or would you prefer to do them seperately? Is combining both in the same meta-analysis definetely an unacceptable thing or is it a choice?
I'd very much appreciate your suggestions.
Hi Yusuf,
I suggests that you show the combined results as the overall outcome. Then you can perform subgroup analysisby including corhort or case-control studies seperately and show the results in a table. Because the I-square value of the meta-analysis of case-control studies might be more than 50% indicating random-effects model should be used, while the I-square value of the meta-analysis of corhort studies less than 50% indicating fixed-effects model should be used. However, you cannot choose both fixed-effects and random-effects model at the same time. Therefore, you cannot put the subgroup analyses results together in one figure on this occasion.
Best regards
Kai
Hi Cem,
I would prefer 2 separate analyses because a meta-analysis is only useful if the studies included are comparable (you don't want to mix apples and pears)
I don’t know how many studies (and cases) you have and how comparable they are.
Best wishes Bernke
You should not put them together.. It just create chaos. You can put them differently and make 2 articles.
I would do 2 separate analyses and then a combined forest plot and discuss the results.
Happy New Year
Dear all,
Thank you for your time and answers.
Sofia, thank you for the examples.
Wish you all a happy, healthy and successful new year.
It is convenient to do a meta-analysis combining cohort and case-control would be tantamount to mix what is possible and ask the impossible.
The meta-analysis is a research method that was designed to combine double-blind randomized clinical trials and controlled trials (RCTs), usually performed to evaluate efficacy and safety of drugs for use in humans, not having a sample research enough to make valid inferences, thus combining research results with similar design, size was possible to give greater consistency and internal validity of the results.
What if it is possible with these research designs (case-control and cohort) are Systematic Reviews Based on Scientific Evidence.
Have you looked online at the Cochrane guide for the analysis of non-randomised study designs? It covers most of what has been discussed and more
I think its possible but then your analysis won't suffice the criteria for meta-analysis. However, sub-group analysis is the valid option.
Hi Yusuf,
It is possible to combine cohort studies and case-control studies in a meta-analysis. However, if such a decision results in significant heterogeneity you would want to explore its source, it may not always be the study design.
Regards,
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis