Hi all.
I am doing a brain imaging DTI meta-analysis in PTSD patients. There are 2 articles which use the same subjects but the subjects were scan in different time point (one at 6 month and one at 24 month) and the results are also different. In this case. Should include both of the studies in my meta-analysis?
The question in my mind is whether there are two studies and two reports or one study and two reports of that study. Based on what you have said, I would tend to consider it is the latter. If so, the data from the two reports should be combined. Presumably your meta-analysis software can handle outcomes from multiple time points. If not, it can be done by hand. See Chapter 24 in Introduction to Meta-Analysis by Borenstein et al. (2009).
If the study design is the same and the only thing that changes is the duration of the ‘experiment’ after which the outcome is measured, in my opinion I would say that you can combine them (i.e. consider the two publications as a single study reporting outcome at two different time points). Depending on the type of effect size you are adopting (e.g. if the subject are compared to the same control), this could originates non-independency in your data which can be handled in most common R packages for meta-analysis.
- Badges
- Science topic