I need to perform a meta-analysis on a number of single-arm studies where the only intervention is discontinuation of an agent in a group of patients. there are no control groups involved and the outcomes are dichotomous. I have been trying RevMan but it seems to require the control group data.
In this case, your "control" would be the outcome at baseline (i.e., before agent discontinuation). Of course baseline "control" is a bad control because it doesn't control for natural history, expectations, third factor, Hawthorn effect, regression to the mean, extra.
Hi Ahad,
You could try Meta-Essentials, an Excel-based tool I co-developed. We developed a specific workbook (nr 4: Differences between means: within-group design) that is aimed at dealing with this type of effects. Check meta-essentials.org or Article Introduction, comparison, and validation of Meta-Essentials:...
for details.
Comprehensive Meta-Analyst is one of the 'easy use' packages out there and has the ability to pool proportions easily. likely available in other packages as well.
Yes you can do. If your outcome is dichotomous like death you can calculate proportion of death in each study and that can be combined by any statistical software like Stata and R using meta-analysis of proportions. Same can be done for other effect sizes. May be (I an not sure) RevMan doese not have such facility because it was designed for cochrane reviews which focuses on controlled clinical trials.
Of course, you can perform a meta-analysis on single-arm studies. But, RevMan 5.3 is suitable for RCT or controlled clinical trials and not appropriate for single-arm studies/pre-post studies.
Comprehensive Meta-Analysis Software does have such facilities because it was designed to deal with a wide variety of studies including, single-arm studies. However, this software is not freely available. If you are looking for a few free statistical software then you can try these:
1. OpenMeta[Analyst]
2. MetaEssentials
Thanks
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis