Hello,
I am doing a Sistematic review and meta-analysis of interventional techniques in headaches.
I am using revman and when calculating the effect size of the meta-analysis the MD vs SMD, the I2 drops from 95% to 50%. I wonder why? I am inclined to use MD despide its bigger heterogeneity because all studies use the same Pain scale (VAS), right ?
Thanks
Intuitively, I think this make sense. Because all studies did use the same scale to measure pain, the difference between studies are may be mostly due to the variability of effect between their study populations. So when you calculate the SMD this difference is taken out.
See here - http://handbook-5-1.cochrane.org/chapter_9/9_2_3_2_the_standardized_mean_difference.htm
Dear David, not all the studies measure the Pain in the same Scale (Some use VAS, other use NPR, some use scale to 100, other to 10...see an example here https://www.dropbox.com/s/ggfk4fa48tzmjfu/Tramadol%20for%20neuropathic%20pain%20in%20adults%20%28Review%29.pdf?dl=0), and the basal pain is not always at the same point ( the higher the pain at the beginnin, the higher the effects magnitud), so, to macht the outcomes and be able to make comparison between them you must use SMD when assessin pain between studies
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis