Meta-Analysis without standard deviation?

[Guest]

Hello everyone!

Is it possible to do a meta-analysis without the standard deviations? I have 10 studies that are all using the same pain scale (0-10). They all have the mean and sample size but only 5 have the standard deviation. I've read that I can do a weighted mean difference, or a raw (unstandardized) mean difference. I tried just doing metan n1 mean1 n2 mean2 but I have a RR output which isn't correct for a continuous pain outcome, right? I would really appreciate any help/ advice!

Kind regards,
-Guest

[Tiago Pereira]

Guest,

There are over 100 techniques/approaches/formulas to derive an approximate sample standard deviation for meta-analysis. One can get reasonable values from confidence intervals, standar errors, graphs, range, interquartile range, t statistics, p-values, similar studies, large studies, etc. Start here:
[1] Furukawa TA, Barbui C, Cipriani A, Brambilla P, Watanabe N. Imputing missing standard deviations in meta-analyses can provide accurate results. J Clin Epidemiol. 2006 Jan;59(1):7-10 [2] Chapter 16: Special topics in statistics. Editors: Julian PT Higgins, Jonathan J Deeks and Douglas G Altman on behalf of the Cochrane Statistical Methods Group. http://handbook-5-1.cochrane.org/ Hope they help. Tiago

[Guest]

Thank you!!! That is so helpful!!