Hello, Michael. Thanks for your kind and rapid reply, and sorry for the inconvenience of characters.
Yes, the primary studies reported the n, the mean percentage change and its standard deviation, but some did not report the original value of before- or after-intervention, and there is only one group because these are uncontrolled before-and-after studies. So you mean we can not perform meta-analysis based on the mean percentage change and its standard deviation? For studies which have reported the original value of before- and after-intervention, can I just simply treat the before-intervention values as the values of control group and treat the after-intervention values as the values of intervention group? I will be grateful if you can send me some of your publications of meta-analysis on before-after study, thus I can learn how to perform the analysis. With best regards. Qiang Yue From: Michael Dewey Date: 2013-04-27 07:28 To: qiangmoon; wvb CC: r-help Subject: Re: [R] using metafor for meta-analysis of before-after studies At 03:27 27/04/2013, Qiang Yue wrote: >Hello, Dr. Viechtbauer. > >I am trying to perform a meta-analyis on a group >of before-after studies using Metafor. I read >your webpage including your correspondence with >Dr. Dewey >(https://stat.ethz.ch/pipermail/r-help/2012-April/308946.html), >who also conducted a similar study. These >information is very hepful, but I have one >additonal question which I wonder if you can >give me some instruction. The question is as follow: > >These studies which we are trying to analyze are >performed on the same subject before and after >the adminstration of intervention. Most studies >reported the the Mean¡ÃSD of percentage change, >i.e., the Mean¡ÃSD of (value of ¡®after¡¯-value >of ¡®before¡¯)/value of >¡®before¡¯¡Ã100%£¬without reporting the Mean¡ÃSD >of value of ¡®after¡¯ or value of ¡®before¡¯. So >I want to know if it is possible to perform >meta-analyis using the value of percentage >change, and if it is possible to calculate the >¡®sdi¡¯ (the standard deviation of the change >scores) using the SD of percentage change. Unfortunately not all the characters in your email appeared correctly here but if I understand you correctly the primary studies have reported (for each group?) mean percentage change and its standard deviation (and presumably the n). So you just treat them like any other mean and standard deviation. If I understand the very last part correctly you would need more information than we have to back calculate change on the original scale from change on the percentage scale. > >Thank you very much, I am looking forward to your reply. > >With best wishes. > > > > >Qiang Yue M.D. >Visiting Scholar of IMHR, University of Ottawa >1145 Carling Avenue, K1Z 7K4, Ottawa, ON, Canada >Tel: 613-722-6521 ext. 6554 >Associate Professor of Radiology >Department of Radiology, West China Hospital, Sichuan University >Chengdu, 610041, China [[alternative HTML version deleted]] Michael Dewey i...@aghmed.fsnet.co.uk http://www.aghmed.fsnet.co.uk/home.html [[alternative HTML version deleted]]
______________________________________________ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
