A meta-analysis is generally performed after a literature search in preparation for either writing a review article or a research proposal. In order to perform a meta-analysis, you must read and become expert in a certain area of scholarship. This is a great deal of work. Since I am engaged in laboratory research, teaching and some clinical work, such an endeavor would always be secondary to the other things that I am doing. There is a saying I used to hear around my lab that 'you only work on a review article when you have no data for original research papers'. That is the kind of thing that happens where there is an extreme paucity of research funding. At this moment, although I do not have sufficient funding to do many experiments, or even one experiment with all the controls I would like, I have sufficient funds to do some bench work, and I would much prefer that to reading and critically re-examining what others have done in a meta-analysis. I hope this answers some of your questions.

Thanks very much sir for your prompt and informative reply. In theory i am aware of how to go about doing a meta-analysis but haven't had a chance to do one myself. I wish to do that as a part of the learning process. To certain extent i agree with your views that one resorts to review when there is scarcity of data for original research papers but then having a first hand experience of doing a meta-analysis/systematic review would add to one's research skills as well. I once again thank you for your valuable suggestions.

Ravi, check out the Cochrane Collaboration’s website for more information on the methods for meta-analysis and systematic reviews. To do the statistical analysis, I suggest you use the package “meta” in R or the command “metan” in STATA.

Thanks Andre. Will check out the Cochrane website

Meta analysis is a really huge work with lot of effort but at the end you can find the secret pleasure of science

to build an action plan you can organize your work as the following:

*find the formulation of the research problem and question.

*identify the literature review plan (key word, selection criteria, literature database)

*select the suitable statistical model

finalize the report

for more information you can find details on the following link:

www.cochrane.org

feel free to contact me for further details

Good Luck

Mohamed

http://handbook.cochrane.org/ is the Cochrane's online manual, also available as a help file in their Rev Man software at http://ims.cochrane.org/revman. Also go to www.jstatsoft.org -- specific articles of interest are http://www.jstatsoft.org/v30/i07 (an Excel add-in) and http://www.jstatsoft.org/v36/i03 (the 'metafor' package in R)

hi there, systematic reviews and meta-analyses are imperative to our understanding and very useful to look at larger sample sizes thereby giving us a much better understanding. I agree that the cochrane handbook is really good but another good source for information is http://www.york.ac.uk/inst/crd/ which I find really useful. For the actual meta-analysis REVMAN is really good and user friendly. If you are going to embark on a systematic review and meta-analysis I suggest you get some training first - even online as people often underestimate the amount of work involved and if you want to get published in a high quality journal they will be looking for good quality - including double sifting and quality assessment. Good luck

Regarding the report of your results, you should also check out the PRISMA statement (http://www.prisma-statement.org/). I suggest you use its checklist and flow diagram when preparing your manuscript for publication.

I agree that the best way to learn is through practice and collaboration but you might want find this useful for learning the theory and for reference: Susan Fowler has brought together various resources, including links to guidance on how to conduct reviews, on one website http://beckerguides.wustl.edu/content.php?pid=382127&sid=3131645

Do you have an idea for a review? If so, join the discussion forum evidence-based health, you might find someone there looking for a partner:

https://www.jiscmail.ac.uk/cgi-bin/webadmin?A0=evidence-based-health

I can participate with you to do a meta-analysis, i have a good experience with dealing with it and i have already published 7 systematic review with metaanalysis in Cochrane, Human reproduction update and fertility and sterility

We have created a very user friendly software that in addition can run bias adjusted meta-analyses - see www.epigear.com - it is called MetaXL and most importantly it is FREE. If you would like, I would be happy to advise on the quantitative aspects of your meta-analysis, so long as you are using MetaXL

I echo the suggestion to consult the Cochrane Collaboration. In addition, since you specifically mention the practical aspect of conducting a meta-analysis, I suggest downloading RevMan - the software program in which you will enter the necessary data. Further, (and taking a step back) I suggest you also evaluate the quality of the primary studies using an assessment tool such as GRADE, as the study quality and consideration of various other factors (such as patient important outcomes) will influence your overall interpretation of the pooled result. www.gradeworkinggroup.org

I do not entirely agree with Nancy Lloyd's suggestion regarding GRADE. It was not developed to assist interpretation of a systematic review. Its purpose is, in the authors' words, "…a system for grading the quality of evidence and the strength of recommendations that can be applied across a wide range of interventions and contexts". So it is more about applying results to guidelines. Ref: GRADE working group. Grading quality of evidence and strength of recommendations. BMJ. 2004;328(7454):1490.

The Cochrane center os excelente resource also You must visit "Consort statement" and Equator Network.

Bandolier and PRISMA can be interesting

Best Regards

E Latorre

IASP Neuromodulation Committee

What type of study design are you examining?