Good day, dear collegues. I am writing an original article about rather seldom histopathological parameter of tumor microenvironment and its role in the survival of patients. There are only 5 from 32 articles in pubmed which could be used for it, because they have information about number of patients which died. Will it be correct if I would write in matherials and methods criteria of articles selection, statistical methods and in discussion show cummulative odds ratio of all these 5 articles for explanation.
Dear Dmitry Zinovkin ,
You can go for analyzing the findings available in the 5 published articles, that could be used, to obtain more accurate findings on the same as well as to extract more information;.
Dear Dmitry Zinovkin ,
I have been been encouraged more in innovative works as a consequence of your acknowledgement on my answer.
Well, in fact, in an ideal world, authors of original-data studies invariably make a mini-meta-analysis of the key question they are investigating. The purpose of the meta-an. is two-fold: (a) to show the world that the question was not already satisfactorily answered (making their own work a waste of money!); and (b) to have something better than the hap-hazard quotations of 'Previous Studies' we find in today's papers as a platform for an updated answer to the research question. The 'mini' refers to the efforts spent on collecting meta-an. material - not to the care with which the material is handled and bias is avoided.
Thank you for your informative answer, Prof. Jørgen Hilden
Let me ask a little bit differently...could I make odds ratio analysis using groups data of 5 previous articles and show summary measure of adds ratio in discussion past of the manucript as it have being doing in meta analysis. Will it be correct?
Oh! That is a much more practical question. The answer must be that you go ahead and make the formal meta-a.; and then you surround the result by skeptical remarks ("Study A defined the symptom somewhat differently, but the final estimate is scarcely affected if one leaves that study out" or "About half of the reports appear to constitute by-products of data collections made for other purposes, so the meta-analytic estimate should not be read as more than an ad hoc summary of available reports" etc.). Without the skeptical comments the procedure would look (and possibly be) naive or stupid. Remember to pay attention to not only your own but also potential readers' skepticism
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis