I have twelve articles for my MA for Gestational diabetes Mellitus. I want to calculate OR for Age>25, BMI>25,Multiparity >2 ,Primigravida. The problem is for all the risk factors(age, BMI,MP and P) values are given with Mean with Std dev
With means, standard deviations, and sample sizes, you can calculate the standardized mean difference (d). The standardized mean difference can easily be converted into the odds ratio (actually log odds ratio). The same goes for the calculation of variance. See the attached formulas from "Introduction to Meta-Analysis" by Borenstein et al.
If the individual studies are case referent and they have total case and referent numbers you might be able to apply normality assumptions to the exposure means and sds and create case referent numbers by the required exposure categories and recalculate ORs by exposure level. .
See Rothman and Greenland 2nd ed p 657 and attached Cornfield paper.
There is a very good free online tool. First use means, SDs and sample sizes for two groups (e.g., BMI 25 kg/m2) and calculate the standardized mean difference (d).
Good day! Honestly, the best way for anyone to give you the proper assistance is to provide context on your working paper/analyses. We need to determine whether or not an odds ratio or risk ratio is the appropriate summary measure to use in the particular analysis step of your interest. Feel free to message me in private, or anyone else in this thread, for a freely-detailed discussion that would otherwise improperly reveal issues/matters if tackled publicly through a ResearchGate question. Thanks!